From 1e61af91d36483615ec91b8067dca0701735900a Mon Sep 17 00:00:00 2001
From: Jessie Chappel <137214022+jrchapp3@users.noreply.github.com>
Date: Tue, 1 Oct 2024 16:27:28 -0400
Subject: [PATCH] Add google analytics tracking

---
 .../figure-html/unnamed-chunk-186-1.png       | Bin 47891 -> 47856 bytes
 .../figure-html/unnamed-chunk-194-1.png       | Bin 266446 -> 266974 bytes
 _output.yml                                   |   4 +-
 coding-best-practices.html                    |   9 ++
 data-import-from-pdf-sources.html             |  99 ++++++++++--------
 ...ort-processing-and-summary-statistics.html |  57 +++++-----
 data-manipulation-and-reshaping.html          |   9 ++
 data-sharing-through-online-repositories.html |   9 ++
 data-visualizations.html                      |   9 ++
 data-wrangling-in-excel.html                  |   9 ++
 downloading-and-programming-in-r.html         |   9 ++
 fair-data-management-practices.html           |   9 ++
 file-management-using-github.html             |   9 ++
 header.html                                   |   9 ++
 improving-coding-efficiencies.html            |   9 ++
 improving-data-visualizations.html            |   9 ++
 index.html                                    |   9 ++
 intoduction-to-statistical-tests.html         |   9 ++
 ...ive-modeling-for-environmental-health.html |   9 ++
 ...riable-comparisons-and-visualizations.html |  49 +++++----
 normality-tests-and-data-transformations.html |   9 ++
 ...-experimental-design-and-example-data.html |   9 ++
 search_index.json                             |   2 +-
 ...machine-learning-model-interpretation.html |   9 ++
 supervised-machine-learning.html              |  11 +-
 two-group-comparisons-and-visualizations.html |  33 +++---
 ...earning-part-1-k-means-clustering-pca.html |   9 ++
 ...-2-additional-clustering-applications.html |   9 ++
 28 files changed, 322 insertions(+), 104 deletions(-)
 create mode 100644 header.html

diff --git a/_main_files/figure-html/unnamed-chunk-186-1.png b/_main_files/figure-html/unnamed-chunk-186-1.png
index c9cb855dba83d492fb23cb1bc1883e29137a9ae0..a246f40070e5e3e6ef4aae03b931132bc3f51fa7 100644
GIT binary patch
literal 47856
zcmd?Q2UHZz7A`tN4w94P%#e{BBn(K7LlzKFaz=8HFd!f~=Q!j6L<Gqa6v;>i5fBh$
zBq$k?%xizofBet6Yu&ZpUH66cUQ^SH>h9WmS9NuFec#?Y2Bxh-fJcJ|0)YtBR2B6=
zAWSF-g#HEx9XRr?mGv74gjNO9G*AMnAdnOYWCxOx0!ab;WRP7mZ}<zo@E80~>x3d-
z0_T7ezfS(z?4;~~k}PG{ER~!rl??0`rDGf95}LGrZEjjxiCS)MT5iC;Mk^60gC;(r
zraoe(zT#%S;^zKR76DRLg&nq`igtiXDWCunlkMz4zoh=!lI@a#(rlN!XxH3q*9`1$
z9Kuzc^zJ+9n>ZVoI%g-j<RrTw`(2G50Lt8q&3|e8wI#YG0;R?+anP-%#;pd}ue;}^
z`dHfd7H0Ze*#{Kogk?CyKE50KSUS!_Ca%ac-V2)G4NVAwCjv&h0R<4Cm6%wQSTmSd
zvz}O6nK-zfICzsR1xl6zOavwWW#}*ae*Kz(zvO0MznHuTl$+E*^^9Pxtf<)Rmip2Z
z7bFWhk`<$h4YL}8Tg{1A<EB-U2o$#(K*2BT25Sb_YX)y>*4Jy+Z)%09YV&+wiqX7!
z^{Uwp)NCi!oD3AZe`#n2w#DW}pxiXy+_ai0y|z|;ZLQW}quybs*=euU<!1EO!=%Sn
zr$5?fFi~qT(QU8>D2anLfS$p@_4UE^o57p)!JC^Aq}~Wpf22zPeU-ucYNOF=<I!5n
ziPX@EmsXRlK_9(UKYFWu^Z|<cM?cMZf35itqlI*z#bl|)WV^*?pd>E>ru0Pp>hUj&
zfbC}S=H^qF$@22@dSU2#joW%n;(D$7`e5SvAkaBLIk4RT`hng0&CS<Bt&MECjcn*f
zj{Ihx;#RTlZcD&Tv)#>L;>}|6&0_NnAaT9sm$APt2f7Pb4QK%Vzi}(sA_1l<?qgLG
zFA#|S*{>fs33Xr#0x^Nq6y**4^S^iABvB3&?BUZ6h~1CVVKI@ujq~+!TZ|X`UcH-;
z9Ee6z_A5psOVE5ufxzqJ($ih1QX9n(ma!NL{K(CGiLv3w_LJeSa3Nn)aml>wAoXhq
zA0dK7wUdZQ>WAOq)?d#w+=6#rY`@rju`N(kS#y50oDfjBTODZp)4RJ+C-|F%pkgtM
z(}2D9whtzZlYH2i8LG=kUR^?r;3UU0)Ww03<3(0;VsPR`Dj1TZkw+>NbK`+y6^daL
zkszgF7&2A?UIMgN`kyvyGz_MVRKR4vB2<AlSI^8RzKNe4!Y(B)coVPuB?5#=fi|N8
zx${?@7$HOa|G65>5_FD(hYH(%@Z|vx3<iu9au+!!Iu;4SGz<lq3i=g{&_sHYqRrf8
zhH6BBl)`Yq+Xci3c3^DZM1X#1<4lkOZ5}GX10fg=F~F_4!wl^tLu*t7ZcjEbVi^}~
z$^zUXO&r)>1c;9WjU1B$1L>y#&%^=8qLKYyCpHp)0^QmgyKFj45Cb_Edy@@De!j11
zoI6{rGNgjRi44c&Z2LY~K9Eo7GOqxVGODEmPbBX?yh*<lyfU7wF9WFyNS)u9HK&_D
z(LE#q<5iELH?p)18mcP^%p|!{(=MJ5vSL>+!RZjpOK-2$&a4CLdaNhG>H_V2W<)Ma
zQew^J>mN(kS<%Q1%(0O%4=6QdnChm{-_<9j;wLL3)=hawGx)N>jbrrj_0Hy8p_+S!
zLA*3jwavDS#saal8b253p*p|xAVqk1^%sz6$_NJy`PpWq)JbT3J&w26mZ1aheRPM&
zySJ_97Syk^OJAr3Q!vKi!NSLD%MQ8%j*P-}59{|A;!=~q=-y}BNatk37O>NKXLB^x
zK(lm%p+%)NHV(AKg_)d<#KD8#`t2>mw&5P7Zq&@VCI+qHPcp>u&%(xBG4~J~VZ(#B
zz@UwT&abD@iy=h<YJsSnd)sp;o+9?P4{|>J9~_qOhv)R4-jl=?AQGF;7@Tp1)9{PO
zozOY{Opf;YcAE%F6y*Iavlau9*WxniS<alyl?_ovpW!RPa0y0JAdsluwI?g|4Neui
zwDU#DHUv2DgXq)Y&S%m69Y2dw=6y|Rn4!~o_jw>*>~6)~pOTW9(v|YD`(^K$Lf3`Z
zZs_&#)a5_F`IsrV6ldOLHm&7960JA&yz?{Hs0I`nrZP)Flh~oM5J>FE<J|84$apt%
z)GW?EMR7=F4|KktH2)0M(vST*R&N``z9M!ReR;9W^K{K6CWRfk^s6f&brftQJ+4c+
zTOZlj*!UR3-}421_xvuru27c6tXBA-wd1zB*0s;~Hy>(C=DRbNtgT&+)1qE%Go{+z
zA)kIO3a<#j#c3kJh)B4S@<yn%Bk-_Q%^+$m0r*Osudh|_ybN&GE!I7pG=@)2xLfN&
zDhVeRzX+{NG4U6FM7o1)^*;+eP<2}ojFnRFt2H&Kk5o!rUnphfT-w+DP!bw}fiU}$
zS7-+@SW?iCcr&P8VsbJL{$QxC0v+f--At287$S^Tt6dD&I`5eI;eq>Jd6In|_WAyn
zeZ>|+v_)FXnC%5E<6SGz()*II5~0%d(ryriX2;Q|(}(+IoAc{e`IfEf2kI31?pzgo
z5LQQjfg*-mAL!`_S2)i64h#GWX$KqFpeuJreevPkb{bF3HNHG05O~RPxpR;ZdJ3Lz
zlJ|bU-A3+1g~&IVU-^cC$oKHBewMlpDrP}J&PFNy)J6Kwp=!<!nz+yww`GMojqnvd
z?9E#cnt&;>uIPI$o1hHiP6}y*+}YS$UhYk|1cZNh(rtU<zQ3fOt~wS*r)*5s$c)Nm
znX-}F#3;rnA}=;;j$#b+gp)?IBc=J;C2@I$LBaQFW!j}r2T3y5K0D^H@;fgTKe{!s
z7&IL<USDGHs{Cyz#|p!BJ!lhaa-iKgd#Jqn706H`r~OkO1s~dkpp*^y$|$x&C`p2h
zW1b3HI-$c9PWty0+exO3pz<`u&B$SE*g`Z!0tH%P!JEtS<Y;*aE*9yTX}vjoWp_Js
z`SaIp<$%xu?`j0&b9b9>lnh~R%e<l0^S91bBE+zz61bKkLT?;tWca&<5UplbB<u;i
zMAYZdIhh}&>vvbby;~_UGszs`z58~#U-;e6LTgJ2sM-PT*u<Ji3YT#1`AQTT;_2JU
zI>J4=M_1^L??i{nt7<^Dep{?sv2Q%R=&Gj(D@LC?dkgQV(bImhKFRO<KFu0LsfLF*
zxH3@WyppZ`gvQ(ILxF7(gy*zFI=NV7LIJ`%aTctYnkwu0^m}ejBJK^{mbmyyo^*rL
z?Td=<sS8zatd|txI`t*6UC=Md360I{19O{QYnZ%Zw@|nRBMCfw(2<%&1`jdg-mZ{6
zr!G+jbAQqFJ&JebV6JYpaKyQ*jP{}4CcS_uqcOL?Ssh+I4I99aJ#z20b89crOHX{*
z>E-)e9W%deDy%qLy?XyySg!J0v+qiCd-lA0bbFv&P3r}<AxKo9TM~p-T^UaJtouo#
z_=KDmSu6?X{IuF6t^NJAV!kETWKCD1^7o|bZmFN?ay_@UoiI$jSDjo>`GlP4X#3IS
zy!|=b`bwF<#jS9IhiO#JTK0#29N)}4Y*5lf$3AhvziJP_zj*e}SOVXIp0kaN7IzZQ
zeMy2iF%1dx>iD+AEO<Bsr9fT(W#0$aDZ=ql&zEV}6*Z~^%|r$hUNX+fJklr$v+E_I
z?t|VpIIxnRFDB#>3C3R@>R~mK2OPBC{r13iB~@tiak3l7Q++VZB9S@k#S1h8`SidR
zC)Ohun(0hFGh2muo?-qsZ2BS`;iG&()etjn#k*>Z^Y5wEujdsm6k1|qRLsO)F&JPi
z)pA2??IEHhB@|C^QhAt+6F(lj_}QXAZ88>m8|g^iIFg2zelmTu<!f!`4Udw;8}D`k
z1z*I-OQ&0-an3G&QFnfTjpWP?Cgx1{%J~GU{yH^l2q&b_<SwRoJx1E-R?ZWd#>7h~
z-!yVnDl*UdoXp!fXfjpACGg7bq;y<aazOF1dZoat8vl4gDK0I4_D?*l2I$PDL$xDx
zp4*cll@)c&$R~Y^Tl=ECPfF!JCGE=weCVZK9NrN4Nr2dP8D*b>h8`ZRTn0De)>T;d
z)=-9R+c7I4)bGi(!RI$>51=}M75dV`262LfLFw>h%V^#jI(ufheY4XE(n(e`YbvR?
zjy#oj_#|VXO^Dk}cCT7}j-b1NbFZ=Nx_rZpba@7Huin+aU=6G~0Hs@wm@&lQE<BAn
zHlSe-SgEWhGT1E{Mt9I<<Xo&AXgmYWdTnj<LjztS#Pu+tXU(nXcM`eRXNZEAH|=fo
z9llX2E16tUiUsvD%)QA9hz(G2qT|6*CMsLpVM3oowYc#Qn(qY$DX4cXL^l(32VKtP
zp0FK!hDRZWZ|fb9$pf$46;Oi87}wZ(SNVYpnMHj2E7DC1bf~QfTd5G9Kr9{d-En*Y
zCSZo+_8n+W!%QkqRBG{K+q1mYgC%Rma`W8hbzZLsjh8)i>e~Zqgu1*x!!bAsR4}1L
zW&DB3x1V2r7*#UKbN$ZnBd3{tnoF9g$aI9CmnNJW>@ak13(&s-LV1Zh4sf_WB@#2I
z=bxsKr$dAml$9Pq{jO<)Ii`i<-8kShU&CUl3)~Zg$*b*@ei?H}guu3f(=60sxzU{_
znFY8*cGBnOyk6RBdiP)2*+xx+sR!I_Pv{WywmI}$m~=|k?VVLJS1D&%I>7<A)arBU
z6s!tVBh3+Du!senVo|-f$@6=Gk&_2A&|tU0LlQ|)<Ldi~Td-;Mvb#Z_9lp#pWvI?P
z2O&J!QwPhdzkx=@R#8qq-!4w&4$t=U<`=Z-1F<*TlLVVVTqcCW6*KrV)u0`_5g!aD
z6}T88ztYc{>m!|17fsiCtt#+N^MfCvwyX%G=p!vdVJ}_JB(Gxpx=V4fb!3Yk3J^@J
z(zmBybGh<ediwH7i}Rpyn&Eek)iv!a<j9w`PER+ppXKm<FBbFW>R8RJ9QTrX41K~B
z<LSzHTO)V>ismjCvDj)}^4XPOU@Tu!^bb#4WApT&<68)ep*nfC;$xADb<>7;9M)?C
zDbf}ifqUo_vs@WNOj9p~cTRL`(u)0YhIv2ImFifP=;x^mDFcgEhoe*t*);pjzHM2k
zFxgg`j4#O}#p%g|(Gu#6wkac);?_Il?ke*Xd)`jdnTM7;oM+J>R|}2rLoDxjzRW+H
zeFB<5Gf|2lagQ->d<d$|7Ju-Fb2*mV((%^<ccEhqT=Si*=n4x7EZ@mzgPEbPxlbN$
z=+M8H6=eK>Xc7Ov205za0QlElCp9NQ1+2aS98xV@?UUfe43(dyf=8h+?8r;VvL_w#
zi4h0|`1$7VItf}M_W!?X+YsnaU!x47Dt1g@3?e`~VZ?|fJ=8rXHLz3|H8b=Hq6@8&
z^C|ie4JDcg8!LtbQ3ObIM+uJg{qL(QI&mi808su7j2o%IG%v}re-C8t2@4e67fo0J
zegegTO<Q0%ywQHBj5EQ219n=70}*`0eUAfPKxc+(Jt0AW@BXEa2?OaUfF|+~a_9ec
zW#oj=ByDX5KCe_qJg8D?I-1B_^$TJIOu6<7S6G`#@7G+x!$Pb_F^MbH>*p)O)jQ*Q
z?Q_d+A<m3^+%@`|Um}r-F=;`*!;+CRoHsfj%SxIFG_?Uh9f=0>ya&tKjp59TdJ44S
zIMLBe9p}YGceq~4a5ubY)&!Y8Ffe%cPK)6#s-OqC@8qbbJA$GGFg5}Y1EGTA^DtQI
z?^c=@xn;0zu6Q6R=;0mimiG`e`pLYcwUq1vB3i>XRuyX258~tE9-d$Bxbzv8iA*_g
za=fSex-aiva@G=Idsz~o7@KeTknHSiCIriBg(sEP#C{r6dV_}^+A(q$dZ?95?Bf}p
zDps$)M)%HjYoRO)#P|9gYnFoZkVa+c_oP%mJl5%1E8bE!nUZ!)0}lmwIQM-x2YApJ
zbL}H#?C@1uM3Tx+onDi{jVoTP@<onQrc_?_DZlVO<i%FVA#=0d8r?nm2YJV&2wOk@
zbx${^Ar*-4!q1Ju9wB%`w?At^neOOiM!o{Z=7kPXL}5^2NVoLI^-R}0{-*=$-`c5o
zJxwV2*ZH77aG1&52MHnqEHkq6$&X6~Il!CO!D%c^1)o}q$`)0czw4kId{R~Uc|`29
z$2KG%M{E=;P<%%~plk0_-Z9PumI&AN{xH`Je`YF2N9;;2M#8pj-B)rkD0aLX&cuTD
z7LI)5_|M97R$MB`b?iW#$YI(3r*8LfwDjJ5w(Wcb!J!%{TwC+|VoQ=sC4rA@v|`C*
zL|93JPD3v<k_O(eX2M#8)k@ZZM-jL%rRA@eFX*-I+o0LtQX=-*rm7-8ks~Wy`r^Eg
zl;XNkE4I7kThUk<(r7nU`w?3Dev;E!?ZI3AXD&2?eBf;k9dK?Z`lgn`0f!qoNtCAc
zZr&xgv&Zy@pCRFW!51bGLr@ieRTGJ}Bv5*P(Zz{2ar>?{h`oiFOYi)9-a4q@<3PPw
z@jUU4y<4NSDWc20r~NLUXtYTE4YG}%8b$R8S9y`cDNJ99-K+Ox*ANnK=xxQ0JR0kj
zZY(`~mh3jEj9E2VW>=f;oeud*zlvgZIEZgSPLWV+si$&k%cH0FKclT%u)hb-wPo<m
z5>&|#$IK^3q+wX4mt6eZEq}zJ#MZaEe06b>>MM<_5Ba&n4q6&1RV9SQC{+Yor8};K
zFf3`(&GPiHW?nV&JXA2!Kog;Q+5!4;``A(QV)iY!mhNs&Jw~a|o<bL3<oW`S;eg}b
zvVkC4Rz!&Bk*>T(KnW|mSp{3plez%zsCgN8zovUadE}M4`DoaiJ06EAO%3iz?w9X~
z`mvu9j33GP6|!0uN|)LSqa6?1tEGipQE@&~lOL(Xe(Hn$>|Kl6Zh@6q^ZrlmS4)*x
z&-4tB$Pj8HCg5c4Sv7`plE)z#N%_+R7-@7#Pt@+HgN9{^D^g#kefHXuU6!B<Mx;mM
zVg%r{`w3wd4zYo3>9H<2e4>u*{P}D2I;J_MMl7W7Dds#m<bCU~@;dpOfi7;_)Mm|k
z*_5_#=yxqgq3b*8%GBEx>gqU78-=gBr<hSW<Q#hWffy46<NMcIACQ(b8<vDs^O_(T
zzCa~MwkHAmDXiZ|-cAoUyR5bZtkTNP1;b56<<l}>p&@1l@wuLz_<6#$qwOxcf-l&h
zj=gGIYR%sl{nCA-8}8_ftp;>g(rHR?E3ZO?Nf=g$f8KTw&i^>@X}7Ud$fxJ-ju_W%
zwPCpc--nGk1Nz?$bnP^EYPW4TImTHY!9?D>jf~oExoN{t$%k68dR*FJ=;@-Xm|w7-
z>R8e*ODbSTZ^B7p7_%&?Ip2$*dB^nSiR?-%YS2nw)~+6sFhp3`<q$-sVE`Dm2fF}!
zbKQm`+(9J*L6<>Z#*0z#5X|ZPWv@IKVYQI@IX)XZ7gKuiu|bcg9x4j}R}HZzFL!m`
zQhc3=Bt{I0VK`J$pdWV{WBio(qUe3w+8vox%u~9b7e$Jw^ylyCH7Ct!;A|Lua5@FQ
zu+#>fyH+9f&$s8gPMX5+B6PKO?Opvp_(N`&(`T0@FeLRHdHDDgc5p?fe_3Dnlrv=0
zeM?CAd8+xX1zF@hdzQ-c8cX!+i?0Ud#cQ#=2DI;*8+h4w4T;{Mi1~yga3^eJr^X(=
zmte3I;^${1dHZ$ez$SSJ2gZyG%QW4U>}x5<&zdO5^u3G5=NMl7>H|SqxjsEy#gVpf
zOlLukE1)T~&f)DIdkV*mvUsxcrA7Pgkd@x9V6uxoPHA+`q6bBOw}Rd~>gC+YrGGah
zSyt-7SW5`Ym_DP=?N2ifxlYeDYyK>V*5oA)L3*m5aVS-hS0cde*Hw&@d*L3^yf*Rr
z_o-V>yi``RLVvz<mllkEl%v?n4bA8Yuo))vrbXClD3VKarp2$VOLcLaab8|`=MsB?
zAwPGMu*o@2Qq7-09I`fL#e!oJ*>)<;FsyNXx7JQeNREd}^BYt$a71Y<b4}QDFQqXI
zbN3gR*9og=bu7AaxK^Vp1p7o_Pn<Yvq_lPoGa91BD^FCMNXswgGm0dlF*wSJ(yJ(;
zD<M|IL`3sw2AN;Pp~6oHz9o?d;dW_-_$WMAE-uIyLR%WyNSJz@M!=DBauFhi?Ok-c
z(jI-AY`A1JKCuBO<7xvbD4#$0hTC;Utu1aEds8Z9IS_XZ$uTt`qWamdQ3@)2+lZHR
ze-k4%3O6HNm=ABV{j<h*NZFC@(KYw8^pxlKAF=dR)qzCcAH1hlq<Q}27Y4mbA!eOe
z<8YV^_p^Qk;383K@dL=oURLS8%jf))&?!t+k?T*dL&K9GJ5+GgyIrN(Rn&m+hL$5o
zKz<Gvbj)Kxu1d$c!8u&z>@`h@-s3QRQLTQFWDJ>VIgt@1&xO}|w<zpaY{FFuqGmRc
zb~2h3D{1S}^_X%=u~?&u%K}Cj_lC0vZ9yt=z+-lW;m{+&XGV$W*c0ADZvs1f!ghMR
zMF|~=&VH-$=nXqdZ+9y7{r7E2Ni<q7smoLx@e8ffgwm}Hw5@aFA#y0}-f|G1nNYSx
zIUCsY16ES3pE$CjU`Dv7h=GZyIW^6LGm+=#fQ0(oy)(Dzi<IoR?wfVRI8akf_S%dN
zoV5SoZE1O!RCGEflGn-c6Byx19x@c?sHs$tn4?j3=d)GK3Z#oectV%G9iDdLsr|At
zZek$ryLFOP$*bJWSbv&?!<Sa!a`{|0YkclKS#pgmhL86_S_LFO=nT9{UIjzq<-FR~
zV}w4;$C?D<hwqLLk<VvkxQBz9&^C7t-u4T-$V~FH=iwrrY8gKp@y<w*f%ZvHwDi9w
znjmFaIN2ZSIZ5L<G;VstK0Kth;94#%978Zp7T-XjLDt~Y5^B|O`eF(@B_3BDl+WK1
z#u4<+=_6OLa=I66lFP5p2p2kCxe|D)^@=iW$&*x(FcOvb@d0V9z7CkCPH~~iPWU*?
zg7RCkMX~LDG=mq43#UJJ(HrlN$po>zn@e?fDV_{DUywkB+qAr~IETX8?PN`cB~1d2
ztERK#FU%Jsx8%#Dl`>JY#_zN#jEqgFa=9B?aosul4hC|lU!f>3j!rft+J<7k$Ld>x
zhv#}mn!erhNxIHR?5UF*Ts&_E**dSzMZTi!Y9Nc@`Ki^g5vJ`#8{q5`o^TrN$HMjM
zu2-Bk$JEAmUl*#13PBWi18-dyjW4$Zt7!gv2w#$PfHFnY%15(Xhil^%+J&R+#S)P?
z6EBb#PgOzMjV47o6r|l%)_U3wlH7W8G&8=gRNG`*M-+y2sinETo|beES&!64n04q`
z?|z9Y;N{2d!v#>|`Z<g`i+h_9dY0e)T};ZV*Ihi13!dreLUr=FbALAuy?R8mkTTv=
zgcR@(9hq9`P*Ph^kIO#qN&=074TD~b;=l%Qnz2E(FC`F<e!+lylZPZI690B$L`CN)
zct$H2v^n?}##oJ(i8#~87QR%3(?oy>7XIIbApUn?BV-+Dr<$8lPXlk_@6O#V-+nfe
zist<rNnK$e_W&5h{2QMEcq!>$wE)nBc%gaPho?6a3RgY<@L5U`P53dGr_2^HL(Ktq
z-6ZaNPDQ|-`2zj80Q(j&lI?&yuaqoXHyRK`kG)wWYqthLo6*97F#=<*0I1!TfVxwk
z#ejSOU)|r1P$|L*YhV%m4|#6r$GcS20>&YmeS4i%+jFk7f||_GV4yu7pc1(VLYX6e
zKI|~j`6Kucj3{!cQ4S{gBe$IZ_90?`aJ>Bw5Z7%1O9jz`7vRJQS)kXE08*XtfKLS+
z^K?(?MFz6|83G?VgbTxT=gv%}+&`MnfrkU-8J#K>I{&GM5T-prB%OE!b;kLR?l0Lh
zK6$?{K|*Kz%j5Z{yRe{aRkA-)e?i=BW~e2hiEB~E3C7pI@$jRnowT{aH~D`dVweEf
zNp%J<_)R2eC-pZd9!|i4rIl3R({%zPJ^+q&Qow<=k$|1_X1v2n{r(2a|7VXJVj}!j
zE@w5|uPf&9eklK^$U`*5UWgj+#)C777vXiljQJD|66JDge=`YSZeZ5Dw$!t~6MXv}
z%$^yFQpP~4=}u!NE(5phLQ$fy{qI-+P%z}jos=hCE6S*FN;CsEMR+c`sQKqjV2%J2
zOhnH3`Kd4)vXmI%$%-~phTK9^*#qvA4?H)MC79Ukz)b=EA6=gVpSPEN9Ts(7yH8#O
z+XEcJ|FjCYku0FjjgiZ5v`K4vr+EL>cmO<%KJny>*u|3d?eKPdFt<1hR4X=@|MBtf
zDO~Se1bL!F+EqCX%;JBJ<+d19l3%XMtv&Lc>ezn?Z2=FrLw_*nyyGy4>A>7POM{FX
zdjT`_M5R(qCL&g;jC@w01h>+Tu)hbXo6w63#^bgGtGjV}>bFVNUreCGrRWMcP{i>{
z<c~S+jn^%3dG3l=R&TtUN17z`4&fqrx=Y-w$i9&vv=7Mv7NMk}y^<6X!syG0;8&^S
zlEknk!K52rq>!9@J3NNy<$AX?yiwYVbOqYh5*g5Xh0<R@3s4Z~_=wps$D5b}?ZW(A
zmK$dZ#QIk`la{tEZ|~f-dv_>x-xPl`PjRNuv^7;Ksn?HsJv*+FsGCNQ)~5~A%m1Xj
zI@^@D7jjEw^_Q$ZdPB<eeOe`{0y#mYEwyX8yl*!5K<rWakP0U;@o5$N(C&d}52fS*
z!&WFHKP(wkn6q&gH1LWuYT9BcC-~x!OM(4u@}GkPd791(j`SI`_Np`M65Fx#m;vm~
zIknY0lm@Yo^$3RxW$auIkquBob!Mv4kW?`qI@+_&ZWW1Rt04jy4sX>)0QqyW`d@Kh
zW3@gOdie?VTI$y2fbw08Sn?+K6d%tJo{guWHyo3Ga?3qdr$37rdQE?Z-)m0oH7=A~
zce>e68i_--OMkS{Z%r-&H95@;iS%q;#-u^C>Ez^1)Q;Gh-vgHPY0M_+%;#*(7Z&A8
zd4!`K+thxo2=V5phev9il7NrbV_Xb1V?Aq9WwNCTN;YmLU&s=0IxW5ZZ19M$qgt7Y
zr3qsK#fpyMAOnfJ%h|Pat!A@&S;oxiYS_<w>x}92>@emFrMRKkX-rI@Rjhx$122Y5
z6@Ap`m4!rfI=9*ty~8WqP!<kHk8{<URxaPfmYHpGnKK>VAC6GnZ-iKoC-B2ZWnxAB
z$Jf&lW`+%-n6O$p4S0K%7pUp`CvNw5eST|l!*l|Sp*iTsy)z>al#JrVrs{VtkZFiM
zb^E4`q3U|TAqK93GrEC{)X1*Ii#110)7c+aqeLtzgK1QOJ%=EcoN>^=%v*_qw8l$x
zq|(ohwe@}J2|1bEulDb}fLM(ku7RO;FO6GT>+i$}Cb&LK4qe|vo2(+!zvXuvB6HyR
zoU8&&Ia`L13@B_~s`MXfdz*sVdph+NMuOmgczA9ViU{Tv_A8Fu^$3R@_YTcakK9F{
z8XEMJNt?cHWmn^_b(+P8B|!nnf=pFfPgE4C7GVJZ{o1_b`>^oSx93lTRx-v@qfX8R
z9%Q@*Ap1J2QYF^wD2=Cp7x^vVi=N40cD2^0SI_d@kZkxdv?2P06-oh_dSEFlbNe}~
zkAEI{x4uiMC@P9pF<`I2?zN8X;F2I*%Eyi3o!D%IT)?%&<OuJ_e5tYAg@L4^{M^6*
zb=1a?W`OP6B%zP*m6IKCav9<6L?H$)JsT9tvBf^w!`_-=cT+4dPS+zi03@ojOdVao
zxvX<LWQcUWfb0N2Pm+HeD`maO@!R@|UYmG2lQ-&%mJXN?%n4y$Wu0`KG*+t`@)clQ
zZMC+<fp(8Q(IY~Ek55|lavR0L20nkeqfbzu9c^jRH>2#+4;6m_cn`SLWXY}O113@;
zqe|`mGHF4jM}hf>#LGQfC5RBpmZeDqnnrL;hAXQ>Cs#P4+UYgTk5w%aZ9};M;d6~W
z5&B3@JO7a;{{E^6x!m{@c@*`KTyAen#0yAVsGZAL46-NYR^3%@fBJ*`OzjtSq9H%$
zMbitNuh+_*D0uY*p|}~c!9HgcQeF3O)JmZ=1PrOK1-yLubgTIY?)&?1;ArsX?gu(n
zV&%OcRC28E1S4|E9vuMwX;ghe7WOAUGRW6T!*>Ms4y#WS(bLV31zPw?Bc@)!(8kFi
zP4tkd{V`&hSIl6}MWuy(0mh<X*KSD8&Yo4Jqb4WMnQ)41)vNNfizlm#X;QbepEdie
zAFR4*zthjpTpxK6QTLo+{Dk-5(yAL5ftG3d6KU1$m<>1bSQ{*B{X}7(38E;5JmdFk
z&j%eZ>gi@3C*3Gwcw+0q8RQNQaf4Z7=qoE`+BD~vo5M*SF)V5w1RPg@^tndaUwG%J
zCN)eU1-zd~&dt-pbECoy;D=&CbYh41Q~OFe;8RQoX9kYSa3}37lTw*wuUPj*wzoOq
z{CqDLNst<3UH)ugKSBx~j`}Wai2CIB7wD^pLTLSp)R&BdY^wYZ1EiP?V5)P>T3QDu
z9b7J&N;R3-+lZ1oO72E+VjTycsvakC&wlARmdKF}I16nYWYKC%<s#|8g1j1_M%lPp
zXi2bBRhH|wfJ`SMbckXFZg!uo_**4|(py%k1^Ys6ar-%>T~Tqx#l6eYctsL1l0^-L
z+|UFhks~ozmb}&a=gXKcF58gRuV-A8k5ij}ct~2S35igI$C~V`1#3UX+C)ESFYi&e
z>P_qBy&kP;5M}$6$xFFpJC#RvEls;Me72*j?PB9zD41q6OhSJ7SkZJj7Nqrp@vDfB
z-kndD@O~N7$gHrnb*|s|eIZ(APfiOt^i&QX-oPAOQb`(-3$B*>h3cmR;FNx+)R0%%
zQKQuKoB88aTzD`kBat6MXWkFO%Cpa_4v$(NJfrq|JQbs+Z~L=Bk5AH5zJMS<y2b?`
zci?~!A6}SLmHEuB)@}0TLc5od$2%77IQcS~$+ize?}_=*U`<F)0lzO!JzHf=4hI<}
z5*cX6N$y=|U+YlMo`)<!%geum)z=1mbV9E4x{Y)9F^s=i5Hq@h)LBMCub(^AYCQSE
zl)6U#0K?S3#g1k?VARHUMRT&=S8C!~i2g(I3b|XZQstW(K3TyRvbITgCLc>P+fCB~
z<aU`TTfUtSL2d5-hKZ(;x1uB5h_*4&UbEc(C`;SAx*n!q(|*aWMMaTw)QsTB*FMyp
zO`R^5$N-CkI5liF?Gucb3t%7@u!o<p{g8zXx0`F^KlUVtbQ2kDHwfKwM?>z~U{ACX
z%}ls1O2WVLJh7yHnAi9T16s)j)tLexlF`V`!#WOwoz%ZI-W8xJfqZvI9a;6v;7<C#
zUr#q|4v+BSHTtMaX_l#&Ud0jQlTYAvt<Xod^-Pq3q6GV-K>#9?P^Iv8b|xZiNfaUB
zLyO3F2-?3OucX4v)NB=bftKF5YPs5LN`w3u2d+-%h=JUo&R&6L3wXiYXPq-Y00N*P
zUuN{)o3wbVTZs5vm&o&>sxJ@<3x-OV%!}p?eM&>AFQ+bnR=rZ_v6Sj2(4IE5K30O}
z2UA!nrZl_5+<LOc!S@S_8|F(&nA2%I8qUG``R%s=HqBO=$lKpGAG&o_InS%zexd^A
zbbKdAk`|{m5)>thQWFd<;7<se_qS7pxKWC)`am?`c$F>v!B|LoEyW4{)TUKQ=#8bq
z5Z6h*b^G)EvR?K$^~#T>*@-$a8)4g2&gGN0mR+UTL8i$QXru(_Ga<Vn3Jb^&*{yLU
z9C-D`TCB=OPW?k){d7FGDWPeOnzsx!F27VJA$d~7*csJ&+tk2C`PXW)&qEagDyF$)
zG0HMHj9D_Y%I5Ncp}}7{{qY%Gt&PYjk^g#76)Gjbpd{Ca<?e7R@RDth5*2@!>paIo
z+x+=#xPUCF&TH+;ZK29}j*hhYM!-`n*67$js7xSfa=FW;U4X@D_N?m~LHEJvL(g<Q
zKy3f;=>HtIVLsKg?>*X{UEX*}C!o!<Gv@V04KGE;o!kO=;?ip)EjQkpo}x)6yxyk>
zA}@xF->}-Q&ya7-t4UR>nI1wCUsZ<ePmLAD7pc~nKc{$oV>9P?rR<*eRCQJ)mm`QE
zW4eNw6+fZ3N&{3|(~8f;5Xo4pK&vPOu@;Q(NE+<EVY|*qdK+|Rdrr<llgC{uO^T>!
z_~dNrLtN&j=GsBk^?@lhjV$?1d^;|zsFCfih{`S4S-h=Jg!&*Wd|`@>{%KhOr|npg
z=IB>J^u}!mWiNL=XE&B$Q6;Twx^&66IDntfRNrrn<1`xF-W&CD)8F{(qaU*1MwYB8
z?NlLCM@i5R%~}xoCVQst1?3yS#b_#3t{KZH2+d03OnNhoqJZXkY+pZF<*A{E_u#z8
zeLtuIzdjfp>!7@D@P8vXJ9q(|zn3lNbH@z#f50K3=cl=%&I|_6;zQK1Q#$->J^Ug-
z8MMH9M@frHBJekmw%!u_>Bqoiw(Ht#SNvaiWEq_9>Uh$b|Mq`8u7ohp#oh*duK$6V
zbm7ZaV>A_V_D5gkUUGaW@^fS91pLzf0iRR=1T^}98Cvk)wxD^R29iMQiDuOP3;FOu
z_d@DH2%;v_$<P1QXbi0pNs<Ic=KYr;=J2oLYjP2n-emuU_ijzR7U()%^K-o2ez<3;
zdE6;(_S_0^r2mJB&%o*#1=z8kRR7udFGulz!bif1*WD$xzTc0YL;i>1M{@k9vA^8H
zGa7KL=gm46Ptd#oa)J|jbNV&U{n|8SaC@%SNJN4e`c&^5q)sMGGK|`X3?bW@F^&F%
z7y*{}%^m#qa{tXF{IAw#wBRK){~%fZ_XLKw<4(v02Wb^E^g3$t{~;T6RlrV#*D*)X
zntxC=e>u|sP2Bv?YO-kL*eP_8f2bcfG=qS&0a_wrn-`Z8KQ}ilgJVKW4puXHXz~Nm
zCgxy2FplieNU{O7G;w_;jEO7GK|kZt5k#WqO#qdlG>`Lj9E0UfDX_Q$A1xn8Wmq}_
z)4oO=#HS99zzGEu4GhsBT7+6VBAJn|h!OWN9IA*B*Agb^$M@|93!1j%PV+i$aX$TX
zW*$u>L@0O_K;D0jNrDaBd%BN6XfuX56U=ChKom!4em<yL^yAqbX*6Jd|NEc|c-W(D
zZP|uqx}L(ZtVBH$M9%`%_{)qy;w_)969-gipc=5gE8S02fD@7+%7LN%VnYDXVsFa#
zfBfU}5cVlSGN|A$rpr4T#0T78MgWz?TEMhgDVKHsXrx<?NoVKmFM9Du-1@Ozo8@&|
z>^MM?D9oZ!{Sh$G;@Nf4!f5|gjnRL9@A0)BCNuOI&V(K_^va`Wj^-&ISV|F|i3i?J
z0)lP~{wVn?Hsx<?fiwie0dSyW$h982c9zAHTIB7&V(qVi=`KI(3MUOBjj$7Y{YaE%
zY`pVi;X!lu{lCKUX+4}+=s0~m(KczvU#Q`)3~!Zbqo;pP<4)+JdoD!%V$vZb5ly5B
zk#!sDsNBNS(RoqqOv4VPl^Y3A9?h%FkA1P2tCvp)Iv*=-z1{T!_#A7HBF;tgpBi?J
zKTf!|@(n3G^bp%H%I<yxl2XxIf3W&>5!rW9{E5y8cvO$<B?pk0UQ24P-Dt0CxV29}
zeU&jH2!3>f2OeBTUms>TCidn~H(!s?VQ+qVASL-ac0=I35PWRRJiACZbv}re!lOgp
zu<Kb~hf*H4L$hy+pgF$%yvs~f+t)Jl7OfjzPVEmMl#T%*0b<1X(G>t`19+xC@%Gu|
z*4{W5g1y~Nnha<Q%=^5k>^5#g1>TR1d0tlRx@0=>2e(oE7ERl!M2n3q_D$h3r?sDV
zn27~MTeKwn60PlyIz~tA$;l809-gI^0aSy{u=s81OY1|QC45S^EbKhG?GLu_T!DuK
z2q|J*RP<LpV~!-Kz1qECZRgUXa68MqJhDC6qPzm(2t;q;niu~pI_%8>mf;zL+X<yB
z4WeQ}2@{uAW<<Hv=)EZ9WKq*h^@#8CUz!OJPf1((7^x#}!6^L1w3YFY#zmBa4^lw|
z#-7Pba`wRpD*g`Tr&>)^u@qjbP4l-AUazp=q#+Gf8}?d0&eRA)^Fz1E*quNXcxKye
zXQ~`MEDRXxS?gITa6c`#>JRS;gzRY<cDq$;j{*xT+-qd!3Jv3TZuJ~2q_QT!JK&Nc
zXm+3H$L~yF)U2rlYrE;-sCQ2<W>fATnb25UFddf>C+wM7yqCt@a85-~-O@Te3&4fB
z-qj@Q+OEw^t8;q2On@lo?^jl|QKyEH$+|I(`k^B_@Q(|3`|?csK;^5e{UzFC2ypd4
zny+@211y_M^%LS4tx|6Ll~=>%DJezOeZ{jl4e9>;;GLf5Ob!nUB>RnG#^>w^jDJYu
z>IPJ84K`kZ8V_Y2d<_0>VGDt0zIIp&OKsNe>&dqjb9(o{k0opM)AvY|bI6VYSt0sy
zn$|2riwo2#GbXzn;1+bp|M1p|FX~lUOxHEOF_%RZA^TZ8NNZ~P>-pi<+lNbUSNC1@
z*0k#r&Ru}mfx%3IlIb6v(=I=`GR8F@gS5;=eYL+nRL0=rvhpk(8|~^}Sdf+~r!VPl
z710@(=i;!p$K}Ap6^Qvx#otTdA3prvTQ~Kgn{TF1t=$Po(|UOMol))MFMOs~U8A5z
z1SuuIdg}Tqi<fC#g<!dM+^$Gy%;&TR53|1J-I&1g^K<v<zwe81qED0`$cGP4mt>_y
zR2YE_r_ydQO7?WoN!PkqgO>Jh=`_bJHjLBCiYve?!dZAK*YNq4>?*5<M5Q$Hc3mVK
z5_it~*lFy>gtSwcynOi+@0<AIX;UJm#6qwT$As;7MxwHW6hS$~IbF~A_8g~kYL-1U
z(HoJZV*DOE<mt+ijw5#s=x^`>lz-ss6j3gz`C^Wair>6KNJs*o*}L!tj>`+3>H8jM
z8DPvc(l|y(x`4rI+`4i|Z}6#^2-Tv04h1Jj>UPli_Wl@&)VU&=i^M$w><LhCt#B@i
z?}v$)Ua!DXfaJFrcBcI@-LLs#=IKkH8&Mey3(wL*h|C7~qs$r!p}Kt4DH}4XfJ!fv
z@wU`GPOQx*bvWM-JUmbKxENST`**l+N3MvZgJ!3^Z1f)N(hPqhMX(&NPoOATe-i_7
z=Y8WU!Hyj5`!voq`nUT>M<go(2qyv*h62hpGGc_El~6n!O^Kv4U7i12rA@{#TM}|`
zR3z><-PW0LWHG1lh+MCdb*C(!EDQxW_Z5N9ooS#T2m8<bwM8+t#}x<-{D^p{zC3J<
zNl#wW_Et{1{As|#!z|&AdBg?#@{y!Yi)D7F$UyGCOPPa_MJ-iolX_5i$0i1QM6lM5
z-t7CSS5TGmz%MNX1S}Z_J$-#8xbAr?n}sj%J5{-<2{iJK8e~daJs<(J(T<vGmQss?
zi>iJO=|SE>Je5RsDRJ_xfVZDOl>hL}<1;_G%YHI+GU#aOzFHv3$S|=nCg)1eAzOJk
zA&adW+ZU0Fl9z*x0`vXt(lKmWs9DvggYyB)Mv%|1kG*753dy|Pc6%P^aY;!is`7x&
zY0UQ}b>=umsb(LhS4)B`avU{nGah?z(JM9#7aAztJT1QodTu9IXJrs*VAa0=saym)
zdu<_`Vqp|@BQh@P5vDYxGS}i6^4TNwYuIod&2PHk=SPr%mODPogpdtN@Lf==OrV37
z;BCC7PPpc>oLb$YztILSIw!4TQ9<Y9us5C|*HrhU@B~1^MlQor^$g=Ck_j=2C1si)
zQi}?*W-)W0+dR3M!3<<2+tT1!UDFdvsN(&uVJUMSHRE>ugUVHLFKaz{{gVBP#{O02
zrcxl6A%<Ms^8|w%JRnu-1E})?a8*+?K&IY$l;46H^79j5eJi}Y6c2TG>7m6=Nk2AE
zglep%4|1EF*jttrl_p(3l>7PI2Y6hXjnZ`PT8)U~^2Vgmh1u`;e<<u`8_BIqM}f+#
zHvKQ2XE*eAf;o5$<32KF&_ud7l17qkjc98{p_D+KfGd7gyzINm)2r<9O1sxeLD=qX
zl>^I4Kc$FFRlu-N298uTq_Wy_R?Zv{3WF?MCPa8JA}yq+2f?PEwstJZr_Y`kiDchD
ztEXn){EpfDI9lY3z=X<-()t=Inynax*~J+m+tvNm4rS_(H`RtOf@AwgamsL_Y9kUP
zqo|WLB2rNuduwr6N2@tazcc^XeQiXb`b~A?V)#w<`1m+CZpouhjT<~+v(9Q5g<fkc
zQSLoY@L@f|vi^;V*BKX?{X3xQ$`%YHU8H4re!qc2|HmQG{2Z#RCFvqmfZAA<vA_6$
z>DtlpmO9%|_sgdJSKbsj2?CC+7Q9*Zqc!+w2sY9JcINH;&xdpddzH>LLFD7uJ~^tE
zhPu9Kqg6xi(`NZ0J7#&Hj4xT-asgx^+VPqd#xv%z_Wm<O&?i_xy42GI`;QC7=8gBf
zr)O7%%td>_=W|Z7WYF1TfEGN0uFwEMMy8tmy?|ZC5oNVMB~e2fRb~mkuEUv^pKeG&
z<&*3?hiv2pMH;Plc^K_gP8zZ^{JxE6eiBTm`r1kyt$Yl%t<g`jYYUOAT3G-nV!4M-
zMw&(AUH8K_2A{;Xq?N2&6M=q32}q`oh-1@QKH`;TpCdd7ijsx#b+SWI>SYO*j3|+c
z2avhPUrC>McgV^J`&>;89-(!6Z(Q4Eimh&Jc5o~_peo2ONsC{Po9mg53koq(#PGtI
zP&CJKP(DYcyNb35McZ^Nur`}(S_}un{O$>8r*ES}&rJR6wF?rffiQuEBQ~QT?SM7&
zYt9Z$;Hhe+452?YJuT5F&75YJ+$?O2)0^94CkZ7@zRm}G7bbR%TQGqND70=`W4)}x
zrymIA*PVoYs6t-Vd)rl{tFfXve5tJHbJT1jn(yvaPL9W(e4wJ59@N|7*$kPGFC(4x
z66!^tugJemadk$gw0`-?nQb}}f9);P!Vg{$cN9%;0-D&eOm3JDBw@QiHK8b%5;8^G
zTflVP^*ic(eL)w99{{{!V)T`sj)UXRI*1IUnY^_%L-q88;QcG>f3;8j)&6_AC+ywT
zj;aLsGWb_!Wn%zeJtXc{UA#S`(H;-fbuCPTZGmlw<z&q&C%frFf9@3_j4IA<CX&jX
z(#hK$6+yl0t~OP5)$RT(tx~KG#X8Ko?;U^-<EiJ#l5)ftd@wEvQc_De?A*VmN$4M<
z|A${SlH-K<>#taN0y!6Wzjhx6^j8M!c2XKEbmc1b{#e`L0pW}Ay{>;knB(278EQ?&
z{yQ@!e-`!+K<C=qk63;NbtepC@?mTegH5?HkaF6DAgv>>zF4_?;ZlDvL7EIyOCv=#
zy4GL(4{T__-u#xD%QN>p##)!L5Vpep=ai>8e&~4C*_=4T&bHpEHvYfCMcTA;g7NcK
zakp11GEvOP*MH3Xh0EH<(>hs(l74#os{AqZ7qB{0^O@$OQ}(3%18B_P4g9dqf$5jH
zhAe-q|0{F&Z(IZ5jo~_E-pAfR3}o6LO>IBIoUC{2tQr3a3)>QUYMjMau5Ksx-2I2S
zqEPp8qi(N|?JtK&p?~!6k1H@SZnXv}4pN#WeP4vD_$zq9p@GtZf?v@S08+Nh2syH7
zfO8^*$;L<MSXY-9f<*58D~-FIaALKHLF<TY<4zj@y#9`UD8P2$!0Llc0cQOB*b=tG
zCGp+s`x688SG&A_Vhzqf6RZ16_H(en7iV!<xi1U^{=`DC%7J35+5cP%1k(Q_P~(4A
z103I>1_PgNj{PLiECCAV-ys-6(7o%!@h1DQWz+gK`Fsph`G1_e16>Ia`SFSU#N{>7
zBl7KkSP5iYwlJv5h%0unGCYL=+4KL7$9M&@t=pYoC<wxz>oF$(6Y0>72Rmbe0dK~E
zeMe1-Kcs<M>pQN$V*yHEpdJ6vYz_E%mq;~y4Sj3rPuEKm(F{)fLQEFV>yoT+OqKp&
zB`4G!1wYKX9d7iG6nZJV3G0ATrS4w8z*e3hxP`L@t2D>&Tzh~me0TA=Jn<3jvI!0h
z!(ki<s*t(`bEVLGZiWN1{{cpn<Hn0Thq<gjP*?g_T!<FDL6GezNJmVLyfR!07aR-P
z$sx;5@dNRFj2s3s&%q1Af2S5G!A}4t5Sa(!+eEk3^8b$RL;A1trEX~HNlmmWtCj=u
zY(P*12}0&K(PTNEhdrFdxhf5V{Nw1^tL%obuS?dgq(kQawi6E~n<jg2B7YQV^^Z$Y
z@8pbSufN$WGNO*M@BE?}ep>_#s6lUTdNGk~f1|S78xTNw{Hw$5coU~F!zMYE;>Cvm
z_43>1SD=i&<H18R+I@h4V1YgXqEBW3vW}Sii;myFwcUX}^XpDWZD1c1<nR9bGFu=&
zaxdh;fc-WFBPWp0|F6D@LDk65`(Ig|UZLNt?ma&4%q2$POcefN(C86JEOJVy`!b@v
z%GH?!A+B$c;>AUQ`Bw_fub>t@JVb<j9DW>z^R_Z(qT}ZS0!O%=9<@?KvbQ)TWqPA-
zhPtsX1?%f~PCZq*zanFzbep^^`nZ79BuvBGAQaVhLY_F48mQg>s6u@LBN5|;xJauf
zS_Md5tb*8Orpb9Gk-RmzEWwE`Jk2FuPrMHG$k4+{`5_SN@h-H^cZ`QqU6utPL6qR8
zS8g>i5_E2<H`bu?qSOdMIxbl%DvDM-G$5;+wn|;#2_9@)nU62A%ETl>Oz_SZMyi^3
z!xDxy4>5m{g-fTdfMN2#K1z7iglIF|MtWqp61@8$0|I$2X*N2lzOig+aZIo%d^x?0
z>)vUTGu@n1%-O)q%i()@x3dqD;+7tj1F>>8#s!FYH+48#<J#}+|0Q%JB@ar;)QJW6
zStP9^T*&{91nEa}v-|gF(3DADdcyI>UWTjX2eOL0C5y6vP;TO>2(h2pG)4gqWawUk
z*B(`r;mSVS^I8zjQd;p)Gr{kE#y}n!8lq(YTj86Z#i)+2(___(>t&)sgh|)FdsHLf
zG2cpE4h<y!w#t}R_s!xZauM~GBhEmt;4f1!O9NSkq9P5*3il8ocGQDz-(ZhNVLOzO
zs>?!fe%=SKwL2AGg?I>VR3|wrX47IL-hi5LT+_W+LLW5d%?EmaX*qevB&woWLOI+h
z?j(;h5yTF%br?w_oIp)`FrfR8ARZiQS7w>J4HV>Zo+*F=3lwwW0&RvD7UpnS>MSiE
zuwj_Da_y?mV{CvreS^b@)9)BQ^r%Prir<C?O!ehWX!*z8)h|rzQ=%^`sgCSqGu=Dv
zrX|g-GV)00HnRC>)S~HJDi)!28nnca^|3?9;tdbL0I6U-tqf6=z1(e+r_8j2JM}hV
zAoENy9VRhGq#=<BXY_tFP^AW1{J1$Pw+X1;Cr_5iCg(Wv#JLyd5bsWH#Wh%A#A6?M
zLKePcHhlc&=}HS9EJZ9%)Dt#EDlo;kGmgbBKO9^dq+>$er@AZ;r|X%n+2#1>QoAhb
zC-Vyml3M-dlVMs2qiTdEt<vE=&)FINS04wJ48P*7oVwD^igT;W24cZ}SS!;&bqK7k
zm&9P=P&juBZx&7ax+GWp)SW7Jk}!&5^G{KhWu^~rNEig6b$q0cQ$tmG3a{^Vd2TP=
zXR1B_7B{CzTga4d^C)cl$2VoEB{rK3Q7f`7Ys#aCTPyMv&+kWyUD~fMl$`#I`dqn5
zoJo5m9D95Dmt`NS5SlznYb)#3RhD+eDRDwso^fr}z~gg{1&e9AJDy^9m1ck1|5H0@
z+8227EPlbe&cx8H3(EWvV?o~!>U{2vyACoKQq29<X#E#Gvg0u+{3dt5WkCFT729Sc
zYQu}Cca{a3VV0@koQ#s{i`MffXGdh4t-mKPUMD!uX+}~}yy(^PS9tF{hoRs|7w8(+
z^xDxChp%oVa0(}b&9bbsT%1yi9=lh*<O%1V$2@0+S*RKN0&-&1UQq6<=?g0injTd>
z+P0)CA!<@5+;nb=bjM27I)7GMO6oRFFRqsCn$(i@Qg@K~TCW*tXUVfux08T(8{`&S
z=4n7A$zZ5$C8rVudvnn#tC||&R59{2iSCglVWCm8Al%JA3bL4C%>3d6RK47wDMfHp
zfMIW<vTh|dwbqv`p})(%IA|=()|>p~UfzZ3Y1^B#!9b}kl?CPQp{*cm`A7I%$>Klj
z^kH_}K1UTj4p&{RT;Odw-g#dYr+mJG{`7Gjy}kFA{n>(7ed!4AS5HecC4Y?Y-tp9J
zifuFqE(OfC1eu5Awe+uq=0=5%)V&-`4Vr%(6++F{VJcfvAGhim6PYLyAx-YCdx0Jg
zniVo<xW#Zy`OUnXRzTPfJn{DPwDB{)#FuDl;hK>VpFzX7^}VfIDrk9+*>1knByy_a
zGE%MI(MU}eSyfy6Rwf#{(;prpQnL6kMfhA<u!PAVe!C<e2WD(oqpI<HvZ77c1O=fh
z6~+t9cQ@SiQbR{#-BW(Z5T)0#vzCJ+H_jQ8uIo?eXd00j!<uFUtc4Uz(F`IA>Du#0
zveNK^&A6R3hJL*3zPWIZiFNya!e+LGpIT=1vhZ1hhmaI$HxxUU(?NI%OZqcP2J_t{
z`pn5XttTZ#$&N7l`ym%I!K|8!Dk0|@Qzd1C;Bip<HIuDv{J`Ruh=Onra=0K=f3i&K
zm>m6h$0MN#c-nLLy$nWVDo>lW)RJnj%>NJS-ZHL=HvAghba$spw{*uw8tIglZjkO$
zI;9)7bgP6kNSAbjfRu!EH)qf%^m)&F&i|a>`EouoGy9%9t{LW9_gd>JfA;JZmW7pc
zQ73&CV#_OzX0a~@tw}^-1j|qCyK&GL>tuQ+-&2E{X@ujayqkD8f-7Ub#SgmV+~)B(
ztE!NjSCnC;dp)8pnhP8zp?%(OatpV0iAYV0K{?_`N$}egDNV8a2GwMQ5O0&;IgDwG
zs3VW8FWnxHD5CO$!%hs(nv^f7;t{=pHBBynPOvxhh{VHwO(HHU3UK_AnRwRYz643s
z2=ZYMZ?rGv!5pDXYsFA!c2n_#sb%TGt3wm8+12AN%fuu@22JGGi)-W0E;DG!-z=I)
z6}pINkcu-v7gavZEL&Sci$I7)mg1@QDG20CweFxWx}SuPo`X_a%gWc5_;#^H502+{
zK!(J56^37fL+nM^LZv(r!H@1Tgd6)%mAK-W!Dx8cR#=d&OO{V>q~$K1zyxd1{cz=i
z-4k#hiJfAP61yQghR>URI2nqDi4%C>b7y3O5<tlc2>EOlmEYp!ZF8bgB4;B|yhDG?
zab<c<7O>%!D-VMb&`Y#31sOc{x9_c9G7Y7qv(h^0y%67Nb!@>CsVWE2mAoy~miT<3
zMFg=IY@{hixPKRPm5*2Zxod66`NtgQ0g>b~634XA<)!_dKIify;dm>f4O9yH)CSUk
z{`#+5Av9Y2>+hefJZJ7|7^rJXpSDD#Cl{&i{oa9t%fjl;^7R{xGz$Ay+*@dG6KOlj
z7u03CkW~CtkCQTD6dSZnG?j1iSm!K&?L_$n1I`voR$PKz`#7EZXvzH>nZ?HK*27_~
z(WvsDA*qXlk9BRoU1A9XR!)>lWWW|ZtfBxuy(bQ=hD2cC=tRjSQpgSZS)^=BxiJ!a
zmKRb0`xkDM2~;Jacrx%H-ij3lBzUTLQ2dI@Z;#*o>;uM7Vx6FiI%X8r<?9^l2=B;E
z_}#BfE$*4}Din`aq9;wvK=Yz4qaV0DHW%cPMkExJ>Ny=nL29%HTWox!-<952Zz~rj
zMjk2aK^PP!sWM80-G*lo!ITb_2t}6*-5)@Sl)Hs10dA)3V{zQpn~t8;ah;1}?x;%7
zc&Qb*pwxy<kd&N)QovB#-%XCG3f$8_vG@d>z8J$hh~V7Rt?O>@CqXT)w#@zlI0$)W
zb{07!__S9Z&6x6SIXj37Ua6AeWZq1mf+J|g7+;w-e>h|hK$gUNWTDwsoRg=@_+v?4
za@U#)*rB-{?~%I+gx_S(wv>5WcZjNkUd@>0dO20sUHV1aowym>Yz$5e8{(N^1z(ib
zhYk6T%vLiO__V>;n%C6+WBOHm?4|`a^&%ItFu#beC@S5dX&Mjq-l*)6f--Rxi@u`S
zG7979W!<aHp5{K)Xm<Gv&mNN`mBcQIk1RhVxpL!{ssz<vl@!es@1DOD&!=y4t%ToX
zrWAPx#bB=6pC~P`^+~{6&T@77w2-Yi&XaiD^kd#8p6Sshy{SwS-~%xLOcUQOgZh&a
zZr~ZSpW#<X)!C8Up3TS^IN-O9sgkk`l}O>Ka6+{VT=RM3++=y^$$x^?X##W1GqI7*
zRDck>O|F{C3DRIaLV&ap0y$n!1C$b9bmvt2K>o(S_xz2mjes$z(^CY~MgzGm+}p&C
z9A)>^K75D`^9}#EnuD&Wk7k);iXrC1r@Ai(^|>k9vIHm!c(*Xa9u;}N5q9F+uUAAl
zQc4I3Gr_1BHI|h@*<>X27!Sng0ciKl*Am?WiH|%Fz3fY_Ty$Je!H_r79V^{0aus+u
z_Gx7`ItJ+3jGlkGd<4VU>>V#-BArD{fjmMxR{KgWHJonez0q3VV->LE0zBP5bTb0Q
z;)%W^a^mu+SEd)lk!zv1>4~aMVU>}}LcZl*m<%Kj$)1TnQ`DPhP%%umbeAgVHZhVR
zYDmqogbK*MDIbBZ`aW6VCM3<oOGIM$p~FIE-KtT-+$iw2(o#0py7Arp8R*`dshK(#
zq^aWl+`=b;lp$e^EpN3Z{bY0CA%l;%ULq)&w_<^uVSfrC0IE{QG#2E{2*dDhqHU)%
z<u#NAAl=^NB!1YV9?CViOiW^!Ao(ZT;RVkM7I@7UEBXO$X1`LZA0X3afE=OvZV@mg
z`<XN<1G)jEtAHK{atpNIez`AvJ!2f)6ZWkx4gfa(%8^JRxqv)?_wOw07pdUaxu4lS
z9dxxhkgojPr2p0WFEzklxtoK3P8#O_S#r7xte!~Al8WsW73^Pa01AzAJVxPzu%&9X
z-#QDh*A2D;VvI}Z|Gp4duSbDBZ0BhJIH)s*81qZ^@xY%51ldMs_kx}RsTd%)SPYc%
zpHjO3q5Ulkv4crKrzrni+dugmzX8yDEI-}q*wOH<&m=zmJ+0>lhwKaX!>`=0bR{1I
z^sE01832=y`>5z^;2*k;l@4)Wtp5lr{@?N@_|cWr{v*fumoM>aa{r&(Isk%<tHPM~
zY+Irk-D?#8;x|AJ%!^SC^)Gtz%eUZfEal;^Q$)ynS7|gdmFs_nJ5o8p=*2z{&I)Uk
zU<b@T<!0z2OM4taCp>89gkMU9gP8sZL4EH-`}+*&lXtG$1VGsSlXqJ-D<Fm2Zvph1
zzi^8s5Wirs9%Sh6I9nqF+Z!SRc+X$40Z@(57#HTHA4D;K9rA@fdDCi{_n@))s}aEN
zYzSrsX`A8iOIn~#_SiC{_y1`IFstLoNZ4$QS&^ms7f;{?J0&?9J$iuie)aGYNl7<i
z?VrHZhsM=8u=}TL)L;klW`I%j>sk-Qk;H6^;@53}JL4}%2KYq3{I;|ViYxv58sNS7
z!1np|K>nKHU+BbxfrJlevnw?<q66DMoRgISRHP@AGpGJRXz@zn6SPx_)tYaZnu}xd
z$^%V;7sREv#IZ3ni0a4h6uZM=6k%fy%J<MBE>*GaNSdH3jQr}k8E!Q@f9>ug%4uWb
z({Cl9{`TRgiPIo;dXt70Bw_QR<|~d$V^2?QxT0R^V^;sH^cLyFW}z3w<NP}eUuEc<
zudEtSwcbVWol4V}tnPhVorfFEiM^#S{h=liu1f8H_hU);Y-Z`Up$*XE2*ar5zumf`
zxyVRO6+YkYvV8s7X&e{AM3qRSMGil*MI<>Qr_`&0c~SDA)R8t{{F#0;s`9&t%AHCO
zPl=&0>D;xut7x^w+N*vme)YH&H7H1{xBswl!^S|AMhU?YA%vk%Zu%S*h7M{&>)EIi
zY~Fi)lP6nzv94)%1L+4FW;4m0cv{7<EzTQ}4TF0M{kGdZpyr}SVHdR-L$UeF3PEK)
zHw};HI30OSxiT4K94etV8x3;4cpp%otqb(2jbNx}ApT|iSGe?G?`P5|Bs_zc{9zA{
zu@vz9`&1mSjRuc#KQq-{x%X@l`edaV^lt8KvckSw5qji)U;Xv4C47I?+V~gF4uZ=N
z@3nLTYEyZVJ#@(#OX6;`m*x(%>pZ8_U<o3&_`*`C1QNXrKDVUc=}X=y#z2sZ#tV>-
z9^CgI+a_1x!$vB~5{JeZN`0KLMY&e|qaR-JZ`FT6po<cep}b*2*wn&~$Py5cpI_(L
zu(0Tdl0~u(Wsm^%gFKb@8KoCw;4SbtPgOxt5qWJa5w4kO2ZGlp^;K&d%U}6|Wv*7^
zgiJruaxZ7%so~tcfd1ljx15PU>`pj-t%w8cI=3dmCazO}$ZVPzWG!GI+6>kx1RlsE
zL<*mHZNAvI4+CMc3ZZ&Jhs0kvT=~UZ=9`Cxjfk78<dK}=vV<kheqQW8=e;(2g_?6u
ze%b+2{4iT!0~&=x?pCU1-#e^VV>vQtg&ZH>Yd;|scf`!z;C6~toVFM*Zon;c^tXDl
z(cHu7tKm(oimF8X8Hq|e6k6>7A8NaZqHs*OIl}0bcN({D;c@qSKCrpI&~u$_a|u-0
z*P~TbM<!3`FCWqV;FKViitc`9nd|@7C^fjWB)+eyXfT;rTqW(mxz74Zclo2Lge4`h
zMRq<&=zP1{q~)CmMc9l3aeTQx9!Jo6>n`8q+Z8qBk&F%hDNoXPXaFBrqqK%!B44@)
zd4wr$=bp7+e@uFLv_?ctB6GbYPN_d8|2)sfW1sV~^0)+EiK{JhV})g=PD9dP|B^|j
z?hXa_I4|oT*9EPuj(7EYCG`a?(A774_@V?7siYLcsomlu7x7M<Be<aaK@a@c{uxzR
z$cDZU52b9upaoF`HoClG=|V0C!DULzHZII7$xICZJ8k~dUxd%0=qu#tE1R=Z|G0;B
zFnV_<a??e6eS%TVjl$;`1>-1#I5~Z^l)h#|O)6ZK9ss8W1Gfw=8HE^{r+5({Bw5R3
zZ<%^^0k;KQ*x2g(W5k#ja0}JqbIJ2J1Hr@eAa$J3TPuXuM36#>_jEPQ559swg&aDc
z_R{JL0QfM+hrZ#qz<fx7vE02w+!Ojsga{u``H0GA)m7Se=%SmP`($gP-?~iDH>;r=
zBfNDjJ+F)kxT`mm-{njNDo!$U#AY3H4t#CO#y(vOD~2v}dky96Dd{{vcTT^?$ZlUD
zLNGN|?7>Hrf4L>DL|dq>k|7J#KfiO{#MyM+o%=f5z)-;5LfeuM@D(J$Bt``!9e{L$
ztnD2oKh<X)UsDJ;s$C<p)2YESY|MZ=<DCIK61WBy5OvIK)KTz(^o$N^q69ET`_ako
zre0klVmy0V^%~WM775o$egTHAakOgMRyAkv%xO;ZYRQQToDZA1{6><hbD~?<Xp?ZY
zpVxmv@+fgBfYWYHfFC@2`i}dRC?_**?v6w4y4v$_RU&|o%GQ==F*p#At8)r<wa5r9
zmmVi0TiavIf_5ruhgI%KXohWRkR(369S5z%72r=Ts`r}L6xVlL$&9~?Q387}t~Y^e
zJr&Kx0Q*JDFIva+$#^zajmJiICSUTbb5by5W9Xx#0K#S@+e^UfBWE)h*Ga~Broar1
zS_|Onu*SBIDf)45N94Ph#&}I|F#uVbgyIHs`FuQDuO_SoyP?GfYw+8#=6Y{&@i(<S
z@uD>FBc@ZK!!%<~8Q}=&M*amVJ-3Mk+SDMtIFR}^MxD&3J$hhzBEY+5Ux`SX`76x|
z3{3CuY$R>l;-#Hv7G9e_p|ga_&TOHB&jKQYsMOF7VVDl8C{DEO!*CAo(@2?E?sGZa
zo8rA;vw)$_@sv*3_dA!y@2cm^u8Sa9gTWBVLhdz_1UL&Fg_=A-IglWcnKL>ROW%5+
zml(WktXA%~Ep^i9j~O#h7>Juf9c7dCTEc4<2wJL2Ir!KUie&i15hWGIW4fE8ecy{H
zT6i|5co6VEPUSqIzF-e(qVm1MXRtiUaflm=!f=ME^;bXFy`55-FERww`G<X&vOOP$
z(T7KGs}V*~M>5d0`+J?ngHgYe6*sbqra9R3eLLgmkW)0J5OVqQ<6ZS`Gwv3}YResn
zvq_XHDvqE12S{zplawDifh+Gh6oXa@!pJh7@b>Y9dcFpYf6Hy>V{-CJxFWNPVlR-3
zwoR7a5Ff{&&XP^!9a9vtAB`E4@*&ja7Hqn?Dk#%8@$t#diG<i!I?^Z*JA7LGyJBCG
z5^Aeu!y!ea*(;7!ImL2*0*)?Ig%`0V4ZXTellB?ukqWlGebQy8J}%#L1VpeP#9*{|
zRte;}(EFuS@h$w?C(!bE^*6)Wc;zl)NPrL9%Z2#t`|tZPfy4m~Xon1zeC3;@xlTrK
z-%&Hi#3+<rD<X|l_3tQNot7;I7QQ*K(HfIuu{zR}HDh0Mj@||_gVP)cm``ZCQ?n#!
zrAn3gu2!5yI0(LagXWa|hYq0RJ<En3l<ZIPqzfKLUEIgX<ZZlNc<XKL8=IU6nub+c
zJWo9dm;-T|Ruo5oT+&R6s{Nk!2LUXiH3a_U&BF2iYU()zVw@A3CP{r%iossXtnm(H
zRTp-mFCEAHV3b(166l2!oK?ADIXkiB*QToi3g_b)w?3)eYNfU9eZ@(*h0)LnU}^N+
zLikwstG37S($^FS?UTyxT%zqOM*jV2EKvd=HJ<4S8Gj9&<-W)lfldkvG_AZ&(Xh%K
z4j|e2gYhP2zi6QsCX&vPN0xF9otvDqy2Sw=U-Z)-^G`r&dxxY=&nVtc@Q!nM34ifH
zMgvvz;A~l+YVh(+#;r`K`^V8atrC6uMw_}#J{~>TDMvfj5?Y*MAH@i~UPtUeN0&pf
zaSxQTs>__8J`XNV@=#D}3M+Zbi}jX#xIVG`eIYEPSql8RgIOTYX!&Gwfz4dr3(gVZ
zLrINpYFRM?|1=a+kb3q<^MvBH;Iso<y6pnx%td6LLVlGD0ONL_9T}q1!e;5;V7Cmb
ze(HAdcC`CFi3Y_kDGg(6Kh2oI_df2H962s#m?EZ(bZcZJ$RWQAa;VcGj&L!KpH1%<
zE%Mq}ZsFZzTzlCSJzX2M#;wuVjK<*xq-<3>>vqNDb^5du1AEM3AeZGMp4=>d<y`Wl
ztGEiZjxt@-mnQ3_BWF2{1H9&rr`jW6caHg8$W`y?Zm!-m8Grtj{~JuTYfgWuH)*12
zIY=ExC6mm|T2>7&NbI3?^2ZmNY0(IZbj#j-tYY_0q%VD|n8AF=3T8Y-8!IKgRQYrf
zS+j6kB|{Q=B&#=BG;e#Xkn0d7TVV8(j>GZUXni1sQJkzUxFeUcyYZ<XalGj}5$3e5
zR9qWzCm5V99jp{Uq#gqGK=L=a+JosQxI=?VOVs5kiB>{_gIB_d%ERf<xH)?w6MM2F
z9C}W&p^#UrZjxLWVDJX8vQhxV`YwbJxM%)mHJw>Z>`m-hJ}-PMKwNJn_#JHW&l?qE
zD&Efp$C@@=wZeZ=IEDKuqDPY$km_8?->J%Z17pz=PoyIXQ{D2*{2-X}F9L@3qrb9<
z-?QldUDCY(OIM=&cLM**i|*H6^3T-sf87!OZ|3p$PVE18m!Bd-7L>(8CjJ^0;Ep$<
zTvRM(W$?G@9yt`}<Ntx-d=vEMf<g0zit*%D?7uno8Wcbk*yh<;a=T~r%Ipd88+Yz{
zP4I8V-GDJ-ApGHo@nq>jNCx1bTs8zT|CZ{KLzbTO-ca4XLN%b*1U{wyVIc5>4GKHG
zL%rQM0xm*Wfr0_QY6Pt1XPkdhX?^ua+MiH6PtO_;a_&N2ERdT0jbQ<x7tHWqFzzJJ
z{C)j<-eJ_Jb%Fo<WmhJn$P<~Xk%K*gk7z(n`j5W_0P>I~b1r)8aWB}1y~Q;Jq;h{-
zUjV1dmWMnAV`S_q+o(uG+<tRz5CqZXo3f|i+CS6!jn$C_N>DWNv<L;2|CZ)^K#$a+
z;HDqo{YGa1zz{SFWjpc+&Bx1+;{ejk-@*}Knq>xrCgH-zQ@;Ljy&C7jBd>%(=-ahd
z7u~tMv5NpH;GanZoHXcBcx6zsr1I(##+400&N|1$muxQWKP@?;Vbm}=d;@#g05A#e
z?~n!DHxffwKR(HuK6G%!!$k&{uJ^-HLv!l>lpZ;||A0D0lB@jyWdLl$!lOi%m~2=7
zwDJH9wGI}ZD*O6|iUeoD=3Rh20V)Ft6KhFa@%S&)>=8uVZPhXxP&m;Nmnj<nu+WFk
zMY077Bo7uFYj9xY(?1XWdcx4~3Dx^1EdMQsP~8Td%aF=%`(UXVs|g@hr3406W&A?e
z(0`mk0M0N92cPWunI5MKM`B<*;Qeud@;TJi0K!G5;-Pju9&bz^K$!Z^aNI6b#V5{*
zS7oWB{xcFX$oqkir_|Q`4$Msep747Y0pM!wFk1!`9>J#<^v4JR6Op5yS)T)*N`w;_
zjBm%_h5zMzWu4PHEtSk8?BUuD?j`{F_`PKSJG+#{STA?HP<Vhz{+QZ-19C-&??3GV
zkwed{YM&_P&tGf+5^6FZHsm6Jt_}eZp?||N#~-R>JTBbkF-f8>5*TBDvBgU<4hkMJ
zOZ>@1P4D4F@lUkkvVhgK_hwr%y(vhF&Zz2n!F`Ln^Pghxm&nW^^IT0@A6XnrUBWle
z9XU(hfikU#_-?;XIy%O!tZ{~l*LyJO2L!&6-m@Zs)D!3y#}DLNfr@0GzgC3q>u^-X
zHe>f^ceR_BzpM4B?PQ#-SRRaaiM93N;Brn#S0;noRO{*8<#`_+pN>mAt$d+IK&N2r
zs5$@|Z2i`sk-;*_1auACTg6tu`-Rw)A-dA*!QCAkB(d6kEcb;^@*U~aKl-ETYlABs
z*-P?)t`YLxeV=^gn6P3Zgpt7W-{)u-y87E|MhDq-CRC|&n_*~Y#b*OTq>#pOrLP?y
z>ut~N4rz?1LN~nPt3yK0${6_yz7q2f^c5vRJD1q4U8QR-^r?r(ugd0K5K*jf9nR)w
z{E~KfT-_eMjF(Tei@|X1qAv4lqZv4u`1aOk`=Pjji}w8lQEAS27dvAvQ;cGfx<9_&
zD@$6k*y4<!Wjad;&P6^qgZs)&-fc#s8?9*^EOa^>xV9LierPt(aKeTM5cwzyw<Rn5
zDs>C{b3^cf-4MLa$VX2cnKle5kiZD2%x9p)PAfP-W<tqWGhSi?`0P5j<!2a)A6Y*i
z)|zq5@uq?@4u>EI)E*x=UGLeh8EAc3CWJIytI}Dv*P<)gw8!-JS(pS>7l5>VPDa}M
z?kwf{#8PH3$A0j|RC(QRh4_6DX8HrUAb+4rkobz<`RpeK=;fCI4wEnZ8ImncWb38$
zb+i=%(uC&SNGcA6upz}L>;WvZ)bO)mXFD>UT=7Yb*a7{bRM7JaR9Lr)0`eE|@UaP&
zFAlNR*63zTS0@U@CqWr?a*|~Ng(pu%SwLyg1_=1(6Kmpw;`FAj=z#Jh3&Hc(8_O}P
z3l}B<-W)m%KL><oVGRSYSxn25;2mzm0l0Mho3ft=rE&WL`<rhB1AT6B`;BNK^JM`F
z9?6!SJ!bPV2YuQHc{jV23}*f+{VIyYAI2Eoo*)mCkcpRwM9DlUa0GRu2EAf-vr?_J
z+7&<|iaZNFlhV?l&KxA-;zir)V$l&;cd1ESSvwNHa%&a8ynJ#;=!yN7_>uH`P=ay<
zm%}f|NJI)A!Qu0-9?{+&D!b8CmiOrvozGn#J{9*fX$)NTCqX4Nb;r72;TWrNt1=2{
zYqTG%C<J&(TL$}8jxv=Q-)URY{9O=%Fh+4WAMqt7eL+(2JG%I2xKwONhJh?fR(EtG
z8tS}4_>vv81@<1f!bnPId!xkk#+eYYNbN_uL~daBBZs=R5MbM<ni*XNQ&0w_%wZ9Z
zB?lIoD%dhqX#Y@9(aNpkDBM}2-5A}DG|MbnL8)lNRC93jF8;dUKDUuAe1~KlPU>%m
z>GL<L(e(BK)hNt(nVckj?U?R2d%y<=v4Z|+U|Skrr}}rPgzw)h&oDUfP0Iul)Sgw1
z$tA(xr+Rx`^c1X)I~7|YK-ds_TN`WllbX}51WWj2{8dcYhiuJn=9={qB<bZ^_SiUA
z&kvL#aya^odVy(;@shl@m&aaa=XT7nrWuPJ^Vd|)3zSwn!m+~fxuoq=Pw2T+LW}QV
zP3*Uy)g=kk6z5sl3Uu`o1D-9m@b!H4b@kE?RyQ@>TQp6eFE0<aN;?FkUR+<^7Fzb!
z-t>blsa0pfW8H^Xn3b;vf;mnG9zzu+eVcfRb6r+FnH+;U1r$@rOG-w%Zjq?1e0;(_
zAi2YnMbvVh!OEsZPxBlGD5D_PBPF04ubZX~wV5RpB0sI2KG(vW{R+0T32oSXAVQP@
zD`4z{CN&)#a9R=ran>b#Ohh#f0ctz1EOz&j?Y8Y}hVED0Jl+}vC4PM;3qlylK(n0N
zFDgYaGZfnLGx^kGExq3_enJ|E<VfAP-iOK;r;1_R+-`BFD>8~<j*6p|?oS%jOYlkW
zgID7h%6=73u+sOt<ncS^;UZL#Z>(xdkAynkR86F_vV7S`U!Q))x})qi)&eurz(`TJ
z_w>Gdn~ByVM~z5Arm;XlZSWIzfyfIlJG@Eg2l=`H&S$!X^Z_aHH4iUAA`>H!o>|m|
z4TVrovHJ&lcnb!XuLQ;0Zo`N0S3ot1)k9{Wh(bhUKqIc^v)VE1SS5BJ?_y*tt0&Nt
zY1<B~-U8OXsvqIvy;KxzI?c}aY>h@o?1axV6)AF-DRihh(+>~Vloq)fu35XtW<LsW
zb>s4co>MukZ#m{J%%fLTXpAECYkclu+jO0?TWU#vqh{GYAjW9ZOWTt`R{Jy?dX-fs
zn3($fi6lhfAqpPhfERwv?<DMI>z%<?m>e5;`k1B)gj$FO6Zv{3FLqG%gEULJ^mUwg
zQs3h^{Wa{W`)983pS?7*aH{gfS?N30RmZq1!oO8o0ngRRFu6g=>ojdp5j^x^kYn`y
z25a`9@5NKq`ps`>G9RN{eC#$JFEk6JVmAnpOR1PSR<h4VD#RYXDj@cHf#aKQQ!sv-
zHt%Y0-dtkTk=5bg<(r<O`4rS(K5rLgF)J*aS)@Ft8(_E!QtzSYx7*rwvsugPl6RQY
zc8g;t1cI6>Qyu+Hd_?*iTv}Q{_dV&B?Gw2v(ky){LVjImNK-V1(7y(&G*!(4SG2=F
z2(x*B{AkpppEGQG*CC&)0!D2j*k=SrN6Oy{Z(@QQj1dvwj^D5RY8M}7*Dsaa51%*h
zF*T|&MshCYk3Ljdd=r|$SsZ&aoH+L{4MOUVF}xBDDY!E&=gEVOqaSH7bupM06AH|A
zckdv%pVEe#TqIFgIi6Lr%Jt;UzML{N^Wl^W?nbALw{o<9K5xlyR_ID?z0eZ>wy!tI
zLWVz`l!lgo$BQ1|vHP5l;%TWv7a4vpkQOC5ycYnF1*2%NsszjZu{pw<ya@G$_pT-U
zl8_P^a|$IgSO>B6Rk@)P!5f~dg|d~mZv(KSR5T^9R_#k(DxYv;vF*)fS+O7Y<pxH~
zuOO;@S~m8_KC|2SNQ<UcJ5F=s+MmAw?5^`B&ojluQ1N-e6K;pe!|a`L{Kz%5(r`Nz
z_Y^9Qr`RB^h;y2mcf)2Up2cQXFy}a3++15RWA5rBhpgCUoI;hv>>>8y=3(6%u!o(H
zRz}3y4?aC|#Qqj{+ZZ=!=8u0GQ0dbkBMeprrn;NX8y|0;nme9LQqn)XKv>6_zI$@=
zg#zOXb}IoGx{bmyloEMz_25}8oGgV20(tlD;RVucubDiRxx~-3{=}t46{WJu?EO?4
z&qv6WtW5I9ISZVL;4aEP?g%>$Ef>|zl?%vl8JV0KPlv`Q;sPRMa*o3)Wbo?pUcRUm
zW?11PRCeHhkmH3US&$|U6_bLMqn4JX;-ML+i0REsMWP_ha+R81Kqd5CRcAZ}7gOm4
zby4gD8}~91I?}K!E$2i?o*eUbhc~qf`!+5CVGMsC!c(?gvvPeIb!cm12NgT)IN6C0
zTk_vQY;Vd;-gus&vdZPmHcd5*i+e8Q6v03`K=LIEpH$0eB6k{e(q2`@uK;>#6MAo3
z|0AU2Y997qfX?7gK*uuacR&ZrGzm8oq2YLGlOr|D(3t~jf)zupcFal{q$HGFkVHH7
zoL*!G`0`*vLa#lh-gG4PVc1b4>0$$~^$)1U^AWjFc*}K%#q|sw+jD{z<gW$XdKlk*
zA}UBUvrBvCkKg_QepHQeGN-7Qc$hh?M14eCre@wXd@?Y0;n4>qS^;_myI~mlA8kDx
zrF_hQSZ7n2JTf6L;;3Fun__o>OgOFT@JBGf2?^<wrK+JAkXi$c1apXSfaMuuIT;7@
zK#)pkie!c&O;5T}?E4f4Su?VaC>*H?2}>WMps+=*#q8N>a}`3iHu*{DS8%o@Sf-Ba
zLxeV?p-x^wD~+?|?WN|>c=yd(z(b}a=FTg$dAS&1EhqtJwk+w^m8_oD>yLPh8h1-i
z@6Kb!L&Ch-(?{ST-tzU_TB?@z1K&RZ|EVi8YVq}{US?0{e`mXmwj&k1^nTsU36Q-v
zd&gDhv@h0owB7AgV1KH9ZnWJc<}6SzQ(WTc*>xH0X9|`+KRKe$RcPG$hG{)gTL0-V
zKg0Put#$kO9!NuO=%s5vr5|RoxC^;yvay=3$@)eTVFocC_cBTB${PHGFaUYT;JaWn
z5;Yv)HkA)ed+ttuj9zCWvpp`eqU+3>z;#-C404WPqA`EVxf?4GRV+6t)~=hoLQyK*
zKD{<k^a`^WiT*)YZPJuC{L%R75rnGCt)GZ0*}Ek7#WOwFBCn<d0eW%zu-4ia!(C?a
z9}u!a3t1EKQ&bkkuT&<z3Mxj6E_+VC^Kmg~?Ve=@rm<D>%@6X<N;^3Ema~z_8)4*_
zKfPq-T=9BOeP$B7oOrm?u`Ph4wBb&l?Y5N#BSerL?UwT1b4=%)<xC}3Y);$Wg_VTs
zE+<sCP?EX??OW&M+t{vIyWYNM;qd}sOXDoKEE}AbM2=GT&kWasTXujO!KAz92lBkC
z7<&Zn2WQJ(0{^iBf$PuQfKGwx47w3e2(eI0_;bH_sYb0@81>UYlN)^b7~A?oD~s@l
z)OW$P6c1ef-;-^i!q`!eG-IpanooA#y6<_41BFnD^q&bZ8A6f!Q$fqhs)B2w7$Jt%
z+@1fl;ou)`9)|g{#6A>|E%5v&Aq5N-Ve>z$X}YX%($N0r92q1d-5I(0=%v68G$2X%
zJ!@Cq0K7L>l|Q;3C#})4tUKfWdG;Ji7W3-<AXHUwevf>u5O9C|Jwf&cS>zZ9UC$Ua
zO{n=?0nC#(aFXimE^`l92W2ZDk)itY&@XDrZy3b?0B`_M1JCo;_kB%T=(G=lkGYot
z05>o5=gJfvgKqP;ngLGh|2pFCX-5k10)VEFO!vkV<NN|dNFks0gQW0)R!Bu+q4WX6
z5P*~L9fS+DUBC*O<M~@Wgr>Ah<cc^_@-N{8uqb87`h4;ZFh2e9ALG+r1b^WPji{^N
zaVC(G&HL%!$Ta|xaHhfsYoYfq!=wnH{savx+5wd_egltYSp{33RTTV-LOoy@sV9%R
zYYC@+!@L^Ml|JHcwe=XNoeEKvMErKShd*&P&FfWJh2Q!<4a}fG9d_}j->!4<C-2D5
zIQ;{fJ4uSoMpOa<>u=r*&_^x=$e)MiK~dS=)s9gG2v`u7-);i%lOEAj9u3&47t{WG
zI2HqRYof!o{d*T}0TH>NtNYY{=0_F-n4cif$ThKAdARr>sba|A-n3dI<mm@o|A=E&
zL8ebO5ggXNgo2=y3OnIa#Qo1c1418w;RV~gk+1QS2*#M4psvQ~M)2$L6a;srM@`>^
z^x)pYb>{)y)?lG81rTf&;=jyh!>1irWb?HDbK+NcCO+?HC`oLL_!&Y!*<Kdwah2g#
z>w7srak!tPFYWrgvK;ODMvDiK>=$bN-+L_joUz5f!d(Q6P5P-#@h1BhBz|ISe=(i@
zxcq;1=N8b(UV5wNmi*3T84c{8z0wDrEIbW!>v#`TqKeWi1^k8nxWNO6>?@Pmc4CMs
zmKg*p*HzI_F8`{>`P0hx8sxJ9<%7N{v}B*_D2f90$KZfBM%wp{a6G?YSQp@J;V7de
zQRC<~`V-%*fSIVqMSApMr-%j0xo56Mpagdk6C5T4h=>62t0r^D9<i%$H|)gw#$qsy
z7vHD1IjvcP2h?AIdn9sXACz<CU*|cweR5iBoCvZwo69>pU(iHbex@ANw60R}wIa5J
zSRuMKge&7KteM($-*T{0hVFFx66GT5eRpmhZp&+aC?(hYvbi=Bhqj$>ni}42X;2NW
zliM|=n&tAgIs>2v?iTJ<$KGJF()~tz7wSM1iOH%YtDX{d>(%*Uc}l&>OkF>WQUN~X
zFGU68+CNBgh`}BpPy4TCtzM{NN^UMtkR4gAMo+F*ec4=mPxA`x{I8jk{uR`8RN7<)
zegd1x>VnXEH5$6<oi3+V=ql~QHMkI0O`AJ5_1#c9d{GerF>4%3Om$bs`r99T^}t=S
z#?d|CZ7XQH0E2;^F`uj*`9sD%mr}E(Nhtp##-P9m>vB6jWNg`Zs$wLFfCh?xxW-Mi
z`uTfpmwTU54Gw?V7pk6dtcl&OCnQjG^tCtPgT^MD&KONQ)k*i{z5NpM$=RhP3JSG;
z%(UhKM3YX~v?F^w(hikybUpcJ+N(Oek!xk>Ky5VoQa+q5W=+EOT?q5>wI<uwaAz9-
zaFALsgz99iZR55H>1mmTjxU$>*wS#(m-7Tl!H`(xpw9!c0xRxm$)sF5-T@{2ME3XX
ztv2ek=Jh1%wrX_(^Z?{ohN8BA$Z2C`@H-Tz*ZqLJrT#1{YW`1I(Q?Mm6HxvY4fl7G
zOm{kix$P^M(J$?G8Dtx&V3`r<VAi%_fn9)>H{$tuW!H3Cn(;B`?^on6DjB)6fAz)M
zEwn>xmjfK}=D0**$dS9`YtbuLBkVf6Mx-Sne{{lno6FNX4)$q<=h;f2{-<we!m7(P
zm;COd$x1zNT1=QM5#Tm&#Y+M2JFaE7w;(l;ng!Rq)6#9}%so~`C1yfg^+`B$-3Q9r
z4{iF^iI@o_qQxcg%S?%|8|s-E>sD&b%FEkS{M&F9I^^O;)w-8>E;#8gFfc}b0d!o#
zGp3ogm%T4L07td`X7&#Qrf5n$voO^b&<zot`}NM+UC{lLfsoRv@QbQVx8x|;1;uxo
zUjwJ%W2g2Y?sCvd-HP4vm9ZF%%CA-Y$LvPqt~NNv;2;Wv&m8m4!?s}B5yf5Z)|$He
zJ!aZy!E1M>xg`n}j+}E(VmfP-m6vzQ6GaZzn!*&z<&9T-Aag0#KXNx=qM&4rTcb2>
zs{ozVEP%~HbqVLfZTbW~ci*9gqR*JBq`736EA2~pM%hsV2dn+u+%0aRbT2}HvG3Hj
z;ZvfeoV8mS#fUoe2wADtSC2;>w6kKXN^>>+#OKuIHP8IDYa2<^gAOyPqaLg6D`pKb
zV?5b^N}|r?n3x-Tv+IZ^%9zo7@m{Wu$gi@Wi2BjOU-rc|xVZ<&U?N;Cl&iCgZjqap
z-mEhRf=f;td`IgtHMLIk!9bDHzE9qstS{VN@p`A{EqJJ}k@R|FLmo_F;lPQnli?ea
zx)Dh9SvA1J`MvlM1-~KauE0x?%i8OUO?D{KrPgqs3i_-_HsV?77qw~e(YA(SA_Wlw
zsnV@)at?j*eI0G~US8<paU^zUomYjOL@ox^${OM<ep*<cV`@@t&8C?5c6k_5U*?Cl
zj3*`ES6QTTYp73$l!K~jSA}zJDuHTl*8YpK<-Qv?6Fuut2{0pQIT7((!$>nJxSp~9
z!QVPcdu#XZ?f1R?)_8{m6t$~6^NsqZGQYeN58g)|sGAuTg*FG7NhQ&TvYwUG2?)RK
z=KEl*=MTr5o=+rhj%0jgVo-LmZbWS$ci{)2UZp~b;*yp?G;kD;<*YwW&dVw+X((+d
z^=v5hvT7*JvNv)lEG%tN7h1^n*xt}hDFiilV36Ijm=%}_l6Xfndv3cGKyMbSsl!?{
z$5e@1-kh!T1sgIYt#wdZ-|X5nL{m$?X`{jaSkYI&ZS+2G7<ZaiC@v??E$}E>k(oj0
zi<b;zXN>6^X*{uk24nfl4f^7wve-1AwV(-<u(x%1RV`M;^?C@&YG%F~{ZIPt%5F<D
zV(u@*3JiR0B_fuO<=xxfw|!erKW<F^;Ng2uDxPq?kDefTahKork)w66a@Kg~nx3@l
zu``CEX?;c=f2$jj!7KmDvS@>#GR!pX?kcE^+puClzMYmownhN?q?1=1J;RTCNnA)t
z5$?ata0AnT3M$EfjNo{t3>zI46F(_RL)E8^5)Nv}D$nyZQT7efVf#^95IgVkO2)KN
znSn~aQiN=(z4WLBlOB!^6+thh@CG|J+aha~XzqgdQIOwNq=m7QNq=CZI{_p+p6A%@
z3Ggv*b2-fZH1$yKU{(n<JJ^s6bz-g}Lv5OER&z{XHeL};&SxZ!wJ)!?S`8~pJ_zdX
zdHaDrriV~(I!RqPx264L{c5^=6hTRBkQAj@Q|XQCnucAn+9#5HeUDO@1Psb@PaYZ<
zVjMYX?3)!<vF%+0?AVP=7u&a6s{>5|cSnOB{2*#CemA0fUWtv@X*ZdAY$n6HPuJLo
z=u)5hXUQh-(=@S%Qr+=;yYFt!r5$}&)+MLvj#D`PreaHYEE8+(O|wbN@;PuiXMiJ9
zqUR3FI-~PwkPE*E`9*sX);?C(<}oXLLCBLr-GPS17m6wxhNpyF%Mr3)!?g^!qP|wV
zdCourW_*}A>m7uv4fCTL-8KyhdT{o4X2-Ut=p1XwuV-%DubW9&5n(rR+wvmKp**(g
z1isU}U_wId5cTY4=VitRE9MWJKaOLMl5q_o_n=xhqJ$g3`goE)iZO(5e7wTxj`RI<
z$2&@EQiv2G92?$n<~=2*_XQS~1-r^|w5LVe`!E_oG5RKyaqzhCNOjq1w<mT78waBo
zYv>xUJ4Irvp@K-|H+<O)8&b9D-rXrTY16G|C2^cUNLEq$2*c}jMW4t%qAS4%i%w4`
zPSv3+Iq=`TL#X0{Dk&IK$ATDP8&oYUPJA0-UzZ})Ud8E&37tKa>gsf>cfdA48=m$R
zZkJl_dO5L7Nzr_Sft%Tr9x|<Pyg>-98!F+j7-!LToO-6+sl}rthuXtbl?A$S(Wu>2
zjPEP5<+|;4Qf$pYGg)_7F;g|DiOP5ulDl5RMWY<NEF@?Atoc3cJA`b#F7^us(4^U`
zp>duGTsy3yv_lojk=zWyD2<uT!R1cUV6w!jEi8!CMa83#>9v_ArY~ER*br|89440#
zm=K5T{9L12>Z~ioN}Sv;x%6PgjxxnyLuqtaJso3MA=IkdaBFi0{I^qRe{%t*`PnAt
zQV<jtv{{V}lFv9(BegXR8NB7U6kD54Zgs-eIH?y+uWq*@Aw+Ve)WF*RxGOu{1wNiO
zFK2js)tmEdt=PaEzNSV!Lss;tAm$%1UFRo}yP!$ctrGN+v?v7~e0G>u?5C_NHl7M1
z_&P`HhanQ5>4)%iXu~Azy;0tBnJ~%OKWiXYem2rUF(rt4vvE$HLG5V5d`U@|LEc?+
z5M`A~7)_KrQ=k)_5;;fUEofk8Ex+AvA#B-a@ve!`(6&|Om_x%J#pk|VLKf)XJ<E+%
z4IyO9;z~}dxH@^r(g^LaO47gI8P%(wPS7tqrrnw-8KkC|U+jp}qHpZ#fQ}?$@kMFJ
zIs`=!E8nPcb8c(rmeNs5;N^;?gCsd*2JH%QDb#Yem(r3^X6GWX>TKPNA_SwOTMp=1
z3Oe!2*$XS;s5yd0Sp!SvDhUH(8EXZUwXR+yh?Tg5HfpX$qg33^FbJxD!|9p1`KN8N
z?*T0*NIQKe3}b_^N*#hqY`MpZKSG}|uVVXe7jvqUiCZ<v>j0n@{~+FHQaqg?a^W1g
zPuat#LWwaeV{(fH)9DAim>6QuKP30gQLe0&=>;0AA}Vh(y)0bUkz+=sNL|_KK5Uw`
zAhH^@Y>N!W&+K3CLbP9gKXDiRg(BSZTsrQS=rMR4PHem7`E@b?mK~Nl{RGSQjskLV
zL`F3zM7ya$iHx!04C;x~?K}%~b?9J@SA8Dx-m9xSW)r@XFqE><FMKd%UX+uqM<jRl
z5rGd`HlC+XeZYZ?tWUIa(uie{(Siwfvd0Qki*E>J{JCWmL5UrLbtZZh&y%kR(b4t2
zg*x=fA;*m$OgD*uTCKhg%!0#psE|7RyBG1mn){Ws&ji*oz$M?*=jBqTdp|zwTzoRe
zN#g-sy}z!eT5hfnTPFoU@$dQ&?2~CV+0dk5Rd8$39<sI87rMY}V2SB4@pIaPK({*`
z7$P^hH@QV3`~-a7^%xX2k2@Z#gsHqzzi3m47U~ec`r^f$X_c0)>!ikcLFA=)eXV!l
zi%J2vy*99UGWwJ!jMr*`Qs!)f!oQ%ACLQ>Y;Bwj$cv`>4F)#Ksu?H;?Wjw{Ih!fN1
zsPVw4OJ#N_XEISuxXLg@+9kIDd8990S*(!%Jak1IWNX0jRm;TJxVcWNd6;aA3;EOL
z1S|y2k%6Ju*QyFVmRJ%E4-(IlTcE(Tht(AZnrj(@DqRU4CcBW&>)xgWrvq(;DlA57
ztcrv5f)5aOT0C0n427X4_zH94B#R$1)p#Dj`UjT6pMOJ5ZH&(4Az3~$69+|sVU;#A
z!exo`e`dIg54OH4rT_PrAwsZ1LE9Os$}+hV51C|NIu;ZsG!{71iGjf~q2VEc_TZrK
zmwXIe<fI@$9i{R!YqXaXg5E?Bxw)G56EFQnCJu5LDlW=X-hRl<*9HOm7oUA);Wt|&
z*5&vN$p2{!D3ok}4@$Y@y~m5mHjf9lR)-r%Y}X({hWPLjIc0w)*iA%$N6NI?OMTVC
z(d&D?i>!!0`>1RV<w*_m{hi<KiL&jJvH0iA=I;L)E5Ks}*!dlroW5hSLuvm#sqC*k
z_4&C_>sk&x&3{G)@F&VCI5C8KqWum~F=Uc;#zy`-zFg3{?U`6>nuz;7QQNTJXC8M8
z*$(L%TzK%jYVK<)pwx}n&w1OX0YCB<>G+FJ4LdSN;s2>e(FpuB0~jC?B&YB=X=*nN
z<_LB~8yO06$mao6<pCmq2P#~-(4rdnI3bwTAIR&c@IQMv{16y`)ryP^St{(v)o%xX
zuHl>%t^L1?W$`GpiRfW1LP5j<h!7tjM*a<*e_H#W|1cLF=Hy$<{nPE_7>1`8jL;U$
zSm0}*c>qDMmn4I?2t$O_h=X=>f!M|b?u-Sd#DuDj0XqCobdCrEf%I>D{qtkk5yb!P
z4*@^lIBmTDtA>K0MpU0qF>FsD{d?a85OM0Srvw=elBiG;-F{X720Qty-e2AQc}H7%
zH>H)h#)#i%iFqAeR@c!j_V+R5e-#aTzQ^+W^nFSYpmXS3kobLYNsb8AEcK1OzgFqj
zysGR$I;q_Vejl|?KUqTE+8)L4EA{;USDl++^tZFKu_0dS#jNEck3!o3%R<xfd2j)*
z2nZgOv(xH)dOZw3iF}z!$c&+?EBy+$<Z9HRv`9!kkyt-Gs3uNF^I_Wci#4+F6POl6
z_2<_Gkbb^lAb6kzm|Z-yAP?@e7k%gfa<GXnO^Duf*O&F=F24X;*DBhuqpq#!eD$Tx
z{@qNiV*XJz#AFtJi?-RBmfl_xSk|>i5K#4@`!OEg&qnI$G3<^s8InPebg%-5u4CPT
z!{<@f@Ny{EP7k7z7p^ry1lV?MDv{87p8p5;w-t4IMHJbj48Jp(r}OlQKi{)Q=t>~u
zj;|MZVeKASQ6sT@!IC^J8Zs${Ts_qAO7O#nuX-=|jLJw0*eC~#S0`R^8r_QlW<JK>
zMM`jR0k;_zxr7v}ht{dVKGiGn>t^r0kDj~sKRlP)dAI4METUwW@N|M~gfzZ7%i-OU
z1z4D_k?%QFN@&;XJ1GVZEnq`U?oo*trB+YAuz9&`Pe|W@**?3V##)_H*dzM0biXtz
zVGi!Ht&Cp3l)ZlOiB8MztUsi|q)=`#vK@Z&*6EE}&tg%mO<Nz2ezOvD;VsFM;gn)_
z{kBPyz=zo#GzE<TtaTE|LFY@U^`#jo1~2xHWLe~zuM@pH`$7E&eIuc4H2$1XFSYbB
zeJg7_8;yfCdj+Jj<0^o)46;YLVYi5PS%59lF?b49?NXQp<9r{NXj31lpSbt5p75Lq
z0`O2iaxXi)wG_JvdDK%ZLodf8wW^DV{L;h(RuWy&(91_G5JEx;yGf6vgxgdqvS+H7
zh?aIha$wr_dgZXFn_4?hfqR+k+B10~@M*qBy5zccSiFe8@Towd$u1}%)9t2qXUS_H
z9p0P__#`baROeidWmr-;tnGUvk2rZyW3fgldbnk@|IKUVNf6J6fcMY|Y*s0v*xiw^
z;Msc1YrT`#&zi!u?1s9kG#1s1(I{rK<ax7j(W85s6kTWRNXWNGi$$q6C`^$U`SDCr
zGJ-j#%1j-&O5P{(N<fbY{224)E>e372K8%kk;jZ%%7u2tvimt1t8#ivt(c5v48IJ?
zRnkI*R~38*B|?fWAwug#oqoVnDhLi?B@_{)wuqcvZ-8|M$`B&QCV?2p?_HnJNBC*!
zX${Vyie9EM6w;U1B&zX0m#Y+7$HXmkDgJJvm@R|f9oIhEE6V7gjR&S4r7E)fKyx;}
zap~1md*sElKJPZAco3VThc7e)PkXn@^WzPatJ$7m?#(md+Q#JBQ33<aofjqPm5*N`
zTf6VKauAXlM^9^+Jw4>{XehO+-)nwjkK(`B+{B`of;KcXh0wF~nMRQD3I2B7!GtQk
zb{{7VI4MK=>FY08XWA=5t$pd2Ou@KsSdA!}^70W4*SDNndM%hncmv<L!z*ENwA0;4
zTeo&Bk=A08hhw@`F~O?Sj1fFjC&F7@)<oXok^ExHX+&+s2Ahay`BbVgPS)JEmtqoK
zw5mbPsj^V9hCW7%r9M42LStnD%ij_EXzElTrwaRRi2c673Tp=eYLLjXP%(iGsJW_Q
zt-EVQ!S=+${vbK|AmeyHPycw<GwmZ&a_{zq8&=Q??A`e{ot;V5RveJfUCByup-N3!
z*hzXZ&OYq?w>QGjame}HV<671{HYQG1cm(Oj1IQI<W*H;(*s{Ym1jz%B~o))#O&4K
z{xgDyYD%hFbZHpsT|Yj;THK{BkVdi9kw<rz>$R4#l$%;=s%$ZBP|zcl@~N}xs}a&t
z(w27?lYmS+JubiBHO@3-8R{Y)E(bB`i?bLK@{zp!?0$}&^-#aZuJKJg&jtn06W7HD
z*kp&uBm0V;RHCXmfUYyF^w=x;5N>$uuD<QC^lc&?DsPqM0X`#wz;%2*XwzREOrH=Z
zJ1(2#m}O<)jp)n!?4{_r*_=AkD{`~)i&3S{>|Cr2A>14CRaoUjKP_V)q^|O(<CDh+
z%G8NNjx2F;AJSCeRwaTQ*d51YGc?$cF*>IBk1mwB3zu06sl6PyR~aQkFNPuu3HH-L
z46mh{y%J*DMSLWyKk~6dd7I-IFMe5qgUJd5Fci!n$IoQ<nPGjfKP`TX-7U}9{qN)$
zv;vYoCl9i54?dZi??pq1iG>t{l0GYgHZiYdmrpSZml=x_Qk;MFOQ*}+Ji#e8Z7QF3
zx~CA>B^XzFi>$O5oGgcvD}cy@1AYSX7Rp}mzB@Es1p5)at?TKd^WujZw0y#9Of-;{
znBV&LjOtaEd4eE)8y`>pSk3lAihC}hnZwG5>jFdmNQ<?YY6aCP?(W4dwljlv-b=y|
z>sKoyX2L+3*JK#Ebl&%yxCc1xLx<ln({5RhIEe?xgB#~wgb1tZfB8swpM3k#Q8;n)
zIWzb2lea-Z7v5O`WjeQ~GTSqL@S)pVnZRQ^P%kapZl!=1Nr2bg*DJ&5uq}=wEHc${
zf7p6Ej`9W-0+81_+MIb2omjb-&EGg<%1-};R<<Nq>7T>(K1T^iUH^>UA=>W-5)e%i
z15I4Gj0A}S{>G2kUOOa3p*=UC^1|UR9uMB_p+x_(oFb6{QX+KV8J6fYq$7IWj-5aH
z$`}U3A1p8F9zm{sksE~-*)KMA_4Q7*keA1!r!)o2A8wW~VoSw%;4yXv7xqBOSg~eM
z)<n>l_4n9pb$i5L@4YSdXnkh7$oSZJ;2{+C*Y`7gi{2aRwwG;Oy%L}Y=fCn+c${dQ
zVfe=nS!-Q{BDyDc@1pc4;fm5^r~9E8xCxd~T*v$sX2+diG^_0CIM@OTjE58>Mft!R
zQ+#P}Yj<F}2$MzKB1`$DFul+(d$Fy@Ii@Ueit0AtQJ*ibjYqLPb1svs#Xc^!8Bc-8
zb;OoseKzx5-+qRtS#h;uzE09!FKyXgqL9`}u2WA`wEkE8HXtbBi;#N(BL2&zL;z75
z{m9dqtq$rpnz=tG8`Wf;jvj_P{7RvOcPtbIhGL`qXyUCp#2_0oGoJr^mD6N(G$rL_
z-{U&I*M&uq5+vaH*ekWDL&WHqj#t>(P;gMix)S_DIh}w3Ol-!12!7Xh(;!@@0ddc0
zzs2B31<Nja{`iy{_BUm|V!Q-0G-W$^R;mh})2+PEQC3aX@)CJnpdnU|==QHCuECdO
z&xe|vDVl09)U2M81Sh8+b(r}lod&FUc`BBZK|C4s$UsHULfNu$@;KusEwi_UDa5c|
zM@y1Am})SUnJ7a(y<-o4XB_IRC25Q_qKXY6$*q<qT>CcqF5}hvp~A6GuCl08FElnG
zglN><Ke(48J~I<DKi6|}oZ0Ii?3DoijgW!YAAd@5?=;AeQ_LtY1)17jxba1q!Qg$M
z2Pb}|8myd@+A0$G4!ea4O3XcAdTxrVk81`&ACK5l(|ABMfWd4TD`oJ$uaHo=#^@0v
ze~r#_!RwKVXW;#QU%E9crSr)J+mO{t=@y)b;d8Uu{S?4!;Zxfrdnty!iu3nSvS=RO
zZJw?px@3GosENF*hy$`H(rWBNbtuks#?&F0%%v9_LQ$*{I8&!|51k`WVOUNW(8PMl
zlztyO9M#P0;DR2{u2|yQqkjc^X5xvrWd_jP$cuW<A7B^2Z6PO%ISPELmIgSy1ZmWV
z9UCDE>ougdn*?@z&y7tf*4j1&Z3+&O(z3R(-n^;{dq-gq^ggnugNlChiqw*jqches
z^E=Vq$J>gsrT6C(1ewT^&sEPet?Xfd9V7!_0=qbA0;ePIFOR5QsK@$t6!9Ih)Ibb{
zew89w-`~hNLXUyJpbt96Gsdc!>>I>=@3|x%0g7WX#_k;%WnYAS!rsdFpa+{m0dNc%
zl*JX{nj12fP*l@t(>qCHaP=}>t41%=PAe01g1Ip4VP=v8V%TtD|4{qx&WHR8r}u<X
z-?tNDvZb_+JQaP~ieFZ9!aZbW?EWj^=Y_)qPpnOu1tZ(5L=r1{>)8vjr`kIhT-lym
z-m?!Y44@^KJsKD^{sXXKJ#v=RL1{J6MH!||O@hLmZh!QnTl)e|A$_&17$wenooAfv
z`+6gCAbzyL=NB=C78)kL+iKx9ACsNzlr&!}Zs<RWW}2{6MGrVEm(TEfM8B`bs+|V?
zkb-i9(;0T5dq2T#H8t>u<}ni3JGjpP<RoTt(5IU+hdGzG*j9qr9otrW)H*YX(=y$|
zE(rYiym(PbfDqv!tsLNtCrc9o73m%oM;jrzU_vh484)m4?+t^9IFquul2XlSx>!H<
z?YAEfURhx{Pw@O7)qQ7JRLj<8cawu4IY<)8QM5^oK!YGT=O745mLNftAdQF!2p+QJ
z27xAnWRM^^2{cj3l4(G44$ZgmoOAE}F*DEi+?jdqO#j(W@9JHvs`jed{l4|?wM>B=
zy;uB*;NZ09D_wFBtkp`0o{Ime;jAJ;X!Rzv-i+v3fB8T%C^n($*Q}TrgP;=lH{)Si
zS)d-a1%?lU3#?V7h|m)NZ;|j1;AUEx;PP?`*xD7$XnNUbtPQdM{|m+eRMr0*02rwa
zsxvRj)J(noPi|mxfm;Mg2apE<j^-gnbS(cXaQ{EeG5n^^{vC4SKXc&)!D@SlV*y7S
z!b0LNVvYw4y#uf?0RE+hzzfpsY&rjNLIK2ycmNFFpDYkAx(gtleseSb<Ve^Nl<->~
ze^ZDM`zL>n`#X>-DB|1SY#<Rb3LvL`(}4f9_J6hVk;jNlb(GYeBj}AD9$YbBolba}
ztcKr#H+_>ncvXKg*R5BeuiGSy3OrKaL#+K=%T>$!M}RYE)a8^nr3Va~woS5K*?9V}
z{@r**Mpo)m*c0kZy@#LjY6j!C7KH%nb<N-cm*)$Lfx!D?4iKnN{K7NDDh9{1uA#k+
zsAs1;()GR=i9Ro}%GyXq=oLw%KWtxD-s)PBEhYK~!1NyE_k%q{%1hoGO1($u9S78u
zc&mf@Tzuk3skbGkA0Qyoff-WgDMj7wMA|e1eC36-!9UhU>VDTWV6QZzMpks}qp6W4
zK>=^bcLKGSx`j`dOd4i_$*Z}$6pzD2oA>OVh)dfNtZ2T(g^eev)5~*Dz96grS{l2Q
zP`RNIKbrYRTWxh=4kas>A!=!4Xt9EN^e!c9gC3#jKwc;{Hze1{ixi2TNV}b8MKaB_
z`*}>Nb#9FIjP=$6ZfnmJPU_6Q*)gkx5h#`-R>#BF39|Ck)ww*Okf_P)*_~#IQQfh8
zov!5m1{=}17u%(<+9RSdFVo01z<t)&`cL7jm;{#ztx0Bcw8iXMAITq!zUlP+sn;~A
zJtmrVn3dg@9V@M;WxY2dD9_Y=gJMSA6}(80XcHpAjbDNy%1>Z;;GWwgcFuBCeh`h*
zjf=d+>3)5Pr&%a`Jb;T3&AW)W$>TcZs#X0|#I-WS{?ut@rYuP{-qp&>5zTEp?4Xiw
zr=W3M*8`+2xpyG7?f`2(oK+d7&vk;VUKkDI#)Z+~!87LiO@B;MJ<YhWWs2WeM3f>H
zxcLqE5AnusMQaHsVs5ai22;Il$x-wk9*iVvx`vyo-bR7{BtE^ikQT9=R3=4<>~(GD
zR`V`}iGo$7Q8sxw6xY~CM^e2)`);gr3Jeie1vuQ6*fTrq>`WlQH)aI&=rV=i9?dY5
z_8amV+91Y;^FeicIT)fTT2KQI$g##bBRe9(-M;qm-WfY?(R^V1qW+~-3bpfvI%$Cb
z^9v$BO`-YK=<L!5_KV47!>N|*1E$%H^1b_5cg{+^n*5(_e0{=?CU4p|ceA0wm#njh
zn|vk{I~H}Zst(jBGO+$dIa|b|FF~;`-lQX<AZV|TA?=DSF$q$Swus?$(5=QV9EZA9
zd5;&4cBUtN{gVn*Cq0MC$(rl{bL~*p;R&sQrd!ZnPB03JRowDKW02!}dE(ApP|=9d
zi@_>*zP`#dZa<$yC=C&v5m>uAeCg3hw9`yCC(02pF9g<?U*EJDD-6)_G@S9ZW1hh9
zHE$O3yCoBS+!gBInodE1ruDQFMdFf6-->(UFiFX#TUu(Jfj7$#hV;E<tB0!Hwn63P
z2hBU;r9=UC7i|IOZyUVMy__;?4RN&J##;iLVu~;bja5Dc`N_TnS;$HTi6_KA^?e`}
zMWD6q_$o!Zu<Q~vIOEp%cizcufj(S`YlH0yOi9K`+g$XtpN7n-8g+-YXi@WCF{+Mf
zTg9)3H5uE4C-c}k!wqJ01;3}nsA<v+^JqSV<PwtEWg6{60&ZN&U>^o6`cy9nRdoYV
zYTcrTck37(%L*$8KQ%O4TD0s9G{tbmbaZ}I0qIv8sfpe2b7J;!#)E$dml-RslJ_XW
zMFMVg5^~+H{;@G3bF^|gQ(ewIWXe;GXF3%4@%7}Mc4=VhOk<{?*9=E+)iJo%c<91|
z$&Vzm0(DADxV~yp&bKa9$4BI{RQ-b)&;=H-;;b3WHGMaCL3e+YjifNtvCqKUy*tm5
zu{}=;G~Lv1ws15c2d}vldJw<SzEiMm{l{nf%jgcywk6Ua6+%AYETxegn3@NvuSW18
z{n@K9|GYn3$Q$SnHO*6JzX-e!KQ!&O#gX{UC_LKU?K_!PvnOUQK9MH@x<WTeb1VJp
zSbwbetgy>mA%{d`@TXH=&SxW7_Q<yTN0@fMClh5=sO+2e&#fCsKGs;CCGknr($+Q(
zm!;4#RxamD)h_CbM$7|}G22huo#2;u8ruSFhJACu`s-;5kiFZ~Xs>wgo{$tjYAoh%
zW<*~U(X!nNW%rjBC%X5NaIZ7-?iaiqaK+Vby=Tw2Y=W_g`el$x^*ka?hIV?2_!si!
zOi3qZRP^@g5G7I7`L$ag$-=0RRnE0lnc|l_th0rlH|rms<(NK<GFq5QD^SxUbAaKZ
zQ3E)Iy7y*G?6xT%z4<AOXYb;t#O$LA_VK^SJAb1^|2YKzw6{Khx|^pQk_;_XblmlZ
z`Zhda;MB5hD0{142OE8pz6-p~JO-P2-M)VNvrNQ7P9~N9H=Wmmp7gN^8vY-5YrQCz
z1)b%jSQz()1)V`emZ6%1-D2r8Zxkvcn$4c9E~r<)tR+!Zc{;g;`jv&ekcas?PQ`VA
zLmB`M$qeGD@|kU~+Gom((IDwFUH?J?uKSZjzybK_Y@bN<-;#{Q3y*t(Y{GrCciW8W
z8Vt*z5;SUEl8GP?-DedAIepO53B4}H6b$!vdplGTHzN-oVkSa+vY$s9e<@avRZ@_@
z<e62+Wmg{4O?R<6!jLsq2kXHyxZ{`~y-#^^f18R(2TPF7N?g8P?T9mSyEHb7W{P4M
z&+U$!3T>IWXh=nj4@z8n&_U(g+#FT)wuF`u<EU*p-JN$%Gfjhv1VNp)mO*@yHD)se
zuqJE^8*{rA$baoim1Oo`S`CWsc~(epzP}fOkysF*^Se@uo-*onWbi^L*p<5nFtlVS
zS65!AH>KVlN&<H&e|edIi`;0tS2{>m0A(;S;eQC`DZXdz?AtAP`-?<kShkj^$Xl~N
za{bgth<60ld%S(zQ>kJ{q#6QAHGAh)=3$khfsPqEnwtlssVP0OMTd9=5?=U7cXt%6
zWm1r2E6=^h%b)2k57RFnLg&)!^2O6f)kP}JaPuEjN{7M~J!>y)pH&S|2X;ZnX*=1H
z78jOG@W;)y%snb1Ks3FgFI86$+e#@Xu%wAGg(RLNh6v3eKuAW&_B!}7=+*1G&l*C6
z8bHf5_ItSR{Mru5uX?6ip0asIQZN{xDkl1=vB3{sN`5!;cIOQky`k`IjSSAP3v%@F
z<NZQa`Dx;wwqNs^%%~iPGp+rpOcy@7Oh(&JOE{4kxh(H-f^Q$zWVvFo_3mNx=h5)i
zHNB_!k8fBcBNN#Zl+x&xEEao3QD#2^h^N!ge4Xc9F)Cw3Jy`snvg&8V(m!LT)BJ*R
zf@8!;dc1?U9D{QVmPI6ssnoC5O)|;K%6Fa_I!;5W<~^2P{8U#f4-nM$3#P03j{e<Y
zTijnuM(irr`^WSUcYIQz*^g3WoN#lY%S1~&^^WQMi8Mg_KKTKGO>fP7Yv{PH?|H+T
zqwW^jqmZ2Ems!tPpOV%Uf%Qq@&gGA(UOj@%wsfmnSNQw#A-_qfJQHvTaiuW5@g$g=
zeUML|DAL932B#O51TdEcogbI|$Q4{lVd|N6MQrYw#QMRw_6^cBQ2JPhhmUov1n()7
zyiY8pSBB_+(6eG$RZIVxre2a6N`!p-5_vUQD1opv5o^={<YRx?wDB&oZ7QJ(wgDYY
z+Md+Y<e1{6#`N~THb~c-dEXiILnO>>cu#`b7m{=EIJspyb(Mg>dC<Y5FY7t6kylS%
zYQ#PtePvd?-o>s_gelF{)44w<cHQ4?GV6=q;&PkT{`AM0zF!>`++#Nz3e;JqH4<F=
zB|X}o4F4t7@bs;G0%$b!%H4zUz4C|#ESz1}IjERJ$CAm;;$r1E;;)Bt5bRf<(B8VI
z+fK&AHW>-pr2WR$o~oELN%lE{0$D2)y0q~vR)F*Src|Z7TmgOMr`K1lneLG~D^~+D
zHYwN4n+x&JiVO1(Y!MKD+SW?{Zu-6cbK2VDF0bwRb%w)TQ0nXjnbg92AXiaS4Q*ON
z;h*viZXiUi6nl4rSVNL9cId6n3mZl~pPaP8{=<0ORsCQ5h^C#N@oG;WKW!0qZc;bs
zzH^s+8rZ+)*Q^Xsw_O5VGiIh&P~CdRaWrx4*m$gV<CeB@Q0%O^P<qtIYK{)<O|I#r
z-32{{>ak+g$hgq)T7OQpK4~`q;{1{ry^y>Qu|JC*E+n1YE3#$3h+K=Eb;;xEUZjrk
zYdLrF7h`S9k}~z7OUw}w{JfMVzKgsf4vIRiv=+8C<?o&at-AbD-pb4eRcM5`Kv6n&
zHbZPE(|_#m@~9FNh9xgi!Oc>r(<<WijQ-lezNS?3Wlwy7E}%H{Ka7n2eb~i=ix4v~
z<GhK#-|{!G`&X!)dkKp8TW9d!R|fp4M)@}+1^<}~PZ>fLh*eLiG(ivZ4KJ)>l(U{^
z%g*7VZy%JmqxLnyW1;yYByl65AjTZ_FA-;a`MSaY(ASKb?(nBE$fE-wFsjRCp)m~j
z6$Sj0-jnKQmWl>z!Gy>`-m=M>l1#n>d0g}=^|<qtkxrb?O$bk@0(+(K?}KC*NR>TG
zIg%X_t;`Yqi4e%G7P`puKQ3xs!`8+W2+8}_`M$hOpHf@5W?fN;lwr@gG0Q+2iFe35
z{UUA@<16m$N(B!Z-DFEa?O|)dt3PfO2q}Ld#@w>@vHL2UO#*xipIGpMVWoP{aMAF}
zCMm)d%v2K;AWUB+TzqFF+SiuzS5Pl+{=DaL9Gg})m=jP}9_a<MyFd~mt=5E`UcA}!
zoH6ONi*C;V<B41OD;x`Y(_*i!=}jzoQHmqwaM5g{H%Kk&So()+x%109U02wse<<Pp
zq?Q)=0`^yxKN<i=X=vkkeS_%QGG>ZKq)z}T=J$v-DiZ{DY}z42HuI&RVnoR#;ai0v
zINI70Knaxcn$q|0MGcg==tngD!-+D{hsC+C_)<P74_d(ca7Q*DxoreY&Tt^`3*6E)
zA@(1jB+g~y*d^EEg>wintiP)SO|bTj1tsD0dA3fx<woFVF{JPN$~6)%VX|b<Gxow2
z$9i28!g8Q%${7%7m(B}qI4gzYWh(r{!S(DqJjej9W%q6mE(i8kU>nmd6?(aWimf6C
zL&fm+J@IQP9au(`pq<%yR-)IJ$6fHo9+F&{MWU$WlefoO2o5Ova8jC}Atjs5YT~_h
zTvvrzsF>-_o0}cfLrNUp-x%Liwp`ZzunNl&QA8or`fnfMTwA-9bWI+2_1xuz1%<0k
z8LHq;4USb{k77~;aYiX~C{e&wI4HYy{`ZP1KKN{hm-x?thVOrnJ?q3D_dDfzvRfPA
ze7D=qbthB$F#G#RYXFvy>E1DUv7)nX7MFwID3PK@ajM35<H+>OTpe2S#wLjm{Mzmd
zD%<@r5$8_tKohdl*k-U|$W;2qA@~$~*KshdYjg7zWW+Qza`o`TR9I||83ZoUyOO*V
z=-u18wDJS&<7DEUq=Ji{!89n5ullrzqN>1QfkF8}1Kri9gy`rs&-+`_YqS^}Ogsq9
zFZhIz(41Of0&L3p%2@*!-G~_rd`;_TWHx#SF54IL@FlSiA-ZhMQ?5GwS?immVt<68
z)vWx(0=JSM;IQsw{AL41Ai%NuhsyQ7At87KXN=-HD_pH$7<_@%xgqPA<;0AnzYxWK
zEKvB&3Az~B8L9y~b$91%`D@1a9~V3D!$C<Ia1kW-R<yLV<>@FBz9q|OSQU_~F3QQv
zP}pMkG}cYSNT+(Dt3j2{o&kR1`>JuB5P1>(qS)q!%JbpcX@Me${vQ{6t}9`OrMjCe
z;EH-!fsr`4r$krJ(~5!43rN~sLG#{-3S42?BfFECfjKm;CeqfqzEn)pHMeqj4M~z;
z?gQja|LD7*rk_V%PKrEt*`s_-2|vp=JDJRR!Ro(_^Rf2D_RjYT-}UA*_ceR5gKWfy
zN!d3ZEG>ow`2#br7d|u@%BS!v))NxW`KN%Py=rlNCe*17M?XAR9c*~z;6eP_pJ9@>
z_MWn<sq|&@H$dJN3b4J=WPa(aL1RIxje@B>6bVS4#WsJ^uB<-zME>5prG6g%iUIDt
zWc(WFEhRxNkZh}08q+h<9&{<KcI4}&yx5c!a-HSLo?2yhHu<YGmRFMxfY-ku-{V3W
zk2jjmbzY{&NeOn{<YW|u%~JBpjOtPkuY`~MB-Sq`Q%%*1@1s~!VKppH5<BbGg)PvO
z+Gh8)A%^zurc-<42v=vc1bYUbN{#Y*kEd1S4A~CZrY}x1y(ps}hh~|7QRH_9;>RY+
zj&TC_QkW=-$!BsGD|$fD-53<`>Z_-DfS*f%fi&mi5E%3B&sno~dLUoTd7WcRquBbP
zn7)1pax0IQg??y-4GlD>q`&^Q8dTco(S2X3a8tqTyOv-lLovjRZjW{THanv3strB5
z%ss^whboSszB+g+xMyI05LkH*)iBfc_9x%=2JB9GmdV2xCa#^VlrY9Pde@z9B)<PO
zdjc4JofKYwBG`ZKaGnuAAHX=Ye%;;;&r!b<+G~+qdQkd&RKQb9&8`+QRww|v5N1yf
zai*~VX;;H|ha~LP;~-l6vn$x-r0Z&l>&LvWOtV2tqC3BBtM^14p)FU=dWk*gP8t#d
zWS0C<Gk>&)C{0zp4Lt*nSykReG~sNllm+G8KRnY8jYVr9UipttO)1(%#`e3120EYP
z6`GiU_zvoeT1Y@^`luEgwa+%{oHK%73W}M~vSI{?U2p&QtyFN|oT=95%D>j>wwV16
zKb-_!ewr#_sw22*d-0QD$-=-@`I*vvrAyznuEzZ!%^**T#t4bSlJk8lNYLc05i3`6
zxnBpFOlN+K%;dIej_`vhjbeD}T31V8)Hj`bB@P<o7hxr*yym^~n5RZxlME!`bBNlO
zC6HHs3}2cVv7z9oU8b7cWzW|ys)xkk);#<o(_kMBwHgILr<l`nbh_XWkSnI?C)i+l
zGM(S<bRO(}-?6LaV)R^+fCa1EU7^ghfDF$YIVTluDTjrxoOrnM^X?;dt+F5RT~95B
z7p-+kYRnIjNbvOaGvSQB_^u0UrQzx36HiLhJR`7vrLQ|@4r@)&>VxIgW$SBcv}^fk
z)#B`rd(Thm{8GEq&r%FVo9o1LL+rOa%_i9iIlQTAb7tcBaAl5{EdAw}3|R%2-a@@N
zWD+kD*Ev1#bL(g0#*GbE%m_w+&X*MPh#K6zatQ;a$I+2r%{4~X?RZlUh-07F;OA1h
zoA8-K0B^{f@7x1vFre792INw5@mmJ?U28y4I#OXRTr|s?w6nl6yf~Q0_2+qEa;>{A
zn6-Mu&a6JLq<EF*^F1j8aDGIou-lkYO9_Jo4QOJKcqBjat$c&bVbX-ET?6D0a6Qq5
zEOy|gWMRTiDQHSEt)D9%4Q=w-a7X2U4sM45GR|x_+cNgyEb<tcv-Z2h?_zK0hWZYr
z$Xx2oWq5N%O2{qiHA^0SO!}+-yvQ0pt!2tbY5!`1#<--E8R_CR^aM8-<Ap6Ak}<bo
zY><L4F^y!0gC>lmUswj_?jk0ZXYL)h&PvCpS$++%DvrJbi;~7&YBK>MAox%}?XOid
z$}yrv)IAJ@jSgDZZGb4PDByxqA`gMct$hb441D#3NUwGMdcN^FKmT&lJqyRTYntpD
z;594yG|+_IPP-c$A(Lu+mFx?(55}or$3QrcYPA>NK87Dy2iuMBtYb>QS3uqFJ`B9`
z>LVy0uX>qy61--Jnl~tbpv%7SB|p#taSB>myWtsIcXs-omg`d>-E<WZNn?=%D^0So
z#filP+V@Hf5JHW>d@J`u@Cn+9BTV_G{1N$5q1WLmg6rk7^>LBK?l$7ll?1m_u}A3d
z0q1i^5yz#{Z8i}_nTq{4?$DrjaL`;vNKG$Vac>RL$~YGWP^CeQ3Kfv!-6ZlkBA?;U
z%6|3SCVVB>5cQNiceQ!`?A*8mc@X$sorWN4v8|E&T4yZSX86-14;Lz=!PTS_U~$XC
z(O#!<Bgn06(ZG+I0wtx7RXn<`awhjF#nq}1KtTE2&U|7<_~@xy+ZfAx*a+4DiGe<?
zJZOZ9&iI>{`Ii$n*vm{y^T*dVKvxYk^rF=~5g8pZg#BGB0M8(~0B%frob90QNN3N|
zJkh80X*=t52*t#&6pI&r>S=eVn%rZuT4IUzH5OYkiL<}UCthK$5*tx8w#MKIzPc9$
zuv)HH&xUF##NvlcNIugePgkOCtcu)(Qcrt|H1plpx%b4?vP1WKpr_{>0{%1v8d=<u
z%GjYd^{y4O`)J0r@jle6>2SrR;A1t*3;(4IVeCPZpM=pab??r;qIXAhcX0MrxDIJ0
zl=GR%;5Z)gS~23jmukE;qt)Z9`3J?z&1m5Sc@@O)_3CQwWaRPpj?AQbehR8k;wIl2
zNFJjw0`mwiaV&9c*5eSQ2a5`%cRQ$`=R`yO=RaU@$d58>WUa)_YExY)Sm!t2ToaTz
z3+y32;eX!LBL)_sRFV-QNJpS9acZ~T)wI%e*eKf3A;f9{kGVxs<Xr#R?V>8%bBw=~
z9!ZdvEPLFlB!*z*OHo7FVDz~SYD={&<fqt8H`6lb$eYq-w8g2qFLJ)rpO8w8>l>3;
z)@5dLQowz=o+YesG&*ux+@(3()Jf+WZcSAQ(A99o+)4d7(@IksJuriZ{Jj>_6AOCe
z!5hpmg+@q$f>_O-r{#q#W5g>%4kmS@DipFsgxUsG5f=(?*Py4hCobP9QifX{<rh{a
z$^Zd+45p*|pku{FY&fD2dg4rev=@J`$UA2)Lpi<H)OfC66XfiE^~C$bCb)!H?N1Z$
zpK~Eh`BT*F)(CUWQ0n)#pgd5Tmimm{`wnkS>2f8*RYBPoa1~nV<Ck;AFWG&K0^}6~
z;u$~qVK41|30%n4WXrY>Fzopm&w830S*7Q0Bq*t}%9WSjfG3*=u8k(riQ5+K8d&kB
z{vy_Yva<)IssxNB@TZ!k-oZReS3Y>WbjMZ}lYx~8=V9X*se^kQNNm6H*^7ff=}3fi
zaM4%{&PB6K)-Scoyxrx5MRB)7Kbf7VJ7&U1Y`I^x19mk`Mt9rlNs+bwnWc9%a?dHC
zFeykyH%+d2>XZ#H7L-gQ)-hyK40kYO*iZ*1u0%qyb^r>N-6Bq06uCjLH^Lg%3kekV
zkF_yAmg_Hs6Qm~_;pbZXVG{<LP?A1wx7>Ni&T?kG^WAP~Ye=Tx>Ehu}L5a(o_dFG4
zq}vJR4#`C(I~$|Uimo><V?~rL$h70z%`97!#S=O+mC!#oh|v#>kk(#V6fR>Km$=*V
zKR&A3H9|<rS}c8+m*3J<Ed#ffVQovF8*}3SVxe0;pZq5~?W?il^7L|&XV<yTPz=j;
za1kwwDd85%(jHJl<<G~?T36^~B)XXQdbx0+pmO^3?CauOHp~1gu$#wuljuN%_ZD3C
zBD}K^drM@}%c6|7S7&u28xL45qm+v<VM9jR_-Lg{mc*$67AURl)??DNE6_YKEXWxv
z0M1QG!&M_5G9dm{bfa#}<+;rjl+LPp)2h6TIZ7hE5C5h3-hP7{*z^WYA7*`fBaD-4
z6({dYu~789oS)_jO;w`!9rlf2!~y(T=tCd4$2ul~{|bUTYSEh|9lG_XSgstPoqolC
z9k_m;a%I|dR8;bPg+0rvovZ&fBOen(UWs+p0=-OHV2Av_i{}3qK>vSQgr!OwuB4~#
SyQKjFJ}Qcu3dQn|UjG+l3MW(m

literal 47891
zcmdqJ2UHZ>)-Kvj2FX!CKys5sa&B^lCTCE|S+e8?1qCGM+~g!6S&*RQOp{bVK$M)5
zNahuLAOGz+_l-N=9pk<`#(&F2sa0#uIoGOXcYX8wrV%g=MLZlT91sYEr>rEW4FaJ<
zK_Ju*Sg632u6K-EAP`D1OifoF_$vq`3If@HL`6ZOz&Q?N)5sZG#T8n`{k-~N*jwNp
zaO3vo?O`Ko1GG3%n?}*NIMF!ZJSQGmD;3?KetWp6tH-FjxTw1T=L+>0p!FMh2^e_^
z8UxK+$i!R7+*kBTPOEjWoDJYo6lj3OI2#+#t<~Ej&L$3MjW%&}HjRxojllWFE>zJ$
z+t5MB$WhnWF)h|9J<h4L*IC~baOPrQcI)l-h;fMlT7^qYze`1hO9gOVcFj!ivas^b
zNqP3f);BLbB-t+VsaWJw@hCTms9g8g9?)n{XmkKP1_;^(Xn+Crn3#%~ivF02<(SHn
znEvIM{+l>aP@E_bA}H>!KyTx_{b>Y##We!wxwtu?-6Z&_BnPUehDW9~)fB`#l`^1~
zGNP3+p_fB&DqPenVu0pS0a(8ct-qpwxuXB3VtKh@`KFSOqB7I_tsqrhU0tIMsL@8W
zF%D=p|1#eQ9CM9xK)Y$Yxp`+I|K3vRy`}PdE9F)zl~x<IHe2;}7yXZJMxEB0y|29b
zW7PX&T>2}37Smq=xa#j;UhZGM>AzX-zquJI)gCI<87k8mF4G+@*B>c27^$>CCIlnj
zJ{f-(@WoT<i>LAzFO@GoYBSH&XM*%+lf35QMCamc<{E((HwT2$8Gbw6+lYYUX71+Z
zYlzXp!oqS+@N$LAaz)H?rR#Ek%yK_4Ilwb;+yKsj)AG&DYL5C^n$%hvbS+(aJyUKY
zPiwcy_omV2ra$IpF79To@dmKCTyY!eZFgY0K*)dx;LkUF9*>-WeTw~5$;bl)qIr4y
z1CCueum*wdgOuf@b)RKzw_PQY^`)=fu{7d=U{uoN4N1cf+&9EhTwmCbeO|?E$d~=}
zN`QvswOz67;Slb#t`<IC{LHxEkB7#O(I4M?pR$7OHY4~ruFr-!Wdkg&8B3ND5w9M?
zxoyPvy@B~TuG!7N$mjCm@;Bwfe!dzu(?<<jxjM<|i*H2ra*vixvdMUH!esIs?0epE
zz-97ac{t!m;5UIVkbEAjG(!fS4}<8c(?hi&q~*ng2nZ>To(>j_6ep~l4GjW}oF+vf
zy&eBQAC|o-28=hD*n5g*S2CEjgEf3NTjs%L0ZH2Uvn)acLIxhbj%G8Z$oha0Z5{Li
zgg1^oTpFwKNC688#sX92lcZ_UWq){0h(IMo;g@55P#`H9LP-zBypLx03g{K+iNYrX
z$A@Ae34vbcfT<TC8aN?>9_aNGFogvc1<^xgfnK*RiBLR2|HmCt;XtKgWY7^TpM@z<
zNOgfF9HbLy?rjDpPlXS4Ty=3bDu8Loz%f|S?81O`en1a>PJ&V|19VIyM0~^s8`A^p
zt%imC_5xH%h(e0af@b$r2Ck0<jzl3r;TNWda=ZY^hhT%Lfb}E;9@__C6*aJsIKZQO
z^#7yTJ#dT1P)hbl!h@1eU<WN8!ZU3G-8+Nk_MhlnlG^3qp@Tl~1ddmx^3e$oSiS96
ztNJ$Yc(S6}wR=YS38eO1O=?R7^u68t>8O0nu7~gS<fxncJ%Nyl1p`!iMFfgXTRO;{
zKgcv3VV)uKWgpL4Xf0&~7g2l6R3_h+F|;~VknIN#py7`L45%*O8kj5|At52r|Lxor
zzMBMpfEgn;Gi{@mFB0#G7P`D;`VwzUq^5a&^=osO=?4_Vo8#s0Rf5ja<WHT-XzyEK
z;w}4M%@W)1sujL$v7hYjUL7IH6g0W#^Y?sQ;gMpj9(zhLH659F%8gF?JT0$^8B<rg
zQzEd*1WoR03ArxtKW-OM-ik17YsDTrDfv~?Y`!yy5#u1S+2ro{=x)6--l5k@$ol8Y
zRczl<k{7uoB$<+<gvMUWTIaQWAhUMiU8-fW`_PpgCY1gNQ{qs~P~`GXZd@Q!+}HXr
zVqW!kVTBA92bnB_iDs9pQ*WV>x%XdNgR;N9P}#M_g3msK*i+TJ&&OgkD&x5|s-pe0
z{@A{H#;n3qJ5qnuZqIjx!%AV_lc%^b^a13zRThu@F=oKjwtwvMbLTK@EZ%f0N7%u!
z{setxXK8#l^u?MY^^S_wAvL1-i>wRDGSj_brkU0-Wn#~@+&F#Fg8S&}<i6A{Gqn+N
zh)gNTnh5^ne)JaKAq~7cVw6x{jRda}Z#vm-Il(VVTjPbvTey&KEI*%taei9Y=%6I|
z@Q?HO9R6%iH-%@~skXObXq`VagvYRdXlOOCT&OIuZmdjEKt;3L)i0tTk6_)9XOY#!
zo?wbXD}Bv>BLmz2A&y2iUy^B$Npj;?Lcwm|=lTgJf7FDf`~C!Z5qfpwlJjmw3s?=k
z(6>ee{;9@&1b4oW){%sEdnlA4F>vT8h~+R<a6!m5-(bC1-9#{zHXx}T7au{;I%v+>
zAt5lUpQ4Qk@ol;;S>D5}emz)Mq>uvgN?Zet1l&ye)Jw?v^CdQ3*ht6`UaH9;xq@qz
zkF5?XJX8}Ro5qxVIdPclvy@yWv-3kV-xn?Ov+PJV#;9&$J}7Fc`b5xL^f+!63w!_7
z)7Q_Hp6(axq-eu?FX#OEI^69w*X4{>A%T?aQ2X_!yf=wDnCsDO>8|!kNiY|FCAb6l
z-upBmSDRN^aiA5Q_z7Gi`0iyaou(YOY!Jn+RT-MTJxJx3lXCdR##feEdzIOfK{>x2
zNp1p8lk`2^&V_nIF5=2NltP1DLS=nI(Pi%~M@*r|n{%Dqy<B4Q&tEF!;^r>zy_hI-
zg#pV!fe4Kd%FU#Trj;6OItUu&F)w^pq*B_HgiJ|{=Wk7oeXZc=({mukbfe#XwCcfV
zl!CU7srbbO4C`^SuXnb0sJ}-4^rC~eVaZ2|<^89q7$b>PhCOh(slAwQogZJO`EbTT
zy+H1j&8Xm#Wf1yAwHt9nq+^O%RbfQ;NxW31oo_yBZl6azW@Ji!m7F_)SSI>|u((mO
zIewV?rtMwXL_a5z4KY-^RbvKL9XSNaA#A{_UF2+))#AXqNPc4o30AqEl7IMWeU<6$
zF!(s6z(Hh9k#25Cds~#jgUPybo`x0F_ED(vMw~-Mz=Z0G()ZVg3KYbaJl$5KaH?wZ
z4M@G3-<wxw-}%X<RhvSRe>!fubRpLbt2OHldv09Q61gCh$tL1K(8EHFSFcihf?ZTF
z;;qJ0ez~n3WgqrXiAKF=9EGh~_s$0zS!vQ=j1~tXGAFr?8B(#BEsm^vfDoU82Q&7Y
za`MN29C9#5zGo)an&@|L$|LaHA(2vFBx#r2cWz0x78$E0*>)DD3f{lx-vw5Qv<OZj
zAB!AL8=tx%zXOj^$Zfm3_;u(U)9$5XhHBzx(ABe{45cc=^DxadHrFQ-h)knjYiO>W
zQJ|y|6lQ8Ps2P0fDu7kC=4|vk{#>i95hbb5PwV&}5BJ2`o4tjSKiyZc2*-hmQq_ym
zpxOo#9G%AQJDPliT(+>vQD-tHhPs+?PSTQ(Tz1uS;=^VREe12W4v2~2k;ufg@5$o1
zD92*j@qG>yxtY`0NA-z`a)l&fO>w>Pcj2jy2~1EU=pcAg?)C1PSbB+%wP=(WE|EQ^
zVp`qoJmZl@JE!|1{E~#L+3)6uh2aAKM=*bOZ+&utChpI#X)20IuR+m27RCFh*K>=o
zd$5wWHwxF)A95ZbQ+A1ziW}ea*|0GSsXXm1Hf?E*f&LgyW|`e5sSOnW*!P2seYY^P
zzQNVR`?^O%gxWllYmkMf28GqoG2$xk!lyPUfVY@QVJRD~>(QPVH5j$ilKAi|s9s}a
zZWr%yti^+0A0e@oL_>0d%xGL(e77rv^m-#*wV(9E=}buKY}JyH5gH-+4vHBfXsQ;f
zOo`N9X8y7J#v{2+bqLWV-I9*0fe!T*2{&yCMJY9GWphOtpApGx9GVH|aw6}4krsDM
z!F0wVExmGX#pOwv<PPucd>$vTK{O3=@L`<J+*>Z5cp^s2LeO+)>Sc!OlEij1M<kE1
z#zJk(=?I&Yl)2vyu%Etx%GG5CLC4vB3InhhY=rKf8e63s4I<y}L<RvHwRT4>%}`D4
z16eO#raw|~`eA2<y2gpILiy}#KbB~G438H4aLL=cHG*O&VZGw3FR3a;zw0WVU(f+N
zwAQyb%P-fPKTGEvw@d>}wL9TS=fGY5R^$d;qqyLt<tJ}X7l?zrCSW*d+H@LWo`42l
z-dQ|%ZpW@Jw(PE;2-&o`iiG}T`(FGqaQ~SOrE4m+u(-9;IjiR?3C-eH{=DUtA`Pci
zEeChY2uKP(r<<HL*nvzdra5|!3-YZbq>-7xAt!*Jq*(QZx&1)z2+1sBqPX@+!Ljuh
z&@^P=LNYoK;ZC7P1RXku&PYlWb77R`sHQf7)IS(uF^>NbiRUEIKtW<(Ga2(1anII9
zjJ`CGnKm7Ro@+!r@%^qwvQSGTyow3(nbrLXk`-8y9+_mn%N@^kSs`lQ>_kfoO^o-5
z<ZE`dXJ33lQLy_4t>rTaW<)+v`nofUL2>hegrzMu^rvY?0D37oX~yI35=V{F=VSMR
zNSyF&h(i?uqA{Q6=bZ|pBC(krv%sYk(NJnHN|cxG_=ksxRcqWuHZVJn3UO$$^XX&$
zrJk?qvo*?d-wL8Bept6Kyh1&$p78XfUH=qZFlObKF+PF@>k^0Y798!ch09F8!A1CS
ztGA;}u^9IE83mxfU!UVdwQJ1O&hsLMvxC%Ab7G~!YjjfS-xA33{^Y;TNafAa5dD-P
zPh!F{q5Q1tuAB-5=D00qHpNz)`D0IUaqV2{+ExR65K&_~*ugl$Gm+q)^J#b^aYRNk
zNViIkxLG$07bct4;NkHQoz1y(G9{=OJ9nT0pTRkPzC?vY60H3Iw4>XV(L3di7g<e5
zgAkhuYZdXu*RU@GzUDpQhbtRT@I-0CESgl`I-iMLMfh~&VJpT;<v!uLgFIz1QyI&>
z%dYr3<;ZcV*@zN4V|>~`Um4cq=uG*Yvr&1}_DSj!Qdc6-ymp#FbB!D+2aj7Iehy|d
zl#X=QwKck9wFmce=Ds-Ed~Lc!e_jWHF20I1KErfEi<)HQjBA&A(sko&cgq8aU?6FP
zV1(~wG*tH+yUjbnSKTmGMOmSe*ryQP+jqW3p5|q~z;JU0SMcoHhGezEHI41HmU|CN
z44k~l`cJ7tN*7czd%LT;xV(ux{u)rtd}0-U+Npc}_~$iBeZwbmk8(pCgEwCvwj2nd
zXay#V#xE#qofCs++Y`VQmn69o7<LsJJ7ChVO_he~dPradXh&xI<7n<oa_|$${{|@F
zAIviO6cj0QR{aOep%!TQ8|L6ZFIN5rc?AamUXc#_zYymAf5(-`2jKI<rt&}~nzEO_
zf$nCi1D&?wX5QZzm5LoY^iVkNjWs>=!s>58Tik|)q!s%Y^ewgLL8<TiJ=S413ba%6
z8{KA{FhKpcf8*f)2E#BgkP>`+b^a&Y{)ftkxuMG!C-<SU9<v6QZjFx1#WlAZX4(~A
zORegbkv_wBw2U9wA83_&&~Eum9lYu3m@?g|(uGnQWKJa<@4nt=^cWc$U(oyg1qfX)
z7Tn@GHq6;pqg|3NVKuqEohSx@{@TeLSoFD2f|sLA1vs{7HLRZq+QC30S*hn5{wKUE
z8g3sWC=|lzo=`8QXOO^2S<xV+D5#}fJ#DsQ+c`!u$Uuw4^poN3^z8JRkEg<Qu8Z5-
zIopAX594k;;1i~+CX*qr+zxh}6W&&;QOslr>HBER9BI$7oh6+lQ`;$AC|P!2(zgV*
z6h#`d^^9YKamwF<HlMmB&RKdBgg~(QZRP#124*Dq#Dt}Cnc`PD=aofNl&R`oC0j2A
zBI@{iUT=IF{+dSC&C${p{UorCT=A|^(sYcd$6$8qtVTY9lYPM6$|Am*L^HZyairbs
z$aZ$+qDn=JI_#Y&u##k?P#SI*_r(WyiE)CQ_3J<T3=fNoFa}FK^9~>PBA{nyhCJrL
zlk_PBF~9eD+PxW*F4htG486(6IK%JxQ5AYhW71H352c-Aw#dq;s&h#wpV|Q%eOdPK
zt0s*vO^&<w!~z%f*c|e3ZA&Ckjz8%z=Jn;4?2W|=wpdd~(>_aXHM<Cv=N{?|lR+un
z`2I;Gv7J$QMndLYKS<?y{0t4YWbRk#jd!+-xH|;Q5awVkWlBDDsL=xD!r<r3v)L=-
zfn#^L&3-1x(?TS*K<0hpbh?f|i0jb;?!_q^6Q8X}r0<Ux+Wf!Dn5HwF^H}*EjG;6o
zL8zYH>){|SZ}x~*T+RtsycL?jp4|i*|LcSfmb>%x)cgFPZ2Iv5{6^~(4=>E3)fgjm
zM4gynZmBtAC3$nEkipgWIGe>BQ#c2(p;-J8E%~84I~9B-jT!slb^T3aI7*{7{<Q^$
z^%}xvUIUt`Cky<$*wIw0)hkn)aN=t%u3J308mG=!%nY4hriKnNRO7bvSC%YNZN3j=
zKx2+Rp`_JB&Z!*-YuoMKcYb!B(NT2D#N&fB^ke};u%+2xwzl7e<qN#mY|~vzg9h&M
zX0jk;m?EIj+11IcWAZ6oZd}6FOeMmvX7C5PpZjTS&&32j8N{}6{dMpW>dxM*<NlBh
z`s-D-R@=>s=%t$O0H&Ijy3&Bbr){ziGjy4>vi$I1RCF%y>~fu-g1DTCD%NMax>S<+
z{M2L@het=Z(`F7*>qrrtJ2^Ds6HGTV(u+tjSB1tc-v=$S-36`ar%ZHCU%FbRl++R{
z3s8`qTK-2qojIgi_88d6bcqW!X1y$NiW&Sq0iQugn9W4?#0UGaL`|DWWt^0N6cT@u
zdlM|J!-;v^Y;s%&1Lb>ZZ#)zl@XJ0FibsE+@_;X-G4L31T=Y(bFJ3)!;5WVzU-UQ<
zIb1y`FcxHl@&v^2Z+s4ncTXd&53u4le$NrJNZ;HP>3r5HCEV#vN4qnsJHi9k_gpFo
z_+D*et2b(uFCF=cMr{2)v2$5Op=VaMB4<5KpYgFnAA}}5m+UADB5mv4A8*Rb^ifzO
zvy-**K?7gxEcNrYKw`wKZaEs_P|0T4w)5Gcsg=aN&tKRIlDv(Mr`=|MEho-AsSg!F
z=c|7`#nc>WeP~RhMl`>9PdiW9o}tGc3NpLl-&?ebJ}+XswrR>3K|?a;8zd>z{)I^w
zwng)uhA4ubQ?y?n<)GF-=@*U2JW+<T-~S@SS79lweZ<3d!8CiaJ@YYR`(hn2v5jlJ
zj%&Myf0BTPsLf4!Yv8TCn@J{1JZu>|!NJZovuNbuW4aW-OGIBf0i<#uvDJ1}`3K7)
z{XuV<%XAcM)Fg~P55Uyo<p&XvQCE&(>kStT*jikq`k*$Oh6E<WvGY!E^xw!*74y|P
zoG5}W)q;|3m><P6q9>Do_p7=pwZ#6;GhZG8nX$Uz@J{8>pNjTN1%x1Dcg%X$S(OD@
zP{xbg{3VS}?FAo&67dlh=pwxWjrXt>xQy7<JNEI^0;jYR7}iX-j7-#dq@(qtPuBun
zuWZW-f&rejg$M0NextR`dWBE2-E)4o<snW7Zwg3a*n^10C9ywAKHhn?Ggaa1O#lvZ
zPf1>KW6R>={Bwu-t#9J_k=n;h7#hUBhgGmpej+$={jM@&1I?2uZyQB!cO8Yg<*xxz
zOOzh3vEA3Tmvw`$96!)KmDV$6uNszV;B~4P)Lt_G1v14Se<&*j&KOmkrDA>k<AC|c
z_d|0cr-d{X%=FhFvILNI*-B1VfJx(1Y%^oTB2lUHqA5zFQKh=nP;nJxz_9DWvG6Kq
znejzt(L5q4wlG4YMnsnChI2D3eI~x~oBfe!Wo#M=A75g&1=G2Zyr_03ZQFU+^Ld7s
zl8P$We1<o+<1?=!HC}I>@qHfYEz7}qpa(oLTm?iR#(E2Dkjmnn-h;Psyc&<KCuVu>
zg;`vWXttkyfrMC7EtP}xaMt44G{1>JofTh_z}7(pQ9Q}IM)2B@9M&1#vve9hPJ~r=
z=E|3^(p{Q0q}R&4im<7dsr4_9v^H##Im($o0N)Iw?9)8vZqAb0JKSs=gfTc8Q37}U
zAHFK|7^{zV8QqRm)16Hm(Zk~^N*iTS?RINCN!GYRvr~~TR{#aB)o48wx4Hu}in#$b
zWap-RjHMLx|KhaSku+H(vb?@dh`vrM7~ah|$ZGEc`M4-47};HA&^)Y3oDtnQw%)Zp
z7`bT$%O*`cr>J_zEv&sG$d>n5*Hb?-JyTsI$6n1Hg-D_pH)vBN{ZWYr-~4^j9t`#(
zvbwdBG<%h5GLKNdmr2R$dZZ<JC{qvm>K+l=*Pk5l1Z*|=Ur>e^bG#3qM57jbB6%TT
z?IJ<Pn7`vq^?2f0{7OciW6XM|`UD2Wk&H*3$c#9Q%Y9N`w@#kNkHGFAa*JW(m$F;h
zU9#mS%=I*uiM=Ju7p<WvQs0lKJXH+Kp6zQV8|}_sE4d7GUofO9nc`InXIEIn-FKI|
zm&jS3FN{ZmVEa+;<O@G+AFC&r!ayj~m40M4gxgrX!0X%mJYDwlI0WIj${L*T>h3Z9
z&Nk&qsh`TY6`O5u2#I3E*M}VC>XW6B`lb|Y9tEV~5n9sM`gCDho4PtwOI}}_R&;&w
zabAC=v+NTMKHYxAT7qAA>+ywKz`mV*AwI(jm_5ok!yQx#Y?Z@6DjihH>R@cxSvavr
zI9~e^Tj3LijbTafd!v&}LRt#4)6y}HaRc486fg4Xcih^76A&Ddf_#)W7lO`u>$;lE
zTaV`4%K_;7I*YEb)iw66Ip?7M@>Dx%NeIeu@04AaR%4D`W&-dTH{Kg{@X*ZKiU7t-
zaJ*ASi#io>rSm3;D+23m^I3Smm32Z^kI0Z^52d@R;ju_fZ8D)wQgxW*%oVxAwcI&5
z9j57$uI9`dSt08iu98_$rSr;X**Xjyq!sgozCdOk!EvkM%VQj_E@Ff`^F+Ao@m+)s
zQAU7SJ30LlJ`PgC{aLoqurI#~CX%lsLf5!HV7YYZF-!5)w8Ijs#O#H1BzuXUHg<LP
z+s3R-!KJOz?_1^aB{*yvvgi<tJ3OBO+Jg3q>lUd-HT`r34}R*b{u#^_(5G=pA!hhG
zogU9(fSXp$6&p5#DQM}JmE#s%6uzISwCQW0Y3QgC=l1yQvtFjgigwmH7R+MmhcuHI
zs|Yd?Gue;z3rdnyyhgTfMsu$FMFl?^+(BEz&|gEcH|!}u-<&XiemVF~cl<4g-IY6v
zj74AGpdr<4Zt^**osb6Yzxp(E=Wr5Rqzy{F39<vN2pO1woK3Khk^n(=k-yS5$hhx`
zhmTOhf|>k8-g05HbTuy@1lt%~KR(5RO`4<G#r~$&rs^rcvzEU}IDs8`_{o>w?3ubh
z5kkWIUp(7?hhehlrPcoduaY<agjXz3S)czmMz8<`I$!0hsB78V6Dzd{{sX2ipdPnj
z8Z7@*#DY=$!zae?fdco;QKg6wE`PwkG4T0I&9u`9kQYGD{iZB^@sV2!K@8COKPb)|
zR8J|Cdh|c+ErQyvn0tqk>3k0^T+7a_4ZOBX{vk8_UEzn^NU;YCP5=e?oBhNABXW1k
z6=aG2@pXm=`}Cp@f2{bApgO=*9U&lR>3@jQWfACx_<6F$PN~fEKL-2a!vteDMqM`f
z?)?$i9>}^4@ykai@8ch&Yyc*Nb^Rr6uP*}c-)41u!bc!s^(+f3hC>9PC4Y0c;+POg
zwL-jqtlrYGRQI74yLBy+zVq0_zc<`PQI)2-4~oa%TkbhpDZLi!F_;2niuG@fIGLIV
z!6NTWAshs3*fDUVgDe)TnF#ElJ>?zp19-515yJn=S8&h~O^fF}s;*b%^Eh}4|85co
z1+f=w#JOgA_Mj@XnjWq6>no4|o5TAL2LL8V2O;My^lcvml6Jvt>7gqMXr;<pbLcUP
z*u(w7@b*;Sw0;_?^?dg>aui(`Id}$XgW0!JyeBz;KHzUH0PwG{71J&ymqY)Oh<Hlt
z-(4TlAc~8W=z8^AGYJvyj3`rur8_8nduTSxV$j97Dbr>@FXE{0B%8p_`j3TSM?s3U
zcP_S{-RBPb&zOP7ObcH~ub*+<>m!^p`wy4*5EpfUIoqrW`#z%oX}1GB=yqoBHR*Lv
z!Trw(0h74IfNY^m`-|gc!pJtopqlqYj|+y2ajlgiNwAO%e?74hrGRBZ4Vd;V_rECu
z9msb7{_^FqvnbF<7L$P1fdH}MW83tL8)fP{n&NT3uxF{b7DpsV++9NhWOk<KMUrV4
z3-Z$L0ITdKk+24N%5)g<55|#tc)Y|<xkSV<9uDq}#_XA)mG)p?FJZHQLKWssmb1T}
zLmPO13B4tlGHSzE&wK9i@-9>WrTy}u;&*KCy6iMDe%9dl5l>B{ukq|p2$C1FSNt;p
z9>{tT9sMFd(y8I<wMMkHTd1ZI+<P%HR@CQ#aC3BLg`?2o4O&d67zYI2g7F-afF>i8
z5j5>unre|FCKuP=|7^cIulYG96Ljsch<dl*iA%RT2T`r8<XE*9A=|KPcG1O+X*c5|
zJ>Fg_JvHpF=oBlUZT?xlF(`DT=-&A_b}1qM=GTwktniRq>C<~zx##a%tBr3dVx=%N
zSf5xrMyi|#B|KqOv}aS=n`C5UgeFUDzZSY~TJ;q5u(kn7IOLB8kr@#v_t+D?=gkHk
z%CGAT)VeA)%8Gjaa{Jj`HW5cg6N!)LMfag$%i5<G*SlQ#dx5pjdjLL)irM%gRJJj)
zn<8gn$KmEf#dME*Hl1@ln`yrooP5zK#^+(k0=){dna+V|@#{Sx1?<MC|K%epW+M6J
z5T?8JT6}6{Sd1aqSvzQe=-vE#?w4j5$Zekxzx*iKAEh=^!-LBfk6z*%A3RVopWa)C
zX+G6r`|=`H<x^*cgBQKFBFJn(Oa@LaOomBU<?~Q|bobj0y$gfFYZA{7*<Yu5=m_c8
zhwSexU?7@}*Vzn*y<Yvir%MxR%{RUAb_kRqLb@YIktk74AyD^_;K0M}(QcgjwNiU&
zv=f)38=LSertX*VEdf-Ml_gkBn^ul@5XK_6<gP7<3tRG`k2`a)q2FNL%^sm@(9;;=
z=U=DL5WwCPe)=dT`J`3`iHccgt-t@UW9dEtbWP*#g>-Tkhih@M;u_imG5AF1Pm*~s
zu=lfmtUcNNkS9V-t8f<CBEs%VkW8ojXfln{<_nt!{ut_2*YoDJkVdhW*x=dh*BaIz
zUj`}aE~2%*k@dh-CdrX|fr+mjngB|w|3WZ2LKEN{Gyi${I!GiMy5??e6Jbu9kR$8~
zCzbgz*8F@GU)X3Ta{T;w-5AwQTZ@Wtlx9l9)9t9;g(N5b2!RD%bHcx`QY#3u+eOO+
zQGJm+Phj(AVNGS7u(d=s<|^f)103_UdH#0iWWjigS->Ko1qeTB;W1Bo@2ckfjZyxk
z;@tP5@xS4>-*-E;xi7l!>x$+BVw?gmeMc6sU_BB1pZ{Wmm1!2}dtAxp?9`i={InJT
zGI+O3#Qg#tS?jO(KGC^lz$$}|R#Y0IjKZJ-6I>ky`R(7|%ghu@B0h$-595{{iI3+G
zicrnc3d=ECcwA}pv*}tJYDY4WNm>y+$dqOZ*T8O2!(v}7>NE8qjQE)w$EN}Iv)hd$
zfl~7?{YbiEo^SDrtSr*cT5ho4T0SXocV(AeAAC8NgJKK@R5qQ8`HcD13EGJ}<_kk^
zeG7!}py$5|cezo31t{XTcSR}P^b9(#Y1<yDVjL92K>pO^Sc&NT94z?=?qFEc(x=3v
zl({d(;I?$nzGiHM$d1@=0(P{LIxru&56T|#l);uJpc)*0!Qq<&ktU$=Wa+l#wX?rc
zA<z|RkxlCH_ynp;>+QA|OfAYdm&l`iqU)%ZMF+-~-oS);algo~&pFf4VYJ~)Hv|<4
zW=%v*P7M6Cjb0445hz;QlnN{}M<3F~hshL;nVSZM#q1HvOhDokp1)t&J<2juO>$6W
zB&S{YIz$}B(rP}t1iW_FK)7mzK)N0boI&5_>ctxmE?-@^EO%eNad>wP;R*oW0A`|p
zJ$#|`R4%_qWY!?nB})uSec`b}Hg?1m!9G)p&E}<XTup8{wwn1JEpub6*my?MQpRXt
z(@9W&N`}(gOF!#tFh#pqS=TpDI}=4(y4_WN@nL)A1t-<B!P$MMG2$_vlS~l%qcX2p
z_7mzKr1g;`iYz@VjZAfDrC1Uop3ReIq=?1%$kqvh;Nt?ikgwA9dFpHM9O3RxZSF{2
zK{wubwnMha@MTh#&@oi&fMx5X7f5WVb<#7WaIq_4f7#aa-oWg+q2|$#pecURLX@PY
zrNp=KeF>bj0^lT!h%s9gJ``OK(J`u>0RX7aDPWJ{dfJBsDIywodbF?KO0UWWe97*0
zl@fQ0<7XyV_d$j{f05@AfVBbt>Jx|WPv}r2tH53P*$RI>i^)Q{uHlU>FU@Y*_ow>y
zY1)SEoLQPGkTR;Vdq$#}>IuB%+ng9j8LaV4{u5mKd`WuD7xZ$yo@S+ZCC3R3o`-i3
z%<8Axh|@#Ypu>?q1vaQJbq<lvr9sLgr_M?J_PbOnJ0k`AWr=A=Y+Ge0Q?L9MG2@@i
zEpSt2q!;7ReD_4$2{q(dzd-~dJTNUr36HH?4qtWgW%tSP75S9UokrMo9(f0>eC?a?
zN_M!@P7YOVf1Ds8&uTnGZ6r#OmXuPdDoQDu?osmb`Z;gbYJXNOYQs_zu3HN2oA0C$
z=%TMv67K9PiKZNespXHk=+K!s$#D`knGuj)wiB3r2pfb*d6LGjPWA+K!y|x!<qm}4
zl0ft%=mALxOE(jmL>l4o2YRpW-oQ;s%ysNeuTNqy-jQ-9$(5$Iqi3EdNR;D;V}t8l
ze6Ms><+mVF9mEVqJpu31N34R))e57xWApQ7pNcZ|q78O%k(Q{yudgec7S4+T;y?44
zl-4InK$S1w{SDw%NOHb8?A-ZYga=i&IwQy>sr0gUy(g!g<##t^HZm-iMJ|i*<654G
zWU_yZhDLr})D$eS%}N?F=ke_kaR|G$C?@>n={u=Rh}<^yumGv2SGJ5U82MsI&yjG-
zUQIx6pD9)aPNl6|Z3C%#5mF!DrNys9t-&fR&C$+DTEY%z(Nq__;tUw_*iZ$RpM)IJ
zDz_0N)w;|Pvu54vVcSJb7OlqZ$^?~*&4BqgPz5~aFzz5jAIPaqJa$8X)(Hp>^sk0y
zTwv$Hf*;xJnKvef=50NXSa}69`bvk~aI}0$7HInb&5olhhn~=qI%}J|fNcD0?1I?_
zA3Y8Y{oIW2%R*3*@zZa=9&7O6aHc=CS3!r^cXf0|H1KQJq=IzW)QL5(SI&+=)7hM0
z_OrI-3w{;~GBhLcb4x}M%(Hw1-GScbYz8A(XW}hu8UCPPfQ`5Byu9$}SFZRuwEhX{
zLgSYtB83fFEB6y9uk<P?2j@dEN~cpxzc2GaC(n`w#F3|+q#+ZPC0{r-T22(ffOn~s
zxG43ZfDC~ZW!~UquISd8r~JhvLPMGREs4y~u*o@{E6nu>&m3WU)jD)(sRYfjZ)hvd
zKBC;gaQ?*&C%^LmmT}E$i#HAAOZ9hPLTs`PXLD8hZTTq<DWCGOFee6{O>h|SrA<F)
z32^$!TrPZuj`XA=k)0+EKAUOVR4I(!P$hPA(fz<$5}ZJ>b&8D5a^WYglbxJcRw*ox
zIL-iquwJ5WA7_M<zn~A`Fx%q#?0*A`DWkLEm?mhyJj8CVO;pIioKXtQx@1WRs`=!%
zEmv3cD!OpgYhd0x2vkpq8Id@m+;sJ*lTOav`;ZFZM04I^=5A<WbVswVcR=-1Upky8
z-ADOz=cK^alaz@=k5=Kz`phx)A~v~`e6432Kq9P>JgD-<aVcAuFnr%%)%QD+U?qoy
z>8rR>UzDB4^|yO>lvkYkL({z*XvMU6o;u4GC8D?4b<PRTWO>dJu$<bfjcQ1O>&p_j
z;Z~$Ut((*LwIAQj7srWVel=x=xu)@!&C9cRHS6&0@nG2?m+l*Xy*FXi;FD6VK=G$I
zfedYRiJC{yOb2CeyBJEuA?45}Rx+#~REaNZ7z(M2&<P)zQWI*zfxTTTvet9M68JzX
zP#T^2A?sFqi|55>9~w)EP<6tJit%wC-o>Fp{LFftWKV(kp*>Ci?zu#{8HGWJ$Spld
zz47)TThK7@{#GH5qDcWDo_i?Jd9qu`10Vz$wDrBkbG@V4`huhH;67c;JydIg$JfJE
z0nGmaZbYCr`vcPz28$*5lvi2!4Cn`Sz1LV^b|D6+dgLtk(@TK<Y)3sNnc@BQYnRz%
z*Qv=S@YdUJd<8fQMTN{w?C!arnEyMx!NEsxqWrX{wFAV!gV9*9#QerP_c}<@;!6Pd
zJwL4$JiIy*Z0D6Pb>hMh_&@ep39-o|L;#B2e;Zp3ie7}ELQX<NvHv~D1(f5HR27h4
za0%H?_J2&03pzw~a2MY4A434;z&oNFqUDnRM0~tZ*N)TgQHuqS*9SZNd8dCZZoZ9_
zAp7@ei(^3af@mRw{{ia$M^t=x5787XZL;%aZ3f5@{fVG~JEemEfhPfZHlRN|F*M6<
z-9=jk;3pe2_~w&|>y>tU|5j?Ye!Y(j{6r=D_gz+~?77}3h#v|+`S8D0(%r8K&o?If
zC)@GANenJ^o3*Lpo(>d_9+3YdOTq>nnUr#)wUhYwyu}}>mH(M;T?P$;&me;HPZH!3
zt@Kcp)^$NOuKAF0v7|cq!1~;#M^UWGZjYZ3o+BebGTev8@PJVhobjqBpH4mK^XW+{
zM0(Hly@LcGAGc;qi^Z0Eq>zJ}1U?$w4o|jl2KN2uM<C-A6fB}8z)?)10JM7TsV|M}
zZ7#wR&8~<Lu_SDS>KW|PpWUz_by<;*3wch6;6<s|#zH>Lasp?Fr?-2Ol*?ct?}Cll
z(CmZ(s6Cyzzc(KR@&eXb0j~d-SrTXih8SJAA{^M@RE$s#AUoDKK#sVV%CW0f!n<P;
zFV9{MSp03{({m-){RTyjn@LsR+w5ML59vRq1fYxWDN;LyDE_&0rBZms0)GP91X5eC
zL6tz~d#wKONOPF_JMsmsf6U@{L?p37dULB};JsK#6?$lwoBO#|bM8$|<d|3xQhM5t
z`yT^+7ApZau95>7iISd+i{CGcsQww%nJa+t&+JHUmHGb_I~EM~&(htJ=1kDBYzhC-
z5oO{JStMZNystBE7ie!t8%!5Yqk+-{*eejT5KnVRb`}<RT|)Dw<1J$0{52XjRP37!
zPZE0R57tLTrSso=pCmO0sbm?qPmr>*!rRY=H{a|Vh-8!AhYpOWS_j@}m*|v0#5A)d
z-@kiyvMrO5oCrlp+Gf5uV#dv>poF&sp%f*s3bD`6<qs=3LoV91x2H~o;=V6~kiTfV
zd-m1IXrW^RbH7Yy=8LxTHA*zH#hTu|ef!K|i4$x`{Nga52})^(hqlA8P+skZG{r1E
z#O60}cP|KSOJ|Ywpp2e{?fMrX%cflsuM;Jp2lfnH5OH{9RP2Ix&m$clEmjm{Pu-oU
zJ0;{AJXxY`_aOtyZ{PKU{bD+k5A`}M#<)0h1T1zevDNxyJZKObtjN}7Qs#JH*W!V=
z)T5F|4;^IB_L3P&^gBXoH9mn>G<A2Wp{YEb=jgad^eb+CS0*AO?zdc15&TPRP;3cJ
zqVNPEPU^K1yUtv(K)-mfr<)d6e0X70RB=ThC&W?0yGGj<qz5CW{8bxkRkO_hJ{Sk1
zKtD-}OJ2h{2z4bZD=3){7rx-<z;t|iM5fFm-Zq!y%J270@6dba4!H{;ZKv5>6@lU{
zg|)8Toz{L^Lv0_=ni|jCcEJR--^xN9P1#eu;O4A#;-?(w3O>6$W}E0;qqoaSdH2C(
zNxt*b_Ov8<l0dYX1I9?Uz!F)KTydLikB@PlUbN=NM^lkF;0&>iAkfq)zt|X9v>ZNn
zxG)X^8wVIrcSY41gV?KNBGgh$6;haxq&_T5jLtYYENM(2W^t<-2sV}!g$6bkH+LZs
z#I=bJhllXH&4t38Yxqp*Ql){Ao%cPt+CEP*sECyO^N;Jx*KBuR8*6Ein2T(zP4PVY
zBr!fQO;W@|6FgzLqG*13ZBiaIqY*tj1>`!eOH&qesGK=e)6q*6)KuXu1E@eQXBu6I
zeGF0a!Ym<*Q<r{E-f94NsKQ*@PTa`Ga}@cF?<km%ElZO)5VI$<m=#^=t0!#IY!Va`
zDQ)^qyRxuog%MLn%n_H#w0jTCreA2HAVa^Gk15iwVib5=(aLEzvap7`nwZ*E5<pZe
zTCPgJ(p>js`Z9xxmX^Y6HAd-C>QAe<ae2O|=j}|#4=}36OnTTBJbN79gj8NuK2tW{
zjLL^4tWRxBhUVURjbNA{{VeQ!dI!Ea4M!=}C{S>7y&Bda*~ori#`~%pyT#COWl-}Y
z$RklJ{%ec4+GN_7m3RH;DGwoLhE99u5fiB{pes-6T{BfMfcQ@iS3PtJ#(yZDHQvc(
zR6q8zXjM8;X%HK+MYGGu5&S!&ajS1H?5=!5{c~pjJzXN#mVC+9Uug~Y-1J_2WWxIl
zotg)iFjrk7_(S+1p)N8B#WUX9V<PV)td6NuT>ZMpI>#&yB;mEmNmH20<;9CHFAI}F
zqEptfpg45e+FNOsT((){o^fy0-_um<;V%xWbHb`F?bWj>p0wsKFTMAxT%4&(&W|@9
zZq}<}HkPe^zCYIZAe~XgLR0_{0q^reFp))NT9Jc{`KgI7iuFNy6N$Kh$o(%tf#t2{
zK(jG=uJ-em)E*sFb$9c4mTy^#b#|-PLpfT+$f~)&8d<j%)H}QTI*~uks%!JG>i*~5
z{B3YifSF-jNb$R`M&EnFCT%?g&?w~wQ9Bz0os=+9VFsy@5yyqiFz?ot+|c-Hn}Wek
z$74Fhnu*x=aU#5j&K!7Wkid=OimZ>|B!MMY33<z&6$1s*SE`&rg>K6(6)deOIfTDL
zTqjMPe)#GG$jI(^ChB0M&GR^d#V|ODDh2`e8>G%QHxzDz+QMs3-aU*r@69bCr_ijc
zxP0Rs!UKVPip6_El^CMnU8zE1n8ClKMN@wE>uFNEC-C9{d@-)xDL_NewKBZz>l>I^
z>=IXCdOfHAw}@bR!oy12)FsN=1sg`>=e~Z3lcfnhrztDdiN2W7mA6HD+#4hr{{Ct5
zA)1+?xYsP9dsbqN>@F2%nv#W(9e-Wz$NR+PvRB)-QWvSY4PHadSL|!qqn|tP7HTNY
zs)(?n3{mTByh;#&HXJ`uQ`r?6ZgNigHVjQA$D#_^d@=62*xzcDXlRnYSXN)17{<PK
z`i#JO+5Qu|D68?pO(!5qCl{v<(k$x)K5}~si8}g=72==NE3?k`W%MvrQS+0-n~Ln!
z^cPkHY4Bq><kvLl=53t|bGkRj*}Gg;lgztqf6@kuj1=rP8IMK`9T<qno>DfWz1Ye*
zwSVIlj~PDHaW)`CN1YCGpbgWUj55>@f5#3zBTr%FBJ(CJ^8|^oD;1M{sD8IS&1~6O
z7iFvIct>w^Tu?N#abT4--ZKgp;hr<mxuJ7svnK<*ZZ9e^_0Ie4oM9cVqrmgt{EN2a
zh~wnd=`<H%VHY#c)#IALn<pj-em(l&`gct~@6k!;`)f>A@1s{1KY;{sf>zM3kX=i|
zsL8GZ@b;rG5X#*eaB0Q@IimI+3~s|?enqUCg|oLA8ooZ`(Q1$o;Gg@6y!@3i4=2)e
z8Fifr<7&;SY<y#~#22ePNi$;~1%Xx_HDDOj7nnJFkhCEEGqg8l4NF*(tAThD53`RZ
zQt;cc?I&9uwjuX-7E;AZJk}xRv3LA@PG-C%OCf>wc*wai@#DB9HwM}mjulY3yPG|~
z@x>`c`%e=~k_F*cg@wOFP)5w%dq7+%8ww;kl9XG^n_e@H0}B&dNqF@<IW|YJ*S<^q
zYpme)P{Fb|3066jr0&WRjfjClTQJ34z4;6vB=0q>`pb74;m?0)$gfc!OVvr`;skMM
z*a68cwy0j~vYh}JP!jf%fMfuqTfnq8#JPz&f~u?Oiis>F(B;V~69ZPJtdk|cp1%2%
z!-!sP?w2x%U6gA^VDMae<BlL#P)KjCd(CR-i51Q#^_FKEyNq0?I=0CJh0t=T=F-+L
zf)~yv7<QD-EfC|yUwkk5D!(;dzDNmT-4CPt{@7xMwV?h`8mK0?Morp(O+d>P-{HPE
zkwgRid7J`DF4$L%n@n?0b^mx@yE&4!`$I|_j}=Av{;69?!?!VR&MY!e5$evih}s*e
zU}>%m&K-6#i9)6yp=s}EB%vldO*EWPz2scvpgDF8XEO0rk*OG)1}bPhx@<hWz-b35
zI7{Dh@T67E0M8A6ot!a-1`}vwhW3afn@aA3jd?L!s}fsfztnBI-_cvwE|8y1X2VEZ
zA7q#nEddpcvrrR{Ove&@<f(=0OI}T8va}N5yu;32TtmL_(VBku#Yh>Ml1q^7MFoN;
zCM%g(HV1grFOl_jcDe5mj@hoGPEVMi00Q^f{b*i%hMsdDXEW9-`~fcA$I<SRt!!d)
zH|am!iN(3?>~x=y@xF|xltD3iREWJWZsdHkO#AhI!sFbl`w5xsHa5(f3gdJtlQE<v
z^dLaaK(RHoQ3|9C-biRZ=~+u+T_B6?96D|USvxL$mZ=Wo*_kifNotmTjdx#UEvBLO
zCaf~7#4Nv<HVh51O`oUNuqzjIW%WwA^2Ua3il1dta?4&Vnrvfrt)8~9Dhqgvd{44|
zU)yxyz0Tr5iqK(`@O79Z?=q?Xl<mh>UD}YGX}Vb~$X;-L)fHV}c)F77*UuSRMGrLs
zBF#Sktu=_b)f#-EW6_rV)VkEoq_KuYH%3<d`U&VK&66Lx7oZW2%>|S%khQ6KiXuQ;
z>ujOAJ-~1fR(9F7DT`Lx_JvF}`!$*pgEj{2)%gqdTMz+A7=TJXU$mU6AXI+0ha^x1
zw7-f{ffshn^_`}mg9(!khbEao5`U~CTkn<Ja49=D@IDU|vjVt+=r`T^_GQJlt3Pl8
z3Dh;%DEs5F7Ore|7$}+ukqjWQAIm7nUk?dLntx!L=~jLI=Oy1A^arc@U$6HkAOSA<
z^9POJ0c!KN1uFui5R-sGm3ek|KD~I}0_#5jpt&ln;|Is(PA0*53+aO6KQ~0K-W)`d
z`LH;j%w6!JdHoTk0Q8sbh4_pnQ`mu6HjuUWH@uXqFJ#vDnCTC6HiMGn6|yOj6Rutg
zRPFx8Yyj~1V}@wRWD7l^nuwoRus>t_8RB5M>-vbUwC*2B_f=tOx*>Bf7)5dEk2U&(
zMiehs`3LNbb`1J1F1E=Zw`qQVaw<xN3)PB7JK_z&8t(Ui7dtNLEy*wm0ky>BDCcR=
zzgXw92gH)mW$_p&$Q&T$b6aNJ?*s1*e)iy(I4rAc8S;-f;gA;5e3vv=X^-=dYjlVK
z+3izHX}aE%_dwLYOSK_r@ZN)qu{XuXrdK6A4Q6H1e@G2FL6GwJfBOC}ly&Ez9{<`*
zyP&%!SN!JqhZ-ROV|@oejVGMQEe`#!FB74Tf4CFJSobHe$$8(M<NS;0wG6O||5cZM
zjdFb6mcAC=zxb-}{*>MynF228jHZuWtVVl`<C5bakRPZA3Lq>!TrRpg<u3;6>i?_Y
z-2!~B!2PxLYQDoFSLnAEz@J(AAmsQ@J&7im{X(oG-}&k@CNEkg;~y(Yiuim7dgg&E
z{co|OuV3en=jZ-#^?O%w0}5>7h(1XByEcBA6^hq^$<shLw?8TaXMkefrg3<o4SGru
zRgBPV0W`Qi%Y%ux3N`T_Yy#+iSLI8A2Ycg&k0fR;VOcnvOlS}k&qqM&?IXAz_{fK1
z!O+(K&JhCj`J`ac2Y@0&3T!-q2^2Mc!ThIt5K2;P1GVDP;s@#5jZcI?1=JmfqzG5i
z5!4mcq^Go{iG&C&<ZTMk7YBCDSsZ}y^_K!lK7pLa-&QrjgSHcESjKND^|*frk%OWu
zdFipz)uGmOWiA<ujS2%c>c4E0As9Xt;<oM47f-vRVjzHg!N2Cq0Zn9OU$f3xc)?Eo
z+qf&3J$RCFOw;MZd*Ht_0MVoY$-oW}*L@(-u=`&Fd7(e%EJlxxx=u5*{<+Er{c(Dw
zxp_K#*^N1<@q0Z0B}9bGVt@i}NASATC09)QXCTAy_e2%}!HP5%3?ErxP%XeC=+4op
z{CSGX+rF^kU)3OXcR_9Xy^0t}M<!(>wq{~DSY<nzs^xAnt$&eeR>p6gOc)kSh>385
z_uQ1I85d#&Vpfj4Gk(S#Il@+3om^05{=WDV@*~8;gH=RWfE1X}t=b02>#^u(#KV;I
zCd5m&o84D%hOA}s(Cp|_&^xLiv<Ofe3)pXD;Z<%Nu5Y&WkExh{JzH1_i0cknWtQ3-
z-8e*&#XUPWIPr<uW937rGRXK9<;QA{)aqi|dx&!)V;UWr2xC@(`Ul@xnN-bfjf5MR
zZZ`#xB%5M}s+z<KY83A)ONVL)vB+7Ol03D1T8MXAVPJPJ15*V^Tn%<s7e?HL@+sPR
zs5LC(!{UFqvlJwG@hCOunEWjWBH!$tQnH)>6xB0;AM5S7Cv|#Y$v;GKGZN=*Ur%oW
z&=Lv}zv~R<M`Hzz3lla85i6C7cUUb?lWOJe$TtwiyY1A%sh=~hl;FS`^l=degPr{P
z=X%&kM+SpyEUuWcoUF|Khg~ZbXFsCNL=`x)jY<Amz<L-PlyWw*5Y_WcDWMNKzW|<s
zO9|}OJRR&cJMZOCIm>#Q%FQR~UTglip3%sN<oeLMHWsu$u~K1R*jwLUaSxGN)&ICS
zz|&iO11-)78|tBG=ceXsg#Nthxe~Sz4~L#w*1do8CsfQN$fK6LuY$S`BoNjkauX;z
zd|i@W7EvOoc)7rv#k(MtU2R?|$V`5(1i6L+oMoydf7=?)__{S<BjeRWY<8n#ub))t
zkVQqL@7cT$n`(6t*OlWvgw@vjRHL7f=VtPtauC0PRFvf)U7^_nrNnX2+yq4AEcI-$
zf&m$1(7hF4IbNFi5Ncy=1$(=X_~iFOH@++@+&YE2xl~_1pS#@ETg3tc1NPHeg9=Iz
zOME^50M_&pw9@1hN+AI(&BIuDe>8i$5;002hk+s1;5Cu)!whY!-qZn&@(Kf<HvbLc
zBDM-N*NrHLg!@V0!-q3+-={dGs}n{Hre^EWU|WykE~5POs3NMvd{$oGKO4ZxxUk|M
zz&#UUo{JPR;dq!dSwF1JNHJZZ#5yg+4pil0z#<wl$Kq|S09@Zo5TiEk-GYOp*1@?>
zK~n>Y1%5`Dzb1j_7_^|_fXb~HM!aTu-+CL|>)>Pd!!RVd{;SF$&B^E{;4x4!<J381
zqW_KOxhyp+pPAfnak66OXHrCcvN~vRfiTOw&2!@+;}7P;(9L;OrZ27T>C?)`ndb37
znVvG|ZCHnDb}kGx>$l?iPD&S&XYAt_EN>2!3i>s{KFgJaVJSvETwP2f)^%w#Na87R
zOSe|wE;opGDR%`*4{`Ajs#(ue^9?+O7DvphZ(2J!+8ht}?WoVHw@HiEmqCGwJK6%M
zx+DLgRyeYni)8Pfrh!hHq;!mRJL+o`=|~E+nzwg~3bU<NAs5tS#)b^I+f?(_7dm`@
z;>a$)Kb8pNu>GuexJg{WI1{jO$V}DL?jN>!xRO81^!&mZmx0os?ezrA#9nt;Q+scl
z)p9f~&{74~JR_}Dm9=V}a<{qETwb)iT?;7e<b-o~bea4B%Gs9|4>>q%nOr~76}c*~
zRj2%N*naZP^hq8g9=#8%$<#H75!#@(rdWX4KH_Jv(R(y(2J+-T-|jhGHP^qd99I!X
z_y9WOvOOv%Ggwi`7QBJ8wP}TSJoHAAw1FWSw62F@gpX7WPlJ8a!QG8DXk=ZCt9wJ}
zoXhrD+GZ}2jAJl)vUQmFuU(<|*RB9wRy+d&B;<2m^UPSxWH%Rrb)VBSJ|K`~TWCB`
zdy%V{EH?-df631!Gt?cgTv5lB5|;uZ%VT}Xy>@t3F9T>4TD1MH_J4I;|FXC_|7<Pf
zZOc&UGpbA#k`SC~h^F#h9b2ELWaBQ>C_se~MoF?+uxxM!xt^{;ojO^jc26P8-{E8C
z>~}KZO|2-dd`MkkF5jzr^#Ro4LSbzk+DEPPB0JQL6pjt$8!r?+P6Fz;*`jl8;121d
z9=bykFV7mI?PDu2%qnGMo)VS*xUW61KBy<iqip}R4liIkn}E+EFR9?iW}*66=8l@d
zmph%X*)HWDnp82xKMQNh>^@o${ZLbh<;caBdv4SMIW|^R6&b>eCw*o~(#|D6Dzd3p
zSg=V=R>Xq}hB#7obQ+J=gtuUttD9`S1WKMRGjjuuYKSl)@sh*l{5ZKb#36f?CC}{t
z${F+-(&yo`PONd$rP`}87jO(_F?dl<Bo8g$1hn~c+JsI#E*jcnmtJK)u=w=R_l%u}
zQyVDt(j*S6HAPEMnfbi@CP2mN)41RCF<NV+<->9(s>DW3YPs?(apTUhiE+HZ^|7S4
z_dP5P<$V7i>fSOgs`u{}o*|?pq(eZuK}qS6?(Xgm>4pIj1qtbfp&OJ^T2W%?MnFJ1
zrKRI+^cVWS&*ypWb3XT*^OC)1Z?1jC%wFqSYkhB=K=}M?-O(@f(4(`15x5=ktU=5T
zUOJBLYZs6uFA6d`8et?Dw!cyKti_$;G+=Wg3?0p*A%V;5Oh+kSd*UgXS&jYTuv|G^
z02`pRPJ9~sJ$HdFo7X3*{UNe7x||5A#|kcarn2+`M~nQcmOk8<APsgGT=n%TS?I_I
zUhLL5b7oP!!z=~&(r-ueWehPr@l`kDQj`5z2n7D^j6r!?L$wvT3WAb8LAIK0LuND#
z<K?KtqV|YK{nlF69`o~FDsAO)gzr<nraBsuA&7ip6f#|%?rPhsn2(^pM3m-I+zxp{
zOBX+R&*p4<$&BHVa1APq8l)($0)H%U5+8+O*ueZiVP@Uo=!!G*hUuHEkpYk=@hH98
zT{gATJzUAWu}6k^m*AsSLwR_GR&H*GoDN28d0S;bp+XPx!iHR~v3}&&-uSBP@l`KU
zH``vy1AD5znbYX7@>i}M+6t%6R?CIHC~FrI=O+~0m)vWV!3YTtVyG2jI4qO%>NV3*
zdgk_=6g+Tb(~<1}Ydstbwu<g@cJf&8V|g!$ug%3F0Q<$<bDM2UY7!>De5L38cvyW`
z+_s28wma<Q=Veurz*c#q5k+kqUrZRBI=^=k39@0Mj$oe^M|*XmWnR$!;4%%V7(<6u
zTWF)kzRIvSr?FTIt5>`~!u9&tt=W#4zjVU9#GZdVZV@mO96FPILePk=rBu2zmk5`U
zpDxcioJy3z^^;degP<)WJRV0<5e7E?`w7eruwbcOCd<3z=6G37u(BzR{UY}LSDBqK
zZPbm*p>J`B32t%F#JptcWa&a~C|xk~b87Q9>j*9MljHJn)kw8fi35Y$CjG}ZjpO<)
zuPL>WCk(!ppB2(K(`OvfH}9uc?7JN<Bj=o6AORa{<-g?LXqZEmBBOB+TmU1)eH(qw
z?+4$RqO=?8_PBlN`tmWtXdyoykKa|{`KN_&L$?M-3r<XGL(_h{gPY!va$8k*>MP3#
z>1ouwve%~}%8!Zz&o(vDR0KR^d55cvRK<Z;$Yo|Fk+ZCAZBW|rr|ZV?2`j32JJ6VA
zN4)DC9mL}#4pa-?&dmsNV@xLXsCo<R10FTTW)Fc9x!70a@+;hD6h#Va7s$>>Q;gt3
zQ5(%VYDW>AK1Xk=Y3IJ6lv-9J6=!%Rmf$3|fm-yw*l{YD=SFSV<HF=X9U$LScC}n_
z2{Vza0+x-;Kt)8G48-CI8b2N-Yo1}%Q~ZhMi8d!b9p(s9d+r6p)55HivKuGw>WGc>
z3SlVobyO2_Pn5u<4$xnX^VrZH7utG4BX(h9UJl_w5$pu4;;I?CnJY87qV=lCUa193
zdVOeh-JZSaWV?28H=KSOr#CnOoWWAwF_`r}U#LbMln{0PrfB*t;Z=dc8fX`>MU3>^
z-?ex;l*md?=W`PZGNj4_b*Md1zU&FDu(V<8)$|^l)wz{@1zJ_f0=B>unMdB+$5Xd^
z^~4<7fy*7cZ3UajhU0Hh&~0$wKa_}6{Ivrt7}~%qRjv`8rv+cRHhhqLH*nMMGm8ze
z==yz99yV-9Yi|U!cW}*aGg9?>SepMU2fTXjmkFvh+4vq0Hzmwns*xFY5n2$b=k6N0
zq9-m4`(1-;t-esXT5{G9`qE1tD-j)HS-ztrZ)`JgR4$rjd9!1JQz9{rw^38t`>uN~
z$F$aFcTB}ZiVAczYB*l3t!Ji8_(<e(N*p?IS8&;^caidGa>Z_6W>zZEzyz1(4BS|L
z$B9%uoce^7=WO-c*h4+{r*`xYX#fp(+TH{5tOa|YssmIZLlJ~p!<8aV>zyOtUP+5V
zsGeOD-iAE>c(;0tvq#3}E#v^q2c6CjAm2ME+AmWL&$7)(sUdD?n+V)x-3<w8zuXkD
zYoLO|b?v~qyq^O4U(gVc{ik=iTl4)OiT)enonfg=VE6{{5s(7nl3%C~IHWX$EgX3W
zAp?TVej%_mK#tKTj!i<r5V9nv^bZc>hjg7ij0ISgUmyDw>h=rV{qoBHy5mn>z+Y(f
zXPDeSca^LI64ide*gsFST+Ig!sXB)JgK`1#;5>qW@_ejHpmfLth*^VT{pyN??4u7|
zk9dFxykDRW=ungem%f&qFwjsmcc=O9?vna0WeW%nfSCgk6ZjVH!7#^zpvQm5+yTe1
zs4G4gC)xv~+25WH0H*B^ZhYF-3M@z`#r$23AP8AHp|C*Q4?16zKZUF`{>CN%MhI*T
zC-L7134X}FDYfHow8sB!iXK1&<xc<4=kNVA%2e`hh-7pwQ~&n#<gmxszuRK|Yiu9)
z18D1Nte!yr&bavBrX})MfB&6_1CDqm6HST_QS_VnmRg1bTLJ<AZ##82k@CQMzoYHU
z9DV4bspQt;^Vg7iL9Kq=p<}$3yVczMnn2I*40(%s1`A;PHPKX`f?A@fZi%*E=aErV
zc+1_bS$><v&(VgkZ%Pwy$Lud+Bm{BY@5GFu7-+;lcb&O#0pwVJ9{^^)2L<CWu!Wx}
z^DnKy55Y)sHKY8O7I}`QV(@DDpPBgt0bhU5>F+gxPYZFOX!)n*$f!e!*_*{7Ed+nf
z&%7Z}<3Vd%2bfm-AA;nU#OK%J-|}ggI52CaETQ-vM--TEs^%stj3I_nbn1<Xd=dMt
z-sg60`JxO7_2lDqO^}F@#L>&bL7O1aJQ>+$g(f)<L={&Pq#R87de=>fTZRZ&TE)(w
zan;Tbk{JVil-Z)6-gv~RSOs49rJluGU?{>tz6bds?;~tR-lTi0WwQ39uCSDZ>vyMV
zS`_2*KP*i`Bh@TK9Vn@cy^z+Lgd0ZOYBW9~YP6Np^?Yvf$&J64vw8_XTf&EhW+iYQ
z9?}}*-4}L?lZG93(JhXkj<brt@sWCuPP<lPlNiAS!4TJKYU1wE-gu8>Lu&9Yr<hX`
zU=&<=a2aY!6Nf|81DKTRK<8CKheP5)7rlfm+{DkZh8#XKnLnwI8H&HWVCd{Ui)Y<2
zzZCLu2)rg<VGNfbfuI>HyMBhF)~zeSKR%r0Rqo|l9%GS|e3{AdJY1rOfo*4m(K$!~
zln@3ZJZWSz%j>~+MC*eO%Va2wqRdyTBQh!RfX)y7&V&jDTSx2)?&N#`r!|y!=+C8z
zFugvQ-Z$?;6N`EKJ_ZHW6;lI^Qi(@AB7Q;H-+}jVI&<M98a++uF5YB3E({sde`Qr3
zl&0vLaj)&1J?cPHz|-Fdp?1$m8M{*;BZ67ImhK?~U*7V}+ix`XgC*hk3@p&r8_S-)
z3nMe<)^sK0E1CXq7-0KA6TTJ-2nwtnrk?@!W9sum>o|0D$tk|qD`rmE-O4BQF?t&@
zo?{NN4&rLXTAikvcOgNHX)w>065_jJ`njG`FgzxJ;on=|Upl0IX`$Qw-8JsX>ZuDk
zw+}IW9geEQ9Rq76*K$&@uhedN)EeGN#8Py^<N+2D6(hOqA^|*JD5o&sF@a{Gjl7AH
zl9cR8AADPSqF$YbNr@K}Adb9BwrO#mHfW}%D!FfttpevpD$dh$={Znn`f4nAJS~K>
z{?`1G(sJFtYhCAu&=vumH$Et{D4+Ih;N?T7EXnMn*=EmCVlb&ZbhSiAl{ZL;ta%y4
zEM|C*b5Of8EOy~OEr=i!r@>FSpzu+2rRS#BZ7UA82L`Q0swXl@WnHR4>(}@xUYA#X
zC-r*|3-y)nL4bUR%sYwW)F5)l_l$Du2rxZndki9E3>Ctf6>oYoNJP0$y8<9K8U$W=
z;i6)F@rPBhL-A?}KO)u2)uM522C`+vR61*YqCens-xYRukdq#i#bdC20dCs;++T!6
zy&)U_q&Pkf26$m?8N~2CBNff#TJRhE3=7k<`96QvQkq5YgTr5DaP#)aurpo1Ls9f*
zTW)r1>l{o|HkSj*AwCY4z)*#v&rw6j$A@G;J2a?7#K;^%U-L{UQv^kIWtCUO*DkNc
zv`EctC1i!ZIdq+LIS5PQ%RW%gefhc6_AC;@P{TDb-k8s%9D&L!D3nbi5$CwGbXZL`
zT7Z@<m2@|o;miy4J+vEQbl`T~=W_9sxl~~|=|e_F<Y#Wo^Mc+R&($%R2z5FLMi%1q
zx{Crj#91LWxtzq`-0stEpaM?AwNLBXs@rn^l60a2?e`#pC~sLlw%Fs<XEe+G6R$`i
z<4Anj@{K0(dL&@q1vlD-Mh>;Ucwcr(lBkKDh$Szb4Ht*XCc8=|6{wtz1<WICe=X8v
z(q$x!!Qj#ujh3H2hLR?SWr2ZQFirdTHDOD87e4Ip&@QjhBE^U-H75~=)|}%g{K#%3
zzdt}N6gZ{%g3>l`UC$K#M6t*rD~P5K*!oqgvK-4)$751T1hp9L;S>g~h;FEESw~@j
z8n12V!0?bhAl5<Z36AWA=*huL%^^kk{Fa7~^W`_q2r4|mU#dJGq=0=Tjv~}%4)bEZ
zTRMA^tJbe<Ri@gb<H2l~OP<KL#s<>{uk25*8fPn;51HrB4p{NUYA2+0AwSt59(C#L
zk9U__A#wUh-My4c!coDbg7Ga+m5oj}x_(;1x1c{D19+c&-yk6wYuW;_vyY6m+(q0?
zBKV|`$k4cFZs%NRo2b6!%Gq-;8V*xZipqbi3D_m3cD<QeuU4OhWE)!wdfVvP6En?(
z65)_WiTFKgZAgy(Hgu01_ZwRqry5MkltpThf3{X)if=1sgf9;trWf)lFKXLEEfxXF
z{?<FbGqcIoQX1P|5b)MK8#*k1AU-RL&|wf?1nM__#}-~cz47H9-G)$<lzL#?ojsT5
zcSE3ED-bnLSX+n)f;OvL5>F;MdA@;lqBfIVU2T2jLDy{J?ZVdt-UAEp<8yg9P@*gO
zP}WhZP#-!fFfh%g$i2tkdiIrCZ{=Yh-N+nkc;z@9GRzJEEFaQ{@6y9~*{NeC(8nT)
zdz%-*vyzJGVT`LI@++M%d2_p-!SQtRi$$+_LP1r*+l_s5cSaE{w4TC8{n=w(PzT^E
zDta-ykKz0F`hy?mTU9rs?>XXk5=D1hCX3}fK?K5?lXXtYybp}=U>l)16)p%=1%7Tm
zRWI*2D-=9axC`qozn86?zxFueE-xLZKVRvbMX^o}%uEZo1}LA#`_FQ%rU^wsSS*Nz
zaD}6B9j#A4^a-Ky-t2#YA03EF&Fom~%&$FKseI7yGfgE><-u5e9lQj`-=y9`$V?FD
zJ(E$yG>@41+M&f@%n|Gjzkim0I2!nnpEG$$N6xW`ec2)b#YX$*FoP2$GjoRrVwvXm
z4C}%c@H8s?7K@Hrpg1uz#G&{p1yK%JA(DP)64UA2)e0!FKsu!qd`Lx)<&g36FRuz3
z>nKPMxmfKavVoohtT+3SzK8|xKK&ZA2&IV9=OU&nhN)agDtY_9-}h6OYprY3vczZ-
zt?d;PQ(-hirO+PA2w$Ru16ytTk55RO<-ny`Etlc$hOxw{ywg$(N}K{7RD5}NE&=C$
zT$})BI!I%~H<BFOLpqWON@Qh4u<;L$ITvzKElEU!?F#YAg}rb9m!{eu+=KCM-R0^H
zRS=Hmob$b+MdOWOh+VaJf69agiRrzeNEh|ojx;sWpx~eAeSp!1Qakg8*tP?7WgJ!~
zRahiJVz;St(OqjV1B#;Kn>`B$0Xd5)p-gH+QkORabUpR&G4I(FcCQ4TX4q7a$;qI^
zO4(E1w7&~%diU;LvEl2_?Sr(LIvAh9X-wgv_8W{s`e6z7NyAf$U90MWc02sb+!VnR
zyPx71xKj0d7?iK`Qcg6CBF-fACX5n@Qjal<(J0+qOM{x+N;x{<>QK8pK{ZZf+s`*^
z2GqhYw!@b7U%e8lLV8dCPgVqpo8mQi?09EvBr{exn&FUT(zr;`XUc>DZDdb<1A$=F
z7k#4foscDi8x3KSGyQ|9^F|TA@r>Ggs}^(JIcEJ9o|_QhwM>Ov%#~P}>*eRs7^wz{
zX)=DC%INkoG&RCxM2!h|+a55iv4Q>uw8^Q=h!?_~nX7Q0s%o@eNMBZLEiev;d!@*r
zAVCPbHvS7U!<I}Ze#QuMYu0yep7ADGaY3eCkBbN%jWIi05n?KcCTc9B8$f@2W60dh
z8j9}!+($n$_p&6#S(yo>i8tf~IyhgQ1DG$sN#d{C*wak`{O7_}(3roxJanYXy>HZG
z4+(~p+ZFxR^kg!wAY(%>huy3!V>>+rIi@X0SYO(FQXCOV?{fohHNzyR)6YNwo{m5a
zUY%26Tg`2dFlae~YwPTEK$V9%|FB$QnzV=|wv>J(dkxOpoU6c7fl$nvk#2v04XaDr
zbNR@o`1V|x;q2Ms7xk@m-$NgPZ%QxXqcc;9?iKqOT4>uHrOQl!9B~!35o!Y@TL6=}
z9nK=gWF0(B&QD*#b`sClIqB6z`zxgFA9@}X>ia-es$sMF2SS(w+vPN5qJ8F0u+|3Z
zU0s$0VXqD1t;a&2C{%(nM4ritsOYXKz59Oho(kfFC&Gv147Ge_(^!f&ADlZ!f{6-4
zRK<BFbNmEmDX=3E&Og{Bqg-tr9H)}d9tW;nddXF0)z-rVPMb}6aeJ9RP6@`(P1L(X
zi9sjSqbEC>b9-K>fEG7R)gx4{TN<i|60Fqr!IS>#huJlaTL$Y+%|%*S4i#D{=4=bS
z4)4pt-oi&7)VEu9F<?Cal^Ojj^k_i>q9@SnzYN~*tpHM4MGlqku-DK2P~nVi#rtLc
zxJ{if+gH2R9sYz=oQngV-ij#^1(3k?OX1KnuBL;*Y3aYVVgEZwJsTuh$NaxR>i>s`
z|9|e|uqWU#$~t$M*}u1H0o;UD=kf2~HW1MF*#AakJvutzyx3oquh6uH{0rfM_zB+&
zqM)p&e`^7V+RTFHZ#cHsghbBE>RBfI{`DZl-hi&oT3FiTZs^<M-Mg@-tMuD`Qz^_L
z9yKAQW*2fCFDdto|K^rB!y4jae6-8JG1tU*r8*!e_!9*Gqi)DTOn5iL^l%j9-hn<>
zC_VL?i~)GH_PEoBeIs`a4Sd;^bK<4Zk@d%E##e&cfpT$2*MMW;4dx$`4>G_ez&+<L
z&O2N6>NUP4$8TVt3|77{;45?@?^OMf_Wlbm+<%=N$RHKDUiXJgQ=!#_za2$T5eeV8
z27Ct6|8gDpsE~Me6^Hi8-&hJD27~vdL_>q9ZgA+cTWI~a@B-*(IlY}!b`^kk_$^5T
z$T`p5xE^}uK)7DCblQ&sfb^z+B_{v}Vx&#@G>%XxM#hDyDDNCRBt1R(B?gZ1vm^nK
zr|_h2`5yHJK)LYzLln6|nXygH#ATenEp6KsbAT$6%Wx8a<p5MC00!dpyk!$FrP3U+
zH4F#FMo2skU#|OE?EnaaxV%mVB7NH_#g!4%{2$!11;2mb5`}U2yT$YxrN~(cp*=S~
z6-1x)VQD<v{%2c|!73VDe;)c}l6i$J<-zIW_;0lZ$G0DfmFTOYZVM;i<3pTp7CVyl
z$_Ai-!0Y!jAmB_$3{}gG@fZKXn8jxj2G9Taz_KVLD3Bxjlqpcl33<RAh=r)_K*$qF
zZRWkLx)6H;x?&HCczNE``5qSzcsU3$A~!7#rhk@sXhl52JliX9S7fMs?nUtDb!#9+
z%A3?Xu=gpWfX?RktAJ@LI<WJpElwCl*+`U)vI&?zet)<@aL?jEVs;#Di?Pr=*36!g
z7_?r;M1vuI@@K^eV5AQigcFPtKh57NO#W!>JjfE$YhtOv*cC4*an$h7{|g2sjBp;7
zx&G8({n6#WPs-n08VuszK<J?N#|#D^O?{uq{@z!D2B{&iR{>h1{z|G-eB*NoVlY1s
zr>j6mR7OVrV}49(Ri_<gtp^Ja;+kL}xk^FT?;a}_tA^tp10ZIWPP|^$Oi-<GYigA1
z+Iy~BYG&-S41Wh<wiO-&ACppL-ripm`vEj=D_X9hX*MHkMSG{Fu<UqhLzZTj*{2;O
zH=>~IJU@WPpo<RDa~YMed@)2nA1QU__BBzR0m04ABtOF}UAZPWJ`bN^%Lmr{pR2+I
zbY9+kpmvGRPeLMDH365sR&KD<o&Fv_IwSiab?_58!>iaXu-rcXRkvr)Fx=KBC<UzN
zF>2G5ApDy5jF6<BWWYsQX%@Awq&3zxqC%wqT&Yqm<GwaE-5}0AQM_#99-m*u2}Vx<
zpN3<pySZI08<XjG3|`&EPkFP)CCR`({+R77o^(aT0AX)=WVMh86)`lTSFz<AdPZbq
z?iADNx@ZATl4o34OpFj`X8$LSV_QPQ5wB&*B-Xny*7J;{wmWTxUny5#*KZc8#xVIk
z&~4?h8<Y-n40HP$qVTJ2s=g85TPrpN8{}N7?-W0<IzWxRLdM3E1q+zjyhSOd)dIrf
zf7F(bX_ZyK01ti7aF4WucgR*ve76`~rhPJdyj;5SCfz?2(fxeNK{lfizu~CuZowlq
z@j`E2<aL^G?`LzUL!0SR{Dt=&xQbx9ej=O>;!sbgge<eBc*A+PW<>neXv)Zohi|F!
z`Oqn>y|zkFj4U~l=9QlH*xiFh^BUmO)nX5=wB-E|r2_&yEiJT}?4gY%WKgS5EN}|C
zyW9dU{)(cU4Vk<}bS4s@9#C2sjQVp%J{2Pxc?}h#<lDZ74Vhi3gK;6UOw`~m`hJmA
zgoM`2EDI2u%`<zxs4RN5dTZ0NHK#9GL{dH6tSe+6x=$8K^tF@><uouqg!kH5cm#Gg
z5EP~O2j(KqPQj=9B_rZL9YU(+%S3`Vd_UHt;6g{fDLfb!!-pX_8iI~G6UxU&7)y*(
z#Sm;nnioH$Xf-~JQo8RRKgwF3p_I!*1F?)9jlT~Y71+(9V+5ANdOoeC&ayY{%FAsb
z1iEkvRS7nm;In5BLEQ25%(7)vb+$ote7>}q-K7*zre|@r(dx-@np<<S-4+%;f%U!W
zm*xJNWoX71qi<YXeW<T4j0-O>QG@OZqsc^J=WBy{iaXNY^A+-5?sOr=Ze_3eIf+Qm
z2?c7vrADr`hhHXd-@ZVKJQkaCXekD3C_X)G$j8vKVKU`cR<^7AJ8aU_lIbkEG*P93
zJHBAJzMy3IoG7xZXiaSl8KSsAKz1ihtDY=VCx-50=C>43WCI!pN;o}}Gi7R^h29!@
z?}x7Eh3`bE^At#@l|-sDlbOs0q`6gLBQxf%zJBb}tS_14xi=oAn%JX4;CvB>Q;Lxm
zQRY$Eon$o@$x-Uba-|au@d+Qhk1z8#%3X;w48Gn6zC*5yzy9^}z+@hQ0Iehnc+J{v
zkhbWI-~77v#*L#W!z4LL#?S3po9>OzEn<Fa>{u56{;*iD_a`@D>q%}Jl=7C*{QFKV
zZd#&MHTOB!!A`KY;_Y*dgCYG31df?yGf_%+VMu<F=73G2;G;z#j_el&^$FFz6@=Ce
z+dJ~pE+5n$s*RX2P@1C3OGR@la+*)`^=y_=NF*gUAS%Yd-LVp$dNgehrBqI0%SGXv
zSz%EnH{(~1D;Y9Z=uIkxA4t`&b-A~^=Wlr5Amd_pWgS(c=nYO8UauiqO83*fIGa%C
zUe*t^%&(1Oa_Eo|vQAE7ys)}bJ-13g(2#%XT^rk?HNDyECvBg7Y70DBXuy1-zAlb$
zmL<(GS#&3*5QBn(Rs?+HW#Kj7=>9O)BzJNl7N}?4Bvd^)Ekmm_*_*CjS^aJ`qo$(N
zX>+DWEMY<(Ci9ctN>ydiLT+q?F<Bp)nM3pQ4_|?9H^_hNfubT{R3oSaq7ma74g`eS
zV267Zg+J$zg`2-DU}}IM#UPoZx1Ro0NI?+pyPy9;vf+iTZZpm-vpPm}(5HR4)5IYY
zq>YB4u2S^IKKIVuk^<R#ob>7$%f$rbuhKMhWY75f7YAk4`A6e7%jvzZeq`S?tk^0Y
zN~Ci3#GTV~lv?SdjiO9s>MMGH0Gs^i1pY(0lT{$N(&uMzB{3vfmqsL8X3*s^iGv&{
zO*ZpLvh=*)tvUF5>Xi#;E1jQ!Z7kZ;<T$x?EnhY@Ija~B^(Tttd<(WU9=@F4d&+Iv
zyMW+Cn(`b)($SSAHut**pPNQA{LvueE;rVi8Y5T;(+OD_+@`K^QGAL@5h@;P1EdcJ
z1ek5xyEO5HZJl@A3uitC+GLRyBzB_y)rSbh>G7}9U*Q*pfvzTR1O;Q6!0Q5Y7KG4+
zT2rVhPC9S34>}=y1prwi-JIkSBd-k1j*a^UL8sr^-OTC}Q0Xn-`|C{8Yu+=^s;H*V
zxISCz)lgA?_e^G{E~_<^51&n!;Uk$8GcNUr3suO_=0+3(?DlJ>5S#^9kdft8w$o7b
zR931EzL2hG0A&skT8XGsZa+%g(0wHL!y)>Nu)^Ca|7#e*JNFT~g}DV%{<Wv-dUchV
zxFnGd;Yo!xJ*H11RxZtb$*0FAlp^8*k^0S|-=L7hQMOV3`7U<A$U{7h4Ma;<1V(#3
zZ3GR^OhPbs>j!D~tcQxZzSeTKQClE_I#Z6L4-@xbddNH;0#V`Arh1V@MKfnXc?N{I
zr{P^bZ%Sb9A!|p+a!}6gHy?8>!s3k3LP}+YkqT6Tw5ZmP+^e_lO?)UiEu>;5SI#xy
zkkDV?rpH)8_een-jEd7g)vX(yc<tBp-rBxTFe=0CJyRz3Se-g`e5|8HN?{G$)Yb0D
z_RC{_(9snTbf<{{3rkI>Mt1ysvui;8JqmEV;5S@iXOl&Zuxoj=O_yvE2ba=cOTcU&
z!;2{27XGMlZTj{(+n(f$4nJ7?07jM-1ums#Ujh}VvXB=JYKK&csGSSy*7zUjTEbWs
zwz$PtJqNUSUEBq;8}uUUbHk5#dpfHIK#Hs8&f0gr_ZgDABD_~eCMepWEhB7~@MX)C
z#LA1mtkpB*EPOnLsBuYZ?FQY<6BPT6GBNu}nPlI2Z*f!aUPA)241hUgJE`3rQUH3i
zF+m*(){}9{y?5^BIMJlj!esr!LAzS>_^wmY$+x4LhLR&S?hgr|Ov0!ud_YrUI=IFc
zD;%@eaD5jx+Klo;_v4IRtt2}G15sNe??PgTj&oID)*+hR>|0gKh{8DQ96tuChk3d5
z{wT@;su|mWrs4CT07x;cDOoD@32jy&cMm{Nb}x{5;2SBvu6|qr_pvMx?nnp2s_%vf
zTjCql^^E1>+!k@jMBrXAbMn5NYMUTBld*^>i(n(ibrWGpjvBQxL8B@xN<59nYE|xX
zT@OdTH!}sj^c=QYdqk($Ae`D*I%vYCWf5Ry(NTvxdndNm)tl|?nYu!0%aI7c$LlPz
zf*!jBlMskJEIOV{jRK6~qXm0W;2s78plR-<eqw^Wd|c@=el&%;t-=K$4Mj*3xO*$z
zFHu+Rzf_)>l`E6{S?!`RAmt@!sWzvG15ssQ(1JK~7(!ik{!wznEmkoxB?})l)8g5W
z+9FM4#sC;&)hbQ=s9vzyh6Zkkmf?YF$2GIAIlDcrwk1J!Q6qLxkAMdQKt1G%6&+{$
zS3GVI#^x@&pctrFtp(Xs=SurLR3l~YS>v<~ujS)Avl$3;Cf1^Mj2capv6D<X=>b1z
z=jDPZvcz(;^ydSa$OzQQrP0})34IfAws|=BLOe*r2m12ro%=9{+sk$a%2=d=rOnZ1
zGbTul;TzfDxa>ivcTLMQP1A(Egl_zK>=aVYo|_7{dzl?d6*tEbi_6Cn=LB9tmo3WW
zI&TkdeD>P{;zSgdN6$ieK`Zt>zqlaYcXa@hx{_PBM-BFC;pK|i;B#ME-(OSTwgi*T
z=hG-(hUZj&S-I5nb^_coKqv^?n4Y)hQE+Ce<Ywm++JHAqo)Tg{Ue}(ySfW?m+=%P>
z<zGB4@AIT0!3c_vv~&#*Vl6$$Se(Z`4pI?3_5k+v9s*Sr8o(MNs0ks5YsfL?VzN)L
zsa;kmFsw5=xf=}iWtYQ3BlcB}9}h|E+)v{kF~8F0=W?>O*c{O(n3{N3@|QV)4>Cn8
zv_uyWO@2XzEvI&=$p36PPT2!r-sZljy+f9aD1SqXCrhoJSorB|UZ47gyww3yY>9XQ
z^-QSZFS7sutb3#)*KKh9aD*>nvBhZu_ihy`;T#Gplv=yPyX|~7&pl6?_5!aX*?!ga
z69Y3k_SJldcWgIa>0=>z(<7=pO#|gs@d7gb{();iGGpz=fYBSpFJXM)SI}9&EtDKb
z{RiX)u)fI4&!pSVFNqwiiWjtFRXY9)<nyN*>8bkMyc=Kp)p?gsDPKAfNM17wR?So}
zBKv7#af3$8KW9+%oN#Yj%0XH2OKABMbqk>sdnciS3(qn3AHTe14*rCk%wj6&3Lcsa
zO6IP(e`3mA(8||*_4P!Q8B?NSV3+=PD1VFfSD%@EUj>5W4SF^J^UUvey#dSNoZlS0
zR2Q1r!CEfL`}0~ckWHzv(BV*a>zGzR7~l-~?b5tApomhsKMws$&-@+P|6<{OZ_ZDN
z$RB|C2H1U>*IEnI#bO8nq96X~F+b?_*Q2J4{{n{rO~cZlE0A_NMg*;|d`lTm^b2?2
zht;=X#4rQp4TS@#5ksE-RAsD!@BtxvAjsVYSZTKf9RYYg&9#jK^Z1R>3PjWp3f`cZ
z50ancXeu=q{p}PWhPtk_yCp&XuJfo00(Bhws{eZ+7C`l)%r8KWe~N4-_s`&Lzv*KD
zsS;UOSN3cB-z`k{VgB@2()-$xWWT$IP>OY4xzB&%C_NX!9sgaSBMgaB#{1|(2bi?}
zZVnI=+j3=S5I&~!FJT3Qcb+V%4XpmlN&!ixzO&uRNVgN@Z%-6JI;A11{CUV3NT4#;
zYtS1F_%n-k9<+jur+b!IS;~#`@Bdas1^n*XJfso7-4rOMK`-R#&HFnoh^cZMPk}mA
zxQPhawNs5^#;tP`qWI<Qa)Va$3h^$G{o{;T16fk7zi?Uc<NARJQwo=V7!F870G<a3
ziA8EzDo|qP2FQQ=<OJ<8#&<ElMw9}!m>i{j^Wp{m6;o$EyBKtWovEr=)rs;a^#lab
zbqu;L%H>P_$BAF_vrfK;#V%n+`nTGj-{TG<NLP-Be{x7aMO=DK`OUdf4FTN$7W(|p
z|KTip_*(Wc`YtV!%^%liA>xk%4BY?uzfolb26lqO$QQ#;WD6+N`Fj&oaX<k#y^HS_
zv_&p^C(3w$GNgZW2|5V&uC08)<eV`qXW%~`8{pICo3yhLeY>z5iw&xtgDMQwj8yZ!
zije;Np7MiQoi&(sP*Oc1|Ibri(6l7#1;ixa#EP$bf0{HExaSHKNm}|sX-CRKl+~-K
zu|6Vx25;l+ba-=Imese-jps$h18`4O`SLmQhpe-6Ak=J!fdj7p7>eGu6O{&BI3QIg
z%4%IS4P3aTcPjXq<5SzOL|I^i;1s!eGfUH6Yu0U$*(bnd+cKW4s!e*57NmI+pR8ej
z0P;W=0e&Aw1{;?ZG|Qt%0ph$;s!pU_Pf2%%r%X#1R|AM|(HtNEGGwN=-<_DFZdN`5
zB|Ll3(>z$RhmrElt$)@F4_b+v%=5{Jw_(4UVVwXJkdV-R5^n%l>^2GgB@&t5QzQnm
zV~W=?*Gu&Saxm#UJxVRMqH?y5wNk%w2Lc9r=+5A=PqlG`u}hIU4+@7j6dgxm^Fs4v
znSF4%x_>dzNb97$agCic?4UQzjH6e6$x=DR4;zIs%4E0og@4C<0MXmRkF4s1u;`D{
z8mVF|6HKMUrDpbH2Ey*H+@*)6<8D1)D)m0O$K=JcLHOD0vUysgKgkRrpcErL$!`Wg
z<}@<`z=YU}8?F!gtxp~>lZI60JXt11AL0)4*LXL~exX@AI|yoA9d05~U{Z!-gqa~i
z44=gS5+#=v=6vmriQ)#|^<^7wc4%ee1x1)j8IH!Sl=QEx-Ns44^CT7{syv?!1hn?P
z_Xh+Lqm2yF)92*@Pkj8mYnvYAcp#_1z-iELH*)i%TcK}W8^Wq{nB&lDOBCG*p>btf
zPI79Eueh5>iPZ`dlCMY!XX0>@X<-$C;33VUm08-$<_jm>t2jP;1x4>kKL%G}wBw3)
z9(xON!^kvXNTa39?b@pvUj@!r>6%Ysls<lrF?5e6nXS!9n9a?x?q~xUf=7Qv3BTak
ze&tEfBo-gW9Q0NSfxE?fgU$onKD=r%b>j8={bAa$1nu<q+qUzdD?;l$C3MIx)i&*3
z^A`I7kFvRHn3-N_Lp2*>ZRdNWD5*H5d6l<X0@RbG>Zt}%4^U#`xFn`!mdz}Qtdjpu
z3dY7B`o)+5p(iy1_03+je^v9`cGX23Q#=N-r9+)K5FtLsgUli@nzrv}Wx!mnfE%>X
z7SbLyrPsIUe(3@Hk1QK)(mXT+<C<PIPcE8vfiSNVVb;t1Dy`U&^~9bK4ZS$m7_aeA
zuWXW<utAFV81-oL+cbkKtMi7E@Vs*+Cs$rQb-i7&@yC9P>Q>o@hSLm2IY2nZ1WIXC
zL))Whks%})_wmfrVcr>_)Tg6u0=~))ml1SMvoW+wFN|IY8$u$xE2}r>pCCk})5%R}
z{;lcNodkQGV!2ArxC19%tJLEhT#8ClHa}7)T#GryXIf|~(@^Uw8Dd}Tvw1*IN5-~6
zf$2bwN!?NDRzs<{!`l5M--^<Q{hcS0kEBPG!HR>%uFzO_qJPhFgp!f2p}*?+h$Fs{
zVbj?)L+2y(Dp10c*g6;x1$#OhQzVu5WAIRzFPp}hLbTLEU(Tc#E>q3+w-QHYx`Ifu
zHMWldeE#uXy*>~}=>2ygP;=xXg9so>QdH>zPaSso0N1{~%4|upObg8fH*VzRYoxF#
zoYy-rN9PgtHhTj5qpF-|Vli_8eD+IopTq@4Y6pG#q$1Wud*lK2a7tUcOUl)YJ2~8Z
z--(8xQ~liZ)=XYn@1#o80+t)TGLiL><ob&2xbl>GQP?fZ<&?nb?GsvS@2vmm+w>Yx
z4ltmAAcG7!8XxN(W<aB4bBMwPV@f+RC<7cg<ZHH2FFx%tAN*b&RDsRg_Q-m97O^Ug
z8MdZoiKg-{rbgq;Rw2?brc#gYTnOIOMAC^W;zspgR?K|xrmdysDlzZ<VGH@)z-h=g
zBiGI^Ms4C$jp?k~FgGc)K^(%+=L7n8rwpdntK|rH<fH|xJNNAxz?iFXU!SjhUx(~;
z_RN2SxcjJVN$|d!;1E7#I^MXk*wV(H)tuG5QkhjBHT-e0b0<4>r_@<fdD1%Y#~9)Q
zL%;=z$>Kd?qUXAVBBSjMz1LtX3K7&HVeeBM=T9%<2ox1wV2z+Drx>aE7zR-Vh)C8r
zro$ZkBGN3iM{5$t3M5``N9Y>!L~U0-Q)CtvWm&-~2%5Oo9C$Hz8nVtOYA|AT>Ohor
zX#dcAU(f$S=zG?@Fyg(HIR?ZMyjHKkr`u=q1ykF&$5-fc1y9!5g%g)u57hkE5CT4e
z<f;W+(5iEXG&r-czim6aBSXQH+=<lhKs>Y5ignel>o2)CXu}Y6gnC1iP<dtV_n^vA
zYpypOYI8pnRsi|RWM70>_{>I#6!KbXs<>^AL`Zvflvkxmc}{+$*L+4sCwEi(1%spa
zObP&g56Ruxl}<ZoMCXKytQilN90;#Z51V89Uo+w2E@nP+QJUzV4|cx`%Uu&yZjtD8
z8?aUce}1^-gg)f;JXcOmwBK&c{>SNwy?9}+R`mVlcqxYonY)=3Ng!XG!c6ZmwBfA<
zaR=@n-*RYbt%KIPdrTcyUyYnlR{5_iTVRR+;`!AS`Ib=9TAC);nS9P!#veWU<S++P
zTOd3)F8R^N5ZVEbJ6v)tEr&t}&2^OuoB;vB9z({>gba8pX8CR}_F`1ISH~=*Jx>#a
zU4)b2j|(j7L3TiiYZ?^~p+}3mdjuJ9xzkTC)=PbDhiMYm?~E!5i(*MH&>0T9>!j5Y
zy~>UrN-f5J6`Hz@oYMZ`BRV;1LA;u~8{_9f0;^fYfCCvWXXg}n72kViJ_eV4!iw9k
zf*&8NFR7c%CbSi4nzl^hdyt0EW%G^(Nr9Z|Lp&g-x-lq024QHaTzcTF!$j(gj*3@s
zK5&g$0Rb%F^N?v#7h=S@tK_!CN{AG?@qowtUdAd0tWa&m373g0gXFDOTw)eW?qt7q
zSgCoe^m=6^Em?|#oN+s=5C-p*n(W+jQL$v+iXj46`5scTKjAe_dQC`5{K*W}U113~
z=1dR1c8-zq`{Pw~6;1UIN7HgkH%J*%N8XUO^|O+_EMO*9li=m$Ew#7s+S1RDD@3<H
z7{vvG2i!@N^OHLvC3awZcT2;~s4~z>(WvN{Nj*JeMWd7%L8CnHZ8j!c$}ARFfu4$y
zvVIvv5Prsa;9jLxy?98{8X3Nm8e}DzCM<WLBYVF30^1cz!T*^)2a8-5MMl?Mj%bc(
z^7Gxc9#BPFlcb}f>bM0atrqok`Gd)PB`vem()+rJBVu(?4yU%GCR+Knrz5*$nyTm4
zEtH8D3g}`dq79Lbdzzrq@^6vxh_2{Os2T&_IrkVl@R1WY1QM7q;$%=BpndsU3&84O
ze6NWI*l?33cul@w|6P*9sFej7a(2ODMly<SjEl_jJ!WrQE;46t%@kc9w7#%Blp`)}
z5~DUXw}f!2xznz{U8SX`uEC$$0(?dK)4e}<8e&EvtyNH~5|DR(S=;Elv3d541+HyX
zmk%^@7I=x-Kfy*2ro7xo8JC|f>;b18un_p{*2ak;4&^>l@m1*m0@#-lOE*CkGmG-B
z-$03eTanOhRv-0yLZYEyjWw)vcP-av1&b~oth~>&e5=#09~gf>>s$EbE?@NC)>gSq
z3}w!riQ3S+6hkN7XrP;2m+jGks8bEf{e%in+e8``G7SDli$qm(46Kb7Kb#{OuWGJ$
z+c~sXK#EqQYa>bpvlfbY8!xjb5H#{>zQkKCOu6u4R9P5+fL7)97@jr;3{>BjsO?+}
zh?uCX+C%Y+4XWd}Z+fowjBdju0=2zg{BE-UVgy?eic`C3#N>ehA)<!%<>hPJ(O!}%
zx-tf^%FMD|oqXAS-Se>|JQ}^(uC2g!*RGNz!-VS%D#(<t67jZzSIZ}XA)?k`Yt9z2
zjY7m?D{@BT;j<!-2r;@P1ZZgqzGkLGbx>{y)eD(sx8<i{91Z+C1W#1NclBllW68En
zzceGw@SU|6JR5xZ@PtGJ3tbtd0Q((Fm4zvhk+%PAe!!<*65YZw2FSOQ@1pw(b-lo>
zA{$R2qX7^3mgGmPSGcR57rg)}^4^6H8cqvx&_hvYjzHtzB^vwbqm{1Jroxd(h(H?K
z5{o4knsXTQd)of4{1R}%0y2-&!?f<?n(Oz0Fqy`C8#~J9cZEcHXwA9_rej52_fX5t
z1iZZq&><KMvCL&tE9ItR68WJq?~?Up6Z4KUB38OzYMkt!XTMqbrUdi{4P+|G<1L5b
zy1lpC-dBC5<7!Zp^CuedDLipWEx)stJ=P^^=t#?6SK9V-XLIq~IB^Xc<^8QYVV}Zc
zJxe~e<H?)wjB>HOMN$!2LsGT~iIP2%rP1_NM(#~1EgN1bHy=}Q6d=%UP%IK_;Pkax
zx%eS!+WRpRm=pLgzS~Sxjt)wn0Ntj(h3-Jyl7Mkny&UBnK5`yKO*zb@XtrTlMiSU%
z1`U4Jo6yo8S~UmyCmLzhbo^Wg!|>GD#Q527(aRclhL%g<joHtQ2lUPd<Q>iwvU8G)
zDM4&lMy2IHV9Iv`=n1h4`jno&F~-u(8APnmd?*<kFcM9Aw!YlgJpi|RD(wo8PF4H?
z)iVAb!Z&Zuvp}^fKeBscQ8@kb;7Y;7BtzH?)>27Iyw3#V01hO84o~Y0MFa(Xh}ZQN
z;WoaK5)t6%(u>AJaKD!;YH|`^bB3}?1F3NP?W`E|&I6=k4<#STC^fI)ZfW3#s6(Jo
zP~E^Spx$Kv*_z_?pVF!SzmEab2NVZ3!M{SlBR4BaLz$TY1o#3$9%BMU0G2LkvRTfH
z=K^RYAOlx#?H0~^5JmaByIPIWKm)H2DX-Rv-eb&(p{i_U0BF2mCpR`mX&}-UaY#9)
znm_?%&Y?~P$Z)aCR0pu@m?8IVSF64N`!0Eytcq`I3(d(B1&cIC8cS~0{YN7_h+xT<
zCPqog!%a7ar&$>PaSv;-?<3^Z;GnDH;_7=F!_SjOMXCPmq)lc>;_JiUSLmWH4AzSO
zd6E%=>G$n<meEsHVZ1+|H7^?B$d@ARBAa?Q=o$$_mPlW1{y)1m4`dl|j7a(UM8q%T
z@Q`u?OqjOFt;Ti+9y%T?exNG&W8?MDerQA+@~bLHS#I%p<l;^a)&mMN9iN8ZU4I_W
zmB)^l5aVv79VWaHDuhs*<PWm(272yi*D69sisQYeNjN)R9qINK_W+@~w^Q`CX9e!6
zJdq_xMIAc-?H~>e3>0>{od&ma=jZIOxgE&cN%?Dh1Ar0xcOd_{+kgMUJS;@_8?u|^
zO+M)YqHRE=lY|5N`=-FTS_tZp-v4}yHt8X%73lw8k9G-K-wdgX6!2*OcNs`R@2viM
zaDT8(L40Z6o4_4dH@z3xucZzYIs%9o{T_1ssbwdc-}_9y;Td)MS4Uta|FfhVnY>gi
z;+vv=?@YE8tOy6!cd=amu9n(@uTNqXcX)sA<@y9c#%lPi)UQDWBCUWvN^(V!<ZYPg
z{xz7tx?k=8sm6H_z|zUc(#N?FKZoOYCg8*BsuNwS^1Iz(MvkP!b(K4U<E{1eQa|c|
zqPfND9>_2`t}kDSqj}+iiCi=<XNd2`;?rg;$i?y?+xV(DdQ0}dc?3DMEG<*mSCIN%
zcBu5Ds-htFc@mH>=240T!v6;2eijvV@8bf!Pm!G!Z8hyt{>@B&oLWRV-($$xlez`8
zOHx<9Z#);C`TQuHgNY@T)d>%Y*&v2q)$i*)rFiuK%h*^+61EET6VA(qkLlpEWav{w
zvmr%8o8p|Q&CRFJ>>NLnX}E^jn3ARA!)OA5HX*c~=i&<4=Vch|;jJi{uJk9BRJtY%
zk+F%9dsR2oYe{MSbyqAE8!%bk0x2>vH{u8}G#e=w8tIR%IB88|vJ4`3Dt$ABbU(ga
zMxqxacg2b69KYn%eOXfx+fmWL_`>q-BL(Fcv&{F20jOz3pL^M2O410}zCJN<W?@LT
zx3{z?D0uvX0_HcFWEdi##|vS)FC~5ul#u$!5ic^#otS%`W|D_E<-QrJ!4jt3=X5$4
z-W<J!WxR}o4aCu#CVGBYMlg&SRLc3v)KOxz)T1YZclqwSJ2ug{WE?!?>J{ZS!*}36
z>VY4t)eHmh(N5}<Tyy7187<d?aqR)kUfbQxvd`zIhsouUUU%rzXFrzM@vE_s!D^B3
zzQ`MXjd*u%%7lv1R;S^Me0XP^Nvt$1kIH)1iW5z;hWe7$8c4AX9}D9P&4=#gL=Tct
ztJlRkJyjaLpD4g&ea{zbQcBj}F^`pwm)F<{cW6Sb_S@y?mg7pO;c15YAh^+MPNrT#
zstWal$RY9yd7U2ni8L9IZ1_<ukO7l$ikNYUHq`Hvs_Ve(j_l>gd<^Be!dMzdZxc?^
zVcgd@8>ALs$jZlQ9j_lYp7yn_p#w=h_;%``F-Z;)J%?kp@$HPKxJ+BiSg@F@Ic8l!
z5ZzklDf!vS0K>D7wixDIkS*IOkYeMW_&1_-$|7NNlla0H^esT$Pyq~41AQQ}fr&4y
zjAoeCR4m9t1`7>(@Bu$WiL_t@>g6F(*I2TJ8^<oR6};B8cmHVdh*2M|_^4gb4ma4!
zx|x^nEQEFD(vGpzH+PmDxr9A2laqfaP%^V!hg8>wNhug~G|~7b8=INnGY8bksrj=@
z`@I@D;znMJUGSKcI4?2>LV|m$kROV7a;N1REcf?v2#OIsbyS!b{Fc&X+}8&BiG(B0
z4%IoT6pK#f)u`@_g&8^tX;?TAeECo2KGC)`8z0HQ%#CMxb$!G({p0hS%pAOFP18)!
z%IWWiahq!6B_S>@rmRt{wfIGrU=18a=qB@q8Kr1%^xNT{Y(u;k&K21fT#|ycz8Tm`
z*Imw&D2M_n4f293Yt;?2t&q8Z9X5D$ej`0PLecB!lew;EA0I7xKesqW5bPc?jO+DS
zGxtqFJUV{v;nME^)VEdS@jmbO_ITV8vCMB-cFeN9U`B9zF9NDN$tWrLGyJWW_%f;@
z1xTj5kzwiFl}mde69`Y}RlFS?&7-=WJ2VC)N)+DDUKXlfPEZrn<&|5I#f?9C;r>3Z
zv$8$z^W&%fQ7x}kq)^DE)SLx?x!^4?4H+-4Ki+gEd7qxkPQKaBko>MNzfr@z)WkrI
zSdfN{t)qknWJzn$vK4b?LiZqsVoq?-M9LTk|2i%zKZ^?tj7|a>Ubr9!{J7~^n~?G#
z|AosU!Qf<kq-2A)7Ep>+wOiRF2>Uvz*?0ToK8(RAR%l$KItipAz=8S6n~9xI&B;=N
z!8DBe%ku6Je=bi-qjj$U-vb1?tL?h+d+v*}2w|-C-{P^NDqTadOqC4EJ*Jm~ASiy0
zQOE0N2q0T1$kloYWAn`uwfG~Coy}qIqy$+~=K{@Q(WeZy9^-`Cgb~b{mWH%UraVRH
zotDqva#oM1Wi*0S#4?Dgz6Xf;O}dd1iTQi8M-ee4+Ys|p4ng;ieSA50+shih-&Hbv
z<>O?q+zb#`>~##w0o&Sp>^w8%dtw>Q`11`A<2M;QStJ?nk#zUlp<F)5FJ$<;r{+K3
zhaswgOH|aDC3nDbr?u0qVGr@@?u<&(A{0>BAbgGWr<h&hCan%^1Tn_og0pJvl^18+
z_e{)3mp3#vkjGlyJ(NrwsAi8#c$21%yj0z;L!xWTq|}@8{DrMOfsX@o`tZ{i!*R`z
zGY=A_ULJL46s&c5fL1={2?+IAEyRcpUVOQm3ThpVd$hH>3kCN|99^#AKR0X7ycS%0
zg@y$2S%<5=9oa4f1!yz6mM0qpo;eeKAZl;nK_9Bq>=8rdmds$x?kX%}l;)EXpBtv!
zk5UkytJ7Y+5kwd|Lf@RzuP%QzR?9BjAi_S52Hh+Sv4B>KG@FoS&*nTG1>+-pcvn5u
zu|G+qgK{MaS!AclBi=ybW_S)Gd8Kn)%XrmbK^wd;RItnh9?(gaOWZ=`PHhU5-T9Q3
zd^mhKDTD3DDA2-+_~Ah?UAO=hQ(yZh3%*SDTtpa+?T`xE6WKl4>!v%SV$01`=`kU}
z>Jb`f>FKhlJxLqlg=G#qMw1s|w5rF70?kc50P?A&b}o}ondo5l5&qwGY36#n_1g4F
z>-D=JU(eSF2|pHeAEzjw4BxLcJUx4m8mi`TOXNK9rBH}a10TeR4fS!Fyx4fKAqdTQ
z(DM;odL^K9O{-Szd*F<3P?#6up7D@t9MjN=_;|T2`~hN0J>{_NiQ=6h%=OeN-2BRl
zCmaUH_bV_*r8-U<fG;ow3Iw0Y&q8N1X+M--w|-38h*|GztKLzsB%<|UGM<Uzsc+Zl
zv>NT-$2x0Rmj!09WRdHVMQHclm4&rMUcqAz9GG?%IV~RNW%n#ss21y+GGJ8|C?7RD
zC9frVl<-v|BqYDS&gnB6M=s8g8|MTyGJ%-e2G<HU+CI6*r3veXA|zlcV(H!yf4OB6
z3vBU5QJB7uuhyi)>tiptw?56BS!Ha=r9wm6F+u=R`_^!;tW05}n~7X_h2^qteA=J?
zLS%u#&;%VL|A<AWVmUhdl~E@xUHPVlT{9fb?mV!XyLt#%YA7nWq7;5RyVn2*$}aR|
z`0cy1Y)z1y*_WFWW=fWqi6`btLIB0n=`Tt>WPx%fpK2+@y1wkM*aY|2XJFMeD_hbf
zA{CewulEV}Sm-E!?KH{nzJn~vBx1WPXNaoLH(`P~uCl$aFexv4k+FU#yP*fF_t!9H
zK^SsKCwGCizYM|m;thZla|N>dQne(!sa7v}L{`TeC^|(tKB<-=09Nt2n-CRHY7TP%
zQk{*eapoT1VXjSA`t(X+YQ`5z7!^EXz9~~d0t1-F0qcdc#_R?8##1QA(DK3q{ew@0
zgdfqBv*Kc19TOetr%gb!_p2*gD&6j0*}RRkjS7Pel~7d2s?Bv2iX)M|ZC+H<lTkL_
zt2?7<%@r{&=}$gXb_hW}g`6)G;(GD=O9YHPa0DrJ%qz2zCIY{Hy@<srpI0$rzPwV0
zhaY;7!1A$g(DH`;A0YCAixbjex{Vi%sUAiAPXsU_9jrodwY2AJ&q=yRkNo))?*R>#
z?A7U%*mbIizl37lla4;E>@}@M33o|QP9BBb>uloXiyY;``4~YV1abciYk*s2Nx3s@
zOm}}-_6}8l=x%#woj%HAO6qhKyO1biy8E+vQTuKYZ&XCx?N1@MAxmmGVpLUSO{|no
z2ik=%Ei_K#vIZ@2J+4=q@q-NS4NW?kkIRAD-B+B-aL+EI(>q!4BlLoFELfTcOlV%T
zLLI*v&jkO6yD3s>p;~Ri^P{I;_L=UWs53Xwk?~!!^bxl-Gv0N&g2}raac>f75<D+W
zwTCxfJ60p5;7Y|zgmuy<!wsK7%!vZIbam1unvo?dmT`*#U_K8{V>i#;OREtz7UI~I
z5J<e1r5;eng>)P!A9T6!!f6Q~V4iLcar3>!la)H2^D$NWQixLQ#4zHhwH@E4;Yr<s
z160hxhZe=vZNCUF9&ukVzt>=S($Tx00m*nI6AF3V&+71o4(3j__EHH0b1I46%_b%C
zZK}A;CtBS3K3gm&HiEizLmdE`A|Nu(?gw6$r&K2$w|`<4I#VZ&h%Deqd_zRx`Sws#
z&W}jPQ_jh~_xx?2Kw;HF|3^3Hf-k)Y+^v{g%G=>N?cN-1L|JT{YzU0A7(wNNehX1b
z4YIV%wjr{*LTOJqB@Ldu?k=<+r7WMZ7&OoIW8;gMz%Pi$U&T>ti)3=DnP+FR<8Omy
z07wmUs8B8pG}0B~Ze*Au!R*q*Zr(#aTFoeF0IrJr!M<4|CpA*9+`_ELVNWzt=JZ|G
zz81&lCsx+WZy+Ap217cMlzj+amDz*(+4Bb}M`gt_QnecsBaYy-pW=w$p6#~3WBVI8
z60Unzi<F>%GDmIMQ}6`;S9jkT)l}E5o6tc7rHTclhY~=lgcgw^Ed&T4O{5nAr7Ja{
zsGx`tia>yXKqvxA5$TGwfYcB=0@9mEufo~9e(yKFGtL?3-23B>aqh{VjEtSV*4%5&
zRi0<h^~?mNetEQcKGTnKow|KSVfL2iG{_RbC~kQ4NGx0PE(k|>VqBk{NnT#@?;&HJ
zK@EX7CJ>A~zO|5n?PO;GkNp3I4FAVq>Na+?m`O*Tk>Fo!${*nUH|=uDWNnb+GUfl3
z3phL_@BRt-S5FDAfAV1Er>x5<EAywEQ-bA`#QH-b{I9qXc>=B`-F8o);OIH?H<S1;
z6bi_hw7K<8?Vp^=sb&qJ>hOmpWC34%^EW-n11I<XM>8QQobLk7_7}eq1v>lf@Af*6
z3j-SdFM<j}CN1$FYx{p%IE=by*audK>2|wUII@2kS9NqA=bN?^qPBxL1CTj@XDnTW
zYmM@)mf2q7chbwXu$T$zj7{FMl}zrs0Dp8GM}r<t?aVv}rGj4MWfu^@AMsnp6K(%c
zO)#n+TAYXMB~#b8+5DJIT9DJQ5&p@$X|0WLtPF<{T6Iq9x{CUhU~feaa=JB9D(EA1
z21Q*2f#32IG5k*v)s#Ji3(j(v?vvkmF|zRT3ue*}I49o{{DA(P@SSt%uQ=eXwcGlA
zaP~g^T#WmM$xVo^iaEXcMikq98?O{v^v5f%^H?_izsTTzAv6N)OA$HB7lgeq3R(Ts
z?KqmfjDVl9dPrtiHK_b{Uzg@-^RV4IIhJ|gh})lYL6t(M>q?BvuhDD`DR7V3hDHQ^
zQNVA5RhEzK@}40dPn2T4&Ex;-c4QK${j;XCTyAV$W_0P0x5fp9bT{|=`e9K<ULhmF
zRQ8Cl=N8OtXeKWYDRdmK^z}4}mtJVHl6(hVh^`ZtkaNDYiL)H;0!chA#|zMif29Du
zf~y>$wrsWj<f!cK^Z@ZFc)qe*H7+Cgt-QbT_2lkTI^y-@z78qMclX<e$}=lQs|*O^
zj#yVSP+A>^u{Jjz$LBT1sqY-0fA8(*+SO~V^A9hz-7<vlT|X}bQCa*X$jt*@U|Vds
zFBPt;pXNL0q7tGC*0$-;VEi@<Wn_^86#*v6112d>+>sYtY4jh<zX`fR&v{t4E}d{U
zkW0`-x}OBlBqCbj+RZ=~Pouu^PRct*unV;B9;3YqRt+#ub@t>q?#J%MJ@CE+o2SSH
zjW5Rg1G3K8=6LC*C(r$A)3&AvRTQ73K@J9DTKZ*gSit4lxIDipk+7CT4TC!3CCFO&
zH?3^H0d487Od{e6DDU$eX`75qh+<Sh<$~gg!ba;LJJrS8t{`ngESkb!nS&l-DrDk-
z9w0J-U~eZ6AQLoT!O_vrO+w@l6$=lk9$sooCKfWdAY7vnL4}NcT8g2<HQ8!2L^ofB
z$TRl!$QRIF(N4Q@;GQ@(6j4HT)`2PC(9|FHwy@B38dBgNK<<Lg^$~4~AlhO@F62wD
ze189+l#p+DkpQ$!!pS!+?xC)$nWNCNzA_-avUK+5s{C32!t*TNl|nl<Mhc}dOpK&^
z5O4Xg&EHhPjmd5iDnAl9-an5di$C`ul`OvbHF{wE4dYufnMaHv_sBNY`JU0PHWFiz
zpM|6CT3fWT3&OH$U*_|`%{~NB_CSg0#RF-QDI1<E5ts5yOYRR41n{mde0VW13ckEU
z*;f?$lPPest97|YKZ(}wo>llL_G|oE&*K$yEXc=H)uA2bs0#ZoLH}UI?XtC*ehb(G
zgHF8kqBL*EkTCa{xVil$veJe9c;J9=vImv-`2n@dOvaM?v@3LTEl#RZ^HhdC=HsJa
z(-WHK3y=G`|Hfym)n~!Gu|-x8<4K2P`_QPya`q-+!svqVY}u{LxPH6Zvt}_icUelq
z!Un@7@7;Z~jBG!%SO{$F+f-w!e`U43@NaM1u_ehGl5q7Lf~i>H+*oWi1(Ne}*M8oD
zyx*JS;2(IV7yUM^S4ITcA7_2m(&!NJ%F`;X=d1b2aJ-?Ybh6T_C212xkG&2z4y>}@
z{4!DeH4h48UR7=h7(qD;41&^V(e-c1A(vXP@aJuFgyyKO^3KoB{9l%3b&p@CD^O0V
zZGFMVe}4aj^Lu95=IbYVfCxpSwGRtR^`^zS&c2|9tIm4pxmI;vy{yKEF84Z$A8F(`
zmFHkF?y<`X(DthSJQ-6G&<_PbKl}vwtkEq;1JNMhsqcc)rF`ND%W(F`@#=x|tYQY1
zT`m%|H)gA5u2dBw`YykHn`>F~B)B2VF8{ShQo#9@v{RPhbM)qpHertZ`eEuDa-V+$
zn8-wc2i-EOEP<{F>~yCK*a%{B`DP;23un76&r6fmY!UQ6{;a>@o4yZ58z-Zq&5z_z
zgeIc^+q%LF!=WIV1n*L)S1?{(KIvk`Tmfu&1|<_c(+@`Hmi-b0J>sd)K8Y&cwFKu$
zDr(VwYed2aT@oReyh_dXrn@(lWp4_+$#_K>=l^Mt;er&pU?(~5gCPxiqX&~?yYgM{
z%<p34jLBM%TEAq@UEO%7UCOIr!=Qzem`F#T5+<MFO2HrYvb_2ynZz|Hy9P<K*_2nQ
zx07`LMu*6T7k1RUf<WNWdj<+P68$XvWxB|hOq@Z3=9mfXIOji;kEg(y4j6~DxV3P?
z7Q{=}4YV~TU8Wy_K<tBBYBx}^tV4WZ&>R7Utf4}F9id0UR}x6sh86K*ii@c+6I0{X
z6?mP)DdJ+{g5>@6YdQ`r@kLJrV`tq;DGO&>eS8J2J~So%AWrFh&&)tL66_5S*~}`r
z!Tat*zftdBD^FNuY~<{@#iNA^0%iK*`gKf4+sx$!M|lSQ&EviE4ek{eVDu2i#9P2T
z1zFt=Q9bOtmV>Akm0m4c19D%jEr%ZBov7)7v&;LJixo_5m}q(kj|nwSx6R>q;-H<c
z>kU^E;`qM^QA0X3G1??Ufmo2z+CrjE*Pb$WJ;_UUe9kUW%9kg-c<tq4JkyHW#joE)
zbtiw-;oRVuPE7-J=@m_3*l_49WftZ`)z`FCi63@VuS4U6+p!k)ZoQ?R8ZzCLrr%v|
z9KA#06R{o9BVNX7xjgub9W({ND1(vplsBY;^CodkJmVLfgJy?sNWQoUIvCQ|yQf|9
zPN&Y36{i@N6Z(LqeTM`t5Rx7=OuTkOqBX65pF??3IhCLB_mCsVcraXG>gu;^;6sY9
zj-ufXN&$OGt4mLO5d^=79Joa;+i63G&g;cCMfniRY_^j2rNEjN4wBwoXf)Aw&nySA
zdfikaPMY}zUUcr*7#RNr#oib!j2!vllToDj_7|)pTe`&Nkay^A_eodHuNW)Cnr^RL
z=?W_fVa*DqNVvLXOPUvGwq%bNhNsU+hN^1W%jGCU<}O<9KB0_wZ;Cmp=xiL$cYKHO
z#;n?h8|z9WSl^SMLs8)J1Nlfc=APSitV~%WLT)+U<}@lIq<q$SVU2vZU6?cW!r6Nc
z(#$T}I1zbYY8=HQkc><{%h$;EnsylSAUry_9doN)xM${i8^xa}5NI+7?m$%vdT7Hl
zEEcp8rRvrGlq34gOUJF9M0Sf;L~c+$7wug9-RIsqf}BYnU#p3pox$P64-bVS>3g#2
zi0bYaW0KDHq$`l5TsMBjzr!?R?1VkA3D|0K<TYYF=39`Ck&L&4dXI;nU>s;FNC^zd
z92BymYg{Cnqv`Atk>46SfT76&E)bU{l>+j4Hzjw^1kP!G!N4ftM1O%Go^)l=C%#jH
zqed?8jb6bk7Y`|leZZ)3Su*cqI-a+;T^~$SK}&QE4$YoUqq_45UyZyFY=QfQH%z^j
z&cSzwgtb3jW^NJI7t|3xDKCBxsbuHOGHuiyqUQq@wB)~M*owG6nXFraWWEBISq3KK
z9D7O-i9N-25!5-xy;QhS<DYV<#ps9Gu<#ZBD9c|n6^5Or9UZGW%g7(5hkak+VtA78
zHT>&p)Op;ve}PN?<dl$1+ExLD)#*~pS}4>U!gVl|J?b_$HzQ8MUki?Vbde}F@oX!y
z{GpHFa6O0A+(vV5lC2WlS0NS!ECrMxR-H1*2r4KSE?Ge>*f?Z2@hETXiJqVyGuq?J
zw3>k1g9MLc>U4y_FZ!c{gtjNt1~0AlfLC`ir6*_esu4SJrz_F57&U+F3-{T;1NYhC
z`LIUOqxCZFwG`h0=HRnc4C#R)&3D*v>#L>&4gkK}wm6^S_k?#eI3Mw&4~UhBl0{U=
zE}Oo{x7*es?V-#rsjI(cl!D$+Wm()P`l`mddmidpr^P3tex^<80)9kefU&o5)gDT4
zSp;q0zJgT~HeMXJkQ5i$`Chht)mAHBT3!9S+j7z*&)jV%4kKn>_KluXSkccO>0px5
zy6MY_%MrF1lAKkk&NsVxE8`utT^@A1ly^rzI%2Ey<>j`Ur)UFdEsfV_>s7uw5{hm)
za+>gc@qW0(>!UG~JzV@ROcQJ%X_F9oI|KyM>mrskYWxHCEX6Dk(?Y#uPrUU2RnPy4
z;O74TIVpq6^GAefC2{P3W3v%Xl~{1ygFmS4H#HzW`@bCP_S8UF+C}oAm%-IY=;2q5
zOysJz`xFgQW!qz4s1EbYw^4XomC{!geUSgzW4fNMu>TA<{ISv9N>A$u<E4mmxGzHQ
zM?aOE{)?t(l4Zz!%+w|!fSbbk0o{u1&%*cGb%6NHbxOZWyT-S3?f~+{sz42Zcr5Uh
zsEW=)LgIl`ybu;T2O$awU4*I`aGvHrxs+E~9XQut`l2V@!}<a}RkyXd-!=0@Q?B#9
zBvb0gcIm;|rq<lP2g5Q%8;ZQj5stl6tppDAtGc?C@N$t1WuB}U>J=kcCGacH$^_5c
zAa_|49Z*o07N%0<y>>J?ld~tgUviCUA+z@2`SZEQc1}QA6i%t$+|cR@#IjGz-VJW}
z?|%2b^<Q}3@8)>-xTZ`_O1Y3=wJnRE(nrfAeEHX<ru%Rp?d=^`<qNeBnllEfItoiW
zJe<U5hBYZBXLv;wID@q<{kyNiANL(KS!7295Mllgb)|Om7}th1`3N`BDfblt+c)sO
zab4Wx3q`)6Ai9lt-ne+RtmB3DYEKjyc&X^fB~1$>&8WU@bR47>=clwJfLx`gYCW!8
zyh_$@u}_4Z$0_xVRW>^YvWJu+2HBk>K-(tA<p>JrP-*L3R}RRau^ksDcqw+yt#?g+
zy*+6l(hgVGTYP_AFv-&rs38`}%am01>uQWLPeEOb(k#Ae?J-aUX9a?5Ff3g0;5j@S
ze|KzSNM9?Mi^AD?BhNRWidkr9i8pF-0M!7a@jNIen>IRg4)?wN?Er0NbZFpI-NAEP
z#zFDsUnHf6Q|7%r6sMN<!wA7E)5JGH2hW?B%cw$De>F5_NJMnJ3CEZg5xeKZ0=Cb0
zQY!;osMXO+qyNe^{tK`9?=OvLjvsC_Hq^u{)IUo<D!m~)!5N`jzgglgP_?ftGV++Q
zL8#&PJlE&c!;{{Rd3sU(d_|eC`aD8~)S@I|{kY_!gXdbshN74%p+fLWtmdVCKx(Nx
zw0iwOS&Mqb!tW6Ddxf^jIVRiRUj^*?4x#BY&F-hJ-3JBXq&x7bvp3cfsiV5D0eae5
zKc&;o;p&KZ&E5QfCRzMPur@ih*<l9Lvjdu*@Fky{8Jra_w)%X|fEJqBBq=BoDPtYc
zCXv<_K*Zer>Uwc}+?wlKbE<_(vr?TR+%pig0Kx7vO{<{Wwm3rV%(|1X<P{OvOpi>d
zbvb7V=mw>04Zd@~pXdiJfF?uCPeA6k(-;;QHT912VvBvzeG4V)t(N>zP|!kUVp0nb
zOP3xwE_cXKtMxM`(Dr3lDP=hw^69O<0M6s9PYFqHwR_oUdIU;~x;C1%?_@B2Tohgk
zxKw4zK?YFQ3UhNc4B!#*bwB9AE!#I;KOJ3IaoCmJ&^JOO<c3v&6g@wdM6JguAT`>0
zy}&yQS$ZTzkBP(`J(`{nq0t|YG>OG#m9)q^=N5}++id1{&f@q<X7$yy$dtfo9h3__
zkPTwR!<#9b2Y#5dJUl88$`yPJduKUz(v#X==es8L!AIKaAnt2xbjDZt-QGKe?^4&k
z9x~RM!LctM)jpu<J-klooRTDpkrIc=?@St53pohE;g3a-A7C83zg>B5YHjC;)RLvV
zqrdv(#uZO@Usf^dJma({VoMF3_iIB~3U)FNekuVu>`k$4;$(UGm(KtmU~LOZ;Hm0J
zP85S$jh##oeL`R3)s728=%(&e*Q3lV1J>$hRvGo02=q`1tJ<|d{o{_LK2O7+*;lEv
zol(m{Y4|-8tyujGs3BlT6%4fzQ0fy6(K$f-km5=DPy1X-(APv9_^IO4)Ak_^mu^?d
zuy*BikR}Qc$sId+*AhzOWPkIi*nehjzLmu7Z~)bIIRg%`I{6?51YCtg5@$X?dpeyo
zD-ah#-B78&TR#%yaRgW9cX(EN+3Q5*!gCCHphm6GQOOI#I|M?26GM;wEX%I+qSQI(
z2b2Y-f?Y*SdR?<aNCvz?VW!W4$lY&1VEnS7nGdTNd7f38WKt1?6t<-;l93Ihm!avu
zh+ENR4}M7D+z>w&oszDPb<fZ;`c~gLVd~MQasp39RS{=D;-zFm=QneIS;@0|ZeGHg
z-8-s^`w{BeS9n^dD@wDn{f^oDZL}AO-bY{C=@VJYJ9^M>am7?>T~;wLP${3#<^Ei4
zG(LV!;Y1H{ODCCQj}U4e0dkP>?0u$*-qG!g_MfnJYVGj4$39$_<eu@P9&^7fXnLx$
z$Z;y9e*IVfTxLbkc{Sc^u;rJ&^Fo~kUs&^uu8a@TiNE@o8^ac&oO}A5AY>+9@3cna
zG_NG+mqb}s6>}d`t1*0t8ym<$)9Tw-UU*=xuARcwnC(rw!;B;0`rXj%!-O0DDoyV)
zRJg|rH_j5)oG&b#pwfrgLnNGEyEV$4Z#_!LkQR{mw4O5dMz<>~dNni(CFnBxOFO$;
zGghw{p0@P7F)X$%wsJ5#ms<I5O^DZ^q$!NZD&c5Mk9>I3dde1UPSRrQfPxF-mISyA
zkqwdyR{R<fkFI5>z1acPz2mdQ0l|qzKLdYf+AE@x1?#b6B`Y4rKV)`PtGQY2$M$aQ
zV1pTh_#ILNU-|W^o4Wcvy*$N{_9*bU>zN^)3@!*>5KFICD1t16+3GPjY+YtBtd#l+
zm;WvkkMO3XD{cRlE7)-y_QskObsdA)oHZ&#losodFh(DkRT6%Je7CQh_{F!zl@~s{
z2a-05Zw3xbD~uyK&r>kzGgpaDBuG*ypLALWCpAFjpo>rY&+;j!94B+|mr><?n#dvD
z)|g*DoKX}PDG}3_Nsu~q2xCo~xTw#%db7(;khFC=7W*9PJQgnwVIa)C21c-5a{{MG
zAXL*+$W;s&o@69441uJX>BRwNm2=6+LCn!Q_Zh@va-|^kHcaSQ%(B<$9Q8us3t??~
zcRUay%dx}c<STGgI%mv)16sdm2?9Ya+ZE^M`?QGM?E-v22_a0@Ux0`bYv(dJ7S?uT
zHx{2Qj!Wn3W5F>v0tp%GSgevowJH~G;!_(!RLNp}Zfbk##}R*9&t9yP@!X^3$)M{>
zb}R@?Aa}-yAn~xH0>u<9_@?w=KMFR@D&gb+$}9ft8qo==D8%qU;l3qh{s|jX+8qn4
zVBrCrm?ce5{nB?QCbb+YSMxycwV2S8yIvs-TqzaWeJg4sG*OX8vXR;M=mh4!<%*1p
zki%C`T?lf`>VS%0l>>futk)`R1l0L*-$mGxs;Anw-8x}w)h(DvUv|gq=>|qEg{=RJ
zG!TK?ZJRm;KX;wQ&Anl_{^AF7@_tZQRL2{(vpoW)IE3%U2YXnbO|gs3WNr<a`cV%U
zMeDK6Dofs9jMy(Up~R{C8&|*N1V@?I?#<A>Z+aP{Px^>Dk9%OKj$Yb%)w<5bz!#Xi
z431A96oz+{02YdJ;taco8G!sm{#-b~j&r!YMBmMrHNi3<_8{8+GFUh9WPJCnT>V46
z7mP-{VoPqbP|5GFg`T~-gqE!$^DQv?K29R_qpH<G{irDctA`5cT=^8RuoFZ@ivk?w
zYMP?;tAh}2?}`hec%d}kDnVU@u+J{I9?8eHqg3!nKLt4;Ly7TD1r>QwE}`&2jX4bP
zXw%P}xE4COWVKRqXFoF|0$<$5V-1dpF-vAL7GwQ(L$_QOT(ZEgRcpnqX}S96Qp-ld
zP`fx(`6ORxx))m0guf;wx{D3u?gcn4-iHuVoSWOnQGx~z#It&ozRah`J$2C4dX<oP
zjc^lZMj;}{)z$cRNyo}dw2s&PHY!O0W7=-mHTE!=uwc_0=VVsf4H@sT6*PcVn%}eU
zq59*W_SJj)>htD97qp#3)#Tx%ZvM{3+f%x*{RFEtdxsr;$NP!eXBKlDwuVV&OJWR`
zw_-|P@8!^XPCyzKGkp|^srtv&^QrCL>q*szhd+WR4wpmsc@j5fM%VlG{72#<qjNIW
z&FHJ7DLu`;4^!dJ7NgaD={_pC+Sr{AK55Fl+(yb%Ulsjof;i43D0Iy9rp?@-COk*5
z{n%aJId2Mvc+9FfG_BPJA7(sJ5|wL~6!0^mO;O%(=$o9nGqWqG*lD8JB^SdO;@?OP
z^{5vsQ$30rr(V1XFK;^E>2Si^bD+ZBePzo3p^Igr|3~njVUyL<{#Z_X_x>eqr&_XE
z(D*JYXf4J%8B~z}Kp-tH#~66_MrBXhMc%Atq7lnc?;5^)_F1HT2+?XRC`E7>HFU=M
z)7vb5Zg`+TNRm@igDdd<Vm5u8w^G1r8j<r6HO_5Z&5Q~BL}pStJPIl&oGu7@nd)M!
zc}YU2la50|kCwT6;=tF_w&vHVfoa*|XTic9TX$8`GizTe4@|5Ihf4*(5qZBYEdce$
z`+eirlY95Rb>`N@=7=@qd#bp%5g$6aUR3X!bG3<(w*a5kVQ@t+k+oi!K1b?KSk^nk
zkwuW}JK%XduHrn>3_sLja7a^{-{c8vI+CcT@NF{`8AO?p#i6He*UP_?chR6CVq;xY
zp0n!Y1piyKg4@#6V%=y&kqgNMD;D0x=4~Zu(xu|Q1EUO>*|wLO_NmKE_6sE9yL>2x
zOZOHP0k`bR?uhRyM&$Q@ioplmtbVzuAJC@K)p&-6Hi5fGS7}FyXCf#9)Y;T{h6UK`
z9$2~$Q7gY@3=z2!bUTO~zq%P)l1&OoetMPs<}pFe#k7Oh^E_OtJ2LCBi)FJOa@mdc
zbo%h}l}zG4A$QXla@GNs^9Zw?_e3x-YIUnEe-|yqKj&=QhV^lM!-S&5&Qfq53wKKF
zKE}F)?WnMtQstSyRzD4a2e$6kAMk35dv!-8-BA-c!G-~^N9%T{jt&<r*wvQ_a!)R)
z;D?p>rWyN1htm(KxgTAFYq71Jy#8LBtk|A{YZ+$4MHj=HmuPqo-ZN|qqIGQEiliOW
z#oMcgt|lkn^BY<c!D`|28-VTJ{JjU0y_N;tQ#^Aa1AB9UGpZ}knvb=fCUYZG1Vw)C
z-aT{M0ewe&|Mn;_FGQ$TWOpz%eY<qe+t75k68A!~jUABerHgRYihr_%jc3D695+8!
zIiR#&y~w`OT*~6>*Y{Wx0K<6$J_Z*>AY=;$4r&6MWz^B?W&YQwV?gQvW?rmj(Pjgo
zMD@d=lmxMeogE<9Os<9p$UzK2S!5A!9n9`VNgKge)nE1x>4BpDx<|{L)|vv@3H+O<
zWqH-n6!Mo5`s|p6k_T+Z0g1TPXtd)FQQU+E$wEg;cXrDUd}^ezq+D|N-K~<sBX`p7
zuN#b$nxzegVLhop+x^^<Trqq<&}@>VmhcD7|6joP{~nP2EBHTVZfas~eyBjV1p+==
M>Ik)B)w|*U1~gmnp#T5?

diff --git a/_main_files/figure-html/unnamed-chunk-194-1.png b/_main_files/figure-html/unnamed-chunk-194-1.png
index 9b46c9bc109a363d370f5ed43f4c2701b6a02342..cee26628979ba08df48cad1cc6c9bcd9f8f70897 100644
GIT binary patch
literal 266974
zcmeEuby!qw_wEP^NC+rMgNPu~A)PM@NJ}dzk`mG#g9@UgfOLa&Ni)DO0>aSUHT2NP
z5JS$K&FlO6eZO=5KiB!=?CS!DJ<on}t$VF?ueJUBTwRfbkd_bx0+A>^d#VKj-3<hR
z@O}{B0C%hs;Bvq}H=UlnbOnL9l&^lSdHT4ifk5{_N>BgN@k-sE1BV-py?c13J$vg8
zy`*TRAd#K#b^!m=_ZIgDI~h{%mFDW2++$Dar7o{%31r4(m5jeI^V*)ak*%#Ecuz0G
zojoV}M11mBIJ)KD2Xt1nq()#?sj;LZjkN{eh{txoiQ2E%c?_}4#-r!U4aJU^n4vM<
zIpd5duHv(C2r|{Tc*|q10qN(+5!-bgnCkzyZ1kzB8*u-}E#QwqWgOA}{hp28|G)Qt
zXAc-kcR`>#qw`Q-!?_z*$B?Nu=d~PBx{Cvn<@7%sQk;TUS@qePX$01;%3iOI&mkK$
zh?JAH6C=D<y2^eXR1G0An7j6WoL_FFo@Zv?j$v#=R{pwRSO5l-vPJb6A_2NV>zmQ_
z?%lio{(d~YXA;!Z)C9~YeOb36#YS8BLmI@^p5w#mH<tsSFwi>{q?q4wr(Wi_-kXP#
z?A+2e+Te5;GpM%FBw+Tt+)^pOSVY6XmpD|&Kf_N?-YDIb^_@X$b@9K$-wm)Z6>Pjx
z9N^knl`Y9f?lwBsU!Zh0XAmY;dp5-qe=qs!Ygs#j1)Hp#nYyoy5A_FBZvP{anJ;hA
zuR|75qjhyei&yRbzHFZ0h#DZ*IU7m8a^HLRk3=VL|Cmbv`$lZJ{o4$*qX1Iu(dk7H
zWLbyJ*!|um*;odk!x^5@Kbuxfb&bAMOCF;@fbRi&{fg=L<Bw<FH9FpD`O|7p{kl_i
zXS3pNSt^YA-!ToAf6PT~+qwVq4Dh54EznuTUT*p(A?*bI@2-SiUfSkWq7h*1r9QaX
zM>EOCr`Db-p7h^&4)Oq9KOH-zV-8ABc;NK=wyZh#i|?;mj<!6TlDx4<)+eI3wgcpG
z{^=6fuoSW%LxeRej?Wf;KM=TI_3DyWS@ECefty$TK1K}f7>2Gnav4AyPCA>@xo75H
z68(Ol1O9%Gmpl50^Md}KNFC?zo51htP*7x^#98+)ZhJhh!5I};^R)%y-CX&92C7&H
zEw5|%r<<!M5<!txoN2KA%NwkFrL6z%PV(EKxP<?1_c#O;$;O!gBbVE_9@)q=B~o_z
zTRH+g+z+8JPR!4V@)v`%>3^m&$N}hwY|0e6bIuYavIIOtO2IjNFx<>AVJcIW_WR@m
zg>^P*O<r6+nLvVUB3@Qm)ie;vHoD%!G3e1Gu;Er@&a9yi4jgx(zyVFC0!y1lYiSIv
zvYgWW5wfJnx7cG6*0ROR3tpo%+>Y&9ciw8C6T|cNWfA||QiGc&pq;IlL)@)#3{&au
zjlpKdJxm>T3+7uf1_ez|V2?7uhYM)~wGC9V=Vv|nZ%;k#QlqbNA7pr5%;<Su18utf
zRV*>6$ANVGd6T|<V%Pf(qW<v8#sJrJ-%yTX)b334+Du{39?#uKrq=8X?CZ<-*#o~D
z0KfStZ!Xwm+oD9|UVIN2*<9K}+fN0!(Kq~~KH5qHB!!_ounSa=V|`=3*NFl_v8liI
z1>V0M$QqLW7<#N&=UB)FW(jh-BvpCfL_c{Mdi~b#7StMHdayH0<d)&we@-=71*BL-
z?{De3e{N0rUvj4RqWTP3-i%#~+&U;n{813TRbY-W1}KqkR;Ss^&SnLVf4ZV}UfA=1
z9bz=2PyU?`$l)1C(RDN)rpfk><*@0&zY^2Y{tj14#3;afbH?siY4J~U8(tjI<ktum
z&xx=9rF?6$10)Im4GWw=D>H<ORC@+{5=>M0oq3PPlgr<oyfV5{dz4W2{+|g?2LsYr
z@DQA1gaiL?@m|t{B^NGS{;jWQQ~x8Rj48ChNKH!Q)53oybO%SY!_@yjGW5!i>q=eU
z-<U9k1zbjo2=2ko4u&7m-~W$ooely;?t8eM$KO7sTLiEwzZ;S@8VY=hqDKCs)+FB-
z&W;?L4LE3Jf&f(wUjAdPRsDehr-t-Bn5d(Cd`)70{IiS*=gY`pa(euLqkun93xI<P
z!2-4Ea@>D5i+25tDPDEK$?~5Q2Mi4f&_@595#pN)$kqB(<>k897|HJyI;^2MsQ@+l
zv)<}RVD8lbtT~rl{@B(37o>i8oj0P568_9GaZ}u2?p>k1-@%o#9&J&D+tGnQ+M_66
zs@lJMQx#}rl%Cw=0WEJ{nli8-jZTz1sr{MKtD?rtF^0s}4hcX?aXn=!9aB?&Pa)8Q
z4?~6iGo8fF>rUYbSJr|yyY6gERc#)7nypp(-;z|&+YTku!~QyB^Y~+A0^`Yk@T*5Q
zSDwp$&bMXEtXnu&Vf_a$f1HRmIQBVsCN$JjwPthjI=;64Gn`EXU`}6!+FD=~Qhzw3
z-z`xdH&r{rkKRIVfcWr!uSnIcXV3VB8^jnmN>%S%^-qC?3+qEU%uU#C7DEs8*ZSL3
zhR`>c@AKUM(R&amP!Dpo++^Am7<udaqZGXapEB#7<D8lag5N`H>i~MqVH9w>nqV5x
zHn%z=vf0%1#{&HL#u8{Qx$V2M`*F=Nh^!22)^$A+x6w5Ek8A*2!+;X`vyAEYph)NM
zz#(dBjDbIJ-K78h&7jWGf3*Q1x`Qs^A_8@j9yoyVU7z2A7Kj5l5zQz6$Oka@W9Gn|
z#g0ed9D5wab#JAA4b$U}?6E6sI0GDkKkx(UV+qV@iF}I?Q2#HDN#!=EJ{v;|I_}n+
zoc(Dl>$_p&j-Qv9Q@$<zd=pZ!b_JPItSBkKW@5v(=7Rqb%AzWpq3q}oQ|#G9c@ruG
zfyEpV89DxuL>p}!QCho}GZDcpc=_3ur-gt-N4d}M1t)SDiT`Ps`Hlw|@6R6n^C2_t
z`hV=`KlcX5|Np)J_xAt*AV6X>zhy`rjF7eBKfGm~;OD-p%N*qU7F=!Vk5j~_?q~wP
z4hnqKrYP~}*aJT+fB+#-d+nmBM(C-}WpZ1_ZldM7&Q?lYw|R~^4(P|%#T<q|&42uS
z5Qqy3bhdWU^k2HXqtQNKh-vxU7TojhW_GuTV<}wW{~Y1wv#Sy0Db$Igkl0a%Sbk?V
zs`eDUdJ8$$z>DpqjQA<;eH5uWJ!VQ!Chhqgo~{3Z9@E%jB4GW6PHg@op?qdH7l3$E
z(d<xx%S202ul;9|(daKVx`+FW#V@vJ^PlWBqmLc1`!gdv$&9=-=FbTE6m*-tpAaym
ziZ;DHK+9FEN4!8yQin3sE9=^2vw?m*tp#rP`-^{7&qw(0?RY8q$>~3i?)&fUVY>om
z5a?CwXKNoYu$|smxv|(ERkq2Ij01v9`crBuJSC1J9fpN(PML#%MMV0hsOJ+UV1acB
zn`z+We)T@EB1%1XTL<PhHX82_)7N&0le!MdJY5p35%p&Y&E7_x#F`yzoA&<qumE0P
z1N|rv<z=C(i+*TgqNvsJG*LuaJw=$Y=rhje{9SleZ3lG>&SPb@y=FIY<JF3a?t)11
zGr$50`a3N-r8e{LWN&nVvwkW{Gwu92K2UN=&fm+){=~vj^oRyuch>bhXY6Z9Gw!ko
z#Aa@7uN_SjfQHs_9%pr{*~I}1l16VlFBCJFB96S-OgXqVym2T>`Ej#7NxMs?d7p_K
zEd9#%7h}jx))WGojM$ERm#!#pO$w4UsXGa8Uh@8NIY86tAi&zW=%nr(j6w9(vS)sE
z2$C)zMzDu=Q>CAz+|75?d8vhmFZ;+rniuMHV7?!7G9^4k$;av))%G-&xREQaG+sSl
zFsXNNO7QBOMaoX}?)c(^WIycqEUhiR*YMXoY^RjjwKKY-5>h;{9noB0X+P4K-Xx|y
zxlATU;eK%fS*`ITwMl+i$eompOwt@B7xw@9#>Tv`RPagk2YCwlLW5~r+l+U2HxN~u
zgBz}po@^rP0Pt9h0=QDt%XBZxJY$u+$q6&2FEv{8eZ0s6>U0t}lCjE=k=NIBW{RjD
z#J`{TzMlc%yA6YSACE>0BB}fwI&u#%?HQO9?*kq?CzTAplc5UOSx63iAXQwTa@|gl
zCk@Sem+PWiq&#VSp(D_03)8-nAL>W+cn>pEt2$e2GFyF`5lL0J(c`*{)m&~Qd5nn6
z*P}Rp-(bFzHXvM|J5M<$Z>YRuR8{oR*5j~@{q)9Bn#4qSZU_iu&I0J_t6{rQuBGJ4
zhiqQY<%a=SfN#(N3X*Po(lv<X(r$iq$#L4M`*rq9Hd{apfoySN0%KTR-LJiVpeY!T
zq&RR?;LRRidbZ4ML*Ce~Z(4<|ft{isayEbnCet~+2~z%)PeJ^$v3WOF6TVC7DU_L}
zdbB~e8n^32!2E5bZBFG}ZnQ#=7++PgkXumsf#gRrIdLg^##Hj`$O7`{yTJC3u#(>0
za1#&0w|>fCEkMUPHu6&%x75W+7oFwWp8gE%akje4P;TUP)UR8kU`|n2><RBV9ipok
zh-472-X6?)qhG$UQ&d(&<<aEVaH5S=XYzrZ4{{2Y&wFc|H;|};GPW<rOws9*adn5C
zN!p72+x|P9xrobYn2E$*%*X1(MMdxd4CZ?BtOdnlt<E6kx>>{G=;zSyiMfnJ*t@OZ
z57P6_V}8JMn{S9?WcNyk62RWrC^iGb6k#+Yud`mfI(T;dEpFdL<5ES8>2p$<@hYxO
zuK)+`A?n<@;W_Pw-0#~J=>_KQlv6ah5`N7^>6TH!n?*PvYW&~C)aah|z{TU{xVQ69
zm%NlN?2FIzLR>R$T>>alHgf^phVj3ZYjSNCM;Emgswwpi6tQDlgUJBo(s65VUA1f#
zwQ!P+9ebPsxo|FQj;T4$a<D|^A}@({oL71=&Rl0Rw|S#v<R8Nve@p`Ubp7jPAh`^C
zax?i-3SPceTU*-{N_zSlNR>^4i<`=dT#kTQs;PSKb46XX%eHP1nTPJsBAFb)#^P>@
zr_eX0%)hL18MzWH_&KCf=LI;0YCF2~93P}4?|VZGSw(6FWQ(_7YTox};k=y6vYhkn
zm>q06Z&1l~3>WvbfrnDv($m%bd>B!3xbdXNhg0u{Dj~d)RN64HIob)>_5sXPd1Mqc
zmoy5SkU^n>cqZ<GipZe9Nb&5LryjJ2)z|3USI|3KHe@<!e`>qcF2Ts2)EuonwnkOF
zN6!~6SyN(^C{Mr~boi?ho=)rf1j@<C8P^<bUu6B#(-Cr8qG<isGhI#(9fTApu;a?%
z%3M6FyL`2U>24R7kZ8Nuisyya+8uB2@98k$;p1CJ7218c^qm*~b+!e^!^5lW0a_sk
zU~h+h&TYC>mfIt{;dQ))Sy=aSQH4@lDjSJA7rW#8*I}XU(xT3yk@4)r-w2q09^mdl
zE{BqnF{{lXGLllyAw~p9X$IPSe@e6@L~}~7yUA(FShHCrS5Dk^Q>oO~or_J5Df1>r
zv$u?>bA6s_gs63XvTbpB)z!ID`3tAC>{nR_i0hK7?K2kSt&Gl-^66dyWxUh`5i>C<
zX}(sWs?}k9bK&Uvb~YeL(DV5j%htQ^JXX&pp4}(|N4U51yIc~5dZX5wZ9fu*kPWMN
zN<)0xjm_U|p>>6I`|qd})6V#r_|%`T8QWzTN*{Z?t5A`8F>@{gwbZNL$WdL*me-AI
z<{usp9~~D=ijYr#^ETt%xM+RRxtqJ=(MXREsWi;iY>s<w!hijRJL^b3X{ndlVt)$l
zbKR2?Io%%?zq;l>_?k*+%eTb(V%rklo@yNM6v7IlFl!GLyMOvUNEK^ZT}WUuJn8G%
zbR@Q^q;dpjPCLGPJ9n|fyN4ZmYO@uIFU?q1*>?!9dwUDk1nVCDaeW9-KwDC(o(A6x
z)5fCRcs-qjWI&uZ7A@x%Piy2ed~IJknBF}qnXK?ZCT~svtkv~v^wp&REy&0B-5qSY
zZq&3tPhx!Mjo0KWY;8#0E~$HzLD&SZ@S@kNzb`DNc@W~Y3Z)|s#Z8=L8{%5Bt$Z-N
z;YJ~!8}n4I#jlm4iq1A-?9;=irGVZV&~<B53zPTsLy|p)4z=o7Ntv5~ljwRH)AYsw
z5P(qR>$swC-RVI-ZEChc<Yyo?r-7X)J+T7_aOB(jO|NTA(|qIVv}o3Ht#vPEG%W+}
z^6`5|Juz>8$n|nd7|iP%^Fp!n6@@&2o_7S8JsCCWJIHw%7#VLd6Q@ZX?D{$)j+tN#
z3er-_eAkPX&l0vwWe-YclVo0qZmxU1^-itaU+I$w8R^O&X%pEaG$*L)e83UbT#thL
z*XV0L;CZll@_MtZ`2gzUvfAIf=OWH9+@?`Nd{lkov5sv|cq`g*G@6iZ0IvJ3wx77|
zbPw@eDCjkWIxfYR4_3pzNsyvk8^@SV6FlPGLZh}TI7r**Lz>oOEP&lCX4(wZ7QQt<
z;k!O*{TV4OQC$Q!bJREHK9ce0^8Cwfv3Jk3#=_gk5ZiKoYt_Q`n}!UOb5`>rc7jME
zRHw$a>$FwAXxiEI#ldr9C?4B;0504e1^mrBbW@rGj)X>*zD8)wiv_F)vxG#2OM0(#
zn(ke8XCOpNoAV+)VX>~vyLPV4t|J$xQ?=3;>yxIJZFD)B_pIkn5K#k6+Vl8>Pg8`!
zsh%H4oHr6OOz0k(HhIC<qFkFFD1-fcJ-zZyRu0%|Yak*rfCslW2cC&PaYl!Ergldl
zpP}c;^`wlg%irCt&u#3a8kLy%`G^k6*#KCWmlZ6dG*%VMh4&GZxx0SCWe$pap^LNW
z-!d9lXHV__5|Uq%%Z`j3Ejl;~nm23^ykUBmvYZg~ivH$8Am4ZJR)hH+<XyF#x`-mj
z%ND7vi6crnzeKxNU_`@ZxkhRqUZ&nMe^`B9l9w)?p^hNK<%c^VBO8Z@Vb)&@1=D@L
zHsLz8Z-0!%dOZ%!KH0Ah%d-~b3Ka+ZJ0r7H+?WX{iX<*3RV>3OQC&ZBCk>F;tSW2F
z@g9PhdFAL~Y!`qlq-j#xkITQEr3Xbj#Tp1IZwiv1@s({R`oO<_m0YL13?^mar_p8X
zs7j*M)tmVgk+?JsfJiECoZ5#_R^;It<y%{1Ab;_O!<N)7%D4etqq<YyDf^!*+F3tO
z%~4T~9N+q7o~41mbXim;(eHoId1k4>7<(FioA})(1ao8XfyM<7QU%-eJiIa8ki50*
zVLTfiV}3{m!6Mwfww7~BrzrhtxUt5H$v1Y@{rX}@d4k2Rl=GOJ+!5FeO>3z5G%l^R
zUZ*eCBel!ESIMi_UXnrXQ*G}vlSr=QFwNj3?^1cNcY~JU#`}eEHZ+Y*gjD&11EJPl
zrV`_r&ZZY9gGIqv;DOL5`Bjbwt|7-kWvuSiZiuYZ_&j4M+r!QryTvEQL~TS8@oD`$
zXWMa9t*DgHL+|#D)C{C4bS<m3=q$q|-e&(N5i><0VIoB&b}U*uN2U+5U%`nq7Ct**
zo#;K+pYY4AI*)*sr<uAGp(pq0%2V|<w3$en;!idjCuS=>^aPBSV|uLoU3W@OiNHtO
z@v%rMhEr^$zvy1dXAJ1hIBz5}0Pr8*MN@~7$(;<1phE^P!wfZ8)nn8z?+48oOpo)v
z?UUl%s&^Y3=}L{Pk?|2<(HCkG3p(H06dJ_QSp8`7ltD%bPq$<mA{@k^!QFHE{;ZH}
zw<IT*@|3%@|0e21m8~^FR-^UH%EDbTROLIAwj@fNUhc)w_B5H?UMSq9Ig>|=67-|~
z_Ky~d@t1Gi<`xzfy5@4~x@gyZy#{Ku&A3dkwL?Q8X=!2QhRui5VpZ1Gb)56t#^?A;
z0=WKLwSE5HPyC%xI>R-Ks|h024R>UTgsg=zUcL9r5<mU|fDMi<2LRSWMA^aMbYmTe
z2s~1~HS>m6By%k)465ip_NU*393;P<8jPs_0Z2Mkp4I5+^CoC+x1uHm{`;?wuA-3k
z4LWbFLLEQz6^<*k7A;~4`l80yB}^+*Af25ny^2<F1*E&Xx0}ccUQ>KBX?XZ6#h+J)
zX-reQ=Bt^>*+BZ~21FwSfPWKqT^IFo1Z#&nJ0)KnMZkkkQE#s1RH*Z_^#H#8*_*zA
zTwe3yUTl}ez#gh!D|U&jBQi45{t{#IWXAQf=cs%fFvqAJ)yTU%#8rU%^8Etd=qeyO
z450D=C5+OSHOa!&b4;_PTXpx|MmEDO<{&@=l2Tft8K1l*7z7oTdha0sGu*CXC7jkZ
z&|Z)aNo6`17IB&Oaq3v#=N9B7WU^Edh78DtH93jz$3yu<{LP<E#H2CPSULOXg4at=
zOB@D1|0Esi6>7Juf73_m<JlxRtaEmlPKIie_BdiYr(Y{KK6pNa*SLGqvPh_VE%u23
zg?Wdk%ZP}AYD(R)1VRkcL8rLHq=Bh}9w6sLOV{$GFC6YKSocRjHt1d^7`d)SAf&`%
z(9Ub?U>S6u_%u$kgKK}CEaWpa%lQ0pPby#7;m+J*lDJj4NI|KZ*ild<J^-ZEkea%_
znOO<i)95k@k?S@MyQyvRfBo&_33*Cq#I{76RIsdASnIHgh_V<MH+`%rmT2(3t+pEC
zKO2L#)Ev4MLe^T8Cezioz+M1gsh{!S<}KN<-gK!(yS8Jv2oK{+zQMXiSR8}xrxvNk
z;|%a51hv02Bf_<z94<}Fl9D#Nd)atQEL;~0(nbJB{zfui+%a3d=|O<Uc4;`AYw3z#
zxp<uJb!RC_u50E(uj<WIYH$<S?7^38ov?lG;Cxfb9GkTLr+Ky>Z^Y9uk_V^AqyZwN
zkbnB=1-cCuWv!tG^HNw;FDT5~4nNC|aZO()j?8Ya^tC!`7eslebHpj<YVF^8RCCt2
z2ld|_$pT<`=~=Wzm~)OdXNNOK7dYZfo*O0-Q#1>`qwO*7GY{jFNB+{Dm#UMYWl}%5
z#dZ};ue3g2noO+oQ5%zcG$peH@EE(M>;)RG(s;O!6@`_)>*yI2##U+}qpre%dLS;(
zW)Icam76+@;B&bpDk+&N1xp#V*7#BHdbuSz0SmMOe8<$qzT8$Wxfht_vBvL@0dD5+
zS^)NaZSeyO|D96R+W4aUj|gNWJfO^IsxdvxCBWomHvJ&I*vOr18||se6c+@{P~~fJ
z1H^^ZHOkSFg9Z%*eojuhuk*YgA4JA5n;U!fYFxPO{%T21=~g>C!Kkqk@xV4)e`-8w
zL}kdFOk_l0$ojaVn>BTHe|!U6N-*^$54I88BNFz_@n@d@L7vzYlRES?MPt)uyA6>}
z>#!BI7RMv%XqH_Xi`=v!yz(mt#9AK3e<C;ZHAD$_Z6KN^8){7?le9H}AO(ne44Y$I
zmGzDfNfJeU^jDhtRDxdyeP8RJ<#qD0d&|MQk>gf9SXG$1+{FzUSPOVLT2W;bTkHXR
za#Q*20s<{(=eUq3wpw{|8&5&77u;2mY1G8Vv1g30J~%=HFuSH<BJt3)sqHe7sB+1>
z2u}M<M^Ef+k!!d%-?h+Rc$>&~pSl7*q1Bx`gy-JC*}I7|$PZnSAM@I)BHZbBO%6S-
z(C_#P<Ba*$5uIDILuboViovl1%illO&tzCj5AeY5x&MWNPXc0>gZcnh7SSnKDwYKJ
zp{c&lXr*KLcHft<@z`1JeY|*E&d}^HbaA&81Vk%|s?cWYJiR<orj-=~o8|Q)m7YQA
zZk1K2Q0bIDJ$9^}_F*YGZ3Xg8maR#%0K7v>=%h6B?fyOoX5=}0N(+CBOab4%+I&>Y
zkDy_cMyF-W^#gUCo{OX2bb_A`r$Yjc6sX12!rG%oHrxcjEOZ^wXnlz^b!&knK)x$m
zm~_)qaU8mjcmY|<>tf(;$~GE8W`B1(YzD8&O)C2ekY|wlXRRELy9hl1f70^QWls;4
zd#2BhTP?RrkbV~m?s2nkJ*J^fF~zzrg+Owj@XHa9_7V3~zG29?6b}ln5po|%MxLW*
zJx>ll56FqDmLqeXmtdV<r~5+gMUly^H!OSS$MbEZiHZ4Ovy1QcbLbx?%QSGLN`n0t
zIvodsoVx0d?SJOF<_YC8yi&zH@@^P*21Bwn6c@*b9N%FIEgjq+Zhc;hD}7i3-ANES
zBul*K0?XuK2ql-BqzSe(KYj1Bvx)SSths+D%9)fkR+^*pCgzh?F$V?Ok{mQBha+m^
zSyc`#t8#pGFNh3JJ(bu@i&R$9d+Fk&tsmmroj`1M9J3mBUb{<b#u46hJ@FZ>03B>~
z>2XlghJtgryyRH^I!D%U9;I<Jz;bqbvhm7f>o^|V?Kr4eGWy(fCER9b3ie)JqUlnU
zp=$J{0uQqD?tKO}pU_pI_1T$%cpA@zduXF3TiAnuO*(zQtvU2O57%?ic5y>plP4re
z?1m%riEcFXmR75)kB-LKQ(q)<Gtp87wq8fF=H0A>7h7CtYj(+yRo*V>8a<(woEoFB
zU6bnn`8Z3OIl(*S=R-Y9%=a}Y(^Y6}h^cb~@KEwrO)qEa%_N;?xmtxMs(kUPR;t}Z
z8rRv||CWQbP6)owVK6FK|7n(^Ti`(-v!uSOuUK2*xvhy*kkf=KXND0(@ZkIe*Sg)i
zxM#~Aj~YKo%{seOu<i8CEhgP$wtY>4cHS*LPV4tirjl-nX{@=GO!?y(WP;h9$?AV;
zbW=6Iu^RRF`F62!F6UJ+zSa;gsiQ{UQ7)r$@tkZd+V8lvc#{+bVheM%j<jYPI6+yi
z5)Eu#ycNALxHwt)A)(*jnAb7Rb3UBQoEpjmJC1@+^e5|QFYJGKSQTZBQMpFXx9|QC
z{E5Tw$PY}xNG+Asz!H>iI6Uqcb4mxU&&`tt#kNMSkj7PNOeic?47Ig~l6j0z&n8P0
z8?YknwwnqfbtS;VzABz;7YXuq9JSP~<~4_8pv-sIod%Ds2GW@1P7MP|j`yaZTSw-&
zxX2K6Fi`PwI>Bw)@C{V}ZNW%6GI&Wr-g@lVHCiD{r0cCO8=Bw!eE+W5UFH+Kl5741
zB|4jJ*FW3u)pMX|V&yvwSi=*A*ZYi<GL__jzm|Y#e4LLBW-Xfh%Y3pTx436#^<z<S
zQH|gfwAyRAhpqwdyiTfV%WottE$yJZ(X%aP>T)K4jC8+ux06h+lSc|$Io`a5hkuUX
zXwD7hWf-(K&xnvA{|W6*IieVN6q;SIBnicPYj8@_8oaZ{eL5$TLB!&Dv#dE?v?{09
zE$?H<)1N6x_Rj0uNnUHZN$K!AP{0ABUyRl_jN9qB$&kAi&ib%h?<Ml&%_+n8o`RD+
zb{WcAoduOAIL`wg91K_aW{#RRY?}rtC473?qZSd>qOH&#MK-^z>U3xEj!ROV(&ChM
zQq}h+A1&rM8XCnIO8lqkGWf(vO{?AlNvh9600qfrN15PpI9>T|4dTmZRHGbXF`~-4
zp+%igLqr(|iWl|ZKssB8z-+bbVSr5P`Axy6nu?o{RnNGx)v{(v_Zos~6RYJ7vep~A
zaooOk=*{U+##DZB&gGV*A=7l8Cby(%0T!fsaf+NtlGQPdP4ILp7v_Ytb;FfK=E59I
z5Fzu3m^tWNJxV&#il@uhLw{+j$>yxpdE**iR_gglXJ=daW!Nbtj%aJRn4p~K9U19M
zY&y}zD0PJMRFll*ZZwpIoYA)tyB=Rt#q)^$s7nGy`{<5b;>b(8Ui?9vn&6igW)#x>
zWnVUKnK=+mPk+_kR;)P$B|ZQEETGn{g-i|u#I;48u!5e@)V<JMF0X*GsxhY9o;o}@
zAU=~TQN+`rr}GvCzis(c+JbQ)b)1v=Qs?$Jg|dZd?r*14iE^Twt=d^^ey{5~ZsNW7
z@#C%AbPQUPb&c3*W%E5N-YJX=W-RwUA)dlVfSUng-^-k~J7MwElMS%)L%#YY-~1v5
z!|I0n<s&w{$ngpfwX_XHEf1r%XyxnrvW5V(qoL`y8C*j&tl4Y%IQxTokK;zom=IBT
zBY%HKd2^g|ntZAVXXOQV-`3XH$!FijSGyU0UOuk?v0BkC{L;NbdQR)l5kw}hDS7E7
zY30!_VGSVVuC8PT2chQB%fUh&j?nCEsV)WcpEUGctGR1Mt4Tn_Tu&N|HsKlD(Mx*n
z`GaRVq=6^_Y{-iHyjSZZp6!75##BToR_m*VUFfnHz~*w`GuaBgl;%`7L(gJt;>_3q
zDOA-yzk>oo_Mg^HniQ-9%a!@rx*dw+6#O@<g(hHs3mvb`r??tQ80Q{do{$Q0ex%&m
zTJx@3wA{v_(~Ic|2KN&Gx4mdr)0c*=1boilYS)^;Z5<7^8=qQ#*<4pVYoYm_`Vd3>
z?EOM@jUwXBgV*L~#E+-SXx|p!mV0rcxHwp`zo9f<HqY>xjpqWRc6tM&cwcaZ6yvp`
zdBw*ti66gI?|LIxj+qeQraw^R$ZVIs*xBd~Rxn2K^7i97jatvcPN#Fr8N?{DzosPy
z7&8viQq~4wRtEHK`&}VS4ClpQujQH{HXdUk@cfHu++=g@)%!{^%E9yTbwY`f@U!mD
zoL#EDlRkLcsdt%ax+^w{gVi5&r=Poq(v(te;Ka>hlzUF9%UnZIZU0Gd%~(7i4ruWC
zSbI;~sL;vy)|7;FUs^1BHvMRe5T46Qyp3phdzs#Hf8fy35LULn-j-T(KJKtpURX!j
z+kD*6e?W?MG_DIU#J;L%9$T_loNO%t0CDClEJ9lKu6O<9ZehKr^4{d@?q>C?=#-(r
zcgI%T)qdja=Xx99Bc@Fl8@8`^nKu$DFtO2bk6U4DD-8h|{M_A#vuqb$wGYo}S5R}N
zhp81mTFJI~E}y=V+E$Meb!MT|vK$s!dY#1Q!@DGUntj$BM};UBOG$_Z7?9XJAlU}Q
zm#8(lhxs~jn_c=y6CuEB8{tjgKe@yTVf42iRI9RSJ0}PuF~h5t+VUt&aF?V;b9YyT
znZ?INFEsvi3lIYstnxe94yZ(7{U|USpK}3sOI)f34O-4RW%}@lReHRgNt!4lBEdIG
zkY6oTqf^FrE&qy~7WZ8P41Oe5sA%nkm<P~=p)0Ax0s%8e93cyNgp#wcV*1IbyO&qJ
zCKM-+V*=Lg0w5x<%~u4nSGT@6&wZrkVKM80Gyu;PhY?TTU8N7kW}gC(m>}2$sU^8?
zU?+;aG&E&MipbB2Oy$D5`G{UFgby&K?w@`T9KNwZH}p(cF^zV}I&&(<?C#KEuu$na
zK)u_W-rDt9+@PG3Fj04VGd<=>@za|A0{)W}JW45dm8pxg3_*>E$PxUV2=av&QQZ=*
z&5@1uKbg^<f*&-yDRCF4i+h))-{nm~BcuFxhPR^>)tf6;BdoF8HE_)DA8Rm5{q0$Y
zC+4E#%~3VZ<EdBuHwU9^4MZ+e;;WBdqlmd{goA!9F{Jo9O2uU~lgmBJmyrgNLpv#^
zC8asC#g;Df{l&Hhb&<RUJ)6+n^Fqt0qZL2sh%<8XzF>OyWn}@d10d>R{?-1rZzA=`
z#q9;i#b4%^GYa_Q?mY8ed&2-JlEKR-Ite>pyy8Vt97<<|(W|p5LPw5t46Nf2uVtu8
zqm^|rBmFkm4qX<#=y~qB^KzOJ!D6@a(df$PMp<+EpO?f~@j^UTw^-&FCt>b$7&19)
z<kD7|q1#T)XLX%7nT}hs7-8aDyHl6HY|s;MUuJ90Sdl{7^_e-?x5N1^L+9^c&)0pe
z42G&Gq24`s`ds7LG3hn}23wgn_lm3XeKe2DDKg*q&Ca9Aq`+1`!_;}ddsSswsO=Rt
zJ)|!3F#pvA$*xkkTd%iTzoQK`dsVJ%81$fcB8<m0z|kRx)V4^nTy<nRbZpuvDOrZ*
zETao<sk|nXY{Vb_(@E!Rmj#gKq8Lyu%k#_9s`H5A?43UpeY^F3^agX#31g@CY7Tqo
z2><4;(rIaDql1gkBHL=a_Cm~ziJh0C*=eCa{`T{B9*(1}>!Zl@8Kyq<U2Z>Is|1bV
zC(3Hw7ETlxB^<Ui0UG){%qAq=Haibjww2V93d3Pc7=LzAL<#1p5NG+vfsg}0o~A}7
zRXd5(7)BUQ0w47sl~)>JZ4b6gnfwnmpKKM|kC(hudmDXon@F>OEJbsC&(26hYS#at
ztY%=$cjOqLA$R0Woen$`A8Q#}tu%qPkC3k|gwvbv^(s5i@PH3P#rrLU^*s4&gYb@N
z5(hCryq@UFFO0A6&cq@9SHSd`oljWm-o;r)v7yw~>{afp?Vy;JZyhg@;<vk5WK`zI
zdVsmWUUct6;TXfcE-hW-vl=n**<N|7=M=Re4qMTbrl|Bs$9J?t<1(u}zWc}c(q&}G
zWX#EUA1?-lnZ7R*sPup`cyQr~Ho)Q$YWoA3z|9v>*S^Fell>RF3G(K}5qd5Tw*r-v
zLHIjwoSc5G7&dxB8KZcJ0LseCyZ;po<G^AL4GqUl{hMAX%Vj|~D~I0?pOE4cATBv+
zP!V?PBS9+C+)5Rtkt@Z;?MZ}MoG&w@(~WR(&++k&(MRo4n?`<fDESEwcip+Qr3f+H
zJ*xqe%Efu7Z9gJhT;`y5Iwr)I5=W1Tk2`dV!hG}+{M5GxB6lZ&yv*`(D!+e~wWq@*
z29+I6o@j<0cL?Jk`D+`IHaFfHz%@1KzB4P}$xhc(oNB(`v7^$UFu^&qd_j4KPPCB|
zx!N=5o?BOq`gpg_b8p@@l!SOx&*6aco0@7oBGoGt2)f)Y5rX^{a%^epa2gu<d{%ui
zOMV)<eV_uyxj^y4ToNgkUPCGODVse_9U#lGBZF`NEdkO`j%`l?Zt3!v<M`o}TOkgm
zNPbf*P>j;F$Q<N}@I42Dt726WdF#P3AN=huA+zQ@r4It4-JFyd0Z0I=4IALKUQ=L3
zm&;yI*-p!kK>c57g&|ro(gGW*Uo<+M`<z`~AFFcka^;L^F<ZzHB;CY1bY0(LRMjj4
zLojhONwrOw8>M%0ITE^#EC@Wuv@oCVFDK?i@|FP{`M1_Mi4bstA-mMY&+cE#&8?K2
zFCkSD)^4tfQgNWtFVo_JY3CRcI=J3qylVS)rkj-&h57_T($Ft}s@cZ}8yTJ+4z`eW
zaE`skjM*_3`Q3G0l+ja0i0)NA;WT+se~`!TNn}%@f>bSb2wsL=KIm=uR#q$RjTFIk
zj8{bvFsF*Ucr3uckC!{dx9|-fohg^vHBrl@72YMW<?Pl0I^w<DFHLFV-7Bxr%sjHi
zzHbl-(p?^k9sG0{Vc6j70^@2pn5uBU3Cgs>5lw8xI~%{u?B>6h#>q6$s4~_`-3mCH
z^TBq99)=LKZFkuW1U0X<<-u)0ztkwS8ZlrOIHghx^K-z~MyKen(o#>#TqOcKp$9r|
z!bI-UfJD!yx5UpCI82+IpF}Gwi12xSvUG7;gSNx(dN_Fbt#*Rom-uHQZgGvbg$+#i
z16JlKN=Liho4S?KI1|SC*7&j}=hL7Tr%TQEbF?`fHnv%3MAA=M*120j1^TIenui{6
zUG@D2<gsFG7x@%fXtqF7@2_7@$F&_vs%(^6fGZkw_4lbwc<AM4>qPz7@~g@o^Hf(x
z&%Ez3p$wn&>BW87LbHKV9TOm%;;sF}&=-FjPk4^4>MW&1^+e_P<w$ss%>ZO-M~oM`
zzF)NN{D|h>4OLt1qd>gnhtg3!VTBwSR+IOq4&Vb|$lju1IjpV*dd;BaD&#i8z89|H
z+(WL1aURI=?`|5B%XtAg;}vfF-zgQ^ndzy^#@V610q*h&yi8rIJl9=uTKN>|A_*#%
zpqGXCVThlPO~um!oD>cdtT%0oEBx`M`^&MA$zk&l4=xEE{*TL57H1t7#pMA7TYmRC
z%P>;q^x|ofqPZG-^}HqCHobE<rduJ)bKp93jbL18n%L}l|Dp(T_9UY$bt2s4ngh4B
ztUWx}G{9%)^h>1L985HTgLj|?9r=B8W9^6}aa2GaCZ<e^=JK^C=BJ;Nf|%ZQ59ja4
z1!Xegh-SL(cl(O0kjT(6B-Iwo-bfsJ=hbBvWK3rg;Qb_5581mzY<96b@%V}=E+b(k
zKHl@*`mt3`#SmH#bKJsAP&qc$t=B^jl<l$GkOM^@Q(IC>#E+ls4Xe2yFIK}_iurK(
z`6sJZsDC_lVAU%_>Z&Kp)HRcTW1Z%1*eaSD9@5MhgY0+E^-xVVdO8whRaI4;iF49Y
zQ`^`FSARJ1D2!bdL57}!Q2@e><c%f9X%x4=qe3#}YiUvhlYI;cAu<=estVc67r^Ww
zU^FtF76J?m3{wz8=X>op3W=7iWd{7xciMoD+3Z(2XrLCLo#B&&wz3%*mG|PIplE16
zeVA`tnq69&RXjJDXPzokOx0-UF<;PCk_84NtG)S|rq%(jsA;Tg^~yEe=desxx7Ec{
zEO=e%Lv%iVt2+qRNzVSnqib}&L5seeT1Kugu!|=to7z@M(7u#ZgY+T+doH8#gH+i3
zLIC;8EJG=i*$C@$S*%tLZlS0J5*?mW-mX@nX}6f4HtJ*QpF)7t2F=^uEk@{%=z!Oj
zC(dzGYRo}#-Wznx)6k_dA93QxhPHMlG4h`<sKXaAT!bScK+^R4zMr^)?N%>es;iLD
z<wCE4p}kkeyJO0zl1>SG28n9(_VS7Cro!QdtFoYx)|Hr5W0iQ%U}N5GuT!_|+W8M=
zW33l)*H8?50p@#lIWi`{O($}oar+{$RpZ|*@sw92%R@9Kgm(^=v1UAc7=YvO=nMr^
zBxt&{#MjgXlL<TQN~NFN6KXnb|6)j-d23@a-Eyut!_;ubpl&7CF=+j?rkHk$8x+WK
zb0O2%-NU0l13Z_VJK@pCLoX@pwKRFEcwUFeEV^b4se^D>Iyv#BN}i6t!!2jx6b9a1
z=-oA_{ASk`L&6-?(t?R~WSESTpmu(R9d|cd?rC>yr<6m^ic!)Q7~Oy42G@Jx=tgWZ
zCGjm{T;GHE)qcwt65lwx8T_MC{keGBcLaA77PrUk%9?n}WefPbvwXT<KC#>#Pf1fe
zjNok6RTHTk(-U;A>&=TJwEL7V-0pqnDmOq8{)S3z0PnHia;T&3$ZSI&B8)0s!t?EP
z^&9Xx+R^F`4-g_;)Q79G#rn~sYtIt6B=RqT>PJlqAoQqjUtC=LLMKcvbBoz8Ynf>O
zgXkVhK-bb_n45UoEZwSzWVx)%LwF7A!zjxtdDLrO!iS0NbL>^nv4L7CAX8oGRJ*o8
zH9!D?892Mi#f{Yt{vF^jQs<+6Q;OLxR)Lf^__$YN=OWu8p_&rTnf#=ah;NFO^mVR)
zX!GEH@b;;}(ZeFMm1R4>^mKg(@w9#mfhnU-j$FB-Jd|vs(<Nun!kK>WQTc7FzT~m`
z?#!up#F3Hvjf|~N@sYQ*)Ug5j1U<zD6@327VE|_vcr*7|j66JvKiAPdDO%qcD2=fQ
z2I83~1e)rN8;9{Pz1zEmhSsHVciwdMH%%E``s&e(OZixcL(O%u$@2o&f~nT>9WA(F
zZ`9W3_EFX>GGGrc{TDCh@;Ng70WUfAFhUn`dO+v8yXxL$S=9{j7#G3z?PF1~Ecr|F
zijfyb^W`g)xYYPSwDi(&*E9~uhhg&7nC;&fzyp=G`;p+7##eCdo)}Z~H($MNz?W02
z8KcZ|bh}vTuQ>X0^GA&si0#oedQ~&PULs7+c(hkjB{dJ$e`GJOTyb>M{tHLu@O}iU
zJlASqHyQe^*skj%8Sss#Op`nIlxGLg;Gp@L`P0$-bWtUt(8IU<dg{l)95V6cUJemH
zad2neh$0si_-TiWEG66M_Bg6;vrqlxQEFOx#FtK<&&Yea-sT6P8+5@vRjyxRq%r`A
z+P*j?YyDm2>0y@sZ1KTBQ-Fdd#c)mqZQ&OVnZieL%B%-$l}Y4d1M*Ql@T%`Xp_Lav
z+W_*WoC`2?Y_S{b>z_n+u7cM#p>7>Yc(fnDEPp$ASQ5XfvB+!YB4Q=TD?Q-Hs4)%%
zxmQWH7$6ZW0ZI0i)}Iwa)GrxM%_VLt0cFOOt+2~;k%t(D{8))5n=gN{@8=Z&MH1YS
zYw7AjV<E841@;Gx-=xO;o8H^1766Hk2V>J3ow><8wDJ6I$Ep<aK&T!~CGaB8*PH1F
z3WsdDv)w+>JZ{O=a7tk?-U!>hc&1+%SByNPs_r#jet|~mjEeZ+pSZanOoin+9tzc4
z|L|f@x2SOD((NbgIwKqwsT82oP&Bj(H!)Tt)t~wydPtLKpS{s;&f8IB*VSM4p}2uG
zjqat>?qK?0&9|o;MBwj5Mk4X$$*0Pm!u`utuwBxg9t&@fc7GGEvDOSB5|crtZUtsn
zMNqo~%pl@$240)W>-TTnfi>?trA)7FdDofHwYX=9PYw{7hNE{wu8pyC%q85~^ocln
z5h|s2yMEG#WnNs-6XF*-1iHaC6lw^s+98JPpn`bNEov&De$HcWb4KQ`koSG*b2Vn?
zJu2uD(+LrPIQctZ)kVDM>uRZ&%uFguouehRih>itr5X5NTALZkdtX&K1@Qr!>7gV^
z#6GoCp)uQMSobTw)%#BB5(>zCc@Qusjc5i$0aeexECnTEYtl3-+1b)-zY-QkMjkLl
zg@g=5k*MU4jKLmHJ<<hA!BgkTM9*2~V}o7mx<A#q0g}`7<4fq}qcat1^p$tLJR#ye
zvS!brNFgr|w$3w^T#n7czN{`aK>D=L@f<LA#j5A`{JgG8KSc#y*S6Fvsdu*;iPPWr
z%*lI`>LBN#lu~@|b%XMzx7mBnIKg#Km+YkH8DFZbTf84W)uiwTP_rElecryOyNLh2
z^<e3XuYie3Di=)%2hQW9Ez>)pTc&)B9v-}E`aG1SM(0K?t+sMJ0jBb+BrX$F*}HBe
zeSozz-JP`BpyPY<E#{PpIY{Bb%Wt(-E+`RYAn7TP(*KmSpE^yqg4O7Iq1NT{pmpY7
zCX*H0ZmU`c6J_27uNGyn;z7!b6xijevj*MbV5mNmi3j`4B#`o8-v)|yjl9NCgu2{G
zbRP6{r&-Ak*E{Bb`dI5dohCh}vAjJ@u^oyE;7@phZhxkJ$%Hx9nKM5BC88K*0Zu%~
z&m9(dJz?a%?5VNq&y899J0-%o&d2vxE?67V`|3mU4cixnLriZApD)!$4yN?!l;jy_
z?l<_x>mJ-V($L%Edhw!<YBY(>|1YSmH2_?<$>}+-rd5{O8%g&5kOY)@blpi-%zgf8
z33#TV^gL{Ie2QjmAVbD;DBBL%>%Fyjkb|Yo1<#N!DHQJ<ZnO4Kq1%gzfpD{xf-&#~
z#>(^v!Er0ITA<4CNsC{WX!N2Z7Sl)8diCA_bC9FaekosV(yOt2$)M;*<NF3v_dd1~
zR*kT2rPmu;W$T#jKAC5`GJNfKDmZrIvqfHM<c|O$o<+a4!&h?jgQP>vRU&c`+P#9-
zOQyX^`vv(~s7?{et<c^DzKq6B*fvD#o+w+IBCc%MXNbiUYrs7_ZF1VxYoY{-vN1qW
z$`v(YXAA9SjOXRMT%s8pkhNQE+P5!yR$!@YJ!CqX-K)z}l=nG7K$NhH=Z(e@LwP2f
zoJnFJVU%wLP&!;<zVMXKOokbmM%IqhcE-2eR2XV|)dbz|(x;HmAf?iX=)IFMsV51&
z!QOUMUI;k(iJ)e(U;{ZGm#1nZI=%uyjmpxHt<SmM<|(<3KP>n;k6yK$*D)*Ig`CpG
z^(Zojs~06E5$3C8R7DzQyo*1HC2Fhi5fs8KMA1#hNdozL8AU5e;N8RDwE&N4Q=ONB
z$=-%CVfLAVMvw~eiX&8dC66Ex8(ANibxdv~KL4A&7)>a2-bht`!{8mW$bh)a@s0}}
zFcf{*<RK8?ce0>}-AsO2BZU8WWSgD7{M<0Jy!u<IOn58&vu%Kmc_Xi072%*k;fjcW
z5G{hYU}Ba=YFzNlQqz;bPK9bdT6OD_5&{Hb16`wk{`B~xraMXR79yU(e{u?P(y*#h
zo+jw)5x{zB8S-<?yn3V@CRG2e=$bos6aRR{oex>XZcO+Z^u>Ac{Lts-zHuof-Tm=Q
zZ*D%;IGrkGcE<g+QuViEbCs9ZRIz%0RcQ+ltm(>M-GFSpXvF=QMSG_wwY@LccE^(`
zwbUxSmd1VI7yw+H>z381#eMTs@f7JpqE*uh%u%OTHF<WQ5`I=t?Qcl{H5bacfR7*2
zpWbC5;GDDHDr&<*;~%51p+*S4x%~D)j!0!TdpEIlC4<T_^L(nnIdhVZ7Btzzz8!d3
zBZmArd=KpM(QMJ8Y1?JwuQw&w4Cpy&ZBg@HFjlJM4pZjqU$#ZkHVd=W$v?g5M!JXC
zlDZvI@{`18`Q_i!UuWrzf=et98n9A1bh3Y*eh@jJBgURdCYz&!G;1<{Yrt#uIxJ*1
zdMzvYcE(EPvpDJ7L*{|WRZ{~@VO9OpFmNi9Z$_1AxmhCg7*n=~uA0C&A5Z)GBUC(Y
z&g)W($(h&~SV}v$^Mip>llvV%AFg<jTDGTwZx;oCIk8C`&@J>^1aEQqH$Hjt%=<8K
zgeVL=*0E<Aj%z5FC{eE?yt1V8L7|Y<pT!oOU3|fCE%}LAj68*ZsVYhyT9g!}x#Jf#
zR7689bdVpErl@XQc0cv?04%~!>1GR!>~<4El=J3-*N*<)$0`}4+S5kt8Rk2&p{^I4
zto5(JArF0m2@)P%;(RY3_H(fuvAcYE4kTvt3&VM2Z`o)3MhFsOS<pS%!!>{>^>inS
zmlv!E0x1N8Zn3|fz9Fd2kXX0H<g4f5F<$L5Aajd8K-BK&lGD!6lN~Rt3`6j9D88s-
z+u1F|18cJ&+dxD06~dw`{rm%|5(B45$Yy9Ms$T6?HAULHI~uIY11ucGFQ4H&QgC3K
ztXTFx4>*3^sRn-bvxfB%v9oTD7~)YH6S>?AXJm$s7<}UbaZxG~b;u@|7SkK1RXJs!
z`c>0(O`iUMT=}ULP=?AJ6gw5m9P|M90nek9K1>64!Xb+rc&X!#(q{34X!~pd*GXxQ
zmiDVrQbEHAM8-UU607BXzNPny>k!(B_`;>bcO)4fL@1m`RfZmUuf50NyBjel(;co7
z-1zE^Qtd(=<0X;C+?wX)!xb~ti^50m_@_J>U;9|yzDvV9PN9)h;?fnIFwt;#u=@fd
zBzJ+)<%NOy8NZR<@7GD){?d2a;EVaB^Yz4CNN^EApGl#+v0NRm`()*rq`2v6o<U9)
zj+B|H%CWw-iz0Riv#I&2<vlc`*j+id@k)n@(Iy|DbS?=-bIeHnqChB-CHa0kZbxR9
zwFyNlKzkFMgWVqop{~&@BB&)*H%qY4P%@HCUB3r}j;shAsD9y*??zfi?J5b))@E&L
z*D>V>Nz~{lwau3-zrUa*YeRQS5Q#ej0KYag47c;Q=2QmW-5l%%tP%f*{Eq1wfB0;F
z4F18<tcZm4j6S^89StwVi@KI_bj}LCFJ5i#Tk&M+W8MMUI@q{Tjv&jk2P#ok4_jjX
zB&A_5=ALMpq!lvpDq>aGe)FV_FfR9u)8EKsf0;DokUgr|c6Va|m+*1%gX_%kA|q_v
z&C=;u?Jd>s$yOxo%tX!$as?FV=S9=eu}>W-pv@Qb`G#d!4@bv61m>5coUOQ?u*Jh_
zwNGulm1e)fI!>BeoMja{(?)dtapZhd^4jUd`fESILqj4%!m`-`wz6ufcJOZgs|R+$
zqd`A7?AXhD{gqxT#vo0e$%j7GbgRT#6jB!7>@0X8`)GgM4kh@CRNm;6ub=KZ=w5Uk
zk={q*$TiK?hbygk6<H5o>&!ZQMyJPEIqN-RkdWxQrC;*e!XnGQ(@G#|aoX&<K`;WQ
z@CgCiTt)LHKc7H7w6&Fy3F7XOTbm|h<Q(xbJexTzFzR_-Ha9F)A5eYqxeV1O@{KCH
z_O%+|4rd>)MnyPr5%5V8#&BgHuOHVba&%>*4kIpr|7{@rZJ+FA*~?=uwH=B4r@m3d
zO-)fHgp0cZ5gC#(J}jceK(~7C<uZIukmQAMEb>E3&&EP{;{f47Br%uHEt0McB>3FM
zy#|M6bDJ~ctTj%kIz!U^Octc*m4nXJP{*n!bvU9i_aJ@&ija89rlXDuGLBWxrqklO
zO{QzqmOW2i#By0Q-)?2Gq(vkha6gB{vxtqeXx8&UDqOxOihkXEboOjj;N4=K-x~*C
ziyk?zdoMfPg1?ulF7Kw+g}HEtZTKu?jh>~*6_CY7pId5HbO2GF=J~Kg4g5U_6#n&Q
z1P>^AxLWMMI;~-{(X!oNDkM}$eAE>&t*uz=?uQ9&T^_zg&(&Zc3fv2A!EN^;ZRY>^
z`c9fFl?U-*gJpZ4L}-H4^Y+LT{30gkGU@X2`fI46-t2^SiIUHOPKo_@n6aJXyynWS
z{5|V*Gw-j8YG>(pi-mOA2e(n#B!sKMMn45xK0?C}*MpPDSBr~$eKgdxV`3yfouF2f
zw1EP`>~(W0BO`yVcJ2xAdHfa(x*NG?>aw`1t)6j~bxf)z=5KquI$G4cJIPvA%(Z(%
zds&SfZG@OBNz{x9PoLj*`1=A{`v9;3$Hp|$6q^g{8>pdK=hrp@6H(-JniNM+p!|~T
zN$Tx*Nua8+^-G;<?6dTOF4#%3!o=99t<O3fegK3b8toc5(j1SoY^9kW`d$ahAB`pM
z)fr)Tv<8Bj=7%@vl&-n&LGnGBNBE=14zA<bE~Cq%m(b-tr<Y;Q3mu2NTuGI1F4GiO
zFV~@G;V#l&1&7-x4IEXNAoTtPe0opL@WPFAVl0N@57M<E{=uH2fP0?)zu0@vsHVDa
zT|CMwqEr<r(nO?4S4!wn5D;mC^o}4MLWj^|p-7i5y-9CULJdTElhAwby%RzWck`a}
zp5Hllobez34|m)z_v|mp*h%)<*=w#@o;l~UdTJgc?Dsky6*cMPxRNxSqFdo2T!_8K
z8TE$=yFx#7I_{uPENU~Ez~s{b9R0n}un8X)-lXhaF-XnDK}uD(fcN?=KKw!T8aR7I
zWB`V8uyZ`UeRZspwNkQ8eSyC{b;yzB0t&5U^FCO_iFONIzC}w6P!rQ^G6~nYj=IG$
zI6J|u(rsa<Z7-v=k`6_9&R^_4nxGs4Fi?yp!8Z~bjt~}aJdlMCfP_{LHF_D3PCNw9
zR%NcfbL<;Uml7A<se`r7-S9fR-AgF2{e?P*8)K*KaOj<fP%0Ail#Xc+#NaI4YM2Ou
ze+aIQH0FC&73x11Ef$?t)NAP}e}DV5$Ury<Md(tjCvN{(e+&VDT@)hDxvLF{=A?Eh
zR_yYa(K=?Iq{tl)*D1HdAV=2L{BspMU$xS@krgon3EhSq<z4>Lre`ldHa%J(^A%wk
z_-RfY({=JD+(1G%-9lYTO$e5^#?~H&ZM{LZj^DX_wj$KpNRf(bG4K81Xx5JpYxUat
z5?ny~iVeXWPPZEj4ax*BI6Z1NXGfcrUtV4sKMq+O+iy^1wWt3m_fkRCOIkWuSswZ)
zA!RI1&mDGAh}Pw3ka;H^TyS#JL1Is`Ode!GC*Xa5uhf|6bR-kLfB2P$a)+s9L#gE(
zRy6V#O3T>8EIuhrm5D;1d3qZ#*sSHO*7<_IeH0Tw(a_K>t7UA-y??~&F%PAQah~fW
z16q8}U9Jevwo|)HudZHUGGKLuq%!N*D|BN|e~aG$*r>2WzU<pXxDwUpr@>nrQ*sXt
zAq9U5Ln`aCR~purM)^bp@Q9<(tsS(mpGyJJDY7I#V_6Kv%11u07$%XR32qB+KO1yc
z$3_h4_FBTDXNWhELgJm5V>5K>Sf`QcVe-5DsY!Yvh8Igcs@q|gOPy*@+RPyb^9i@l
zI-vT|oXC9zx6@zYnklER%6KtVKANE0{PmW?=W>Zqp5bCq*3Ktq9mB*l)~D3M(?~?k
zpKpfRQRR!rZv@nDDn7<?B!v>bMd+r+eRGfk{$vmJGt*jdjO7&O<&MQ^=OX~uRXJB2
z0E0fQ3>wtjsmQi;+A;Lz6{ae78;wQO-3@zthtapM)LtKL?0BuYwnSWpba#sp9PD%y
zu5)wgeeoh?+ZTi)l4}4aI>W_b1_i#aUErnx3z@4B#+DG~eckCA(np@D^wl^{H{o%x
z+ok7oZ0gl`jC@AaEhuVi;aoZOgiN|)Gct!R*}#%YrNd%~Pb(VlBc+Q5E$wP>ld~R2
zJja(?tr`7y_WSmQJ>fok%WOd`$_XzxJ0Gc2XdLadXw!Tg;ZipiYtQ@em()xeL*3j$
zoWZF|UYqwWL#VmaaP$^XN)%rhbbNfwOEz%N-2UnIV!-3*hH!+2YJsZqr=>HtAf@BD
zBq9pRSuCzUBl)R$7B`Y`Uxxd=)|pCF*{hTnyR0yovxodL!$&v&0$I?j`w$SR?{ckl
zcF*O$h~_mY_#mDYDP;AwO`*V3IxXw;()#DgWr32<)rYOgfrmZQkzU;0Nn;0lO@81N
z)LP~0ZG)jG*ZyI8Tk0+5!CCL^<BU9Nn&Rn5&z0ZB7r(-%@cUbuzI<khVQ}g9;ji=X
z_4g2c{l0n5DE^PWxooQ{-^55tOr6sszo|%>QnR#IE3@_^a0Z3E6C-mV_u{351LTlS
z-xKnND;1F!<{g0!X5F3k<kY?}f{sQRPYOtX5vzpiz#rFpO|;fv{!BTK%)n$LTbT}P
z-TrVF)}*Db6$&FLiTV8wv#(|-k4b;((|XTmYi3LA0d{#=K;&2lOUSdOOwoH)U$F6}
z<%if}`Fq1L_|eb+slnOgAp~nKeGWTBpz5u8x0v*t(IFQepH`ek>H1y1K6@BHI%3n{
z-IGyw@W=()uv3(joF*Fwwl6GT2o=<in?Tgr^{lM>UHaOeCnpZPqG!}mQ|^DsWTQ{t
z<M|ka*vF%xoT_lQHps-Bex}6y)aSGgrhfz19)gs34lZ^F{{`b*^I-!5gi55x!GiDv
zAt64<0{_7yeGq>sI!1JqD`{9ITe`i_Z*!*?!ZUO&q}+8HYd;Y7m^oMhNkfFwVg-jp
zD(c|#SAnFvHWNwbqea|?$3_qRhQs}V)xm=@7Bw4I(_ZZwxODlW<9(Gx*j21InmI+4
z%rP1RvvUE&xnRilrF`vDQ+FoYhUb+<QpooULcGSi`%9?@rIgLZd(lnnj~4E-Dkt1G
zjMCJleH7V{eX+H_kyWKE2G83`A(Ubpc#x67c9*fjukw!V1MZr=17-L^Ibezt3^?jN
zkTEh`qK|ivy8T+?44<PP1X?Ry5w;4NUGV2-v-jVLRyiVH4pi~E70PJv3G`d&-TtA1
z_qp@6y*+7GE%Q*kLkZyV`Bxx~df?&{6J1H(RB4<*lt%_cdEG#i2Lmhp5kH&_1o}tq
zCV78V&rjG+18pRXR;beA&#_{$eD6Ar+<K>v_nMYC4=2HTbw!>#u@{Z>-J53T&BC%^
z`Pom;m$tqip3#j18gu(}S2(^z*|u(T%hbHWc6;2`Fg2mmd+<(w#MJjNl_F}3XkoaA
zj>mondItj*t1&ln=*?H$T-_Fr^a20yvh<dt-~%VelqU(7xfx)Y3Oph?j&0xv5yd+g
z(Lrs7`HFEcxv#UHljV-{36p5usGE3=`QBIc-moSg+)+YpK7K;e(UoI=%~yG-nA2d&
zYAWn}`t!1zA&{~xY2f|d2ehcJ&hVOe=;L{x<X`>3y_zdh<PqB&*7P|TC~YjA5d(xh
z<+k0#l}7$}>!`w&y?*nJOq>3rDd=Yd-SYJ_(NuaZuQLueyDr5Rg8mv8R&Am|Ew`lj
zDkyd$L2-2Y^k^HeyUGpPVAqp9>gbu^BU`VINBpZ=9J`%y^*Nxn7EH`(x3S@4VSsaM
z3*LOR=5JdAWn&zK?0yc_-;7#TB&-FaJ?GLd`VS;UMxcLA%idhaJ{1!*fJj$aI%&*6
z%<X!^SSzpAUP5<&!w{0Pvh5+gy{V!(<t-|f$sD_e;3{+MgGqP=hg%kP>ec0EgooMR
zaggyS<hXE4fw@*W(dnv|aHr1A(`xuD)c0NU4f{cn)xxc$@`HE1VbfP9nGAFUd0kNx
zC2L6|9a0-X(6O*EZI@U<TXffX$&bmymio!4odxuYvl3x?&Ow)pSQqj0qA{m}Op}od
z-Bnkm#1W^|ubH22>_@LZbi>IJr0-s4c1EWQfu4~2YPUqts=G=Y1xiV;{=wG6`F+1r
zzKPJ|R+xRG20LD4Z%xb%6J8yLHdj_wE`5HZp-=`Ha5Jx7v50}tIC)vCdMPUVBk2$D
zgEo0;pF;tQ(HUPrxNGKE-+85`#`4u<%v?B(=6WTJsSlDEd3y%0E5F*e!_;}B3NurY
z67t@lZXgY1Gf(Zy(QQ`#Lg>>OpuT<oU40i$4h;L)+HFa*3_{%G-S$UR&Q1KxHAxwL
ziEENyD|2#otgq?N+peR9J%po#@71oxdd0j}cIN7=+nAVYPUX`D&gj&oRRs=+q(w~Z
zba`GXiJxMuMnfkKqj)BE5)w#~LmFHwjAoDC4EbmnG{~?FGBija{_ys1GfwyJjS%sd
zoPzbvu+yI5S;fMxrn|_U`c~ak*Y#|426(H6F(;~=-q7ItphHg?_4Fp`k{+h<bN%LO
z3>CXC@997Z`T3rJ)8N5qU9HoHS1)9bceZBg$}4?7Jzqqw!=0PwixR^~I2pWQgYrC3
z_WqNMVBA`Mr!*9XjTg^TLZdeFlqr^+lf0MlDCy*Y3V8UUU)_v*y0hi#%w7!#<o!Mg
zsAs-d6pi%h5v>8jc&F)v`%u>Y_;79>kLE$(0nO^ae=2J7qpkL=Um&qWH;N|t2W%A~
zkXXzyqw)={@2nY^_-UoW1?D-5cWU^q7ZCG>qReC4?x2}p&#I!<_7m?qip27S(z~<m
z5;G58^&xM+9ic$1u39@Udut#4mO6Y}tlt-3K>{-3M-01}vVwYYRvX5x-TcU!3Wr5C
zDXx;sD0;;X6$s|@3yL2(Du4Cx^b{3gPfaBmSit+*-p<FGV)LN4VjNpeDMb*)s`zZ;
z0B}j4LZ76hmIL7u>Bp&eoFj}W@1qrNymjP$i-{s>qB<+wLgp{j(Oxm&Hwg_mFs8TY
zoxPZ3y-azaobVRAH^2dT%MD+NN-fE54#MWH#VYK`Czafr%|nVsxVJk0B^SRIBoZBO
znOe;F>~Rg@zQEtOko4%;yYm1^f@{y0Zn9`EK3)*1`)R-EqK;M?cH?H>-GMrz<<_2y
zZJGyn-FDtmpUpV4QT&dWffrg<H`UK6uso1ffy3bhj5Jo~S!R-xzCxcC2UAaM9!R(Z
zoP1QeK$uePKv`~3mHVG!OZ>Ge7CL3(es_4f@5CxBYe<_2DCIr*l=-II(d1!wcWxPa
zL)4UW47rowmMIWg{Dsl_{A}pV5Cx$B<dzyf-`A{V4@9VE5~#~1lC;Y|oMIp8-(rHw
z8l*_yY5nl`rI>tyzP}5iut*le+QQdkvH&!6?T{4Y)u=-FYh0!caS+_*AaLoGJsm75
z;M%tSa}(FU1V}!&^`S`hqinrZpLsrb<;52JqZh81;i#IkO*6u9zFRPg)`#K41dM9b
zzn?3Z_yHY5GhWlW>f6cEt26JObbNijX-z|UG5YbiL+5uJj8qNds`6B0F`E6R5KvX;
zylieCO<(A3`D#oy#wB)58Pu+I1l~kVt=$KuDQyFR{iiOH?P^$S><n2Ky{`M(+~*U?
z_FJprs`@X*7vl9D#$Na1`;BVymFI@A=PX1&d!|2ZuPDzkAvf>4_oAOh;B@R3r6zo1
zF`X#xY2Z^`P7<U%P-sl8AuUTAyKXF<rCg~eqpGEM9ilkwGdZg+KfTHFP07eYOm}1M
z7QxHS+#ii9mnk=ir29#Y;uThgNHi-~h{9|-DrJ2ZZW!4do>MM%g(`Cvl84?yK>87;
z9e7+Y8)L2NqVw0MnbGp02=9(XywHIkmht10;WE~}U1~(Qz=_PsS?;kCff#a~?j+mp
z!~6UP(mDFI+y3!V4C_0~-Ib6UpVf$iBBjLo3y#BX@On~tkyCH}PW|n!$^?h?vN`}0
z6ngG`N=cD))in!crw(=4q2xZxTpomo9YshSc0xSoz^PX#0}mKHUgkYd!{vKTAG3o#
zSv;BXjNRf+`93zX<kM-<V!MpeSe`RG_cIRbnRTAY{L3jC`iO=bCc_Ku(5~xmAs`JG
z{^{Wr+u2!!wG8=3A5{W>g<zx@7(hl9LLv{06vqF?`-BJYLWL?8XHLjk$;|ZLExD%M
zA@_gWVodDE1-Mp`N~ToF`3v8Igak8w<qHWeL=y5a^lOjHtq!LSR<kgPV7Q_3{-1{;
z0jpvoNMo+sGEUpMsB1B{>u3y<mJKXH0sh-!t2e~+IePA!`4{Pb!`@P&%;mjX6IV%z
zmu?aKKGFC{qEFS3^r|h@nAAxDGPtm6j$HlGX`0_VXF+J|ptAr#F=8$^s8}y>ttY>h
zO8ppZ;3pF}1zc&n3V-Q#8`ndYPa{V1qX%SydUafh&t9k<r5DpwMeFeCV#jZNma!h&
z1lOH)_>ut*|HHepGeCLcvHkhv_&ewD5l@ner1-?xULtW(yD`tFJpN1~+l7#Ncl5=4
zaOn?tPuZz)?=h<(ZdX>K!?PsMp)lT!CzaacxX=T(`<e8>P6XgbqMV%;CJJPH$#=S^
zo_bMUV1_*Y^Z5!xx{E?wt>szsBzWu`Ir;3OyiCtv&6EKSGvk^*hlH99r^m{;UBe4h
z37xk!Z=lIWjiJzGl^ObkuAhcCR{wG>fvgs6c%>0hVwMdz?bmOnHhHdRqJP-|oKMAd
zU160PD^7DSr}*f#B{+fM@uq)mXOoRUx7d$wxp+=46m|RC{w(T_^Z#Y}%B`{^PI|fh
z0O-r(^(UAEQQ>9l@1KTo)z<@&DFD3n+}4~GsdAe5t>Pa*g4y!<a2~6lLju{-64G3e
z;4eA)>z7b4qqp^apNNzSzPo&lk@*WY-%Q%i2mD%ZL)+GAm~><&-u!WR<d^3rEyW#q
zs%HGyG0UPc(`jPcy_tfJqnKTo)oFlbS6e6A$sZ5Ymz3X`9@X%?lIVP9;a4s+mCeVe
z?Z!71a`pJeeu@c6g7NQ{fF>WA-oNzdy7ei09?R2~bqbcySWS?p*VQsPFP^x)u7i8M
z&2d9<zQea<809$MV0L_*`JhX%1ByC(Ltrd9k6G4uw36`od4>}@TLe<qJ5_<qt_Y4l
zEQuEuyJw!vcgyyrb-dVRzOoW3KScoIHT);hN@pKg4mUp?N4@WT_wrD;VXa?e6bYE!
zOhF9;h$iF(=OswddVImwr5M&+nP_E=LWKQFUzTSm@n#0tcmT>XUJ8XoKKKX__AmI&
z=v!-UvG?1LZ0H6jGZ0yDub)Z+!89==kVWTFO5ta2;63AOJTQBnrJNwRe3ZwC6|2u&
zT)l0`4bjif7FNm>#1va}NCOQ!FfaKcp$r<~JEH_OiQIJc5fvd->RyIG`F9KKiGzGO
z#SM^!(nF^;Dkpz8XlaQ1>ROhjk6P)+>GRisJzfXGr)#qoGYF0pd@#WhJ=^%#r_#n<
z!+%)S_^EHuf=r}VyOIl<d#rDx57iag3x}(Ux13KH`Rkksni^8;sZy5&mCWBvMBfbr
zR6>_5Ux<#loR$m>fp%aQ!PrrO-TS$ym4ky!I3M{;Y_rSCLH+`UhE?sxW$C;Hl9y@m
z0+VPGDSk)K;Eys%aO=)NLlTDoZUbu|W#+i$Zpac9cnI!lw+Es%;CPm!tnb#2-K4q)
zN(h}r{d1!hfb9Xbg`i`wyGfq8fBWOkX;on1`=yzgCT**$z4aMR2_~e*L~^}<EK}l*
z6a}!H;ssXKk{ko?cL&UQ>5qvG#?%kk!ixs&+F{HN&TrxcGc`y-prM>RS68Zxh%wzl
zgL#SRK5XE%?48iJiW&m-INqC`Tb(D8wDtF-`<W>XcAgRcc3;5q{Bu_c0U(l3&p%F7
zpoKX^SNUs+jI`nf-^<W~o+HyA?&B+`HoBA3e<^jJ*j9xxTQY$Q04mP3kn6EB{dv7a
zov)#mYZdAHy~j)n5Pd~ED?(dFUjS$!YXXLo;B3C}(rN#k@}Kc~Uj0pYXvOKT5e;^=
zR$7fAp8=Rjc@rL0S5*TJSNQ9ybx-nk`y(utwJJLkyyf>V7P}ByV5&j<0#U)h)!6uP
z#}w_e;{VR`4V<CUW#O;=XPp0dwRn2{qyKtT@c+MYF7QKm+&{Gd|2%>JSN{9|)&E;@
zEORa4#DZ5N@uA$tg=Opi4FC5l6~liknLq#SSHM^LhAL<4a{8J7^()@8|F54A&rP+G
z4^K|gNUp7|SvWI>h}YNGrw9*qd6HoE{iW!T^Y6gB!vDHjMc8$=>VH2T$z89-|9WsK
z{?EDp_8An__x}f>bBJ={sQ%#E{ZAU$`9BFVaiEF{AfmBzzVlN4oB{6Sk|__;nHUx1
zw)p!@BIdXQP?|u0#6OQ(riD(vQlR#)w5;ix^v&$$S!?HqFE>5X&btesM-BuQKM<sD
zm#qWsKvdON+i=;37%g!x>sJRAZT7p2UX(!2Tr&$mI-|RnmxlbcKz41=_iyq0=kovd
z8hXtY`(KYwu`K^{E_P&Q4KNx;MCY3^sn*VScr`AUy?L&m<9|9nz5j2Q-~V4*^6&a!
zvBhu(E14<ZfgR<5Xda!haK=!mYzW`^PXmGaw63A)|NY>Pye4e@uZK>;|D5}uX8;2I
zAAbfqo|)I|<`yE#w$)s|dn20=#bRM>W-74Xie;xbalwC)*H8DZd2K;d&HgFT5u?q4
zi{1}d6$7u7!ND{wp{@3y3L#;||0K7(I3QvJecI2q*VwC8+pyC}jaJDIFxh2L<ZgeE
zAC)reahbDj4YClBy4u83d%E}^gzUim5L|m{?}}DPD=i}L%G#n$#a<gfk-^AO!^B&H
zqm)xU7Tjn5ZG1U!SYkcqOgVMJKJhda#AUq(_x|t%--5p9&?eg9KdC|f@7HT6*XCOz
zg-fqxp-F3%!Ifnat;2%5q4zLJYi>`xmsPXvM;2ktsA!cjDdhS>x5H5zFo?9QB?73d
ziH|1i(*kZa>gd61qCVP$y>THuYA~<F3ERJxD{sOzVf#z<$h=Sce(Z}}8!3ZY?Lok^
z-TsXemvej#^u772{NSVRg%_9o$S6<iJ|ctuer=2|dq8V(OHtt2=#o@gsXuO%wM3EZ
z`Aa8@6!S&<k5j)-`qUU&|K$wp`2PJ*9>?FM>Xqb-vE83D?2!N_jV<YhIyR{-K|{gp
zR{L`6v-_xx^N)F+{P+0~GBC7h>w4P5s22wwY+k0UN(rhP^LEKZVP-#~tq&Oc#DGC%
zTR=UTc`2dIi4J&+QhZ+I|MCzlo?TloD%OMSHGn1vAo6+pg)h=<|7q3+>(;X3EroVn
z-znc<U}}33`L%#sIMeu3U>hB}lGK^d#1kY~oEM-<O}l9DUa?Eyl+2<!{R(pjm?D;*
zQyW^3HsTOcMU2Y7JoPk}Yq_ydfjK-(KQ;|oQE@kvBw)BWFU32@CjBub*_wTzT1xIv
z?jEzTYa=$e+p^#AVxQtc#v1U(6kn~#=}}ZSYZx~9kE!H(%V*@n#rXwMQO2Nr>gO{o
zOpQsM0DBk9a)K8@2n45BnSF9-aPgcI_lpwlnCQ^|r>~lH9*`W`*bo{jhJ@8j;BUKB
zpZ7~&#w*OksPOL>j6yBYr8_f7Hr+9K0>+?;l3AqHn_D+S_xt>dKR`*gyJKzWRw4vl
zi+Ds-Co8jpY5Ga~XkEqYov{%nHKWoGde0V84aG}f?He$!MVvV){5w{I|1?P+cesF_
z=b^Ev&b*y^xj#cWD{--rm{WYReTY6?Yv-xo7$8JbuY8_4&zVUTyX4tLo~_I;)y_)n
zCWWV5GQhEA%HAFudnEWFbfOFm$H`C5U-D5*x$o`~KgT9NF~mlb^=R>XN#TV$o;p>v
zl$_^Y>hj<k-+9z-m#yZ*Z+B|i9WfNwBFEzTRmc(gu-={80H8^p`g+=Rl=`cK-a%hT
zN3s3M7a9VX0C00d5fs4MmzzBTGzJ%6D$1hQfm^1vjg>DRN&Rm4kEPfzxbl4Sy>^)I
zz;jw@R;k#|+tijT3SsiZsn_zxi#~B0Pq>>2{;<D%a@K!EwlhwkSUi6%<hkQ3(ry?^
z2se;YaTsm8rE)$}!flmZ`qu5f+Yrj3QJ>MukkcQpEi6=@G}g;$4Q_vl%)Q+I*!;k<
z)adeTfXNR=S{yzRXR9mXaS7^4H(v)I({Z>;t+Yg#H<Zqg2gcdp4$)>boBN?ODZ+Xe
z*(L#4@xxH>-V~?KmDXl;!9M$4@pHFcxSJua8#fFlMV+l+Qzu=}4Cq?N{Vem4sh4;r
zbUt>oJN7f}(94C@R{NL1;||*^Rr1?-bYoNEV3G5{TTcGQN!X24QPL2_-r0evl5_Pq
zh~`1}8i3WO)Ge3gUbbt_i04b+07VO=Uptq5Ah~QQfIbeNbu?>rm0=9Uio*tLx?Y{{
zH;^?g4dy5%@h=17eRLTFh`c^3yOh06w<YI!_;3YtIFL1a#<^p@@blTe{qk2xq>2)S
z=qvqa0s<tAVu<<59Koq$KtU+sT1SV5<PHm-HT>)X#g!K1u`Fxwo!Y-91bt$Bi_=j0
zmeHA`tcYVyrEi}Kus2%bAC`3hf$}kHUmT4o7=6iJToUehRA{%=)&(gkVJaLVGE<Cg
z#&}EA&Sahd3Cl)eDS8W&(XUymd;?NNsZ~CUxkq^XfjN{hZ(81%DsAcRh2|t$7N!i(
z0=&`B$A;&<2^$9A%fns{id65H?mi@LrLsS%(VtDca=RTMg6@aoCfvG6$t4Nqd*C&0
zgJ|<`5*L1>Eq8ajJb=?;WxFPkf&T1G6-hmzkwVROS$5bgTU^6)wn6x~)8cluGssB6
z8tQzbZ^2_HvEEg4*$n8;H2!y}WuoFmvNNZi+f~vN;sn_kEEqtIL^WMkie@gb{2>Ud
z&yfq?J|nmMd23g*=?xt8r1LsA<nY_y)7a5G!0zV}SGBGZDT~Hm*K3|gMj8M$B}>!$
zfN~N$(JUVv{mC$U(jT7re1T+fhP{n{NJM|bl!cEmC`f`$4qrB2*7o(ohQB`S3uMCc
zkF^wYJh*G8yN>JuD|E&f#*o~UL!H?SSOFN(WeI7~vt$a*ljkq!ss;TeeDOePKmij_
zzk6%33s@uSFfS*wMlcw9xsf3g4op*i6Bu7)W43y2r!`Si9fi+(Cg>`zwVl|E#LzGp
zR#xvs)UvmE5s?f-px*9Y;~mF%`~fLhvqGxV!Dk&`c8_c>fptM;jR7CmYb^g@J?5Gi
ztEs_ZmC~q@*e#P;_)ZV|qn&fyC>nP#$|IgR^O?kMs^WauiiQEVlbt;8KW><$vupAD
zN9p23&+rbSXS#OrdLg_RY0ORuBqb`J<n}Cv862rhGAtRYtDm(!i|<Y`#4K_!OX7hP
z0c<Gb4iL$vSGjE$bETUA;G?Fi$PZ#8a(?rj-(*=iS219UoGa!BkCbV5foBg?)t0gH
zIe0X+$i0tz8m5J$Tks7#Qs_82L#l<fbSki@-D?I+`z$bN!BL_MbN<y1k<duzNc7!j
z@LhH@dxdWqKcjvgz6k9mjiaE{6FSl@xvEepeJpolS+?ZFrgxJuMLU_qHjp9hL_3;a
zo<ep(-~~<+4r+@Bj8}SA3j^v<;u{6?SE)8)P{mUe{KjH)By@!VLW`(^8a!k*v@tm9
z!6or*@?8e+khlbB6Lo7zKM~6~4zvvxOtPwhb%M)T9pT0uN08d5dZ=`C{^xsDM&1$N
z09*OR_ywA}kyVtivp9JrSCRG40Kp~H*VihZlaWOZP*U5ijozZC<;JhH%3PMtZX`4>
z&W^irViVRi^#G3AB(MJRGtqMc(TLjZmWHf45r%}}wNgR54*A({EGd{)2+G3%`K!Pj
zvy<c=_Gg~Bpg9uZqCw0*Ghg2e2PZc;{H@e>y*#V{j_Mi6_GRZ_9gbP9$K7T0u-A2n
zI%r=Rj=fxUEEZZNML934zwZ@VRfF!*N$j2R)w$(lyd;5R$Kkj?h*XiBlzt%BJW>Q*
z-Tn)a3Oha?cj#&hJJp><kKgJ-dT*omwDwK33ig$Njdj{GXm*S0SknM5l*(?aPMu=3
zI0V-$Lii20mog0_PC;djxxK`0g*l1%RBs&E;IJBTSG-8A?+Z6Tyk~#e|E3zOv)&HE
zQ)0Wt`!+$)&FCAuhQZ)cRh!5!cyFhI*}=Ln%AKY0fx}zLq_TSedlO4!E0XxWa-@Pe
zH?WGy${2lZ1OXiLD3vIEd+b~`xJ|2NvyFxqZP#H|j@s~K_1!y+yDBXz;v(-x0jIK3
zpp3pvpa|84q41w`j6pdeu^vrNVNdq?`}>plPP3S?YVj;iJF*ZJ=eyQ<^alW`<G78k
zuI|BK!6o680hfwKdj2?22rJ-D0?uSb{z<uB{7GdHv5cgX1dXoXqRd?y(8(nsRWrSK
zgcHbtfB)F|U28^<A@Tf<EaguUN%RrqQ4<EJH{FqSB6UY^KK1=&(X1&&kVrWlBjGIS
zw2;W?g($w3DMq(!=^-3EyPTZ@s!P!%MYFPu6t*ESmZM{GSq`t@WbH6nUjR^B%;suV
zfDS`bS;WR|3le2WA{t3xu8*0;|G){<%sXwTy7}uX^GtX@g()&jUONA9agl2G@aUgU
zT3*-W)J{aN{}`@u)i57Jd~Dnyjoh3WqiR|J>Te$H&hgc$@|?@#mKq=pj~}`_wB_(z
zUgU_H*-S;7n$zNT!byGh!+COHNtss{M-4*@FwLpB)Ix_gVLl0jdz&mAyW8Lf9XRNj
z?xj5$TA&sP2Rs~7hK}jmUBs<F%DrC^h%%%kcJ*{15A6xSBf=Btqj%<SXFu&*za9Wv
zMe4phDxxIR(M!SFI-gfLHUaHYJCss5x=vF~KYJ{%=3US`(HpdVPer=u)ZEwl(!nX!
zfc}|rFa5hq;0<C&uf}P$OC{GqD5f5@*rMg7_vw|uGD(P?rIhNRGh!um@_5L9Qpqg;
zY=hS>w#A_34q=um*TP?+LZT){F;D+|*pn1Y6xX)dbH0$xMnQbZ0_<ASsm1|etE_GW
zR96^ayc-1Q!6`1<*}s}NqRiJrg3WLXG1GZ#p_Nl#oc>f>nDuWC<|moZ;u{clE+?!g
zjEE2OiWAmayj5Zoy{hpd1^lp6^;uCy)5!@IK&u_ahBIW|^LSv9-<P_QbRDyF&W0Ic
zF3T*w&i*)H=(@n10^mn5&&9<uoQ<!mof~I!2q&U>5J{Y}%CqYU)L=9Vj6fAmNsZ^;
z>*eJra#~%zyRiJV*UvDcdf1ZsybHp*==uEh!(HjKa_u}13D<pdDFPtSSnf5%1vHTa
zeEb^HpsLF<L!dgQ^ZqQ;{WMN$Plc8W4`@5Vg|xQck{JOHqR7epwSCDXwOlSofb!Qr
z8FkGo>Y>!ua)NG6vi(Yu1jhX<lTLg>FTzI>X>Nx6!U#i2fFd-!SZ<gY>sO{h%)5?G
zWNw%Ll-u`@t)5zyh2EEEL+|{I8JdRyXIqe7u0B_}Td{?TcXOkMk<QG)biH`5udvBA
zlYeA!Ru20h`>KlBI+z4(J+JNw7ULs3jS#+YCcrHFgF;>Bemiw89~O5g&QkGWzO;h%
z_Rq4XOWK0Zk0m_EZviH5tT?H;9_8+0zrI)nw2(hVpByp3Piw1C*lDmb7X-V}*ve<s
zdk$&~DiZdzy}oU;*^}2w__8Y#$@8V-F2q=1cL!k?mmB<bfKxbYIDPMkY3*PYsD$5!
zG?e|4+o_sUu{)DlAl*6LKdjOdJi-4_|HOB?`8^Zp<>fUrm3DV9TX9nlc2zsvx%;wS
zH)YMz^c84P&$|YX|Ly$(m5*PAyff7n{|;{(?(t;55DI#euuKd-1yXh3Kod!F1Onkl
zc15^8^}Rakjtcpz-P7Ls7t$<Q6kGuBo}_N!Q0;AiZ)2~-3<9{1LCSMM<u#G`hB}VQ
zwGK(t0Odi-lA)I(RY&VLS>@tCx}^ydlI#Rb)SHwgTewz^I=~-20#qo#RCSn)O2PoH
zXV}%lZN6zKRR!e!<b<PZ9E(!I3x|PdC&*bKvQC&2dQ)yqxu2NcaP~VmDG2J%Brc82
zKJ{f`xS2V(%5x<ZPvs{>nYMR^GB~YKb=s0AR*{eiI_QG#rBCTtp97+Opy`eJb95y#
zc(r+ONp8!B%vuSAAY(iRvegH(LpF9><6@`Y`xGAV{ARTK-mj_H2!Wq6_?PWd4|YM$
zQfPh)0eM;@+A$=X`Az9O-e58f{Zw2LW7PCiZmCbQs4~iiE*Kzsd--8Ols_v}U$|uR
z!#YiAmb~LEt>B25n!!MOStM~)<ucNnlPcZkQ+8kZ(cYY-GrH+XAXTY0VNOnL*mjW#
z7Fyl5ahl4CPnqEav<m|d`iI-PWnNZ8djSWz#fx3>x^u%wKL|2IjZKE&J;VjW+}Siw
z={Z~A@)JBu@0EJCt&KWdceL$2S$N36pyzQu-JF$p(_|{l+UfXnP!_bsZ1#r)%5kNh
z4ioBT?NPzEv59B#K~oGq`LS0(0_c8Tg?0wcRv+wqC~{m#%jA%EbNje6sv*Tr3^H=B
z(6qbBA87e^mG{K`RQ#>qjZ*1ONvW?0(;D+T>Fga3G8RHU0cZ$#<(Kl4BFpntzmEz@
z{8<*|58BTG(>63ikN@Gti_6Q89zD8Yc%lm@{y;Dx?ttXkc}lktQK9feaN#KQ=O>}H
z^ba6esyAidcu)Xlv&c9x!2;OdjQPib3h*lcYXJl=UdZxoUb)q11#Gd+p|I`qrDss3
z@G-8ROfpF+(x6}uL2=ty&5O8ApuPWLXSPysNnn+!qr~aarJ@Hx4Hi6u<Ks@yZum6s
z@gP^=$lQ6wB&S!>aE49cUdf5qo*hvLWB^Qc3y|jCvA2a5+L_09zO&qelb5g<O6fe{
zRl2sigO1@0IIEiM#LQohT}V*)p)n7t?;*JIdChdmB{Rd9@Yy@Oo!IDr{d&eI{9sZr
zCY{Qy{QyDdk9_-3M}V6`ridZ!*~KO*hXQS=SG+#q)bz#F=QI#*JEm24%s`_jv>p8M
z{7yNAU9x)R4PJcYc?pyy4uY%W>k_y;g2KhNL}NzwHzs%e-wpo3U}wBUYrN3QbTK2z
z9xKp*Ju5CekjbFvdVS?_S?8C@OkOwplI`SL|49dXd*s@eaTm-D{t||n*wp%ah^vz>
z`%|aO%yaP_Tj51%>=r#5of)~B3WWC357@ys654=z=UIo154(-vM>=<ZS8K@bfUY2(
zSOybK*cdUc#)&0x+lZGqhOuo(Du%fNF>|mJ*=VWqV>y)kXNS)BQGvF`aYW;%V|~+D
zs{68*iLEU{<660O_+RU3>e}{nC<-}$zP4Kb9oa_IU+~irNLGBL{o||UH(}K78dcM9
z-oBqCjKzl6KvsPu$_mdHg?kT^7@W*@#Cs3LOIO{*$$Q6m#yIym_Un3V9Ucc+c|BL~
zvm^gY_h*!l4`xKUJSvKD32k0J-r1?Q<HhaPAoMXa2~d7%2cOyA4uG#YDfljnjHX_*
zM9H0f7puO=w_YUr7Wtf3PC`18p>)x@8}7aFnxKa41>#6}U52(<W%2Ha1$ydqwS0U*
zDCGJgatn(~2L;7k@%dyZdImoA#9-)*g!#?|jFPMRj1(w$x!t{@U-8(E!-vf6nPtI=
z`EeyCe7X`>dfwo}c1qbLTS0Dp@|5iA_14M6K_(RGvvSnwWUl=1YU`)E;#137_sQ5z
z^+t!W0f`~i0U+)pwhCcqV;j->{aMS`U*O}QKxK*f&S(y<DA_yLdOViC)1lOLr!K6Z
zAQhk@K|#-0pVjJdt<xYV@z#8b;Wz3zcC&r@1K25Xjx*vH`O=0JxkU)i1`#3Fgxc7B
zdxT5@1xREk?<mS@@PXh`GVP8<T20+3q};gK32W=V(}0NyTF&|s9W-X4A^x&GH0AAO
znvc8Xa?_?ntO#*^Rn2=_fYW`w-Kkhx9SZ)G;&gCX71^$-=OMk^Bp*AKBqrf_)Q&+o
z?{igMjUP|OdK<4dDf`NoXuT^s`66GOU*eMnrh5wsEH2b@SDeYI91g9Bi^G-{MLha6
zbC1|(GXQK`WKcYA=l+bxXE6XA!||h2w4h$Ep)KyP0nf@R>3Z8ae9`F}Y4pJqJa^L9
zD2z-d<Dk?)|4F)Ns=HU~kZoh6rmP^tgb;00%UC1cxgi~0I3=l+6cf1EpeCbBzh%KM
zEKHf}p})U>?9ea09EA`%v5T*ogLF!@NE>&RKnZ19<kd|5A|qlusA_OSk>fD7$VI-=
z3Z(#ilY3*KZC*NS1V8PaSX{$#6EXk2QA4E73r*B#kpdC7V>JgoNnmpsJ^^2TG>Xk2
ze~6<d@4#gi^DM5LnVMl8NAz07^`MH;H_u?nu||)Dk4*C070+m?u+wU)u=l~A^{Drk
zoyb$=fW_B=%%fg_Emm$mAJ&B=Z@|QIxgSV!V=5}D4^H+IubfXN4Ic%*XO|@4Q-1!#
zt^dn+l;0xp{ZR&4pP@*50FO}HQI1@W;EbHXb<)eNX2&SR|5FR#J)~+Y&n(LAli(ZE
zqFXpcs#18F4te-OXd>#){aFyQsfPK@ENCKQCie@E?Uhfxd`yAwgurvJ^o?VvN}p6x
zWB=H`MjZZ`Aymqi;Dk+69oTV^|5uqK#o1dLGzg^MjwxZkCYcJ}iAkZR|LCgzEICl}
zhk=oG@;-&4we@x}&*+rJ?KV3+moA8~T5xU$L|E)$(=^xm5FIRvmtDrDB8xuFZ-cP>
z<V3c~(>y!nV$*_viOyoP2Ae}D9Y-`eh?fRV7TZ6}3rIa#QYS2kZdh8>^xadsIB5}O
zZD!Ey>F;pbZ(@c!cSk~Mm&*~(oH0Z<|9oKh2|?)VaNS!RU_&NyeitQgIuJ>eU-f<+
z=T9a`=W#X`_@ypS`fHp)E&Mb{qyWG6EjjVbk(T)vb^{pOE=^)0N67T_na<H1Ms9uU
zD;RlDG%E2vqiOTdil2%Kw00M0<jYZ@6&1nQcP@yu#f93U;N5HPGxCdznTxtIhIHc>
zCv5|o5iD($P!2sd-~H=@E7LPTMzGyvrnar6aeMNZ0t5b8=e3$08Y8>RjCV31A8uFU
z=Ir7Aq4Pd)4Xogcl@i>gs#S@Cn}fD@51x0S6`70)-vyx~+1s~Y=$nRKpHZXZl@}h~
zH4t+<9y-0S#*bIXu7=7jz2?+@8r8Yvn2sX~n^Mzi^W*M`7lrK1mnbK*v5AmLJ2>H~
zsJ(dkGi}8Y`qNc|C91P!pBB)8UDQu%^ti_@90GdH%N^cI{|E}Oc<C4Id-IlO&QG0#
z50Mlm@3o&YKc%LX>qMK`{{*XvL@Z5>zZak-<x(LdysrxR`EbTc7b}v2cU`Mf0Z^<g
z*Wi}VD5s0iHFMTdW77|86~Q`&8@KyzSZ{DNaQh|~t#?!ih`MF>(tV{7(If3sUFQvc
zM08~*bP%jm<+3z&v>2`@j<ImIISdcGSkN=*gb~K!u$%q`088=o&Pia?E8U0sUacSH
zK(f|!4Dj+4Q?fKo+x(J~AO&|vneD6ti46ONV{zr!N4fEU80~b4E0Y*_^x{15BtC@!
zUD$&_;6*0s4)`xfsEt7ZfDbpWWv$;Tiy}3+a^(L3+p-6kl)IY2ODu=7(XwB`0wv`j
z5Z~iZcpjZC^O{kj4qOdoBUxcPD7Z7S<u{T{l8Ry4@oM2d-s;<sat@unnapXY#Ju}V
zr$Hg%JM~~r2{kk1M*=7V5-wlkLO<)dd)ykJL0|Aoy$x$8T3wj3-T&K%deR5#h5a?4
zkG#^-$NmL+^lm4mnX4A-B0ubSwFE!yoUK|w{~7N*+_$Y?jO>~4a7b>oWe#_$(W`|z
zoEoN{RYlUFm(i3ct>DjWT}n#+d7lEG`ZEEuyvqV@ea5(cyA_=*E&#~EXJ?H!BL0t)
z=f3aARFgHUxJpN!gg+%@vU^f|0a(Ez=Bp<WN-r(!9!YX}vssi1)n^znjUucRHHGt8
z0_SrdnGa5J`>7_1&~*$16^5E!o`A8YlFIgpZa^MzyaiL@2r##{VoKa*y{qyt>Tcs7
z;}A8{6Khp4ysx7p2%AbY!JccSz_RG-!2x)~n%p8>Z>G~&U%y+>!;#G#n~>c4o@Hcr
zQ%n4%=+9=(4EZi^?q}Pxvqb4BsN=+;kug-Z!PRsGS<~1Glu`!Tf~JKWEM&^!?B)k@
zMutae`_yM(SoedIe0iQ?^ZrGY7<JQ^Y%H7WR%`!H6GA9eeu2qL9+>EEb5BrhY>&@7
zH>Y)ExDBH2r(%lx!LbxpjN>aX&F1t4_uc!gcyboOlFox8R)jX*+baRkHy`bNZCqc7
zxnIgN69R(a#;fbA;Q&y!(lZ0X(caQ%+qqViP7^mQHel=JCFaY^@3^`V!?ED$su)5y
z6&r$}ne{ZXk;dgBOScCXxBO)Up*k>Nvp>p&#2-|mzii=R_T2O_nyxPm&qHie>Ai*c
z%V2|D^U|Fo@xbR^Mpr{cjWSpUJz+h^?{g<JT#J6)$_lh%zid7F2^Vi2>b#kaC}@c-
z{5f;^!eI}$-X8#)rWK_snLn|p=?`SCKUoMyV9Ns&3>dvLdSub9ahEfJi_jEdr|jg=
z7aXQ5f0E&V=feBY#z7)6w48SUFkl4pfXV9oYR^apC>PUfixw4#C$KQ~4RH1IT6+V&
z3!k-ZI<24~hTE!g5mPUn=gVJ4Hr|s}YawRVChGj^{7IVjzRJ57Llm(QY?Pv^abGAn
zMz#0s{AL7cB~;au8dT&v1<mbS>e=nM@0eODgVSWnx0JiID@C3rl$m#{FZ{9$SF&Bj
zpD1#Oh%Zqoq4lgE=&g0O&oWy+PExFOz3pP#{;7EJOjEcRF5rAgSUVYz@eW&cnhcmB
zCKyv5<Y@1Y$Y0I@3gnVe%CGBo$DV7Kby?Uq^SM=_=+g?gk{`U4(XS{iGQ@e#43^e0
zh^xY!`{Ea6N4!M0Jkt(S?|(k7UMTZpq%AS9GxNCs&nBk0c(e{*m8w5yc)45i;;yXu
zUf%Y$NL)|X9B5;Ij~Hh7h;$?tzYK#&Y%YsqNGQArb>Tm0-f*A?fwqCvdhKjWOUq;c
z=@wkTr7kyO*kuN0ZF;zvS85aNuPKIm{bt9^<!J0RRZ20`=6pz~Dt$rMi%W+exxYUj
z^BzpV0ZA(f+ONlQ&+68X3J^nS&l`uT>jvlyN*&xCwCGEQPD-n!*d}Mc;8Swxlsdc@
z@-P++4G$ob5rmXJM-(+>%!h_UQ}U7a@*0}$KGdS$iSBZ(wPegcqzQ=QDt7D@OuNAv
zhJvj(6^Kp~dT7zs1EL@fcs-Hr*Bh5oN}6rF>R9F$1K!Fy0uIjB(ka;sl!VKZs7_xp
zbdm(@D0FmDDE!O2@M@fCqx0>%DJ$LNVm#W_dMVNu9`~O*oT`t`0FLv4?J#Sxpxu+1
z552n{Y`%RhELOX6%X5LXH)RLz=7&aUu%lG%#^ZzeUUjJ_wiJq?Hq(!Hug#0mF~Dtk
z_nJC2WAwG3VVx#|)fb^ZIdTTwLNmx+w(4Ul?^NhtvIYPgYSz-ql835%KcO#4Pv0an
z*_7>u`}G;+OGizd?V&W}XRGI?A`o=}*ouwlI~6@m;iK_R`y+v*O_9NgGvSL6k&5j%
z2R8GRcBh$sicq|@w<^=-(O(%0b8{{RqnLH-Z8{SkPG+-DRqRfG`4qktm;?5!t9-~q
z@NR7xxi^%1j11AF?^5pxMz7epGb4}V+dP1-J}<zBkRRfI%p>T$`oLwz#oE+q|7Ng8
zA^o_n)jNY1Nl;U#d83A*ro<gD4B(;yj%EJCND|WNnmoOL0Eqbgg-H{2Dkd!}K0oPh
zJK70Rfg>l=!?yd~N}2{RcN0b8*h!wjOip>X5B)*yU<O?Y@8DmLyl{u9C3<tMl!io%
zM8w3yDwYC^i;F!)-+k=oU18hejzC=iD#-k2Zz8xPd3#_J2&p0w5kBIugzn{q!cZW-
zB9k8PT-u*IwGo8eWQtpI{<$bv78NC@9{=a%#ZFSqVQ-Q|e0MS_<9o7RmGpubLOq{h
z1j=JG$ARdz-YW3<;4Q%RxB$B4Gn<ZMPfkKi_xG?FHlO<B=d@Kta88LkYFygTlQ(`M
zy~*P0$Ovn_N@Iayv+s?s;^GD1s;*1Bv=iCSJS$C^CA1KVIh9U$+Zy^mwONx4gpWP4
zQeD=E-U_ceqqwJD@rkWoEU`i?t=vAL0YNb?hPQSPdtVY-1VOhX;uuh+tMg%)TWcOQ
zk%O#dQKFjLxf~Li+8EH#4tV`2bXK?0xx2$pnXfmrVQ;#5H>+2Ee6db20HZrGE^1UM
za<<Xdzt~x>95UrBpIrJF^pTF;ot{fu_j_UKhsgx8rp95<8AWLq)9@WaW5sUI+eb*6
z1C)SIQDs(a{l%OqV()s(tMRR+ucN4oB8C|?crTE6e+jyM6q_C)3brl8hUqe;atpuQ
z>uQx^B%kK9^0+Ww&*J_;AE$}cFG;X<Urd*xDR{_UUk2+c+{<W{&yMA+lC#Q1hT+SP
z#6J31Q#R!jPfu`%U&`a`a#>TK26b`{KVjYfy5!>2|3I{%QtW4?`j#g2vO+KWccit1
z`n$5&ZE9IwTJJ^7Q?n;Z-@aWb$BCu(nBiyC9ImwMvlrJGUC{#{nl|P=4$;ds94K9S
zmDK(`rYnMwkq_9QmRgcKiuz5E9~mGG{c}b?y0$u&9pw(Bkze}fuBC-;?;eEyTgKt|
z_wMB0;gL-pNs)!MP#~=S#z49Hp(Td@``vr2{?XR_+~38L`P_Vx@b5D`#Z(Ueie&x#
z+DZB5{#^>_d?d53!$e-K@gluo{WK7-;Qnw}PTJ*nO{oF;z#@9nuvPD>Hf?<r#a*{m
z#xjCm-hcK5*nesJ{9r=K1%J(DyfeAL)y~~cs%i-Dooa>sdWs7yRZb&xs@CeMSe55!
z@SxLDGf#s!1^t?ZDpTKi7=yRV;vpo_r{Ty64x4UG-X!7O<V0=$PXAf}ad&$`Bl(S_
zu)aW{=HdCM|FCCD5_+yfqW&5AYZnl?`*J$+A~dmbq(aT%Oom^`eQ>0ia>MJyf68=5
zpQoI3X0NNt;QUXB$J?s^a2p!?`uZZ7!9E_#W%JU+AxAViULUvpSi|V*_bi;d4-)Xp
zhhd`yc}ml#ugA}W1$tUqUb~-fH$zSu3xG!80%EelhWL$%@596Ir;qHk%c?IlqJH1t
z`mS)!bE<KE{$7bzK4QnbW++gwfcf2H6_^LSEX;xZbtg-ZiI8M#==T|vT!{3=bv!Uk
z^)aGE@0YZz<BLoFJBx3++UWqsW4U=DYbH@)7_fTvxG8UcF=cgfh$jO)fgc`T#?$3d
z4)L8ri<z+`%FNzL;S3DdqIB`ZheINEp-Al)BC7pb!Ox)i0xj9XtmVLQ&P<jZF!AXR
zohE6fsZzF(R$I|KKlz02o>A*iKUEpUSxMu!69IciE+0$RG+$SXt=$DIdmDx)%`a#W
zyyArkrrRLw;~0Ser5l9G+d45R+7Z(*Aa9>iQ)de-WJ@Jr->nbBwV3(IIs-8<=3<Jb
z*Kc0kb`wYx&|dYYhi2yKrE<jDUTrO)U2gx#iY*y4)NQ<sli^T)np#uVPU80qsh)h(
z?CZxGU?xYcbk@>04^ejX$kWp1bxMhH4C3#|Aub#BSC^7R1?Lha9%v2N=eUn~;>F_<
zd?-V~xsnDQ>Xz1r4<EMnJ~JRajx4B3L}q@y$%^o%i`$Cr%A}w{KBRu9B&J(uUC#BP
zbvn4Uq-D1<r@)1ohkTWA=t(g|NsH~9<u}7;KYSt2ap4c2-0#|+y<Rm^z(S*Z!Lt$R
zYuGYQZ7kBwL@CJ<MNRPD5+JH1a~$uS1@?ajn#nR68^3q0{OF>VWBqI@*^!mf9?rMF
zq5RSQLa=z-tMc*l>OQq2nRyYG8x%n7x1*X%C8%<+SMW$l<vVGlsWLg+Oc8T4`CT1O
zAxTLU3Wd;iD-NZb4$5H-`yK~xr-rN1?yPMe7wDoZCe_m+`cIT*bF+RDk)%CqsJJ|z
zojS=uP=+yF;23b7sEegt8oSccTj$#Pb$Mms=&R(eTV`cW9N)0#va46<66H(X&|tM~
zs~Z1}X$=g0l#6W!i(PM<;2Y6l1l&Z4z&opSRtPlRg(+dV-Vut&tOzkZqS1L2s$aBD
z)<-vj_&EYVTa>P|rA*?`51Is<q8g-qMH%vM+YKtJib&zO&u^!-Fu&Uy&euHOvu4a*
zOj~FP3@kO=nBD>+1Un6N<^izXRi<B0KHU$5M#Zn4<v1D7nPl8=%qzmgn`WN_4iCXr
zkHYG<mXRL;D(r0A>o}ko{tWZ5ubXj74bE@_lmKkkz)N!tGW~+@=9ap>By_&=-FC;p
z_LnXU>oJ2mj${cl@m*Z8MB2IRDP|0#NlM(`T2F7SqEp!I>Nl1yOs}#Nzr~$l@*X%u
zO5;7JW%L#L$`seFm<1gIHU{=pU8l;r20%Ogdra{M?$D{nDwunk;$I_@K3+Rpc|zhp
zWjb9OY%fg&%${)8th`TOTT@i!TNr+(G|PsDJ+pYu{Z2l%9k2(wi^^{0z(jtRC7AaT
zOhhvTdM4TgoxHmEpjTKR2~V(FIb~7rb#!UP>+h)*yB{o}vKX@VrKXSgPX$jgfDkc%
z&5Zpk;AHpv3V;J(Jws*D3Qa6J;+<F5Cgm7$0^nG-lfd?4F3HqMWM}V8P!6CU-T=w{
z@p;>hfO=Dr6anMeMy<FskD2v}<W-e`nn5r{1;TurDe6@I&Fm+V@fs&}F4l~5L9zEo
zAGlZp^rDnr4;Olx*^lO{Ik(X<lDk!hVn70u_0~NAnTF3U$7U6I{mCxW{~cG-LPt$o
z?QmJ;ie7pC+d$nCqB<WH<+_*MblxtQLr8(fT2vx^^?o>^p%|_^ijgz7Y)N!?#Gbye
z_cjEVcvZIjW&@!A>Ki;3uPO5<)%a??J8lTv10yB0YW=+~fpP1?T4(0AHOZP~$xqb#
zAbe>4$M~i{VK0CrQM(V7xclXi=1WpPBRw`ypYcthPXo2fv8D*BNuS-={)M;zfB&XP
zW_b$u*<g1u@?fvh6Z<HY;E&+?pKl-JVp|UD=&hQUmL{SogWf-6FDRXO$MQ9kdw)T0
zKsG7uo@90#kZI;-3@d(fTOmo<`$e26gSiz@sulm)`E(30vgnWBq@FKa@x?3n{3Uzw
zl!xMJF_(esHNcHd5lQVq*I6{yl;=o(D~m#LjSssI1MXcCX{6FD0pp)*fXIo<TK}>4
z)962bcizuaUTbS5nd_MpcRDle=rxyX5(-l9nEU2t2!wUe@83l>2Iwq|w$0e=#AkNb
zA6|n{z||KtYtvJ0-yh_N`Y?y-<RZ^pTCF<f)+}55X-X%;Db^mZrjBm|pm~oN?-z-u
zkBLAc9e_s9lOXrA$t)+&P#CR1U_`C&Hh}?WeYfK^W5wocwbZ&%Sy5Ld<-{@SCjYPz
zyo$B~;d#nXGb^v9jj#&Lez-3d{lD0I%eW~1e|>lX5u`x{X=$Vzl$34|kPfB0yA|mY
zB&18aq?QKh?(XhxSa#X};P?Fg=lsrnp4_kdet71Aj59m4v!8g!bzSrR$P0$D5SA4$
z9o`#Js2M!jWzf(CvwiF4;KuiD($<!8p^H9xI0m>_Gf1Ek6Kv5ZLyyS^3DOh)!JKcC
zvZ1a}uW<bnD&zbd*{EE?)|Q<ygQNxZ`30|SkvcSwzKIL=auk!IN!{rNh|>V^()XRm
zz$_vdr#rftR8$E=gk&K-LWqQjsNUhK-$a8&TW$8D!bngM^1FNK+|{m}Mmc}p=#6P?
z+XrCn5=9hxEggq*Q9M7KyvmmhzwFeoh6{*_V{D5Nx)3P$4ub&3z=V!Ym7OH1a5qeu
zBL#4}NyF#T=dat<4n<RWWAKAHQb$bl8RX4uUR1|s^u-KF1O%K_Ixr%Ah_fzptTm4A
z;LbA?dOH5$wM>meTBJ!bbsrJUN*H;k)Ccc~<~yYcgYG;hdLk?KWh?;qbLYnv|3EUV
zGWx|;wrA^Cbg*VqKk_q<Bbk)~t8=pnUcJD~1j3*_oPB{WZo*RNOB-?vt>mKgiS;<G
zI)^v#Qzg;6l>u&b(IW=C(?`3G)Bb4815K=iYt;azgjeWBJ!+r6e#NTLJLoM|JocWm
z7-q2*Gs8`Ocai7TP5Yzyf3!(=gE)y6I*Gdsk+5!rOp=+_hS&pxLBvK;O-`7#5~>v=
z{>nt8&C7f{@qqgl-FnB&@I&;^Q<qghf>*5R?H1nngHD*_m*tAKKK3YDG0QX7X<SA2
z>vF+ye6|#(rps|^|Bce;<lkuG^Q-($hZ~2i?l*-W?%+PRJB0*7?B`~F9(p=E6G)`B
z%-`ZCMv8)TNlNzg9hfP6kmwKApwP(Ds~KfEtV4N6Q^3?K0fWG7Y@Oyzx|R(RP!ZW7
zh%+N}9Z<LyOqD{uGK)lND{j+hvOYU@?ec6F|KpA!x3|^SdeTD0ZT#(0Y*ifq;{v7D
z_*N(s3F+1AgI%s#wFpTZTqW`1*Z{uCAmiO*mi;_m4@joXs#FO{JXGdLCSbxDLqUac
zk@Im8%k*#Dgi;i?+;q)z!AA)|<yrvLb)1RA*e;NGr<Jq63BTQoDrd&#>~*6zXYZ>D
z=}?WWIZ%U$ya&s*WyQ020R93XsK|CEtE)0T?i7M|v5w+5+)eSwCU=O0zPRN(`zo+Y
zp{I?{R`2w`d3uJu=VKr|`6;9KrnuHuJ5~1-_~Qvgf}HtEfL(bXxb$I$wPhZDUU=|Z
z22N<e4L96Iz-)yF`;G4x<TR)Vgw=k89=tx5sk+uf$rbig(@*1pYe^H}RU?ctmHD~Q
z%(VW*9H;tPHPt#iLM#BaC#a_JK{EZ4luB}3)>y->bkwh4T1D~Ayu*ojU%zSZ3Y6w{
z72I*QqU-K`oH!*5KVMN*@^$+~WrO(_Jual<e&Vqi%94~zKJv*`$9|7aG(?pRSP1x@
zLT$L!9iDhX9nIQo*oHCBEWe`ZeqqB~T8sxr7?dQCS`$AnJQA-NryI|h=NQnL$bDm%
zvWqpj84tCK6PhUW1(5A{5Y}r0$MQn92KZ|K+RLqD9&WG?An-1I<fWRd3#U`aoO#dp
zDidpIv5v`^Z}jb6JzPq^BE`%aNA0XY@>&96KxFk)2Z&OFr~i;Nt&=?Zc}2nkHk<Nm
z4-<zp?lMuN=g1QXlodepPV30@Mk9tE2mod{7R(0VwwjV(9+BO2YzkwJBplH`|LEBD
z`PlQm7DhkeBl*S7rf{_K7d?VXj*^yh^DnW2T=&llLy>jf2*slF*rn!<52vK&)A^}S
z$Mj{Wyd;|5mzlIt-8ds8VHS}MTurBbdH{g>@gu=k59?NAclm-8TBYIdiWo6}pC`NC
zkuzh`hrk)#c|DrxrjbFO0S5+peOIgvqlq&|#l_GL3V~RGzB0_uN`dVw(gye;4Dr@V
zIugJdbE^%4SKVJh3bg?F=n~(`|BbZo#~~?&fNIw1aMKU5+}Pa33aV!4G0yZymJDHD
zqXzq`eFSa8B?#n!geZxe)Yopm=dyW-!n#n|{l=@1d^h?d!(2ZY%!FYTr6gELwuu{+
zn&JzMDYiOb7*&z8U4;LMbcr*t@B65VK(f|5M(*c&My^5ceYq>ww~;Hy5MNS4@3Vmp
zzc|-1#oIw}j^@50DJX`XUncU+O;H?{E4v#2l%*_<oL2KrU*BHt$=ly~j%NXTYlSiX
zuQN-UPE50Bhzhekt*zHIL_nZ4p3yxlke=yS*p#Zr$%fa-k`T{_qS4OTYi0t;OU$LZ
z5~m#O3lc%~%JR>K*7wa5R2P8nSOSxAPn_oGyAJFosCd!T$bphyCk1HOt{V!(vtZ5B
zyDmU_Y<1wXz}$0QnZXu39U3c!f6}TCiY;8d8fGj`MjM;e?X25DW)#Aa0Ia`&V8&24
zT$CX<TbDexCY)b46|DSQdW^B=slEIa8v!d`xoBW15hm7SC?N8p>$&m-TohGCS-LOo
zd$sA3QBxV8KUqIk$V<Cica%rcXl}B?UAVjF^Fjjw`2})4XK0Y0v{~7HRQgh|LWik4
z;)&vr;H%7`)ROrUOGz;(<E=N37hn|DV@aW18m}La_2l{puffX&IZu-dwEZGUcjtw*
zO)2|$-`o<U0;}on>B9FF%lz2Uz1d|2`+Rd$#%eqQ;y6id<seT7$qp_QvG^HZHXE7J
zWy|Z!{cuxz(^9N!HB^MMi&d+i<avI5EGxlFt|XXraTZnMauuwqwA|R(Zx%%`o&g4I
zf%V(WaOMZQMA~S-YlCQY2FOMRoXG<kI>2J_yA1#a7=`L(AWtuA8pm_w=@|3UN{8>E
zP8Q9~dPz4H2YU>;H_a)Q%-T&&%hUcX4r0soh1Z|o-Web?>z3XQv>+ma8p1(XnWLxu
z{K0u-0C#uuaD>uB8;fA~{J9ZI?TVir$Fz;H3(CpEOHovSzpg(u_daR9u~Txn5;aI%
z;pcz4o%`7BP26ceU&)=|`4d`oku=c`GluWTPph1d2Ws&0+?Tx;+5BP`#RA0g<Y^gh
z+*FEN^t&=7gE0eJ?P|^vm97Bbu9WZoLSQKI8AWIOq{*@bIuH(;pNe67J5^hL?EoM*
zxu+tX<&QX=T-lvk<U)OD6-f&B+bWID52%kv&dc4StE)J(80q|mp#?upf)z|`eu^)d
zx*WB%%+Fks&=vC(Lk6sX8w1tThST@K2HqA&gDK`8CO<G5iZ@(GCp?XdfTI_3H7}z*
zNAN%1foNy|gs|4Cn1@kFYZc(AOIP~H1)IP*HZOnKi95A0_#k;|J{Sk-nVi@|)!xT;
zHFy%wC(fw!9+z<4z@kX+N|%;^B7!Ux&cl}*hiA<r@r=9;Hw1-d7iGl9nmSZ7gK~oW
zPPc(*%(lLl&65_pMmm-UbfY$9b=IHHZ*x1~?oGP&Kj63RF`Iqu#6c{!1vN&E7vsDM
zHm$I3;e-|hSm|p5Bh`Ax;PiBP!EhG+4|ZY;=jkSP!fIKE6O0!fLg0AjXP1plc2Nh<
zx@VK&xl;kWY7EfPE1>g9%bH0+{?i4{1Gbq;%-UzS+eQQVPpc!NkC!+aYK_|f2QWL^
z8~;0xNZ(^;bT^RSI}#Wgwex&UIX$%P_KAztB`h9|xe46JKa;e#SEqc7_i;o8Cit7B
z7saP!Ou8_;%2WBh7``c|lE)0TG#+nzS?=jGKbim_H}Jg3zzQzhOWvYxBLuuXTO`}D
z<@nK~pa8gJc<)t6_U3z33jVu=!ekBrgago<QJfc_|D^@^_WdtygX#bSKpO`&sY76{
z0ztsK{P%u_ldXaBhU;D8@{AGFjrug?6WWFPHxq~(%(?@(Q%-K&C(k&c6h|+D72YZc
zd_m&eF`hEzw%x~RALG?$z4NR+3Pd9$QRquCcbWVEk@~dlr*geZ&K!SbUW<@PSN|{-
z4}kPHM-a#A?HFTY`|Cxj6)pG_>oEg;bq(8Tc13l#X-&;_3@|0LW&P7D?Zm1r-{n&B
z*Bi|w-%k@PtKJpr6`w0%jc|Q8c5tJe$IY|qQ!JaWzpuxn`6@7dxR_+iB%ajtz*S6^
z$zS5RYP6E3Sa>@W9^Hb|;&@Q>RWygSL=c)Xc@ZthJ>CWm%FuQ<9=e3y-%JZY+IPYa
zo~_l}oyekih@;!t{F)FU&~#sN;f?k^$-V9459WOsqFE4bbpQN5>&IOip<H=mZq{T3
zIebV>j6!kcHm5zv|Dno80jNNKnGeqKN?)$NE^Gf2E~Zp25OY&_G`DqCTfJRrKGHbB
zoaXL$mF0DvSg^Bt<aI7&jagR6e~7e8Uf=Y*+iYL++@7{U2*@+!u$de0DpdN2uvbo1
z+B8IT4)!TQB8<RWa<YWW?z^E~3$;h2q^uf-=9|L`#Qfej<+Ir@qwR!T;R)w@m||3J
zWe=AX%xAOBXGV6xqnI$;Kd0SvJi#_w>Mpy({E3H4R+^NT(G7V@Gu*u85>BV4eB~v4
zNR(K=LoNXEw}XEdki`K6;7XX6!3sdokQ(Sbs+iDB9$|)K5qjZBwxy%G5dbX;oeeQa
zlAZ32x1^3*mR{};#6fY=EvT=kDSnMbVzHDuBJM<IHtw&#2*yf5EKL{3<m&D#;_vdS
zTDrNCb@Se<)B3?Xw)iTBOxppJc~JMyFxl~OP^`Y!wZu8AmvSwSe0t;UVFQ#D6SQh&
zPZ}cTC&_MOb*(G~OiIci5SW_=Tm_5`_c>>FMWwlk2aM>3RL}bK46py_fCLd~-^f=x
zeQYQ?+$<t(%x=3<bs%e;IUIb+gP+~jZXannh2$KnW0zEQ(B;%Oaij0;JhQ^A3z=--
zo3ppRsEr2n2Fr*S3EZj1W52-D%xO2fF`N@4D_&<~dMkQ$mkBN%4j4U7Ijr6{r(SpY
zS=cM<)K3Jh&vvBnTe#l6Y<Al8B3^-Q5+uNabrN(*fH#{0$l<%>`mbPf&Jan>+rR??
z-tpl+Kkqf^^UWVAzwAH+i#61q=jQvW3uA&x%Vc`jhk>)<0mV1X{8|9BINnHPY2H;e
z_JdLNBfwBQZnc>Ab-CGu8}VV0>cZZC)}d&y_7jUlfUU(}{-9Sa)0D~ZVDj8+n|w>h
ze>Wa5BQ^GEyQ0J4dk-Tv>wp4&dc5eRwjet>2*gb3k3?@Y78V&@uKCvW^v{DvbW}8*
z*z%)jYO&69E3B*`bo#WrCj&Tp3!Ig@@M_rXR#I|tv8IlQ3c_jM#qQ?8a)Wl8U}~#b
zehdzhRJYT9+hjgAdp%xjiDX)}t{JwDXmcPR9@k{&-86j`F0<b$@`4%gX`d}+_4j8q
zoF2WC)QUR{V3?6U04hMH@4xKgbb02qWNJ~e>A%?vrI3bX-@<P#032N?WoS$EC)yq=
z1=gJOW|ut>0)M5o{4mh=Ly^;BBy~i6Wuuu&n;=olTAhsN&*6r#M29B;2u^o~fIwB>
z4h)t6yuov+j1{}ghuSW}o`=7FJ7y_YPAb5QjODI79V(mcV*oQ`oYzhfPJVmed(53n
z!17^<#Da9R^xEcClabU~-{+~Qw8tw{-x46Z?r7i%x*2>>><Pjgi%PqzqOAN&GsGR~
zaMk%5oE85KSBl9rA#NTX_-lh_R90kyDh-#Zxk{07-bRPVO|nAZ6Goa!&g9AIQ&oez
zt$C>nN}kG|=%(A<7y>zs7u(E@P9KwM{nLM+Hl(ziR>99_PUHpscFg=>(2<7*`L#zD
zYT#?^iTCM}nQA-$Zaj=^Z#hG6gM(GykO?nbA7xC<w*o+S;xt0lJeWIisbyygs$_Fg
zws>K@H$G-E!oqsj-3I({B&%@8chifUbwI#-$vAg7_FSoKQMJr!><jwa*CKF$(m1$i
zL4-_W+c6mMCEg$puf0q0Jlbv})@Qg?e8*2n*7T6;C?_7kODy{Y<cthJ;aVNU*rfBb
z$+AW2&E{Pj!9_lLmFDhU-D|YI`5YZrrN3{p>x#D+S(~jfiL*R22GqujU@9&j#WXAH
zQo4f?nN>IvvVUpWi(NV1L@MoQRe$ET16(1H*G69m#ZE?7Q~TuTnyn9pqy4a+IJn&w
zCdHz7l-+T(ki?=>JT3^^`|AfkV9dzf^2&b46H!sjd!G&-_a(t_5+~3;(&dgEcRao0
z9wbs!U*$vb7Xs7+Rrqc&#vve~nVlCjfDI>0q1jDn$X1^WSn<zKyIE@C?Y=)queX~i
zBIQmaNniMFOUo?iv=YRRR>j%R%Tt#nKPxq{kI*0%DiF>AbnF*GNfe&qQXODj%+E(_
z+WXG{g(2(nO@zzu;db>fr&j3oP5<zfJkV(d`fd(l3MBXCChPtnQO&c%`FxGz!adam
zjOPG|!HjWVFghK(j*9JHZL&xAA@2bV+~Rs6Hq)wha5FwO6g`!4h5#C+1m58F{L|TX
zf3n0+CW=axf^Ea(^w}POA#U47*BKWkSu;EOyHq;o@#V&Nmh4w<0p05EogOwRK-Qp{
z$l6{!MO1Bp<;GmhS);DyT(CY8{v@@yiJ=D36TC${dMdo+0#(}~3YQ;s6`xPVOB*&0
z*`8bd<OUJ|5qqU(Yj)6IJkK<q_4s~2y(U^-ok)@dbC}P^s2f;fK=NKBQ5(MK1Bfp^
zY;Cf2Tt?1v-q>hXIhl1O0Ly)>`W)QU0D3_F#fBY5BR65sP5C>#vD92I)8v-1NwB2t
z5+?`vZj7q6Ky&{lp>%~3u>YZ52V}QZ(eNQa+q>zWgotiOR&Svf0F*FJ^@f*9dOf`i
zZp11P)TAMzYOoFh4u8W@7ii?Mp1;`~>(2luPvEmHY@I98MEbNrXR+;kZ-Wwxe2`yc
zzuzMUiSXk7K^Rs2E3;|J?!;`N#j*_#rf)!ab(xmeqvE_n(p1)92L<TU_{%jww95=r
z55!cl+wOYg!P5kR1&ZN7hE7K(h6*T?zkUDa@C|3(I_ha9aRbG~FC}NEbs63&h5e~C
zy3?fZ21s0z|7yTNfZPZMWV}S}(<dcqgZNm~QQ2>kSdTsH(C7Q7JagV>@|JpSx8S7D
zyGRxO==fv`YiUfG9&f=U7M*Q2H7lztY?!wM1ifyreD@hUK3@hlVz+AGMB)zJ%`(p{
zH54VW;}2O{xesUrb?nY=lat<c)qkxOf_e><&2fy!b>d`t>3R>CC9XKaDo!3wrOKPV
zpgSkJ?pMz>n40!~%=q}4pS2@%7Tj-!X7J{O|0w~l&(%<nctWn9(+Jft*WZ@e3;>S<
z=Dzu6Z-3SDJXh$oWcbWtO{s+ff?6%0h7m&rA{)oha(_`kbV=<Aux9W64%=)KUaTWh
za-iTYAZuCnv%biS?+DM6Bvb?#6(N98*$rF^7IQ0eUc!b*9~i!AaKh5(l`mtd9TKyd
zXtf$?3Xw|(;$WF|fT{$d6TNK!#Q@MvoOYl)Qfeqk`?&U=<0+v@Q(MFw3w+^~_r+8u
zXcN4u>;jm@*8uz!Pxi2${xngyai_S{?qJ~Hda4n1)ah6sE*~IQ2DCS%|72nGlU!*b
zkbY9t3Qc_)?FQ_1;YImF_X0zKnda!z%S7aGySu}69U`ssbsgK*;=`~1;1H2FXuISC
zC!XLo8V!L)aLf7q4A(5)sLr~3c(TuS1^Kk@bxlg^M>n_rdC0uNP%#E5vkNFKarsq(
zQ9b%_iGxU(5YQE#m?wRuO5u%f5Pror!0kqv`qsThS{<4~46nZJ>X)1P=@D}na{2AT
zy^t@OiU&sm$PHUh4uVbVD1oJ?ei-7(B-`R={2^r)IeKz4Ie-r=&1<9!f0WWZZkYo>
zfVqna*lQcC8@r^25sf!uz?a#O@IZ0Q5uHx8i>{B<o2bG%_zpvs4b5`9q};ztkE>Ok
zKVfNoYVN<9?m1@&0yzT?_MiI!`!9j29)d>h8ymsuijO@ApUT<T=#%cu%F3d76lhz!
z$5aH1I3)$l&>W(gvX4Du=3pElj>6A${pz)E5)PQN_khOzKW;ZF_}CzE5`Z?!nnZtW
z&lcnZAZIL}r9fAX%b<|A8&m&}U&IgmB2Xri(`myuQiRC?kN;8~1O(dWc*JY|+a^Ny
z47lXqx6JbYJ@&tC01Tx6-^T%_YXdMO+2p*PHZoir&};Uy%KkMQ{|-SBVhrcUPyG8f
zDie@){CEhq1@JH{do2HUdXsS1`9BM}{Ku)xACGTQ0A}v9!TbYE(d~;V5jkLH0R@;#
z1J%D>UOgp}&;M^n9{-c}KNJ{i`5y|5{qHf5e+$p-f7+EnMfiOg|GOpR%oUh{(t}oh
zBj~b*S?S&hb@vy1qi}&AI8*Mk6AG`G&;vo{{cdXEI>2W@pk}H6Tbxfs4!YIglXi?a
zI`iOtlk9c;c63i41Uu-X(|`c?$rN5`-Y%bh5KcQr9&LTNNQYZ2H4DWckXGG?#;@Fy
zKa42eAnqDu-21^372Oa(L|p%C_x@kZ<$vbrsN91qj2Y+y<{MUE@(C6ZN2S864;q0Z
z|8=rl{`#T_Xy5<a_U?ZuzUcEm6kqhQ`8`1jDChOBGF*LXrxc(^5S@2ZT%(NrY4rc~
zmj3gY|D#!WEZ+0)+pg9B9{b-m0POz%r*ZI)yV2Vuf7oXgMLmgzJDvB_h`|+5*My-v
zt-H$6+Sy9#Kd$Zh5JbFg(;n-~{l_LE_ZT_tsv>i~q8mviOLhAI^=r`7ea#|(PRolw
zvJf#;IrU4BL{TK!99ZtF)#+cM9&oui(qlz9pZrCP?v->;VGoiK!l-89hmpr`hs3Aj
zm%E?sJ1<yvHx`~Z(k6qo9IO?ft44me<mj@l3u{lR<IMe|e)x`Kdu}QH*ID722$WLu
z_f=yHVE2V&-6hJWzZC@JjH^xI<AEE7z|Aj&p;HFVQ{vN}W(yA;1~7@hyClq#ZTQWf
zeWDaw`7_<)OC8|*RL|OYJhq)#QijAo&~ww&S@~J|<KtqEz8P17yq|yrwR`^W<{$f5
zBknSBoiO6@{+me?XB&K;M!@@ju~j6EW~N5WRW@+H>tkV_FvgL4SZyI~IL^0Ff9n@0
zg4Ot%fz$zUjAe9eTAe@MkGQ<<%dGg_jln}8k1=l2MrDJCT&cA3Ok74%yTdVe_0zo}
zP($H7<k^?ovY#bg#tPlop5I@IN>WB7zjTPhw0|aC))mCE|KPAVH8u5dYO1w(;Up1f
zpFDr~t=>nyqvLlws@P?gVCMB0v(LrU!HHV4p)K2{4W7206@Manzph$GJ!^m0sK10|
zq3yoU)iX<UpkkfpZJbht6h}4&GLeGFxG-L406+G>Z9gBwaL`W?0aLd5b)<1lKDBi*
zhx^3HMbicCiRWsEAz268RTK(U>b8yp_rGflt={ddf@qVUMlKwBUX`7Nls=np>=pw#
zu|PKue_!Mr@);)nJnGF9>hDm^f9rev!&gOt<7uH<r<pBC*yg{Cp?@Aj{jvE$@_b;z
z2+Q-LkM8@`mniYC2d$mu#FPi;z(!Kb;Z2aTySjZ3&&#0CulfR?e<lHz#8u;@E*+4a
zZ>S9P|EU09ON@-F5~)JNDu@cr`}oJfwRwZMlRZG}>si?u&Sq-(qNA(-_(q5x;vP!z
z?JVilhcc|z^%jcW)dF*^W;eupN%Zr4=6{~&KmYwB{^;ZXJc*t9C0>gWb?zuHX77A+
z;OCZVQ>>qeM#v5n1VRhsQo!>d*|u@=Y<-(Q!yZT>;G*q0aaO@U!<vezl$M}lFXAH4
zRmwqVHsjSf{loxhChzS-wj45(9^A$_Yn0-YitzmaJ|Wkhqvu-E!tXylIi0}&&u92&
z?~~ZyXoEqi`)wE}CHljPoF7=wT+*&!A*igl4zmltgPp2!6%trQ$#D^yzoh-)hwsFW
zJW+#?2Lz)>5syUKq3AWZTZ|3pqMgH2bwpa!o`86dUT)zzK`e7)`^<>1cNTx`Lhnjh
z`jvxTtzdzL>aw{IPBqx|sw8|N(KTNo6n~8dM4wbPESXOwS(m$IpQU4Y=xqIFvev;!
z{i@`MYn%~JVFGv&O>62&W~ZX>>^pBCrYm-v{=tSBeM?kr9Xh`%+gq=dNWtF>knr}_
zCm>`uo6Ai)zjqf`Q*sq&DBBXIGXk)V1`}@V#Oy2K)~jO9L~(KBx$V5|KwtcNpNICq
z=;7tM7mb!eJggjz+hg@g!Z)|7Rbi7KK)EiLA{&^&Co)bcyv?ey)5NyaL^C`$y!R}r
zNjz?1GWv-AsihG6VoxQjHJ=_0{K{<tMQb<6mZbPZJL8S_i*voPQ1HjSmlnzR^QyC0
zfO&WU2!(;hFQC>^$gIH!)7lm+M1{W!qA+}Qc0cpr+|70z-g|{^1`P3`5GMNp_20v{
z&4i5wVI~-c^xa$oPWRcdkfM*}KL?YLoJ^J*eN^!pzucnw6~0CosXzTu)e+&TMuWA@
z^MI|sOh_W4^#uA)uYb>3zWR_Tg09y}e59Wrh}}H)A`^@OuX+Pgt1j!tnxT?x%E@_;
zepCB`cGS=&UrKI#bSx`S2WD(mH$#<u(BRDFY`!-ARVo<NyOK1$d&iznciA`-E!BvU
zFW_=o(}iGp!1^%33sU>GKz9#63Ro)!OR;v9>4*6^?$=f&!VY_e2J#4Mk%GG?c%cby
z`@bV@?)D4UIPtVb!N*da_Mmv{%!>Qug_~akOZUXpTCkz3t|JlrFGQ5Jx6P;pv^}->
zTY^4@UTIR*Rb@TVHl@#*p@w;{uySms*k>|go#(V~vho5$*MO#N6a72A#0k>hF&k-r
zD|Do*3CdS&bi2>%S5{Vg75s)fmdg0SZnq(M)N0#pwn)LiYp=^1jf~{GkdDrW=mN8(
zk-Z-_CPYCLl#&7uu+kIp&H0RB<HZ&p1izUlY&(=YsibOXh_s9LY&YM}KE+S1a2CAQ
z^1+B-<>alU()^b#sz1;o7~)FMiY^boCJFO2Fs*jBvFUFl2Qj=i6O-Dek`$9N^Rtrk
z#jNk~{>3#lHkw3n(!uCMm75pUJw4rG-GG|eE2K-_d<?suI!QMmfI*ZmbN#EQ;CjoX
z2sp2E1dlRO!#rNpdG6OrDfwLJx``>oc-ptXcTx07rCp9j>|p&;+ERyx_vN-wDEWc<
z_cjEzUr3lsgsm}pR=Yx@ty!9_6D2O;aQidM7zjcpM#PbeM97N!+O0ad@M1^^nJ^fJ
z9F2h$eTgYY<LP}+^!&be^|5&Y6(pL~FId--*c>aqg{PgJPlfL`6Ya&-(~XxAQBl*A
zam@tjj&ouOCw$ZDRyqQ&7zG0Q_nG}e+l?#@ybq(eQ$@wC-57loGn3<r^-V3OC1XFo
zq73rjeIfgRzblwO<IPiC*lJrmxhU^DUH>*i^`&CQvInQ@ndzb7AI{fbX?=!@)7scD
zI@qWy;ERnuPUn||T2s&IH+3^Fi&#Hpua1F81}98hpcDZ)+9r!|xzo4t3%C(VIzW{}
z<ek?3unU7W6|H9EuuM0fZt(~QNCU6e<9zT3d!9JFSIvaDaGT1nyvpZwcJ<gXbieq~
z5@XJ=Jo*M^>p<Qz&Wp$G;#_89^Y*Zo&p-R!MDiP4sx?SM?FfmdjRR=@-9@?E#V)e{
zr+Oa_F^FXQDV?2QCFIIv0(~Wr^<h@%u=UbBI^#0ZK|}a<D)X^V(!X>^>=AEV2Bm*_
z*oL&bWtsU2G$KDOJN3d<rUCk-C+sxnI;AD6&dmN+c7@G~A^@-=Pk_R?A}+5;E&SYW
zzO#*Df>wiVp5`PASf~Bo=kainG=6U~6b?pHO50|CGm}|=r}r-{z^F?Hn`qYe&mq4g
z?W8HgH}7>;LggH#+PZp<b?+w#qY(l3!p-k$YcOe~UgQZiJzTH%8S7v2rZLh>+V{zF
z4f>uJ8Zytkh^R}=43=bo>LI-l`}kD;1vQa>fCsoqc+Q5wd4{3n5{g=*jo2c^JWurT
zBr1YN?leg!UWyHbpphkI2cjn#_Cr8TR<N57MzdNwZA)Xg%XWjMoW*5CU)@6V9PQ2K
z7C4xYYu=1)BfY&w6|pZpa^^F<v{8FQtQVPfO(kiRc)5k&Ha2A;tdM81raZ8Ame)YE
zQ7AaZKNDtTFKHH_Dq)&2<4Sa`=;kUzR!;#!%Y`3Z7!^hjYC$G-7}d_Zl4e$f{JMx-
z9B#Z1`npIN`<JI@;7LueKSjH&M2?5J{^$4CldV3V6jnl*%yn)$RjbtcyN@pNS-dY7
z!cFJ9a#w&g=;+YMOxfqFgu|6#7qE;}S&vWGl|#a1Uha^Njphc}5rF~b!vLY&uihza
zESSfKg$-<i>~fIo##Fi$_=I;Y*pyViRE6j1^tG<sUGgsjD**<dV)TVTB#V{b91Wjg
z&F05J3MVfrg+a30o!1T4jvP=9@J#MhQD8^b&O#3rk79LcCCr)16DpokVp+?4)gF5J
zV6k?Z9<5HhF9TziiMI1=vqw7zCM|uv%+@i;9~xQRT4!?eJwwC{^hdnb+-lOnbTy(E
zkvn`UwU;RU-4K(7;v0;1jEoXfo_Vz3d_`5(31xQTXAC}KoyvqH;7d$Z`WKbVyA*;U
zZLJbIIctRz7#*kPiY6jIFDhVa#d$V6i^ZXT-nDAv+I<7A9wXk%BHlmqs<kU}i?qRW
z&s^pf5&4Pm+%`j;0jH=kO3jwMWUmPuRp<iwTYyJ$1R;r#hP4ZSt(`zeJD6vZYJZNy
zWr*NPQtqTxC3$4ANZ>01YMrdGCz1$-{R6tl>E)urU9M&d-qLsJPMny$06S~G7t<v)
zm+ne;3-czWwBnGb%8}M*#M*T=+%d)!l*ISi+V@9N5yOkxmhx7Gtu!$$z(L;Y8u0bK
zMSsFZR^YvkPKl4u;`sQA<w11?Ppu*n{V!~k(j)>7r;pBJl(=Hz_^ns-L1J=d+AB>)
z9pnkEr9WzaprpO=1C(pZeUd1CUfdpbdro$>F#5-vPh@Cc=gh#*>P`CafGKZW{aa8P
zee%@#XoMV7=U5WI6*l5Bgg$hMiJSCr+TN40pkB*d^CCQr5K4Ixg-D-<l&IQDH?;Ko
z?c>d^f>U^JgySGv+mc`W(zO=SV2AfP2@CULfQBHxP<oG;k~Q@=+@AAxyk&!N;Ya-|
z#)atedETFpiOAivkKoTI-Pd)^c|HNGz#A`Wf~<4<V;gky=RS|2b07|j8;Wu%rEVBn
zI&C|d%7W|Rs}(5d$@eRU^}V|nKw#ViI8a%9>DEqLd*i0=x*qF=(@Be!7056fhKx8v
z!DVFU(S1TC1upvqpJ=ja=SGwZQA61|(Q(fPWso$zkSpB+dRB@C4|sY*eRkgOtm4b$
zFJ*XIQtxC#Wzbk>&#@-b<8{qeyA2GVyf&U3>8ZZ2oY&JDiKkrf4@E;OKfC=D-^t!)
z#sxyTD4E5YS*97fvh$tS&UhJvHoxV#Q4wx>Pas$CZfyUe5kBSiB;~i=lgxFswlU%Y
z$>(c_PtVzLuW^R`K8t!yyi*6pjE(D0=8CGj(Xw=xz7JiHzh*~kfurOjcdj62=6UWn
zEg;}Y=-4yi0bjbg@MuQ8iqspgWpCESXAYg-*A<hqs;lZ4SId_UFO{~83z}Y8*Bt9^
znH~;`GM11`l%Do^s;qop4phm)E(vSd#4)W{&kZ<Fap_p0J2tQrd;^%A`DsApo(=ot
zFh2oe=poRET_cF;vg4f;NYyu0SGW*nA%jgN$?YGXW+)12R2iy!*6VtES^oABP&T@5
z>Yb8H`fAkLY=5V6zWfyo?je20M9Bdw44M*#Lv~R@0ozY*qp!Io%#JXXL$^gdw}_cS
z($1uAw=e9OxfTQ*_^_v=_w(A|XzoOyU}}Ncs`fxH908Rch^Ru{=L3<^)$3y*=%cGe
zt~;@eHyC+A{=0PI>NFX($R6g8v2IvEyw}rcyV5G?lgqCZ<n>XRKvF}G`xkg>Qajk>
zswIq|P*8P1nvTW$HDi^wQ}R%I%_HEFObrp(8iN}ifj#^73l9i}iQcsGYd?1rOV~ue
za(^5sz-Q3t*(2iI^*NK@yHLGABMibt3&hZCb-UlL5l+L!jUV`Qlrl%-B6eg7xT}{j
z;+K-1Jt_r3laxiK!au=BLtx**msUfo$~vmz7<w53K0#LwRo<h^+_s46PBBWuSct`8
zNVMk>PkZ-2YV0!fC-vxm20o4Q`M@lj8Ac6@erLsgWqpdh^j6uAul^`_|7^8Kvd*Sc
z#dmw#^~r^wrmCO)1CSbm@zL-1X%C4J3i!6i?)U3sigzT((FgU_h@m;WKz;a{J-rN$
zjP^IUVsmpP5{=9OfnHV7&Gg1PDJ9(O=7a^JdXaCWF!C=kO^m{)V+@i5nfE%<i(;Jl
zesN1zo@(Rjlt)YoHS*47ht`bvdZA_V4OjdAXvz^hM-LXqFzI65#&9!`StP#CFf*0_
zR}chR7sDGjk6IxqO#FziMQk&miJNcBOyEPzgQCXM8>~!%KcxHXJh$7AUIMXrCd|t3
z7<%;4#c6rO7uwD%wZMsN5AEUA7+SdSz3x=iH_0!*wL=%0&2ax>umVv}U6%ko85Y^+
z<#>GHt2}!*jyGbh+{!;^!*ftrQb<Ql1Fx)vI{5T!7N^OWM5I7*!{|M=1k7GyGO#7u
zrk>)%e!=-*9p<eC@mLE|a9HXDZ}-77tDZq>?4?`H&Kx$N!DN6X%0Qsik}$e_<j+ry
zgz#-5AKRHhyhC10Sq&(K`ZqBWtTYYR)X($W2f}jsbRjVnG4__$t2DGF<wE@Zppb$R
z%nr6eghC1h&h?v8X21&J@iMj1Y;;Yn7ct0CJM#_y;c6rpuNEh#=6=$tC?R}nV9^`4
z(ST`@oOQrpD`rE*>0`A0TuF|m$&Cb=ZKplCYGWu>BYG9x@sIcec}V<2n>MZhmij)`
z?eb9yB_4U>kCta8@xO!%$VS}RN|c4L^HrV^lBl$}sXj=tosbSfRh3QbXNwqaZ=#;v
zN`e@o-g)}@9RwL&24$xgh5Q{USJI?R&>)SmfaH<^s}^^Jh5hEl61kKn&ifV+>bn&!
zd`h2Wu1zOrON*nEeQPdf?!FH0y&Q9CT*UsP19<hKhjW$2SS?wMi|N<yrSgIK=_J2q
zt7YHS*0SCfR?ye=>6yMi;mAs?mkf|eRdi!_XvtEpoo(nQ`5%Y-%`DEu$)Vhl<`!$n
z-NRtG_YVQuA$_QLo{)fmQF?E#0Fa50*lM=M1g12HktyCuy5``NVsp_h%s^5OW?WBh
z>Q`S5oJcsS6J)c3WH-gNAwl?Lx^IeO_g;G3^W|QTUJ!-_bK&-YBahw)YX{D!p(WK+
z2UnNI%CAa4`x~5~|1}!AA8$6R8B(&1LM7ND>Ksn4hX%a#8x2Eul&hSk1|<(D0r0~V
z667iH0<F*;i^?{n_m@4NkI=7J>0_LZ?<aOXye*$7V~te!tijm{k<m-+2kZez!$;xT
z8)O(A7`LZHSEiD}e{{08J5qF%@fOrRG=`OeWs|;<Z>FMKt&m{1o9w+TQu^Eho}~!y
zy#JJoh0D+?f){Kh!+%;twD{o8$yhl3cF-0@MQ2rQ?|C9EVnzm-Aw)sKQ|slmR}Zh$
za^nrTfx;~BUuIy_BJnDUXRh&-I9dD$C&P3&36qY_8_o-C<0(wg%d7!UI#p4La9Ywn
zD#;va1MrZI^;wN{e?<>P5tWH$(DIvge3J1bl`%<S8zlPVAGOSOeJJ=I1j_=d;rdE(
ztq%>oIwn>F$y@Br6veh6)HthJ$)SZ9Lcb$D_q9|L$FsNwLGLYZu_4tvZr{h0aN~>|
zeV)h?%FN{>v9K>;;O%k@-Qw)P7UW=5J=4OT2IiewYR)wOywP%C3iqC(FRpfqTglq|
zv6ZRI8fHqR78ecgW*W2sAv3l56bLRPBEX{R4UQwL0J_ob{^UR^>0DVDukb?EChv80
zoTp=G4e;JJJsTn)jyTm3>oFM~ZYWoCZSC@!TmRS>8y!z0O3DE@602gDA|q$xRq2&{
zu?q1;2kJ`bTsQYs&2p<}R1}7{D`Y~6wa2jU=cY6@#|?01WODlUu8YfPMHMY+NlC}r
z`4?<SsOrAx#zIfJEZCu9nVI!E%x%b`e9p_H=%s}oRgARpan%Tm2kk^ZQ+P>j=H2<w
zmC=F=Wz1G&d5Me`PxH5nmaV4aBb^mHn}lmADIvK7LmivQiDHEBjs7F!C2@N&iERCc
zLArExD-pf3Wi|}lXMv)WS>N*r8_J@WT3CU-z<fnc-dv9_XTUDk5nvH4-&<i?I^eQD
z=fdc~k#h_V)aHpZ_Y_EOYZ4^ejOjg5rg#m6Q>lPgN@{U{wa?w3)p3`0L<4{3a7bRd
z>Ngt>AE@&O)IZ0s8Oug<AMSb#(9dSi3de6yJ{C4~fQy1M0;Ygxa^=^&!yt~ej3BeW
zmzISWcaL5Hk@nNmoFX*$lRuA%oU;#*8%Lp}BKn@(Zufe*wLZ8F#VYOOL6%WPy6OoV
zZ(+;^1kI3#Y0UK>{{8KDtiyg=Kzf`C7k#Zs_roenUIxEV1GYd@StW0yRV7gUJ;Zu`
z%8B)eC{&xz+1^Z*NAC&Ft3s&=`H5PU=vR7OVb$jH^?wS*>blqPOWW{}$4BR;R$`aq
z%etfD!YZTC5tKFk?u0zN`d3V4E8Q>4HzAtq!&g5j8T%&_xCe~-b129C&O7`_6%z<>
zV^Fr4U<<YCOv8EMAlc)E6jQ#grhIZQKCwL%^2m?~c(46Niq*MVHxwP)09xVW4&4Tl
zuYTGABS;52xk1J0*-PJbzmjKEAcijH1K!GD4jL_iUospV9efOx@1pa(u6uqc*bTD`
zbR-jJ3sO^-lL^{y1;K8zfmbL<_4P8O2oPSx=LKpU_x+5D$%#Me@k#V)(<qcLtT2Sn
z?ITk;T?6rs$Oqg%Ul_RFjXxY8XKC&SRLm+{-zeJ|-$gaEP~oG@uN>8|<|agRO=c^U
zSDTPbd~7tX$G{AgKAaelJV*<VP9;;}O<;A{3$xM9%Du5n)w19#>pgpkN4(GM0{GWX
ze0hT_yl~8;WsW$K^j%F#uvY<ZBD){e(ZZpX=G1lJLezfshgdyt+Qw~!I`ng<p|F%?
z1N-Sxqvl<oUc9K+4DlUooo_Nkb3aK_c2MGGvy<g!Rr09yo$X@c;&!lMP}z=5U<T3_
z`eVEB0#NZ_m+~7j2Ia>SWVJy)-hKy-!qCbOf#?Mi$2}*73{_f$092keN)1PsuTI*W
z8!pz+=``D#hB0SI2E6|HmcUdxQhv3G;>E|zc5vmphbI@(Du^CrB^(KTmW*m#SHa(p
z0_17pkgtIu%6I(7luft03*CipJrbn$Bx8^G?Te?%dwlSZuT$YFa!wEEz3S>3xGyd6
z_OQft5|ZwcUmfHuDZX88AWL)^vt@2t`=tJ!0)+U<#<hz#%=tw=m5ddCX$7J7EQOz#
zxOTg?lmH>{zVhzwWP4UiBtuY+l}-L<bacmJRFS?*zr>GcU7TA`&27UjyMfp(_A3;F
zTFJ&VSzkYsTWG`Dg;Rp9J7aMs4v6ORXsW~)E90Rz#}p)$F6nhmH=NGwQ|9sc{l4?*
zQQ*KFbp#0Yn?@C*K6xw*+44>#V2du}VVfzj#RS&8D`|A&+p`kB&RfRqe1a`peP=HO
zJT#i(M78o+HgBa#b#71xn4my^Nyr$GWM^)ou8Pz+ydpP!Nj%?dqX^++4|e#36@G0S
z^Ol8xL{2tJr0G09WWv1~V*^{*%FPF%B5O&9C2@aTk?Kgt726{r1Q;xowuzq@D@3i^
z%AbOdT=lM+b>??Sbey`3PQtEQ_)q0|7PcK%z*#@<^IVrdEb+>gxD9|A2YhO={zU&#
zR4ne?gWdgBXsuNs{J2NF$0<F986Sm9-e^=t=p<Sb0w8V^Y{$i48T_f@@Mp8g*fDU~
z@hC_W%z*NwrXDYwv(n?weqxJ#G@Q13)qPXj(&%~Y_iH+-t2`@9S-Q)k0M2O(3FV35
z$nd!0H6j>_!<Pc*52_<_C#|&l@94%gBhklYBFjJFv<?rbN6A28?0JNwT;vpFHI^9X
z&amlJr6(qJxJP`1erLRSI{-)B1|t;<zcm#)b4K7_-9DayAGg}OwBiB~yK*X=;l;3D
zF4Mpr>udBI6<SIz6QKB<Q6Ty(U~R-^(r8sR(K{PwPACJA3fLt3`Mr$P2k<><1Dq^_
z7F>Q}*H)JuRU6xfl=b9hb-X1PN3Xp$xmgFjSgvpETbRejH>)}Hdyt?``4o-_HW@m9
zb-&$kQ>hWusAD3A0$}g2VSj^Z?y7vVs@<vb)|W}~I#snwU_VFMr3fb9#w>SXgGj80
zY40yp1i~>>=ueb+4062@ibgE8VCus|SX{Kv)obdBD<d2ghA{hXy*@y@AG&i=9#Pn@
zhFj?yTgRed&(GqZkvmKMcejN7z4ih?=t0kp$fv9)iL$H^g&ody<euvbMf{j>K${!=
zoY+DsN-vyg<*Yw>p6Ft3Y%FFZ>XL?qiF=awcOFj%4yV(T_0hc;@cd<)?eYe;QS&~s
z<~{p6j_-$GYwo$PyrQSVd!_HgEj6`S2(lb`-i;={+QrvezU)snvmA8cc=zsnt}!&m
z(|20>?4==kx2Icv`Q6wTsuNlQQgYUx<6p1BbSsvrsIDG*jJmt6tSogoTQUGY|I>=6
z84N@PePh8<{L<FFUcK}n_~Leoo*?h~?dplne%K;aYg)|tReonnt4l}2(?ym$FJWIL
zK@-LdQ>a<9IGI$G+Hp|U0brWWycnW7ysy8Y_+@@2j=;!g65W#_C!?7V6I&b%VM0CZ
zoM|7P9P-eIWtV<9s3cR319+CXy!NL~9Ibxu53n8aPoqI2Vh*wuvCTNbdJL|94|u;H
z)bMM%1=K?rzVESqdGWHs6CPVH9eY&M2wO8wV7e7baNzv-ZJo4D*27xf&gwe*wLBBB
z(od58L-?a7Wk|J}DSKJnGH5!~?+}H}uxPBnF=Rk5p>tDthy;w1DRc`;WTAI1+}cK5
zOdNTpB(@guWy9TdH)ks7wGDahAk<0;)y@8;!LG2S>3XJnHp!<b%FGXtm^n%f&cut$
zqe8#?Pf*^-IbnmF+YXPL{-L`E@xkB_@okp%{3n^Rf_Z&gO|3D_^G?i!t!JZoZk6S0
z#nyhTz?k>vH1bNj^3M(wk|i8ZKj)0|6*Z36Fe$d~$%~dy$IpyLMI0VN05PZ~A@Q;3
z5kb#8Pke6zy&vv&0*kNu(mF`v1+!a;Iz4NfW)I5E*Pagb_2N3+uANUZs!mOUk9_Qg
zZ?{scsTv1yzbNVz78aJ;6{k^cP(gKB4@}ks5!pMgTD$q-`Z_k1@yYWL@Vpsb^N8zD
z9Ysmud#;4t0eu}EZtj{ZuazI2{m@IGq!S4^qg>FqzV#sdr7ilrTS_tntg-26*wsUA
zvmmWtv#?AW+P`%k|DkZm)Z|y$avZl+IkU;h_56Epe&0onYm>nv4wHHY_rRa8Wt(;i
zq<1PRDpYvFWQ}(lxO3Xr4%-0V_S1ssG0V7=q<_JIAqKN!W?AP|=qQ<EJ6-{(Ht{mp
zp1kyg8r!#;U$TLy)(}a6I?g8hFtR=E)N?ShfcC4_YGC#E2V92fcx#=CdhReoM(cKf
z0G*<S=|n?#IuqSAD}<lue&SXB#IH`23NO-6h2stTd8|Wc+yhJm;K*^`wr}#iRm-JS
z-_Gwl{8Y*c%x@ZVe*M9HBIT^HF!0iSgw{gnd&_J8KLZaTV>IhY-(!LS_MiwUGKod+
zbvALD-ndGMFJkYI0ZhSl>*{1Ugo#B;huk+!UBqj%uwf{ixcek+Z^iH7miG;EcaYyi
z>pB6Z1v(;X#9Xv#(l>U`D{${+$wK}0&hSc<l>6ZQT(rgu8?2Su{HF`SLtb495RkIK
ze83zF`)id97fvZ`YAS(ht+#=OBq902c|*wyx*5IauiR^Pl9}nnTOC$E_R$<7iCl5a
ztCPxwuAr7%E4{(7#J133enz6BL3kh42=jC?^eQ|r0n&abg9fN4w{3b`?<7M=FRnj>
zNw%~A)aPJ6;KmoBQ(S1LyIlM<OiRi&NlqeJ-+@@#Eb7fAcl)5#lkDiT-~qSlx7Et3
zt!_Q>XAZV3Gx&_Ton1X&Ge7B!denMs&q>=cB{0RbSXVj<BdKw6na?SxTFNaoHoM_~
zMEYI@#azo}t~u&FqLmzPNE8=cBNl#3xw~_v*Rku%NKgC&J^}E9L}vstE_or&mWkxI
zU&6keuI$0XX7ToGnwnk40Ek&&|KodJl2YNz6f&z=za09Yb($ssT?P3+ANI?71%x*5
zQsIszsTT$!Tk|H3<qDzmvE0d6O%#;P2RRM{r>3yu6iT}n&2FHgz%@Q$(P%d=XZLP<
zXBVDMABED<EBAjm!(qKAYz}EnX`Y!$^$ZXd%gXNcAfRhE2M@+BG-9!Rh3gE6coMS%
zK#11WI~ryzg()#z%mbApC#QSi!Yv6Qci$=Ld_V=>C`7i@8@$^*JC|119~OLUj-TpO
zO2YVuC^>$E?~`eY#99}r##7}2Kt&TpMrV_^k%Hv8m|<}GUsE*C`NRQ!Fpz*WtS|J)
z<-wQw(e>myKC#!GZ*+dRUdIsZ>@+3J+Ty4y2OrH=@EBmv3-ZuWuHT=_9#i3PMhi!9
zl{ye>|0LN_Z^sQ0Ypj(~-2mK(6`E5#n38@H-+7Lf+i9CN0nqx)9g!swIhPD^^pE#o
zt&bFaEN((45%$#;r;G`KzvB$+G7K!mdH$tA*D~}bl`)x?QZ-+(M(_b1U-`V?NtSZ&
z=+*6P8h}*;N!Q?ujL#u}>|~lNxM=}fhLX$>z<G)XQLUacG*kY*L)^KR5TxXp+yDZT
zk)!a0bD9FM+R7wEEpfThrp~`7`H$M^Hv)i0l!4TydnNNwRr<4@u$>tx?!jmV&Jh+x
zJRW{Jt+fY!b<O|A2M?rNgKDtbKNNe8_#ynBpYt}?IPs(^JD|)a4Qo({HGlUi{%Xv2
zKzcV_!7zLu<gdGX78Nu+?Hjxsae0Z!Ok(?1AbonQ2J8e~HJP4#%%s-ww-G=r>d^J?
zsU9%DbU37Jm*YbDk|xnHd?k#<S8(3O)P8pbnrGlUwyHw(M;u^-f2eNon!&x70O~0}
z-mdj_!|I-SCQb{oc+e#W%;mdjZQrIbvdUc3Td=1peC$h2ZCRHK?ihsE+VPEspL;;b
z(Q?M%>L_*-Lp={#X3;AffCC!4CN<bobKD|3>L{H)+_|kNb<yT)s`L>E+7Mr0SWprH
z2K<)~=O8Rea>W-hc{kS$dJ%=vlI6OJTtYh)?UQ`(7z})_=JqBs^vS|ZxSL8`6;zE2
zcoTS3Mtg2M%rYDdZ07Vrn8*_blx3_7+-{W1L!=*}0kcE_RfCWOxyoSU8f4BBp1}DE
z#{!mmw~@3BGZcjH<%BWFmhIr7Ly7B-+T<p^WOM(=<OMo(r;!G4^xJ3j%n0kIN?Zfv
zEH@ixGHbs>7NdJ4B9zTznF-f?q3P1nq8u{l0O|ES29WaLCf#*T?9~wd)w)b2G%!*E
zc%t(X8^(t>#6d|VF?bc9Uj>CwG(6@=@_WRxy=)j`F1TXUg6+Kg3_czGx<<K+oKAV*
z%8c0DkyRWy%f5P_4>C=qagW33#o)I@!+V*XPxX!8wQ!cz5$@EO<wiBH&U~xFL(tSz
zYJsq%KA!8sQ^2(=!bV0$D6lFGqrAZfz{I}i6$(enTMo^t%V9M(XK)*-lR;Zm1YTUD
zs`p=A&+(Tp!XPW$3IlaWJBv?1^3JL7WO_b$*Bz@J;BNdc45~~WNVVjd_*1YTi~w3$
zDu<OkeR701r}aw+p@4%KR$7SdoEv1dOA3@%^C=u7j3h0r;%WyY5@Xkfgp?c|{w7IU
z2Z#`<Cn)~brJB&`ur6P;mYT~ie~AqrcsX6}XNj{0<kzz;w?_A;w?if=2kI|&^1s<B
zVYXZ^6!+afQXEfs81O|vmztGq)k8kT7X_9su+M(T*uuZK7o~j}8y(fCGh7kWkb*BG
zPc!dxzeKkdOp(fA&W>)J8R;C5^jjutR`R8flxLvg{S%;sBJn-dA@m8sJ0;5AGI4g^
ztfI;nZua)@##bY=O!wd!C$5SX85DRJwOcXLXu*54e^OFo!=GTVEErcwW>}G>X<bmO
zAx=3E;I8$%3H+(7Q9;Jr#RPig9<4oireX|QyiL;6svMT*vOTyqJ<LYyWpiG_0LB6p
zb3Y}u;q9}|-$U;|W)V++3h%TjZ^G!P%P36y&Yez$yYo5ps<+GC&w7n2?`tPo+^1OY
zZo+)qFMKqr^_<`6>ip@No+w+0vO}v?lGBY{tj{=k$xPxLREvg6V`&Gev1(%pCwRSG
z-<YE$0Hga+A2fP($^GfxiJn)>M^od{V61DlB=>IltT&l?g~a1z+0Q$^!dQat7f9p|
zU`T1&?#dBYp&FUM{&N{s<mws<<q#|o=ru@2{Jk3IFA18rIh5s`uRshE4&zJ!(Iuy0
zB<n1hFD+LoqLJ^iETIjN9A8VA?vBEm@dT-19ig;Pvh^=chqb1?^Y&$OI1z?Z0%{v&
zW=m70k?hMF;no^q2UpuYry=z4F{=E_NN^81i>xXLu~;a)QVN@kx>d>oHWYXgSv7@i
zu{0BF&~F=Gqz6LSN^KcCu8}q$`Q+}Uw=b$#eHu=&q1i$TEZ9w-8T<6nsT{ymz$mvo
z?>Ov#mAe)AQ`M$QzJ|Q=`}tm`CvoaA*p#_%^og1oCTeBUe1u<Y-nRVw$^gZG!d3x@
z5U$>~#r9bs#|X{L8#!7_{~y}k0xHVp4IEu8umGi$6cB^%E=i?JX;!+V8-ztsQb4*(
zqy?lKRHVD5yJLZ+ciFvze&O%>-T%4g+;d0I%kIAO&dfXW%+t>^17bAvzQC}==c^GL
zs(n8)Q<0gA#gWB)fb_bqdJ$>#$<s^HGLg+6ZI1w6C^NPTJpH`KajLm2`4YID;9urj
zdc8oU6;`iIxbN@I-ILVd%Hb3wR5-;%@kQhwf+T886o{-(p9q~@ZD1`!dIxV}?mYO4
zoHS(^mJS>t(C&I;C6B+><pn>GZ!zr>xs2L){xvj2CQEOG%4^gNNzj4Z)(9_<ccerA
zT-l0{Lk-NO^7gf0NNToMKi?6OIIE+sl91WLbcSHL0bc_!6<>~7AQxxuC(h|LV>P|9
z%rh?@&L_zphuDvd_RY*EWdjp;hS18gzkcM3Recf5#OU*Y88bpQfm-eif_Uu{-~N$Y
zE=S&x^QrGpZ&VX&wC${z(1ohqR4hsrScC&-;%l4ZjZMtal7;WhZtE05$EE(G&5Kso
zn;vu5dDdghke?T^`gy!(Rz~5=WSXvxOO57uCzrF1=v!=z&Ij{^%lh)XXwQ#M9PH^k
zfG!=<Kxk<0_uF%_XWBeCvo5k&=B+ddQ9T)_@1QM+={j5ZD3wdL*U<5KFzQ&Tv5E~!
z85r-)fYu3cP#dG>Zl__W=fPI=)F-Km46wCfsI=*IQ6jlBDPWA-Nk6|{kwS{5+1J@7
z0Wmdo!VVS4d!UW~MZz@cD#`mc^QLaX2UjbOf6zl|Jx}vA&-@tX3|+42y2zEWHcR{6
z^I;O46C9KhtYFCnA?+(t{4M2)De}S=g-WONW|bHHk4Ub(9UQZ*b@ogyfenl_xGw;`
zFW8;;KAOXhjMLq@V@o)L*b1j;ABp}nQ6>Re^45wFt;zF#e>0~%WSm;}2lE!gGVo|>
zu29j}ekiZ_=5u}k{YP`(0s(6m5fAm959u<ED8ulHIMIivhqz_;7f@0VB~oDy-pA88
z*ScZU*}RxVUj!6l&!!oX*!ZxRf<dRfd}2S4lPH%N(nQcN(GL4=TU;c${`sz{V_@mK
zX@#z7v#@JgoOT7+q5(|tH0t)(vySo_E?a9`@BNe$<a8ldW(%p}n8LL)n>VCE_5vN`
zn%r(ybkHUYzCN*<b6S^mvm(aI+Cpzd8DrLyrl_Fl)fHE<R~79`3YPxLe(XIyKUFil
z6J{V5m4&$;qA!Wat|G~*yPD5jCgJgZSYaox)h?udtF?WZCnVZvdg5Db-;>&U^?o^-
z-DL9YjVVN*%#9mtGErneXqr*H4A45PWwAktCZ_P6Og^02j&^hDN7-Ye<C0*@WVcV5
z&Gmb&QJ-20My2nZkik|)5BnHmN;V}Q4Gs<tR}2kT1hlTKw5q2hmE%gi+o!yp!?OPL
zOJLbWpYOE?+8gcaGz9{O>QSfkuSBoRER-4;U5SjDf6;hmlCIv0W=MMjTyTNY%bj;S
zS&V&T07m+K?sn$9Tf+b+O!c1APzisj)`#+qpM|V%9Z2g`61Q^r0L^KrYgl5AX6mY?
zTp3Y6=M81D$3g`vc&|Ag)ny~rgqB)PCWs}II#Vcb++27cX~4?U`%|hA^aav`E0xN<
zPkV<oQMUu_tf8DKb*+jF77<mFKL~U0Kf}5<E~cE*unqwAMVbl0GVhZux%W{U9)8Cz
zg>oArw;-us0M8i?c+OXL=)4W^x!qpWNOxDS**0)W#=L`Kl}yx_m(d{3ZHTV8uf$b7
ztR{kni~M_)jhC3{?&aH9N}|_A-p^0Jn?AUqV7KJ{wscVni3J||EGUSi=I}P$TeAvV
zO>VM*A)GGrM`c*Tyw-tip-CBgSTa<|g75;W*v+I!8>8Ee4=?RjYiZwkUymjCcy*Ob
z%XNAuN^=C!*Dj=|2uswe%>G&gSVZASW^bPzzXx|(^q&D<KbKmI(y&wtyF3_r$>KP~
z(4seyk69@i{zJRh&M}7;pGx)`<z78HX8EXi;o_HM&&=l|6YbWWTJrK%@}5>!p5tb_
zx!Oxy-R<eF+N`4C9u*^F%p*8YQ{69G*=y~0#qGt0QY7=s%SWl-z*{vQWIO>(USAXI
zwJ3eyT%gANMyt1Xp2lFM+P$}YZ6b@^S4{|TiQj4IjT>vNj6G#K$AS8?d7h74AucwJ
z%vV1<oKN{H70^0GWc56l>6ELOl*^FSn7`JU#lG|{QjcwE2q}kO={5U_X`3!Zs$?yh
zS91FT$K8xuwLI_LdU0m9w4{VZp51L&xGIuA@lIDwS?2anP{3?=NUhVlc`;+NzPLJC
ze9^)+0U(bdPu9I>wG(ll%Fo#_C-lSMv90iPsjpEV9h4-zrd5`A7|*>%Kp))chS{MR
z0ua!907I~TLNx4mR!uyd4|BOx$rR+#X@gC#EG&Hu{u%eaWetBzp;~U)L4{EruXp<a
z4_+(Jbm(+_f}I|61Y=&s;UMWfUd-#lb7+i3IN8Xt6p+0aC_6pTPWZ)ceSN`39GDLC
zvV{;W$@Od=qf1mgEO$N6XWQag*f*l2oM}~8$%bu=(uSH=Rnk10;F4c1r~?n7&g+`6
z@?C<_D-ej15#809O2-`s_&M3=TGT?sRT8x0BeeMTa!*gql?!h=%xV|>O9;zkqdR)v
z`y+4V_w@+Nzne)w{5TR{de0R4>Q(xA&k!b)=RhI;(%5mY-E{z{EG2zUE)5S33tVR@
z-80xq+-RXPDPgC$WHB~0SYRG_Z9f|GTD587HCDJAocK21P47)CwRBN=1%|1_wtfT$
zmA0gyt70skxz>ZQN_|}=`33F)nos0e?73)-dDfe3x0iUiDzGf2);8bgN;q})Iz^R*
z`LD4Ks(x6P@-kmC36y?35XcrZq_y1jLo&+Gfa!UG!&7#@wqv^39PKY|`)G$NcZ&v=
zrp?L)adDrf@Xx}|hkXrWo*!#nn~Zm0pl59JS@TI-FQ-)ze#r6Y>bRINdwm&VIyg^i
z$KLeEmE)P%^IvW}B8viCd@FEd)o};h-Q8|j9N6kJKOWxXg0HMg;<^)h4e--wBIq?c
z%ZH_FB6`fpaO-9#EeiK>&nLbrHaFkw$wf(p;UYlx{r+X;N{&VU*{A(a45j^yj|<*m
zQF`?Yvzz+<SjSQlX2`T`93g<Wt``Q?E-JfXoH{&QGS%A*%J<z%kur1T2d%z733b{+
z4wLrFAQs-n6f3gz6Sn0|EOPc=#%Id64VCuowRMgm>TUgxWv7phMTYZn#29fP2}@n0
zp0wIb{Z><bBZ1@x4<;gZgFEw#*zz~0Z1mz4O1n<Grkb!nLB~q5g9h$HUPVRMr@e(v
zAe7m>1h~gk5Ikc+y}-mQ>MTS2WoMBjYcg4@czHhA$_XJeM4;x~vLta+UBLcUFi(H|
zxG`g#@w?%z1G?4o%i;4S(L9-WJ2fwj)hsBQ_k%-9ZJvaz@79XN=o(xw0CqWH0vG@!
zgE}~_U<pq?Ed4zEp8E}Rt$7T;I$yqA$A0ESi+E<}5@QL?i?5I;lQq+pB@O+`hdh_O
z*HmlyA~KeT_4i4K2rVs(eK!l-zhCoyFJa>MK|U$jio9)Ml|$`%^vt`BH)7Q5<hv5X
z&7MUPCb4wa8<iAuuqQZ#$zz}9WZ#gk+Qci@m9@gD_%@tn-#YW&pnH(1x%T8;o+Hp#
zBs{#8k=R){cVBQyDSZ7>F5@uQd#_4c4<-8T8O;Zpbq)4j94DH+6U>y@<m)Bn^3kU4
za!w5__s&_0#Z?G8w9DXAJUu7bp7e0^8IS~A0rW8b)Kyo_phy5SN!}HI7s0s4BRPR!
z9M0;@YSrhp*B^`j^kkRt^%U(KU;3Q<`eO^LT?I_tMI}9^_`42ldkSI1BOii~DU!tE
z9~Wo1`hLdK9`nm68*~@+*wDoD4)<79ezK(bvJU!PsAq29hS4QcP^+MvSU!xwrNw4_
z`sIV&aP<UaPbeK+a2u*_)TaxfQ%3Bm@Ns>zh6*B^LX)*twDwTP3vL~5p++3gm&5ag
z7{t8Sm1H>FiVFUYaeQKMW9qiY3GjO!uci``1g@BL+;8dPOm7^b_KmIDT@w6?&5ZW%
zr4-%q`o={$uwcJSHi{J{ug7(uQ-apxG%aiJf@I?!XW4ulOnr;3Ue#jpGBLhk=OuAI
zAXJ2HAMIV9w<WXq-S)&ui05-4;4!R&m*c&A7cxyoVLpBEq2)o|7?*t5PJ(z1bRt?u
zNlN+>qpX$`*1hq-ee2O-VRj2<n)Gzq^PA&(6X7kN3*t?agR`qWPD4cG_U@J;Hpj!o
zR}!TZ??nSWxAM4nRet9}yqsgR3f#=XPw+(bO4B63S;a0{wf(?ob71v!mmcfrvzeu1
z-|z}>4jBKW0OXKzIo%G0=HoaJ$1|wrCO%??D4I8P85H606u)Wt9KcP_%ehsc1_Sge
zc6SA~)4=`d6hs<*G*5fK&m&f0_U`>+ioeW+Ewf3E><C?0>WwRLu=l1hhk!athyjyt
z3-MJiw`0)2)J6MPlTQamB6NtjoV+PQ*rB55Qp3d5z0rdH*PLq+J(qjJTW21q@U(YW
zUFXf@`|;W^aOgf^WL*1#Yp*I5*3())U<8Vri&ip`DkWS{O5CYbzJVPr<|#b3#&`yk
zsW*YRiu+S(f%$p8&~LQPqUs*HzVSvkXN2QAS5#j9`G=DG`!#~38=5g%eOM#mkaPDV
z?Z2<+tUoL0h&C!>8~4S{D!uD&8Y4Xm!dr$_*89l#>qz&ru|-#z;=^n6VF<F3fYkdi
zhbDWQHB51(jujW`i#yzX{YmI05uKEq3p*9p>#HxPempa^EN!Y{c9(4?+jWr$>Bgm`
z(V&!Ay$Ne$*0G52`l6#vZovNL?rO9jCY&#<juLRcV_lUcqyepEcJsKum^|e=n7;AK
zcd+-pi;nEMY+;~6V-h@4aR&anyJ7nw*0nR-DpO|KQ8)1@+RLedve*D8rs9~+!@ZpC
z=re!QnSCWe@_13;6Uj!d?f~I!QmScUvb9N}K}&{Oac2Q&RW3RzQ;~pn-k`3gf_IO)
ze=m7|?0JsW=2W;jVa&L3W+GhyR`iNn^NwHG<GaK`e06I(%c>rQcMEro3l0~qbXgf|
zjeN8;9o$_#eV9LI{S}gvGhH(;W6+tMq7IaP@`(XKfcxE=tT$hAcNg__x%(Z21!Aka
z>9uzmKXeFjdAXtRcmYVt%Qt;xdX>ycY3xNP@!pCJAJ0<D6X^0>YmBedkH#~PztI3!
zacdQbqUECbK3hg;Bx2XeVjCeY^Y%7n7-h4*u&|x;(Muya4&@pbbAEygOXH)8PY#N_
z5mgWA!q^)xC4Cm`r?SThlrcU1(f%Xiat#%+2LoUw9s<339WRD!in<#{M6ul9eTRSB
zx12e?R%~U`tS8f~L@3)py{CH6t_|BIx-g&Tn--@gP^}kEmwn9<0gCq=b@rFEF}?VZ
zMf*xg%azZyBLe&*<>ln%L~Cm=9t#=lExrc;_0j2ZF|p9pPd%(s-&l9ZV57`~43PtO
ze$pB(3S>kYC4_Edd#v;8%KmpU1aiI)kM1N~D0^<fZ{4q<Vj^X<=)+4ZD=!O+J(R}W
z)DPRk_2OA3KXjtGd^sHr`*N9%m^g%Rys^Wpa-$+8QN_i;RBs)&z3Ch8XZ5xq<B}fZ
z>1<$erfwo!(NRO~q<Q`T7w}tfPRr|)Z8k<RHF&kv`VpYJl~=xGIn%ArVlrJ6;!V$#
zEJF1~u+6?MKZ+BM-XBT3Zy6_a%E8OhnZ772$*}ok(UXxJ6;ED*GIRPM-@ndwW~w<I
zvM^RS3#lX=``9UY#Ss0CZWWJNvm|`h&sOcs#<J@_uNp$TtrGpR$4a?uPj7rbS8z;*
zv9#8_C1i2l#X!A&mp)-QWyjJxz`tP~CWH*M#L#xQeB2kVyRo=Sgb>r<E9i_rse^O5
z6_5llPR~}c$$YHTW>vBnuRYw#T(s`N5;cy$p~*^HeCMP2uwc+H#Kn5`2XPx_ZVli#
zwG&?p*+t{Zs{xN64|FnKb876}kAwk7VflDMT#@Y=@g53ZwspTCQ<f9x{kM%dmr2u}
z%(ktiq<Dq2xS7Ro?#R^E){cebE)T=L$?3w11BYVGZj-NVZp2P11=vJi?iO%)!fPPs
z8dNStVilY#Lis5)PkONAqInbT^VjEP&C7wo1FP|v#po>p+I>2MiUAg$;CU3u>kK!U
z_R!svNwdQfo-W3eseK#gta}Fj#NGV#wW(HI6d?0xYD>*X-Druq`W(98?0#X;7z2N-
zWfoTC6ItGU{~PcXvyBaziA{@*4YSV;ecJXGW`-%76CWvoGs_Fkgz$p0q@?Q<6z2J(
z{ahRSXJ&(t?&bm`5b=^RLL0C81KYdbJ;yq5e`|lfZ)^C}^rlNsE!z38dN%TIP>w8=
z-cmbAfbvNd%(&9ybe!rSMQ%dGZE3%vW<*nrUqT{NaQ|6B(j_G$E+>(Mq*@+So8=}<
z@VII7`|0?y;^|I4RP!8*(LqlNgQebTnFx&Y#gt~R6d<`t++FggQqEN#FvA|vxvei_
z$)`E1z_+S2N6phg%?1gg9f}&#mcQgl6C8%(e%c(Jetvls|3;rlm>-VZ9hSt$m%q2N
zd5TMjdAw-Q9P+I6918fZo4)yWetbP-6ei8T<>V#<uw(z_nA|kCtI33XRvz1J7vG7z
z$r~%LgXiDo+qIrnnmeZz*1FrQK4xJl@!jEq=9xyw&{FDB$eyPKGXuC~N;Wbm06l(E
zwy)WK?q-0*Qn#n`6N9+0MRv}+H0+M<AM=MX*#aG0Ur_Hv$l(|dS@7sSE*-SWV_9<W
z`B%QpXeT=RMTQ$CFGCngGQHqMrc%<v0{w@piyFmiJ7_^wZGzZ!Ga`qpmt=znWP>H5
zdFqQZ%QH<r3MH&>VTooaaTH*-QcgOhforOaL5QXM@oTv+2eG9##<IQmY}XoN32}q`
z(?T(n+2QOGMrZb^Cs|8_13>2htWxmF`nYGCcN;5N3|u*_%5U~_j=Q%yT+eoE625Ah
zh?f$kMYhCU<}A**ex)^TbMw<w!rfZH5>7kl>3V4u-J{A7u@oTnpmMvX)_$!&YRRXO
z*2m|3(Du%CNUSYj1Yg*lIW3@nUU5>-+q@acYJR%n{!+JeysDH~xN=}8U}{jNR9x9F
zG2~n6e)A=!c=sw5-3Erq^$*=c{$_qtX7(PqqR%O;FY`G62)zY>v^%5BHpSt+q`C#2
z7R*vhL0UF%xG#(}C~bWw<Wfwg1n{cil^d`yzwR=&%jGD<T;NIE_buNOAb<yHz9Y#e
zBZHNJMK;}qLCG%6LW?;eL?Ma*W#$+yrDcmNNb8S{)#`(^p@EB7rvTa;CtLiIuH<#s
zmjBq~gBTxf7gJ$S8+~}4R5Au6B0pd|=VQJB=k|G`67rNP&*Djyw*KX6vcO%B-ebjX
z9hs?`ALzq+KMoM%Bd>+6#5ZN0JQPc+p0++hT74)GGi<V-@BGr>PT9?Tn0TK(e>{(0
z&0$M^=*r0=8InxL%lT#>@Q4l?x(zY5HSo}FR}@8FQrAr8Dj)YM&OzGdO~;pgX3C#G
z3euWn8@G9;y)NviD}4AF55Tn8)i{E4L7V+k8HkE;sXO(kqCREy_@!6Hx6NU8hwuGc
zcPUzQSYnKc?JFg2&%_iabv4Ylga)+&V<>y{9t&}hB+Zn1^!bfJh}l8nT6*z)l$?rF
z4KOye`hE)GTRm0gYH8V;2%RRl5%xZCob?^jBo1Ih`g4t>6=@&X<Z{0Nchcz6<Los$
zod(-yOa*9mytjiXE4R@yu$ixZowrw)boO4qd6xZDf7to-o2Fg-mbF3@k9xdTJ86`P
z{O+`2&@x`b3qqJ`eV!|so6Q}>CS$TqTl+pTdVyp;#S!-e+I!&TVi?_K6?|;oBu5)B
z=qcsSh7gVPTCm%&xVX81N>Z(AC!=U0VOa1!k89t|hRxTCh)ws3jugYNYYHY}zDWQZ
zR;1jk9#~WprfP?0aqHZbTvJ`O9-Vf&XW|ksFWC}i=kG1muEpK75}wJsOWzKIpENCX
zNw-g4C4`!a`xVmqihpY6rNSboq3Jj4nX$EDoLIYg5T+~iq3ob$pY&@53`2@0&*i2n
z2-Q2>K|6&E;onHXBEM+CP{)S{f_Ts$zN$Y_3&Bk72YGg3T&g@?y(Zl}Mk*>8GcFS4
zDSJag5T|vU0S{7)6B6P{mc|sNFW9O@<O^Yt%}g(86Ck*pj!2}`66w#XWy2cRtsEL!
zdL$2)1kD_HUbx>MFSx^7J*ADnp!P8IE-Xq$QscRXZPSiF9j=AIS#q}I(^|`i*a@8G
zSmd}}Zj7k8<WRA=j<jNwO#@P-eG+&b*%Nl_d=y>E*C{bVt`6StPl49nb<(+TZOCe2
z?>+EpJb`*g{B*z`db=>uLL8!3^~qXdN$}YRFD%h|_IsqfdfIvqn=hza!aB)JM5U^K
znayjdLjA<j+6GL4j?%v%LG}BPiUoU>q0NvF&~_Vwklx(yXW_UYu9S|hhcCboh+$!`
zDM9c?e92-H`@E3u3AwAtZG0l86d_`|netaJ*x2Y49$#i3M^1_hGju63oYC`kspdw1
zX+TC~$fdxJ07|UbRDhDifK*EV6p?0@`S>c+%a!2FFQr3{tn>*}8};i#%~j7qm=!ay
zT9#i3=G3ZeXcEF}tf(IzE=!tX3h*bkw&K(Ac4z<S3VGlcgDb@$p=G9;%zrdC_ocL~
zlD4`Mv$-|zQg?-@pa&9|pLv|3E;5F(p}GcjddTWA+T%4*`@;Q+K-AV)ejocGMD(H6
z+VuN}{bC72(>l;%Tw|5#ibr~zqcFZkxa6xf4)1W+v>C2~365AcpGLMS5T%>?>0VSp
zj@!6%{R@wky)c3pnGq0YEroqDc+~iKfV7$F13pk#CBh=8C|oOC1mAmXkr&Wl-#G#^
zcar!>H$G?gHj)77kU=aI@3km7wuj`;D!LLwz@h+l{GMXU8GplZfPtfPYyh1AOb?#P
zsqM~rJoF53fWz-EeKKqNU?FNw`{Ukq$^2lD5Mt?#o8={Gz^;Qev%-BjgsphNDzvOj
z<Ks7?yRBpoX8Nb$&(?NO>hUShfJ5cpnU6Vx;fni7*jZ*VoC_d&j%E%;awQl6&E}QI
zPA9{C%f)nhTR!3Dlv3~Apl_v4Gr`yWKInH<ky^==em+|o5OJxE)D5{cYis$apC_`B
zDgQKIza_JIygjmBKPPLbin6cN))K}5^vtYGTmII(rH)I3rT0r8a;P&+{%Uk+8^l}o
zWi83M-8uE=aBA;-+V}0?wrG(&v`=Z|G_>dR<?4o8*g`sj<}A-ZRRyM~0Cgr>ZOVJm
zUAh;0Qhz-GNo9|g;yz64!jda?&2n?V_-ZmkX@GO(>mQ~NNmHXhnXaaGAs?1a2Z4z<
z&<OyA;B`}NACl!E{e)zvrmgF6eynabP)-ahtG6paGC27AlW*N1c00H&^@bXct4EM*
zHMmJa!k|vOTzB-$y)yp0FxQzsWLW)F+Rm)sJLii^Y+zNB$4Q<N<PY3eO9+j5kz1<`
z`=*)|w?h!|Z5FxXF`T8QUp81b&LbW`h5hupumJo=X*KdRpDZLO^7)nO<LbL?`;#wa
z>bCcGhmtJuONF@W&QEzXN#H!WynE>iZ>ZzDbCg!7(mCbkoz{6}BCujg>;P|G08A?3
zwdDq-RcCIhPP?{DpT6+2sIe`wi`f@Nd#zrqb~v@w4{x+nS+^C@B{VW!%*7E6$al>#
zl(d#*-tsqza}M*+ez;jL*?vh$-el;LS$5m%4^~e}ZPjG#s4UX>YBE7tMwIfQ2<Ab9
z6htj`D8c?z>e+M!^>j0y&5uO!8ovaD6^Uro+p*AwNcm$BG&q?ei0-d{iJYHHf9esa
z#sSoLY>}Ma;!9TRqYg%1?fBmhQ=rw`9kW&lnK8W7Y^gEmgiz$}<BzLR?#)y*EL{7+
z9UcS%&t8}vMpMVU?r`~BftAgg5x1YD2>WTqBMAez>ahiluT}FAP`T~KO~H%k`-uc3
zM2BG3P&EnbX6CVHubl?uN_BivH;cka1!t{jUSvY7Q)f!*rn_27uH2-5&ai6SdFSKh
zWOlHTy=MT}-vjWUhQo^iX!f7t?fuXS;@4Vh)kp;^+B-gEGQ?N3Noi0M-Wq+@az{cf
z04O3Q4BC*_L?j*!mQ##Vk2~WvO$13q!(B7-qwB7V(mm|@1Q$dS0Lz2~7`s|wK{A=0
zwqT0m`HH8?dP^NePMyPz0TE}*J#5qkB4x-4?%$>ROqNX(369-1Zz{U>5X>J0hZpQk
zVCg3b1ZIGLMt622QU%-A|FLe%c8dX(&k+i>3DoViCd-QI-6=X5gpq&y0*Kg9ir-Kk
zUFr;e%y^IHW&op&k?5@}ndMgSBCd4W6pB^ko;{;Uh{QVy8iy1s9fX|K^%_#|D@3Jj
z{iEHd%+ISIxnFOv$>}KRetl=;3;pLp7b&B!Kn>nP;WYQ=<700g{M|?G*T8^YhXUl`
z4kAp#fRg9;k$zPkgM`S&FbDXzqF^h>uKEbXxS}=)-aI)S=xJyWatff6jW(L@j1ce>
zILD@VgFx*kYi~8mq&n`*I4;ThXy}*KFQ$eJ!b+Xvcz&IBY4UQo^_R|rJG6H%5@7wB
zsRw5<z-KcBD9rEu<M^K0o6Eal=u^fRKF<@!B{643lLj16XQoMxrxK(?I^kWtghS{o
zsm7*a0)EC<OZ+E`>aUhkWg^+V0lYx6UjVrgHXU9&yfJ;M3cW3G9zuZsi0oIb->(nC
z&jR!RpNf*g&m>0wdHcB-I7<3I?|>f(NPlLT`On+2!20x`H^7fy-bg{taQ%Old!@FQ
z$Kx@0R4~*)I45TV;hA#uS9$&#h`xV+`Sthdhe#}P#L`e#g8H|Ee{OjjfSvTux1S%L
z8mLtLnR@c?I*|nbf4m9jpZW3r?vOR@ZO0^5f~=3veXqkB7GVs^A#5N^F;t}5sO{zr
zx7{|btFUT&<H=gL#SNgW`>V`QEjRrA`PbiU2-am_;WFPI>8hFHktQF@PV>)FOo=|M
zuptSfV}Xwz<BhM~@;`&GpT+z6A2SnW1ZL(=%q8s@-=!Ydl+#|C&j8#8=H>OEwA5)^
z<%S0!Q-2jBu&w_7{Ohk;V(}&0>>|IO1~_a%KX(8wA|g`i$l`+!)H=TkKF%NT%O1tz
zJrAT1w|7Ev{B=o>{Qs^Cc_0FEcKq}0?1NGfFb1;ys0QwyC~T^GiBrNUY`@i~`lEhZ
z`;9Ml=J0+oofC$k)lGyp_Enwx|1>#s{*M{7&A^+v^@3@ktl`rXEV;w`pZ-$^{}&Ad
z3ev>j1ZjD`j)0RC9t?~p+<yJ<#NPa4VxQj8=^tj^oQ!fi0%r@eh*T|QiRj)zEcZ~s
zRytT32&;~2rYVk5rvp>!cmj2GZt!+Mi>~x){xjG+=l^&+A{o*zY)9*xl#SRiwY8?p
zA}{hBNCkLprUb*PtE;X0sqM{e;e{?%u<88F0ci!Dv1S~Jf7&Db6MS$ZWPS|Kz;#$K
zWkx+MP+S+@P~p-9^@@DMqmIx+Mmsn3-T@SEY537Ujoi8V$9l>{vU)c<p=LX%V0;@k
z8~XjV)mDvAFH|ui0*(o}@!N87|M5ahG9+K@R_B&OLK{$a($PM61WQ9smDe7jP;(iH
z7QjR2uY&q>dee=t$YT`9`8Kz=JE&X+7nSom@V|qXW<3Yua{ek!-+{-{|CC(mwsX9R
zb*O=tzEpw!gQtVXG|&I3Ai8Szzw`p)Eu&Tp31Mllt%Upte~UmAxc*dACZp#%Y8n3B
z854gzx29YS!If5Ij*W4lKoARRdME$W3*`6@JVx|TXA8K@K_F3UhfGccpUwZR{O_mS
zo{z=l=;J0?w~@0cJ5!eOiogT^0a$OwKSmLFHu~eMF&{--Pg(<BO5f4rw_9*be)_Xf
zrP(t#uZZri_J?QT#Ia_7!2A9KO-M`aMMZu=cXwnVB7u?<z@EZ9UWyJYOQFR-9YcWI
z=UEH?mQ?pQvcvtTx3Uub*jMW@FV_`=L#ij%$X|G(J+oS%X!8)Fvj?WU`s1drkzUYd
zBqjh>1C_0i)ZOm}YzJq_%buV6wdCHIlsaT+*c$)#MOff?y7zhNWC@m|vas2VuW_z-
z8BLqNZ<;Q$O_4AygmpcRHTQZaU1ZF9TSULKC_NbVNpkMfTrH85CAQg{bn53Excx?r
z*B&Q3DJ`&3CbF@sB)*|aZvH0JOl!)TcD++J$`VsKIaHCGclNK^Yj2s>NPUw3FwtRM
zeOmi45&jof#@}4}#4=}Dmcb(kq_|=%L-(%ct_ygYdFCE%H})b&FFsZOi@rVmeMR`2
z7cSaHQogz+hAF;S@r(m!XD@BPg82d43G17wDCk;Fwac#=R08&ef95h(Rn7n+y)t^V
zT^)``xe6`66IJbWI-KvP5JMW^da{?6uRSsyFT<oU4Ety#zfL)2hNAxUaPabPRM64C
z;E9#h>UqV3gK4Py#&h0&7$yvjgoncS4wedTIWIhpi`v~2KK{~wc@O{B^Ys7o{F}Dx
zGdMzIRC+F$Vs!fzb>TJJ$d13THEY0c0N|~_YE?|8Q$0UZ?fbNTYX&#wFW`92&i5N%
zyJ^!8lCfx1&ksh_UC*vKY*Y26`UOO+tG~DAJn{w-g73~<uLEcEuUp5huPw?dc>1fR
zsPXSL4~8Q_&Co}Jl?4hu_#P>O`^7<j4X{_@H_RhSX88;vjgw}&8WN*O@zUNP9p7U_
z$c=t|>@VuZnerQ*>n(`C^#vT4DJ09Y<;v%=Hd8&7>qmWXP$1w+kKD!e!TYrh%x?eQ
zgMfuSMEcWN2heVlvpy!)U+sfuJa%VtR7W>DSl}8Ke-n15*57cH8H;Y$mZ(@ye=u2$
z^+YA@8{3n;Rok_}oG@u(Mh$#1#n2t!U#nAk<98sDFI9m^izzCGV)^~L?Fgi?U!TA=
zHA>#n`45&WHSSNwk2S-ItL7h$FKHDsQC2>x3(2hB99QSAx%ws2#0AH!0E1SCS~-&d
zxnzV}+DMe+v{0nJ5LAPBAN<QNrcx_l`Y=Nq%n((r9+6Y3!+gxT$>F)!>g2H<QxOal
z4gdKKeS|?y+Z@9&HYFZda=L{m5zILB#gX#QZ?Lk>5E0B6`e#kAqf??TkP~^(OsnKi
zUam_m1`;B^t}2-?qNZF_&0zKy2blXi<&a}?I?EdSzs1{`?hhu{`{94&P*#lx310ti
zC6n_{)=|#<PtGs4|Gy^t&%r#GS<jWGlXXet-xm#A6cQpL%By+p{0qiX>3@T<FM}%J
zh4J0#$q?%LY`;cj#{Ao&Kc|ly-uPp@q8$I#C&8D0k@>HtzC;cWZQu@q0E#T}mDN>q
zXAzJ$_sM-qjj;{p-b)IcHRnqo!pSXE+XR+R;1m|FXSel#AKu;APDsdr_6Q6f^Y*F;
z@=HiC!EI(&v(x-(v?Pvn2>ET@@~!92egRRx`s0lKea}(f<M5szWI-a4S4j$x?IEmT
zSpHXq$wKz1+MvUjd2xYg(pSHDFHytakqU(Y01Z}UA*j}<fXOV2LbA1(`mgq%fBPL(
zw2gq5lhf$=CUPxp=H_$hWK7Kf8IQukHv@38MD2QmOZKSlgSqeLn^=s;<LfS_+R2<s
z0+J;2yEu7ovee`+g#D|FpV#=D&GWnpHmA;*9~shKZWg;Wzfcu~VCgq(ELUH^Kjd)I
zI}ATPidpPH&5<+_dwHCmP>oD%u$&-62W#1XwHf(8&`Z64WoAWFY5w^aTIrL}TjGCB
zLD271CiA`der~quJXB~@yRiM<r^Aw~@n>4wvoaR{+a$9(4KjJtzDW+oE6jsh5g`1o
zdi+}vM34P?1YB944zhuY8F;0wl3#QnX970Z>HNRXMW%1{CEL<=$uGk~wS5=A{Hy_q
z<^3HRWwLDm(v@`~e{q2Sed<1y#@GV$uX?=?e_xu1sZ&_w)Qp}pZBy0iT18V%*ih%c
z5#(CS-=DyfE3W%^KPi$m=-1t2ghFtC7gzl+T9aDvdz%U)L;f-`!nWIO+w)LjW^@zF
zNhtY4ogd>hTX-1Tb5@=QB~Z;u+pn)PgqV@P=1SD(cWx*tpBjrC5R7IT^*R31de~?z
zmOt!h8<%AB*XpL0{SOpc{U?I@k8=(l=3A<0uWxQ94gMv)riT4K?hw{EEOJ?w+FHk(
z3BQoCx3B$g94i0!{{+GRND2S{9(4f($y+M?bnry{|BQP6z$|9!|6j1)e_^2iI0WMV
zHCiq$a91mKY{^l@Bh;K4Q2jqs1e~<2Vm25NIJCx_{yKW52>)AD$$bY|Vbu3BA9dZK
z)1j95Av(Yzzch4tm(`;iTkn`-IqI9ae>v6j_jOl&l5{o&sS1JJx2{n~2KL|57LGn0
zsVI<8MrGdu6UMKXLov@_s%XLyxA)a*cl8(i|Gc)JWu+3$d%$7HdT7>n+sEz|Zx-!x
ze}?QVj}4U1qxiX7an%(=;oPnZSP<{b2i1n;Vn;Kbg~gbYXR(LaT`Pk|C5KbQwQ~P9
zS$gM;GL^np;kZ<oD5S|LLyOz@fT-vOON0Hv!VDfInLudX9gyFtQ@She4%p>!?eJS!
z=tifsy%4*W&!91T7S*L)B^>I6w(lK&Ui%0&!n!KA{k^KIb-(q(`Sy$^1|}>p+#JEr
zxKXI`WKlAU^w}+hvemuEKDAYDPA@@|SRhZ&SBiOsei{#W<l)yg!>aL|2kn#HChV(e
zS1B;ef8i%E_*CP|GZ2Wcy#3g@{1>|{<U(P$edZNDBnZ3c_q97j)PK0mF>i2@jf%_Y
zwVV=af;ft@`u#?Ehs(ef)GWk1S_KheXBvyDbFsj0kn4LF7uO-Kzo5bR(829N_l%ho
z{NL$`T{|Pl@{hGs_qi6o^$m63`edOp4Ay;mqPBWPd}pw5;a<!`ozkU%o;iknHo10P
zQvQUtm$th<epFeon*Zu;GsJG*y`Pcb$(~{VcNW#0m0*p;t}x2#?WLum6tBI}BCZQq
zXBoo(Dx7ivKomgWaZ2FYhpenDz%qF|4@8z;)KXPtCKux4xHKQ2$2Tf>FQn8z%hi|}
zugT#{I2%=j8_^D)0wn{&j6Tp_b#U%_n+|1su_|^~jpapE1fDc@!y&)x03??Go&!8-
zjF3E%h^Q+zj<Q<Gww4wWlKDT>A-wv6ExA_1b4{2KeQtykc8NeVcMSogmpLAs{%u@_
z-dh@OuyquKu6K3iiQ|CPNprj+`c(^z$w-7zBe~1S-l>_5yed@OeMM}fbJVV=GN4kU
zR)(<~jx%p;`tIzpmL$Juc?`4X&gpxZMKSSyaS~x(w->pd+<%VueqZS2b|FO7i?1pw
z;C$&_nF8UVb2Db#zY`9N99f_)>@Cm|u}U*69RUNZwRN1B`=O~}=O}r6E9Ao}8EvR%
ztUT<N7<~xhF`jOv{eckx+*XqTaS~quxA<bMCgSv5C(}k{v=&}BuwmO8iFo0GAIzFN
z_nHtH9VX)3ESz@}+T^~D`YcUyji{j;`@X{KaHlvSsTtp~g^7r}H=Px_AYf~2O?;RS
zMR~Y)-VVYSBh+=8uL!A7)DF_aEM<h{ruTR*j}YWtH@hg{oPQgcbVE}2aH=k=n@Om|
z#G?<Kz&$d}l@?fWc-S-~rHkB3(s=sCNDDbPG7sG|(LEH=+lYHCuWjvS_qnxO(YU?L
z>ttr^Yf6$F$sk47T09~P`oj^L&vcGyv0WQ`>iAehhu>vxZu|ARX;;-20SuMpP*9WE
z&7|wPjEZYGj#0JI6?8FhZA)AiDVq_2y`0+3GgU0zVrXgDIgzz1gVv(!3w4G1PQ2S3
zo-gMjBts&W@~f5uPPmG8UoO_j^wkRFy>N#gL^VnqrM&d&3e0_WG2R9WgZ4o!pXGZL
zc*##FyDB}E8Z0Y6$&8SfAoD8C9oxr(tbxSMklqjGSCg`sGI=VF-loqF2B;A(nrHxM
zmgF!o0Wi%ISm3m{(e}Q}cd!&+Oz&EYZKM=^JwY+`M-^(MK7cE*T&=hwk&h?*ZL+GX
zzdb&Y5_Dt4{g%BcSqqkyJ7P7SZ!2d+4F^21Q<ggKW(lCvOCb{nv8C0bhWvZ+yvsSb
zOIXm91q)=Axec{U_A6LK@$F~Zu{3)=l<@&aJ#R$S-WU(bo0Z(2AzI(`J}U~?f=ZWz
z_ABDO&iE|Zm*z(7t650oT8wFtvj=Q7T5NjYRC%PjZrw7^Lp{ajo|$Z+S9+Nu9+eJq
zPEmG`)qzc@64_J;-#m#>umAkQ-s9rn(A)FpVGZk(4i{w=bp*kJy&~!EJf5q?oRskq
zgfxS*XSchL8x@mX=Ot$69PUpExy((*ICoXgZed&pE*S7Kle;1^*rLwo+hB4h;Zghw
z{3Ra9cK1b1RC$`8oPat%jgNC#iH+ha)bij(4ei=VU#`bmuR?hrlt2KPi*p3H@>uh(
zEeNDnjw!jkm(6}4AxOa~D5izu?Y%n7e1nN6-c_H1)d<1TH0RKP?LO+r4>m6ERazlO
z{f!z<W@52&jeIcCjTm1;UNM}i_aU=%IRyp0c7~PVgRQ#{WK8r24`^@!(cJ5O+unY*
za`(O8aZ(hIz9<Tsa{6eoJjqolpgKZ``e$7Z*pGZ#=HLsj3m09Vpv>s@-y!p-Zo}f^
z;^{AYCnN$z8EX@h`)R`BO*ScG=DLH{56^7!UQ}cEIy>)2$xOGleMBZtan1TBYa$$T
z_JWO6Pd&#|a?%dC6p<#c%S#udgYr?ULarWdd%?<adx|z*b{$5JKmk-yk7>qh$+Wqf
z=rP-g*TB9#{9W69c>;Z15+vMat}j~uj@MvU@a=~F+mP!h+mWU*0P<$b?J6K=RNUdS
zttq~pXv#!aqKy2h$#$fp=<QRN$e*Ku{}>S2iGC(u8(1vU478nc#RWhx$V^Swu16is
z?gVwEh=_F7SnaoK6T7(Z)T?scA}8M{D7PfC_DH6~J2e`Zdp)!~U1hZ%IQ&ves!ZJ`
z%}{=Urr&A9CvX4RgOG<0s|5l1dh@<YFtW>NZk+m!4vr)ab^RHc0hE+V*Hu-YD91eK
z!4+zw0jhh%nY-4ee)CLHjl9L9m7R|naw;_vE7_{8ojE7d%SqS@%h!x#C4(GuWtJx6
zG-YTVw>=7dc&?`gf1fTjY>1e@7Gl?sSJQB!@wSCxV3~QNl{O?7t@XshEPB;}1ow1p
z8K6T7I{B{<Uhncnl)6(Q=5_hr2@lItwmlw))iP;f&n@j1?XvuyS`%^($M>f6-TAG1
zRxp$4I+vk|{&+h>VW|X}Ln}PwR(|fT7gWACq<ms`?s>;c_biOL;_Qbq7If)c&V~;U
zW@9D~s_|@E80)&r>=!a@z3hbu^wukz&Laf8J5JX@7M?iXFe?`q8>)Q28?{ZAc2eF=
zHHETfqBtLKfytF|U|?rB+=2nz6;SF_u8-8sbFoq6$+b>Rb3WM2+0$MLS$0O74G9sL
zGmQ79H!PQzb$+FTur6Mn@s8$VWyQaAiBhCipcA^=>ytoS(GMyAny#Jfin3)bz%?^A
zOy@D!2|wy<OLR4HcTmZB9=QuR6OI@^`;yOREDD(&y)+IjHR)yb?qoSU-E7_|nuN+Q
z(y9ciem2xm?6r`v{OkkC#MYqXzz?;8VG5>G6kYwlh>kb+O>EQ1F{+Rl_3+*3&Y4@<
z+6U0-E7Wo+hu9Qx^<f;=8UZ4v@;r$jG@dcrE2~`*^#d}288npfRtX20lar@W-_q=l
zp!ra*dkNjsvt2swr+v2bOV3`c@7s{JUZXJDJou*cE<4W>e!^=!AICM0XkcYu<r4Td
zwa-;nYN$RNBTeX(u{A|7QV#B}qLPiA^RK(WbF$Y=bNvq>TCJ;06}3+f_k!gMC-mH~
zDeI5(8E9QTysdyC>tCM(7;(mBHapKl{xB0cml!gDJ%WqgFxo}lKOGoXWbv?*3&igI
zHpO>%SQEZVR$}eCH^4K=DWp(eX}d8|)F6y#4hX|mOgu)O2kG@*y2~@^@}Z5qPMY_@
z`5Pq?+VDh?gN1;A(*5B>#ad<69&ZzOTe!zX4{=>}*zV|}U3)Wn$YZ)gV6d&WVolW5
z+N~=;Nl}fss|JyL-=g4Q4|=F(I-(xS%3{bt(u-tlge2?5#)e11wo6*Qr#=+gl~0G@
z`A|@{o}5>eNpk|<9><6D@?WqOCA-CoI`ZYn=lEf>_w8BN*5xm}_grtRo;au9cuSVn
zM-tL4Y4cq**Pf&BjN{<#ne&Ng0dR2DB7~M3fP{S?8DHBIozXq`oJ?_?oc{+SIVny<
zlasmQos-xr)Cp^A?U@8w1mRVbh6ch?K73+0QFZ=R%hW=S)J7tmnI${J!IPXoe$ZfM
zH`IwTvV9YFdi=pD6L9NXd!KdhBBa~T7-Z9|_mP1DFW%;Fq1Oy;to4!rZ(JqFxm6VE
zL+hObg11xa7~O?nD>j2pxpy`V`j()6_#q<2clajwBG7U;4c1lVffrsvtX`<2IfJs-
zSt8;{TsG@?-cZz>+<eQvs!Khg3o=q)(bkF~Stvpm)KS0LJoRkLEcdg8R}#ridviyT
zc=Zu#TR(BU6Td%M6eeQHa<4(}!`l&07e#iRbkmNM`@p3AD=)xka<cJ7HX0IaJl|yF
z;1%Xxe4A5wcEWkpj*6~^t@u7<iA{XTLS`r?3*O*0^LCbQ*|`oeyn9@h$RlvjEqni_
zOJU04@^li1K+*F9E_IgUtz{ej?PfgbbC0?=-4RDrlb=`OmBI|qS1FvcGFpy9&3{BG
zy(8)=xIFu%J1St}UFUHU^@OUSC3d*Hu5RI8@Y9;B91`9Keb6IIdA$eCLda)tKNn&`
zYNy|xh-;`DS|Y>cGuGnkZ=4()(Vv581we4kJ{&JzBErSRtv^K(gi=xZ!*Msf3@}Xj
z>=%k~+pRtPG-_Axi;NC#D150TO)BqY+^_cEi+sAh(PR^@gKoOuX}hbj<;+e&o6z>=
z1^3WURF$CGb~*X=kyfMdyzg^vnvDUJtByL10Qoy@nr1?so)+ZNoc4I#I;44H<S)hj
zF%wYjOqO~osdS;@#wMweuI4(@@s!<W1-~H65X-!oC*8RBeHQFRcjrY?`}p_qupyj%
zl`u926q1Z;lCTR}>A0#gQha?9A>hVs&hC?LhHJR8w-;>CLP}k6*Q<0j0kbNd!W8i}
ztdnkIgXC)J03T52>TyORpmLg_Xb<#*rEe1N5+$ekqf&yx+=SkcY3Jrkiy-1(b@nQk
zFUvFaC{EuhLT`Dx-6oIY5)jjKmXdT{gbC#{=rz`@SC`*k@-k|^jMS*{R7Xc?ygIP8
z-R$vlFCJBY0!7Z5SxBt_Hp-e(dtpW(K%nmvYA-l3S-uc8&STvz>GUyFkSKamy(Q#a
zup>G}TAaaTGCHN6&EUidTC-DLp+^<a<9fJ^T~K9xyhl$~ZI3FPrl{k&=SgdNuuB%5
za@rF1AUG%sz1sdkX6%dy=xW^WsRAc~5Rl|;Q1&u<AVOMq2Y93y8<vl2uZ48xo}aB+
z<l;IUH*#RG$ffD=@o-7DlelEHS%sQ@yW7trwU>`n*1IH~X#@Q+qOR&NYUEI6a+#=P
zt;(wY9mIRMAE=M&){|czF;qPox=PGh!lHD45|o5;>-EWE>iUUiw|<<W`iK&Ud8)KU
zP3a?^hB2_3C&g7(LnYj09#+~q0?|?jU+Ssia~%hOIP9N2RRJQ7j>A`w2PEIMX+f(o
zjjSyNqnd-SqRU~bm{PM1L&r3%B#U5xK<417fT#OlxbOvzK5r+@n{TPXygsz2`Z%NS
zyP$L*bHVWwbjugK05le3M%xvQjY_EdsAG++cKp(#STMMdqFl7WGmXj>X!WaCuiD8h
zEO*i*`j#!|C48dVsBXJptopwfDg(A1{O=Rb?OaR?Wv+dAs6Y&9Ap{mA3Sizy)ip-=
zvBC0mMkS{M^^hhXQ0df>k9*CefapX;VMgI38CR;`)w`6e8F84Q%;-s5j^0{XMfb%K
zkD7{A=2d522^ZdC(zI2iU%Q(AcEneyRvm~~VRcxWD>>>^Qyx1q=q#SVZ$_}>l#|gB
zf(9~$2E^5lz$kt@^F8OiRhr8bwART4!Od={Eh=%KXxW_o_6p`k>#`h#u~QBM2ZirR
z!VYp`q;4{IpEe>%qDew~9&GM5B(+W^JB^<%K5N)gT<&t4?X0371EAD4yE}X0hS~37
z#@LnsG=PUxU!=GX5THL(Lzd;?fsVW8QU%g72qfwr9c!`3z5_7M(`AaVM>XT4F^5cQ
z;QfXxn8!+(c4~UN$L2Vm(P`(+U%Y#`j~pU}nDGGxoJKWm!lbg~@qphji^ec?;nmcX
z69W~xJ%*SOM@Y!fFjFx_*mDhz-fHQYo`!wgPNO|Ngdt0^^rkd86G+E259gv>x&iZ&
zOl!Ba@mynKR=w-oJ?%$hVz;@V9_5;dly#>Ys}Hkhbd~X=Z%l8-^9iW2o)>voP@=Xc
z#TfFV)r1|eEvUWpo~7s(YSK3H_cVQEMK8F$_5fA$f;XTfZ@HxbJ=hUoKSv$EE~={-
zaxkjs>b7k)RSb|!Qw@x&(Fe{+w_J*hyeV?I?LHHUba)46Tc!x3x}=4cHHX7w_UHXe
zaiu+#ZZ1p)H_^6oI-V?;sbB0*GhN!HmB%|VKL$~OWFDWw2w7~xuC-5~jb84>XeVFQ
zW2`2!x&e{C4|pEZW&m27<cd=P!o_BZb5sG}GUBCicuMo<L*EgH)Faj<Lm6H3tEaG-
z?u(RWIX;lW(YGNyxyiwlYrQt*gmm7}is<^@f}o2Krr+<3-3taJ4~{#I>?-mec0QS%
zBq>~AS_L?$0pA+7lNBML{<350JLJ9DaK(Cj%#KG&p{bQ9Z5_;cuVdM_2dZ1JB2U#o
zT`Q0{Y&W-yq*eot4I8I@A4Atg*b0sMrGcntDO^8uV#|!X!Ml^d0q=a6bnT`X4he4U
zaldVKOI9?@rF+9rl}?k2|I6{htn5g<y<|%8xAs2TVgwgg1!lv{qi<r7Ylux{2#nQi
zD(Ub|-FOc#5CFq5HNCmvi()v)54J55UL;qj-y{Vbd{BeRI}zrUv)pLdNeBGe=IFD+
zTr7x%Ixrip5mI>msM(r<Zfa<Gp`I%$D)ux7%ko%H*qzxVJZqkl&xIjLCP?2s`gq=H
z+I}20x${&8yStv^HI~0<9bJ|RNu8W^w$n`~&t(zyf*T}Z$^P6)Q-X;aY~feJI%EA+
z@8>3cC4$cJDue)q3{Jr6(>4R3jRW!}P{_4e1PS}?$mPF23UJjg80Vu(6eJdV?BrxT
z&_JdeFY}i5W<$a$Yn4-DXvl)j$Li{Er{?^3U8}A8<trRY!X<ZzBqetxnn|}*xZkKf
zmJg~S6j6^5>G}LH#SSO{9?hd|2bCxP0jEHVVM@{@eTya1Ugyjivf^(qOM7K*lk49#
zXA>5@Ms36q$GcXP>H$Csn{Bz_r-jYRq_0AO0=CC7I;*+d(!$^+_~v5%k73%YH_yi@
zPHGrpzlshgA?2Nll7<I=8%XD18h-P+Vr*%6TDOnhtwe-AXwAR+VA4Y!N8hSU_gp-E
zywju$bmKBLdE$caNIh@tQRb@G;*!e6!Up<lA#&Y4wZ|4X88Y?m%*wO(*GQMKID4ZG
zr`so!_mPX@Y~O}3t3k8DBy;0%G@|HuWKzwtWW>{`b~|L1s)6~<tXyyq=0E|*_{90e
zj9-m!LyonpBigCV`HGqE`8I*Rvc;=~0pGR&3=L5WY8N;vySY&*s@1uh_uWr{me9P}
zeonD|;ojH!g`^!X5-p;)Lg~|Do)<0E@4`t^mUH#$D(i&;p!p*Od8)t<kAQm=i#)4h
zir-{UA09CjLbjn<ZP(Jz;+bK4cOJmYCur<sk>j_zH_`n<E_yhNELhyGju=Ewy`#zV
z=XVjVXw)&nTD#7^>1gCCnP4?0ik5|Lq$}UEV{WT<2>Z}x;<G?puHa{n#uyhrd@+sP
zUPW%E0+V#ojKHqjdiqA`OFxC=E!JHlO3Iw)XcS!@l}=p6){mP_DeJ$4?f0JqY*x1;
zo|#((9AJmmvu6e+5WA3!lV*#N+bTJVbx6=quL)zSJj;L@Q72yf(K|_YcUsel+d6hn
ztM-E5lcHa-bSL<y3x#75XMA`NZhps*mRrp0t~7VV6RHVf^lrdbq3{9F?Mikm?{bCr
z16;SIf6$1p0swo5{PX*i^{|?-clUB^Wi-A{9<}!vbJwO0P=6GA+!S|b=L1mT24xix
z#lFcIj5xl{11}ijT)}JkK6JVZr$9}u&&a0>dYBE0I$kZUw`=&4C*$`b`$(2HT409W
zmGd~NpBT#)c$BJj9T(Drv^*%1e0{*ctLv51jh}Q{26Pa3ZD~K<$1&4v3&2yhsLc*i
zBmA~<ML<kr`n^JaTg}PZ<UCR6hk}|z?#%n5=OuL~E$ct$@I4~RqEbNh-e#)xR9*eh
zp_jN{lE8usV=`v;GTdB+Z{0>ZbOCu!M{eU`7riQVuNE_AJRmG#1ZFVBi$y*=?_m0t
zmE373;VZ*z=4x3Pzv<RHyAHL9;p4`HX^Q&eruCS?#LMS6TWlhvjnqS9{Oz0wr#kZA
z3*mnQv753>o)YSQ_>_X55gW&^A7v?z45^FMyp(K;lG0V5M!YB!!qSpJ!^P(53>GlV
zr7Umi!87-qC35h;<$9iuI?h%o4MvTPxF}{g30E-LF_x#D$l*i0ed8w1EYFxZH0{NX
zVTw;Xxw=BYpzgj)a&?zs16qz|@V(Qof2k+=b`_t^RxF#bk939L&e80v*BR%}N+{Ko
z5Ufpnr?vD05qx_0`xaZ?QoiT1Pr{hVEo~ug)6QERKskUbQa`)Bvi<!7-I<=CCYlW<
zUGakP<<8ugp4M*f@CR}Jpt`Aa-P(-mC2H#Abd~82%@I$u1JYt^Zz-Vm6TpW1@n5;L
z?nTL5(g#>*uJ6{dXGlt88uV1~-A9{9iw8uKppIR#BoO`sr=J5=3~FYfsJ|}AkXMDa
zq69z{w?OUP1g6I=x5nLf1LzSn(e59vaPan1MgJ%_ud;ea)}=!kn1BWkO>7ngbDPmS
zsS3|8O_qg7p=Ps#F6pA{e5O4pYRw6a@qz>JC)tO;V+fxS`LlAR7||;~{UjA+^nbDS
z)^Sm`+xIYv0!k_hNQ#IcNT;NNf|Almjndr>!x)6LfQYn!fHX+wNOyM*T>~>T3{&ro
zKIc5=^Zvg4F@oqFSM0s^UTg2md2BCU(2>9h6>Uz6s2|Y)Qa$uQ8T6eEXbL`4TKXL6
zvyvuO{mifgA|&t`B^gibPf+CG;L(n}p206sikNUjss{*o%p2BO-+KOVm|-YDMxcNj
zV^f|fCP%Wgxa(wmr*ZN)$fYI35!X^k3z3i_mN8drIzBces-JXss^i1WuOBs6ew39=
zREr`^Y+7HhJE67y?uEmOjo(V22kY?OG80Ud^m`n*e2LKsjBjIT6_c&R*y*X8q`#~4
zF0EVU)_jW&oKao{KU>1&&(DTO{^<WP2(5NKJt;0g@~{-dgk=^-hc3K3smiZgJ`F5&
zklVG0gUL-dNzaA_kXno^MM=+Q7L%hpD9<*bY#M9P`E$$!?>dsk0Hq0xTM*-XFYBtG
zXp{e5LLR2~-7&-uPN;0hFhas(vlpM>I(eMr=g@3srr5g1@F#oloVermopW9D-Fb$V
zIhxukJ)h&Ll<HF_|Aw*h$AH=US}Xh7r>gHzx)dFXt=1O0vfo$EKuP(<7P+^=un`5G
zk8+%*0!bHo(~6}TYr4Q+X)|r42wPt)s-C0unPSZ}LvA2r5O|Rm1lG>;dg-bG&G}DO
z>KBowXezI_@IWkp?OXz$B|fNwo=+tw>vcBv6n&!(F>Tvo*pDqvEEi&Xd8Za}TOIM~
zBhRnt&q_h|B&-Zp4sDg$A-v|JTc4FY`ctG+urd8Pb#<-+EG|3kZJ@TABEhS%lR*PU
z%suRmMvwgj5e|6|Z8I*hRICSdbf$WB{;-*LLjJqr?oZ2SS!Nn`<U3f^e{rE!F}N>q
zNhyzOHnXba%K<BWnI=oOm{<u=KH<Qk&)I|G`}zJ56z_}!I;_u(!@KXK3@g%9J~8ZM
zdHtE2)+4p9EXMfq107N1(T1^3QF8Z0CHGQd<t!<mFTKaSv9Afn*ss^zji9K&YBanb
z3tk>iXMZrLND3{rZ@B%n(c7;EEc}tpeB@(2?ED;29Y&KDu(1nsG^rwaRpiB+-aFrf
zU@Pza!kDGPsbD+-#_ycHWC1Xn%VN;kdf{wUKeV`Nfk5@hbK%lZSpGymrfPdV)`ul=
zB%v%kd(a8wvAD3Iv~E}E?WsD>L|~u_n6p#uN=I@iDf^*72bQMiNA>Y8cY1Rfki=is
zgIB`y5kRwGc~w(_&aEOCbKG|*f}6Tge{!UxkQD24%v_823lXg&p-VM&##7hca5{9V
zXLpjC8Rp5#T3YAx*=&xV^z@>nrDK#j$50P;w9YP<xuIGSQI2<*ip2n{Mp<k$ms;TK
zytm0cC@bm8Q`9<$0M&{!6Hhjp!qs&HsiGWFu2R>FzuK9IJX1~9dh_OVg)j(CNr8ai
zLWgxiI!V_nKR4W){EgU%F8=9Xv;N!J!K>J$_Fj|H?BXz1eYN9BRGxz~zGHYe#AjZr
zPIjT?eQUM?ha9<@)T?x8wmekI;ANsqvgGi$oDlAk(GK08+MQfJ%-ai*E{1(MvML}c
z7A|j@UOkaIU(sY$h<;8+3c%7?Y%^BJ8ddjtK4OsL0s_=PZ9TnD)Dpj@#a-A24(Ost
zUgiR9-*T)=nfN2N+s^@%nVqS#%{_qb3a$47$fgf#*}|M8W^<Q4(jRRvkAIo0>4kdz
z$$S^Zjd^ionD|)>YmnD>Qi~xZJ7*u{wxsTs6+{?Xei(GtT6nCI)WLLh(*K1?PKGZy
zTf1rerI4&7Cd)A8=>wODP#_Yq1^QpIr{Hb-P$j1eekH}nk;u4o%<BE>7v=1kymBNO
z+unz>zDGh?0!+g?Vvde)qTlHzgvrra1w5;~pfT#K<Npv#@L8HC;@;@?HZQeFTBT0R
zeICF46ASgm^Sy;|y&ivJg>u|#d~#>xA;HXyh|qeVvm?fJr3l@Eb^`Bi|MI!Bfky?k
zHCE%WawI{hyT}wY+PIaK$qj)!%CxhE(+uerN46whV~lAuUaL!V5fp#-(CPz})S>`O
z!<F1zwq4s+GGPqlJ{v(5QXA_8)>?5u5Oa=9<7>d^F?F!eFudQq_s0E0b8~Yw$Oku_
za5LeP!7mA?#kpY~wyfc>2*MqNNrMf1Ii8J{gqzXR>!_>iN#!!Uc&ftNk&f9y-1%ge
zYFoC)`ORrCACkFbC^m$v4x^71&HdhtKmhHm$*h7qeJIERslG~8IE*g0ozI7hIe@Ky
z95`Quu}1xIm8%&=p+xnxB~x`gtHpv}L9@dfZmp`M62pz&?K$n$eM}p+&SuzM7|wyc
zok+<zr%>0UgzCFzo(-!c<7y1RN|#C9mixpesvsx!t5r;MwB-etBHhd8pS_I7K;q&L
z4@Pq(CL-QmIE^bEuKv*DTFVQyf#~#<>v4+oF3p~y1j5Sy(Cq}*jq2QO&L%jD$yxoK
z*Y$GK0!jf@h~-HhwQ6emrBwU}TXFntpj3)n27k5xWYfx|H@wZ-eCBz!1COut<IgC=
zNU|%JlDL(0>N|v3=f*#!T>%Odvl`CJ!?S>NI#B({w{r9q`Y^fahFq7C)UT7HU(x}-
z??pubjJ2)HNgt-sA;JQ328)X^8UVZq5`qC3UE}<nAVT~>hxj;<H(yO$5|4&&U?<l`
zxsfD8nVHeRt#DMj+FAitbuS9iQ>rc7so8GJ)SNbccoz*ML@9&)(&*~%4Y}mwjjgRn
zPPMg~pj|g~#>OkR(_(faQVY_QX@68$j$jMq%A6y?*ltjfa*#U#f$->$tIoF=AXEH+
z<c;eJ^uy>m7kwaYGGwFbgPnIIS)Jj*tL_44j}q^_fdtVVX3mBrZO3_(myEvce0!s^
zoWFElKJmC}Rh`2?Y)BZH*G|(~6{=n__&u+WwN{LWB6o>vg=HC0y<VZoullYx-NPDR
z+a-qqgC%bTmGTYcCwSl}?#uQSh<mm3?bdgaRz#Ju-&{B%Aa$A~AOw86`TSHN2YljH
z8(NE=PKUMY0OSJjI0L5HEZaA&Z?elnJa7om`E<y9v2Sd-(CWPV8V^W!Mu@As7rz6Y
z?wa)nAmczAj(y(&$gri%p~1r5a&V5Z_RItv*6@X{3j=>n4yWR`xD7qGc`y$<t)Qd$
zzA#~;;=kbzeyeU_Cw;sv<_RhU_+`C65C^vlX>fH@O(m3=Kk~u*Oek|Gn7%U+FM}W=
zxx=nn8HtPTwHniQ@~Zp7(Irsflr;f(6?#KEhVZhAG4XYT;vwSEiwXVxwj`>S3r<7%
znTPqn4*6TvRzG6*Ch*nMtm4Y6-SnqY)m(w4fNMF(g{z#`Qt93<oP{U~2AJvSJibdn
z5N9z0=uxn)2OTWId*N}pt(t69Tel7q=1_|{A7VVRRw(^?jqlmH9~rsM%`dCUqPlbT
z4H`WW7^<1&FwNtcl$+`nr*l4Rc3*ve+4&bMY3o%~wQ=^wXPfK~zyBPMa>DOI<N5Xg
zW~iDq5OuhrXR=X4L`1Z0$QtREInDI$rEn}NSt*DL(0=rr%ZV~{I;>t_i+xlK$d%-$
z+5nAGJ;3+-Zng}S1OPl4l^l@yeA?f-`}H%hk)_=;#<FT>TE1;Y4$})J0;;MG8~Sd#
zC`}aoer(Oge(YQ@1b{tr){OMA_Dq1VsNL9OT*v2h1KSLUd~$(odUblyw)vYkcU2a>
z+Qk;=kUj5;fb##JiHD%I$9R##nM2T!R5Z<f)4)wha^?D;4MmDH5nscL>s;JEN?ds*
z2h4Icc)rOmd7kpSd`f`0Kq0HaS&+r;A`u|XoOH}#=V;;yb58xsf0@F#-1gzs={Od?
zbL>?*yORUFcWW_qElcU(t}iwZayE8=E~!N%Ezw1h0vy+&82(^{bmfOrpcx^t30>3U
z4TL&hU5>!~Rf+xszckrW$gvYl*-DON_|WTXQc2VDv}Eh{{H&soxS%8p*;NfDf^oRv
zX%<)#{cB6*K_C^@dbjT88e+Oy6i5So74O;x-9E^aI6t``P>JAru@9&O?!?jj*3iaB
zQ|{{5xtn@0p1a_uf_N_}5Py?};2M0YkBb{DTu!HYrODHzh5X@wj<VRg>$SyX4m~3y
znG+P-6T`-r`dDgcXgK5$Ij4KjW8tASzTTJaw;QOeot?C;0NV3$Q<k+;xT~oLXo82H
z=|bENw?BU?04Ye7;YXuo4D+A#yNN1UTyI-$VizKLpvc?%W+yJrP-fpG5{&^QaDs^g
zRGpk{cdqXrlvSI#l$rN;J9Zi-_zTmHem0D4Itu{u*-11ZuU6ZpsR4yTjdS7#Ya+2m
z>FukA*ycpgbGD+@Dt7R`fAJHpuwPpMt_9>K$&vD6zqVqiDW@rc#3f~(_OAjLtV_i6
zpc3taBJmVN3=C!eU{u7-6G*lv9z_}@{9*TK?pJ3|tK=+pld+rBpD7Fz#0Wa|o;}BG
zwj_t0aGvVs@g!W?c08X8%5Fl1(>PjNeUExj<3v-i!fYU~I)KxJ)72A;=o;Ov9*;}0
z*WVlH-hCORx)|2lVu@>JsqP{j`_A^k6UBM^+oJi|!_448v+@XReuQlDRhGWFToJFk
zTJjsp;EmODe#iTuzcQ?Ss|-Z?_k-IHqF=69WsPO1a$-F$i2%YcB{!ViGSHs=n|i=O
z+;IV!YsUjqcK|Y@9Kh?9ZBNz*!Eo=7zE@OSypZT#3?N)OUNhdTfb>Us_)V%i=Dv$M
z8e5ESYkcg-Wnk0G<Fmyuifq8np0cwbUsvT?y(S@c0#c*3*A8i6zGr!97*hh)K&oPs
zgU#H9DGXxs12vL)YR_>vQ*2HeDONv_5{@)FjTsJq5_A~iCgQ(;e!DkUET`a!qCrz;
z%M$UR2~&|&`s}ryl#Bd;g2(l5a_5L?z1PiP<6)*OD22r@wA>~`?WxhdWVV^o622i+
zN;2VhRS~gwSrlwk;Zd@Iq8jMC4b_9X(h~eW*Gf~Go3l!!AgkFzR>2!@D{8MfsMW@R
zRTKQ0d!r5P(|fA~u2!^QHoVn_r(SyXhb@q-^aCAYr{5aYPQR*UcbQP=N`EIT+Ho)K
zNj!Iou;)ubuo*@sTa%SR?xfXTf<ZHHc#UnE8o-(t8VvTq-!+HJa`yy&+;#gAXx|k~
zwVd|DRoOM>&edZ*oDzF(2m3}LAb+yBh@!}0B-1gTy+`}4fNDnjR;0}Aw0%SN{@AaY
z>O4*^iK~t~vxTuQgMmf%y>IGNA(4#Y4zWs;QqF*yffZ+dHtvZ!Q&DM%XmK|TWE<E6
zTZ%_Kt?bjKO<hW-4*5{|LeI>i#i%Vr<NK~mG%LO_RcgB%{K5^r8%e^SzjN14A6PWu
zQa1cHDR@&`H(bF0(Apap{#8%-hMC*wtO20rpe9U&9>P1^Nvxf`M1D*hdm%{&3w<O6
z`tLPz2i7x;(D27IVv>j9H6hj$IZ6{g+g2Y#!MOrk3|T-#g$^||>AT(Z>vY^1)-U3D
z%bMCJgYI^{4K5u9-U_Mb{mLh@fnZ3$45QIrk>J&yeWw&}v$GQ*0lvFUUq|2`ahtr*
z4>dlFI8~tR37Gwz^RCK6oe+*te(ooC@4eaTL=#T(Ce#3LtwUH3XoWOXDHchvy%P}j
z`mA-<PXJ{igrl+*{F8Y^$jm+bI4)uD_TB7arY8~?gu2UXmxTk%$}<36BH(g3WWGQZ
zT%HnjPPII@a=zWLY-UDZXE^{6W^_@BU6yf_LE2ni7VCmsl;{`O5O@xDeb=b7N0qYI
zQqfgxStPd|AD&*Ieak{9@yzpAm5U^mQs2F(^smDM7rebJqm6{lS;dzN7Cx3NmI+}(
zQd`uOQ*H*TS=~Kzpy5K(<z)V}Zc;!PYYP`-xAbH*apA+PgX|{NRkNG?m)ZBTOp=0a
zgQP1oXIxJF-p?#7tswB)x6?GeYA3i8?OBiv6)mJ=ci!@<G|}Lt*1v4aFO?~f0CG6M
zI|4wcgMHwOYy}xB8H;8R)y#eJ>)Sw|2J2V~$+u%~lL=F|&e%u~2nFcBHEHMo;dCtk
zWBj)}f{#e7FNpo;ycDlE&wD-mIv@3EJAV~0e7bh*Jh@P83=SoU5@C(l<$ofg#`03d
zE4?Wv(>>xtUlgE}DpM`#JMZYH4T07Q(pCq)(&!+51=^s5uqUW%jF3kM1azwgkq)Ld
zGBV5704n0YT7Z@#mkIxkgW9BeQB#1P#WnaXBE_JFnfb59q|Ue4#Z)Pg5xhx4UJCvz
zE$_uZ5&{i-!HOFov84fm<0>z7&vlBVIz+?iv;ZmhwOVh#*r+Z!67F?{*+CT(w<oDm
z)w@wjKythOE$mxR(NX;?lBG253Z7e{)k6Db8lNVLlPeYo2*ykSC`995B^Jo`oOu)E
z3Aa*j1S-wWxzwE+E64U)7Y2EMHpy5kikLjx+5Ac#&BCiWi$T3zXz+3K^K+^_nbAv|
zDy>%~I`WuWf1=2bul>@U(yt>d&IW+bWVHl+Gi^b2*2f>Y(}d&UZ(ZGb-0Sb|ur&$$
zHr#*hAjMv|h<&wv`Jk2~fAhF+WetA_#k7aW`u}`WKSDdWfmpx%D46i6g2XpHV^M~i
z)E%*-;+^2^uz>5X<<f4+>DE35Tw#FTOQ}2Ho9afpLqzaLbGPKU{+!X;w$C4euP<0y
zEPSo?(io<{PD7I#BRMX!Qt@FG_FWKkV<Ayhdwu+l7{iB9>jdcv&qS<=k`8h&U-U<<
zVX9r@m<oC>?Fo0o6DF$lmQ}l~L9>V1Q6iceW8e8s>>p6Gz;y1RepI^B-_jlg78c0{
zRBp4T|Giqp=CP9r7v0mj(xXV*S%WWPUnI&GE1h&ll-4Vq);&ctFO;%6o;R(7wQqH(
zSpL-hLJw8hoq#b(&&wwLE|U)QW0+<f+E1fz5cJJ0sJ8#2Fc7MoRh^|CojmPwn6la4
z9OG2%W+y+TrAqnUvc|F_h03;ftSiWL?9GYlp&sTeiL`S;_GA+3v(n})HIxFk0xu8B
zfJaqFtvXM`1Hnd*4=J-=%Q@NOI$g39fMjG4wR$7JD>{4<c3N3rhY)=QnN-MF{~W$)
z=CRu&eiBk@DmY8{KoO4P)xL;}0>LP#0dm??k-fmOYCc9%I9wECtuv1%B6!sTL_$aY
zHWo;%Vp!x8GDQUFvoEguBc#&Brag;Foy7o8+j|~9K_f@f$PLwQeWzf}HaU>W((;7l
z?c&ZXtF|P;{^r*?=SCEO!mE7Ocz|pUt*+~leLdMEytMyeehV#T_hdmyA__Hl^?{G5
z-XAq={pP`ErFNH_*}rbP3MPce)a~WGvM18RBt@YQZb#a@n4k*h07Wu1#&8?0Bc^&t
zCE}_%lf&!iI8Ma8YS)3l*OntB!50s-=hB{g-ve+I@nM<v5=?SjfbNswOaJ<aI-q6Z
z5L^-a@?xJe>F{RGs#S_EHS{?*KJ1Na8=5HF<ilLW+qfl$-1gI<8Ssf;Zd0q9Q?jOH
z<l}B^q^|dS`djOo;|91*epDFZq<UadZGQB759B^6g}3r^lC^yL>c^Y_m-~}r19}rM
z)?t7*IuRAU0mxYgN@y#=hEM9JspVjsjTd47`+Y0=xl}1s+>?%7h9{4S@ef>z5?{~v
z=9es;%E$&J!b`4yr0wJ*s)(|aIS4f(=rqI!um;|1UAxIkX2ON`*2x)cvKfd?UVoJL
zy-3|%y#!SMgxlv{V0_kT2YKz3?11<<QEbzt)U8^on|Ej=<!f^7_fyFcYIoKoi3o;b
z9{@E|M~^VA%LxRetbxv}I15q)a@V-}V5!94rg<bSvy~Vi=b6;7^Dyig3av~jO~y(G
z>Czqo7%?ek9!42g_tE8cog#lAH{#vMPU$eOJzvN3vsE;=rpIKjx!meOiSnz5hv^(J
zAp8m6pk0flRqD<eWsSdXAtY|v9r|Tltta(`N^8yN$2hMz<48LOI=Zg@cKT#ifJ+9_
z9D$$vDTtka<OJDw3o+(#(w~NM%?y?)A8@_Qc8y#&mFbAT(3-6bv$I`lo3>xr&I4rB
z7p@ADz+5*309`g2m;GQM(|go@y0Wnc@ck`aq`N|D)a{_lW^VpmWR#gtxl$Y$W`r`n
zH%`gh1r7v+bW*Qhl!s(Sub&(lY+!}1z2#jA{#7Y=<(lvqP%WYi=81?`VX%KjuZ@ac
zt?6`;DOLG=2zsRm^s?Lrmh=FS*8O%w{@vrg7O&un+Imm@khMO@!_LlYrzb}!kxH(@
z5!R?vLh-6#RPkml-LOc&-Af96XIQwrI8c_m)V6loNxJBb-`Pxf`B7e=I0Y4AWLO@6
zPf`%QDm>bm=`7>*%rl^2WCLKJ^421^Mm|+@GeusnC=bQ(j_Kv=gU!LDmk4$p4=#(h
zj1Z^o>_VY)L(LcUb&udsSubdHwOF9@k28-}B_Z7Byt`h{e>$H{W4UQdI6YK<0<0q=
zh}jO8I$PC|t)Q64v&Hk+$jKRB9-=HgVE`;;rPp)y_Zejv<!WzyQcQ>KUVc+Sf0&-k
z9wo~u+YV4XfZY1_^V&~gE`*W~1lb=(71Isa0rIcP+o|k2OIKI9Q@-d(fz8Y*^hM+D
zME&MWYl`Uy68Pr|{?L4~sIu$7UI5$uK;pAf{A;m5?G(2dAdK#c%5iFY$Xt6rreovB
zM&q5}RvA|&hIJveQ6K869G8ua1Zb2rwH0Mih6I`k2&?TMXt{%aVjRswURn}=(R27f
zpk18L7Z)7762ZWxzfF*DV)JTEHuh@hXA9et_ZD&h3nbx+vwIfep!N{Rh2(VHH+`me
zA^M$#@_c_jSMLbYS(^;mb4RFIs*?Ei-dae*D24XKKeFr}#fBcIUt20ZSUcn}h|r?@
zUI;-OFAt<hS8DIqL&%nzTzXP!w3YjxNE<h<cBX?gpQ8Yh+wuN8wg&WI=Q_wsE3xGw
z!#TIBj?Hyt=h#U-RZpd>1!FnZ0nu(%+4{mp0T)y%aK2VD@q+Ho*;{B}`{l|W{<_i6
z&)0p|FCC=SbcS2y9N9F}5hW7t%_o=5&-VwBWOXVgHc|Cp3y+Kzs$Y4?J_#DCwwm2M
z9}Z?I;sW_w9jIvsI7<;cu}f+rh#D&2U-CHRgO~UR<%ci_Q^_IYAeF;s@)DrauyJM2
zPI*#dPp2tCui2b-v(CE|%<~6ldo}f_>UqJFeIYAi`IzyR#bF)>g^*w5DPYM!O+rS)
zm%v!QZzE~ztUihYH&XM7Rxk?bKX{FBZSH=paMdC>_eG?+^nKC<Jxe9c#A|PA8zT4x
zQCw<GoFc>3QM8|<H%u=W9y%Z95n58S4lRu@s*rE`pH;8uo6EmoFBzBetmmSY?Y_#V
z$fznf_g-8v>&MuO_c5kyYb(>G&nHw>NvA38Hf^7thRFzvzPz6{sJksUL7)>vQQ;)_
zw7xV%*Q>Iz)zIoXw&p*q5oABEo_K%Yw>Ek-J!MbUbJndFJT#-s0A_aU`9ZUOUAXx8
z%nnemzMYFTgK};s6Y1dCUZ+t$@&L+N$AB&}W&ldM1vG>g9pl|%=ybxUSzKj#y=P!L
zb+iG<cEak3E5hot_>;n8rkPtB*~5#wldUw4$3j-0>&QQlR-00s5-FP?xCx}!cFOu@
zwD)}{LAQzcC}0B@QVOvI9vqO#7MnzZo06C>tGdl^$lf;b7Hh(G$fki@N5|R2nf>sQ
zbGvXW|Ay_HVn%!ioR7TIwy(}%mONak)7I!L7nZ_ME4=}f&7AD@Z-qCZQP1OrJ?t0r
z`db-pH+f^hAnrzuC*|{_w-ww3M*QSCQ^U(7kaT(5lqZ!M;zGX9s)Pdldg2$*CYhDi
zhyz<g!kP9u&WLqW^E9BIidO+74WZqpd}~}gGMEkUjd@Kq#Bi<%+C0G>LW1uCIjXjH
z$}HZ$s>^uj0=mY^yw_3>6&+au{98MC(9skWQI7z2rw-B!xT_mB%DT}eY=wS^h9yJ7
zdNKRCSbaP1N<OEb&i7}3E0>T(f%2QTs~_|5?tfqjkaAKcCXRjpvj485BSQMq1c-!e
z)PV4{76Q~1n9Dy6-#6+h8+|BGvhNYfofxZa888Xh?-gAj9>M{+4H@ia`8<dxAyuRM
z>S{0T3oegJ00hPXCg(-aXUszZ$pe`7i|h;4v!;;kw+`J0%jOjaJ0Z_W$&}e$#pK>j
zC~$?zaoyG&m9sU-FR(iU-HOcdh`0g07w_J5*N!dWeyj?N^QY*5Lf#Kof-ajLoF+G&
z*JS4Ndp679TA=tKz<o)z7Lt9!DY12xi=9?TZ;wwUvLC6vd*HJJ2%c{@_0!Yy2!IhS
z+P46<@|N^eC?$XA1G)9Kkid&tygBFrap;QYgD+<<)+tQ<-`Ly*^gaP8_^`(pY6_nG
zy8zyh7zYsRO9xVGr@U1kSYL5vcF!*7i?*7Anc$Pv%3Ff7=?}HM`$thlMI}`t?=*;E
zFXw8vPq{%Z)9DY2HNMrJMaF_+SOZJfzCP<y+}v&~Iq-aESqI?Lxm{Y^+L>f_50I9h
zM6M(5%L#CaVOJ*CfU^gV)%BEU{5R%cNcxxfYvnYU;iLTI%QIb-K8og&ERS>$zX#lx
zO^(WLkBc5XTWAEts1`j4qePaI{b${=PiA}Z#y_Fv+U-p^49K=8`0Uo9(L$pNtVz%U
zjqGyTsf}fSpPwoNwuYY{S1<V5aSav$e*ixiQl`qx?@;Kj4xN8MI63G-8OGz5OGbg;
z)&;;P<ZlZPOAHRJ+JrkehO23+8?Vsj<)%gQv0yCnlM3GHSj;ykkjt8HZ<>@HXOx4O
z?|6zjd!Md^c+v8R;|Cm)X*dvkCO#XO*4BhJ`V01ZpQ!k@$!qvb?V&s!poM!s&(PE$
zdBOA?#Wg&5F#1d^*I~7BG2Z?mfUyoQx7SW9vdXet<mduFC?|Bf-eClyS`%&`pOb@r
z_L9d^j-5T~dHK7S57q!i+L(&(rraknp+EjKz+l-Z7Fd^InP<>!%WxgRNQWJf3gtm*
z?$|sH6+km~k^Z}>!8C3EB6~S4Aus#89uOYwKll;CyUO(_{{Bt%PM)8hd&HIM>4a~3
zlKQaOIY8!a?1xP_t@VzhKB`4T_y&~GWm*4MQB4(%PuWkUSuk@P6Ej+LfXB&N4=A1a
zL#z{%WK&QtGsF$W@h>Qj5Pu~C&YB16*Gqst2SZ`ZuU~!s<g&6JWld`B<%ftoR~Y35
zYR$Zgu)X!Cw{Fdp74W5=w*RUQeim_@x4AWJml~{6=Y$qQBY1AE0>x+xKcc9zucHZH
zMN%gyI4MqQA2M`<=SRJu@yTrIk5S95>hri=({6eoJ?FxY>K7^Sdp$mXC21w;F$B{t
zL~Qa2iNWv8J4BdGgIP9Piw{~pti?ytl~t5~7_<BJrQ-Ul?}gGQ?1ARGFQ|S=zj#bH
zK4AS|^wX^v`rb>Arz$?gOMHlNCo7bA@}lj43^;XDeVu$4HS~GpM@*sMwT6bMSmV}9
zGfV~D_Pqrpx`F0a<f`_3!i{s3JNe6K<=dcwG~%GAaSf*__{P)EqsAM1UeHs9!8vry
z#sGA~7|ej(M`C-N@Eu|lEVwgu8Q+vw5vQ1==n%;H-S!np(D^C;1PZ|upG#g}I)!;-
z6P*G!_Le{xr(k=+ln3b$S37)A79<9I7{nKFzNLOdY~p_$j&?f2gzv;FS|PW(S!R8g
zWSNeSIu2Q|+Z-U~kpVmPhM_~sU%MPu6lem`E^c*u@M?Q8jI>c7yqsqQS!^#Dk3Pbg
zY>syV|37rMh!(cJy;a)@0$b`!g8VV<3m{Ty0k<h)e(i!Z=$E37R#3!>c$<q+5smz<
z_+leI^Fo2wk-7x|4yztvE8<Zrwq<mgM!UZ4&p%XpIc`Z?a`m^*5k2A(;ZA%mOa_N9
zSwRoxMx#?$%G*(jX#vF;eVA)~2s1atbH`)}!=?^7DUX)&c19ZFGrUgIQLpd=FwDet
zr>mYVt<(1@Wzlor<fZXPqtOw*<objKbx&P3(pE#NIU(qDT=#A@)XZ8!-m14%l90tx
z-90URU7;=2OsXN$Iep-;7hgQX0)abnFf8v4o!d9=?uPCXm$~nvPbFt7V7}aBy}g^}
zo_+4O<ON^(&f~MhsAp$GcUeeS1&T8~B?F(C`qe}8pJk#4SF*$?m`p}PTYtG}!J9ZJ
z?Zg}MJV26;_sD+>6mBDx3o6qfAL(sZc!%2V2gO+hx#Q%Hsnh*XdsazM(ef(MO|17O
zzjqmu;}IOWD!X5s0I}Zq_;Jt+kG8YnaKIRjFnnzn!e2c!!*O3Z1cmfL%4B5rW~I5m
z&?bE<>uY*9y0|hc9p`fjJC{4|-hlc8O~TY=K)b3i62n;*SzSdIzdWajJ)B`QWJy_F
zmx2X?%G-b*ozdoF8_7wx2)-G>Wxc@pnBVGy6;21XNaRcWSx^bKKOL2TUreBMQk%}l
z9zb!1vd5JWTw_{4^Ldiflix%>gM*Eysfy9K6}@y+>Lm2MgN0EF_YrnJ0EfV<DQ`*u
z0ZS?iWYRBca+DadidCd+I%Q~N3WDKRVCe0}$uv|4lF{L?mF2vaMcPE~Hxa7M&kz6n
zq@OkovYb6h!Q6B*oREgRcnewI24tq}i4qdpwcxvg!X`AGZIv)f;kNBqz*{J6@#iQi
zxvBHAgy``*rH@^)P+H%8V8No6n_Ky0a6N7;TX<Dfggz6Xhgne1Ks!mF#^~j%H~8+I
zjsN|<PZ_jakI>n*$X6znKCx+-J1TryTLJ`fS0%F^KinE~pL|$ssTJy8E8?U4Cca8<
z$m-XSH6r4rkpxkqY>F|Z^oc#H3|j*ATv>yitvC9eQl!a1a2P+?qhV~P!u<C3l2z~E
zE(9y))PP|qIYk^TP(CO217l<M&f$QB$rp2M#e%47o|KP{Pxm{OJ;U@G*9X9514~=#
zZBsX$MSJ91DULBHMVa$0YzT|*CLI6fOm*>{M_CHPE%k<ZFRPDe$dM%)hrL+vo1#wm
zQCXiI!47{jhsq4H)CE|>;c=@~6$}@p7Vr%CS3`~=LlmsTFMD3tJ5?iLNag;`A5LVd
z`*gPFe%4->AK`lEVG=*h1D`E_KYd2Bn_=h(><!8&&OqYkJI2?TKDOE?y`6qo;`SQm
zdq^qgl-Bi)+Vz01@^Wr>8(#j{p<&2W->?^j9=Vs|99Xgo`+jcUb2SN9@)rJz2X3jg
z5X~Amu!%h>!I^Gi>%DRZETVQ<W>AYiz2@pth4mjDg7_8Jp>q6q3tYI!VKu(gYI+Yl
zhTl)L@tU7Os1Kc0LnOx=ulv9gk|36E!I!LB!qevh4oIazPBjpGQ77~jQ`7Opi83mK
z&%d@hrbtg#<n!o$_uK?+zBG&G5)NCP9}RhH?f~o!h>QE$@7LpX2U^h6&XDM;fHNe1
zT|IK)437(#{3yFmD&%>*9D)y_JYBb&gTQC`FbgpMrBj97{g3L9v*|mMmvrIANDL_p
zeh`1A9D+PTVh0kM_9a2|W&7v_G%Yi>Iim#AtOv<0h+%;ot)t$jM6rNZ-DGi+kPTV2
zaSakAehw|WSNw3BSag0P9aF~PYsQCw;nmT6QnLQ=A{M{3;S9cpiP^4+$XsGqi8RdW
z3F#no9NULQ!~B7r5CeTOU$?(%wxUF=76hba12FA47%5=TH-dp&E_QoG9&~2v#AXOR
z8GEAENj+%acxHs;D?vhUr=h#LKv;Y140UHX<7MBm!@QkOu;ZwkdgWWB{R+*laoINU
zqj}5ry+O;ST;SXZXLV+MPyxvz3G02Kcm+2n-f*fcqphB`5A7lLu_5W$9=&zR=Z8bB
z(bDL>LMW5Z&+^GlrH*mbQ0w4V7MYHxDSK0$I{|&IKAU9+dJj@+yG5d{&CjB!lauvF
z+T_WX8j&VTuCLz2uK6byxLK`-CSO=wV(!N;@4yeQmfA1|XjE~3TzQ)E>{K~qmZPew
zYT@K0{cMgbbTI^kvxh<PcPX-zbNAI6+TTt(fsv6Js_a$uxb3tTQZL9}rZCu7Y1slx
z|HRPAL-RRN$@BCJYkikBPz*m0eeA!rXGG3J_vU9PJ|uQte(I+a^z`up39JzMG>E+R
zNpkG%mb&ShQ#1XK`J($jnlC5fmsN!D9`=8hY5O^P7adnU-PZI@YsHKYysZo@ucX0f
zVBw+4;*O3+a$VI=tehhd_L5@vb_n0tLNSWQ>5pfT$&K%KCS@>@>8P3NT=eS93~F!u
zwUEraalg(iB;?2tzg>WLUk;YVZQ>6GmQD+jMu-RT<I(uHdTO07keIV{pVf~XTi2%U
zM>ifE(Vl|7ZG+4?RNA~zTe#9_0dS)ie+!{_s?7ej)_8P-_;~KnHEID9aK7x0nF?t3
z;|=ED0u?=aRw}T~PvLnOHN%=#O-43w7F)69p<l5k09#eE=Zg#LNH)tef!<pZ;Z(B;
zhR1fD#=+KGtj?DH=*ku$zBwQ-s3?Puw^E#g?##l@EjK7PU|1M$9`74V<ulU};cB#u
zO2THn4!WU7{4FU*!+T62m<fl1&S8sAu;>tVPpErKi8XGq-fIElh0g)w1H4ic6**cn
z>%WO1{lV63^M=gZ%CJqT9eW>q^nyL~SeGQU)j`1+_Ata^gA$KJ`U@A1dfB}v7wpGp
zNcgkkB4&s9T+{sl!;^75T}WDS<27qxh6h78L-Wgzyz5Yc9KfuQAH3I^p>@vp$Be6k
zQQdqI*z0W*C)Y;BBuvh=5qTR0aZUDRmDs%K=!Q#NzmIS_lg10x3ty#m83u9dA$IEE
z-E{S&cAw2HN?C3&q3F0@04^85U5~_;^I<!Y5mSq)ko`u;(uVq1K+i<7#1!3xFOhLa
zrejlMn1W8RXCYx5z^11@y<-1Gea`#5^4c(+p{!@iQ6!Zd2{4)NXGh9|5H}38Ykicq
z^r<Vbd1*PGnb|v}^Mly}saiNpz+6)MJ#N$ie8TBX@2KQ{f*8m=Ke=XeC-7OuhV&{K
z6rC5@>4jtMwDZN_@e%eRL3Z(c<iIdORf?8KL!UYSrZ;>Gbf{ttG#Aveso+jkM<t6n
z>aAR}yOG5?GY^B+TQ9LXi#Fm8?9i05qN_w8SitUKDtnkC0TIKSVUUUh6ELw0<g3K}
zw>HYKCteY@s438}RA131aFaH36`1`gBDAQ8s%plmar`COOM3?o+)%ZPsUJGTQ$~C(
z$mDRJAcZ1#1^?<%=L3q5LLx0<%bcRz+*jU>c3^0Ks{jz^1P9x4+^Ti~U&`(81g3=~
zo=+MzfyG6~78UKbb%Oc^i*jjY6G@kP)umy~JGKb$rM~AU?UYvXKc$cP@N4rpr!WTx
z2k2q9EQc?lC`^)B8b6hVJZ~();h-oOZch?+w(!vkcG`;cA51^%PDjU_Ho%EDCWeUz
zWj)=#XYI8jG2xJV$D^_b?<oh0*xr;&o%D>d0{B|JL0}bRjs!XZ?xPepCPo3nZB9}G
zKFSsg#;!QQ=hXw258dF`r|zbEALmh-6v@b7PUUL#n9~i_OQiNyNB|Tae}a9RE*Hw-
zTGCg<{XL+32K`LiR`fa+t`ys?k`Y_$&6>JR{^6$y4`_hQzTb#E5=Q5lC_x&^n}9Bf
z?Y#oG8^6RzERlW?^y(UG;E-*l90{v%azB+ESG;d&Wb}>22;WUB8FyTe-zl*E7Dxwg
zH62f@BV{o|&0MWc5HuWMi`HmOAf=c+%qW`8@di<fK(b#wG7k16U`W`7+0e-ky(H>}
z4_kjr(bcW>oFY#00u<Q7pe!@H0}%OpHz+d2W=@g+*x_18xqHr&I1v_`-}&|ke8wAk
zg|oUF6_<Y7CgLyCwPI))R=w~$eEsFMl$iPTk@KWQq$qyprXS(gd-(l?sM-x+g2(J<
zp=aG@C&C@!>3#iBjW_$hKxk*r;Qly0{P7r+!e_q~f~UJ0`5C-F*ssY^`aWP!Iywz9
zSk|<Pi&BK)NriU2F;NfDizrZyPk?|bVfe`=M-;m9u+zlHw`vy@aPUn!av#M(yEiEH
zCXsot(YH@p7O;&%j`%hKzogCa+_cx)qopneUQVYnH+DgcDDHUkCEqOtry6<x+&drG
zABK<Bd1T23m|4H*d2gPfQ6JZJvYOx4Zg(ulUjFU!dun~4*vy1jnVjeGl}+!>vS^uu
z7MNe-;o<@uf=rxLv_I>zl9k@Uz4@J3zEzD*TOnViy|Lj!Hku&N^A@2169T*z{@SkR
zMj~JtSp(Z9bWQExz;^M@4C_pKFl6r1NXXzYil@A~a%YL)w?1IXe1XyZ+imGv&)Kuc
z)pqPpI<JkVb4|0?6NkCA|5^d3xu|+ipaGB<1cEeu+ZWL27g7V9up8#-b$ZIH?)ZkQ
z$hWqnjkBkJBp~<w_7YCx%Pdr|WP5uu42)vX*gEezii}dp{#tON;H*-k7Q&oz(E2eE
zryu1;4@af__}uy?2FDegAOQD4!X5mhx)sQb)jl4REfRc(wGIZXm1pdoQXmdR`T+2X
z-m+q;!cd9-+5#P<Op>p5hJgBRZMfXj$_Z=m*?`M?v3Q8yoSnT`dstcD^0>|$G??2k
zk!!-2L{kTe)x_9<7GH0afcO6UMa1sqDZhEm`={k7t?i`Z)hqW&BR${N<ruDEvQ0&p
z{aK@c=(C{R^b5CfNyLjM0FY1k)p{ah`(2{yeV$Tb0-pu?kVlTh#Yoy3B6rp{B)C#T
zZ&lF#bIe}Fr&yr?R(K*!y$ee5Z%b@6(r#qRoS(;%`bpAi<=G%sgm%I_SbBfbs0Gmb
z(KC21p?m547ThY~@zht1?0*lf9RY>6r2@<WRhW3J(?7T$*5HM7X6FtMEcuTXU>T4J
z!q6Gf2z*~F<!e8Zlm~9m_x>EeGtr5yAwumwDMMvi&n3EP->O-y+|N<Dxj%1q@#=5x
z{kLHuR?x5Z)o#D{Cm>!C2y;d2@q+qPVrKcD6{M+KpySwx+Qv=ueRC&@&CfxJwx?=@
z)ve#GKb47m9nE9Sy8+&{DiBk!R9CTP<;@xKa;@*=EGh`E2i9gH7laM)|HEl{xeN&-
z?cS@+ZSC3tR$W0!oi;nJ8wS4Cd(ZXmYz;Lq^F~S|$L+bu7<9Cz12cN_?!6;?lO=WZ
z7YCXZ@Djdxy?%SynQmi$J+;J37P?9hZln3U$;D?C%09_8vOM+j-k<4|?)y(S^rJ2#
z-sj`Z;vdC}JEm;~0;w`NtD#UYmvjE8he9M0`O`;13ydFhA4_8qth^>%pcLuO?`i3`
z7+l7h_jW>U3_caCEZ@%=Y8Zv$5d(7f7?y{Z7$~*NQH}3MoexH*%2hoMJfg$%##R^d
z^?DP8I!UUhfCqo*<~R;?X8Rp`bX6bIwqX*mS%O_Y(BhFu#Q{IbU|zqx!!i8uj2PB9
znov3HW4S`{zddH(fbWw|Di7C?FV`46vee3Zb}vS4qxX2?>|`5-Bm2OqmoGBQ)Nk-`
za?;SKK^M7zHvs*R{~|e8KX3olmGF4pCaT1`j=o-Lr=&Egf6_1g&WId{fiT2}u?UzA
zbPfM=Oj{Wf)6;TMan5`(^gv6u`pPFRrIjAKs`205QBHG7{AwW=Z;|(>C9Qar*5~{G
zScX4?TN)4>q~PXH)cIb#n5G<fx>T0it;1ZP-Cq8#ar$_<2R=i3Ij`=JU;u~1BB}?S
zu7_yXnLM78ydM{&R3zKEcrrizi-yQHLaARNyP&P-H)^xkSwH@x!!54J_i~elgE#IA
z1vQ`hsUI<_t?)~)SXkeYucBwvDz`GUZy8#<HDls%TBO_P17Gi#Tsqms)udBTHIDo=
zU!T0_ehqx7bpP8hl;i6o;Uvj50KbfTC-tw`it0oHw`HA1%fTli7l<-P*4|ricDKYE
z3HV9=v0^8lnt9rSK2vHrBaSA;VovjoJ{ujowheyS(D2XGy+O8KV^0AKE{E66A|Lz4
zu#=@_s>iV(X;{><WJ&nP#yw$R)yiATpT_-&@(}T+AyRXXew@<58XS=@7Q`9YRBVe4
zPmYzW2Q$oHnAR=D_6sYQrE}G}VjHEbSFo?0ekL?){G#CR%kank+qrX{Gx4FeU{Egr
z$R4_h1CbPT6rn0-y|6_pFa;HVWO7%_rn*x*w(mY^q6%kVk8seSHNUv2(^RVMdHKz{
zteXa2XvcV=Wqyul`i1eEa<<E&(i@Abxi0#*<<(Z-RCSaZgtWQ$SrP{G=8n|fKAPsr
zrOw;Q)cvQs^l#VydI*9ytQW{)#JYNR*RoufYc11yZerZNX#wkfH__C~%6+P(&j4@*
zDeAf99AOo7BnEUbK%@O3zs+i7m4c|iJ&0Lp%&`bOKB?74%Ja4xlMdy@)N0R|c!m_Y
zO6XMl(@FcCR?ELr8J?hGc|Axa0{|%%2FHez#S3(<t>3oC57tzc?&iOl(GIq$1pdC-
zaiP}5FQ7XA4v7H}i#&4)V_K2yr;3$NrsJtCXK2}1W|_nO3_~RyEA=l4Gta(M9Dl{Y
z@BTP-fTz0NcVS!p@3^Z`{?{%NJTepn2A)3Bvs^dq+$79t&bGf1s@s^N&T%tTo6GQ#
z?x1tv(``9CstN2>H)=eXJBo4z<0ga4$c!y3fRZ;|l&wc1_HU!$-}nA!6w?fVg>1sR
zak%9!5yPwAW%t7J`Xuo^@%5rK_1s_W+fQ7d-dc0bHYl1dUZXK;b{<s#^=Y|URU;E0
z5~^>D6}TAj0GT9U2l7oC+Qqe%g^kucf=jr+w(2EY{0VY9{ilWh|GrC~YC($H56wLL
zD*Zy;74bB#xwm(3-fJ3_!nkeA2_IBG96y|1Ao$3sXD2$l>sdeiQ~P?Nit7Z8ANd$d
ze9~^7@b&H+su41M6Z5ki++yXEEYtk`&$k=<)h`|;;XzPApCv`EOXG##ayJ0_uVZs&
z|FOl+zTf_XFbLiR#L}Mn$rH-~1}wdF^vREB?yMeTrjSRiyY;PREZZped#P$0hik5l
z3${#-W66v_B?n+zR9w@#ue!Z3*mN>oeH*-<8n@tN_2ROki{`hE&$F0kXxa@1Jb%b&
zjjrh1O;;`+zIUzX|Cvtef2Xr=^vUKEl5u-qqdIwMNS^HMTd!Phvp09YE_D>WEviH;
zM9DLDj@y=f>(9~5d)7pa8D`NRPDKn?9&TRdU=ZKj?thuVMx9LO^3Cn_UR<rwb~u{E
zML^_$s}gu_&+n6S@hvHFch{ZiCwqy2t)0DV;UVmD_FTi!>$D}PF*TL?G@hOj`_J~)
z`!56`_*T-lrm4NFJO(V=w#bWgNO<DvTWk4(aodGD&r1DBbaL0OMIAFcA40Lc#xASI
zJI}vtI05$xI+4D-|2_WwZNdFiOA=|x>K?)ycZGU(*5oQBGc$cF{U49&nF7;(TmIHA
z0s)$ut^KCN9@j0PN+#Qfd2_Qv;KJXCX~vak7U-}9gvx2xrQGp8_;K+DJr6F7Q#K&Y
zvTAb=RjPys_0=_r{awm$Nd9L`U;lDu*JF$b{Z!CyUn$;>BOIpqNZ!$|eXh3an_XU)
zJ<bKH*WQ8emBz0QIFd+*n_sxzkJ$o6y4acQozL$=?~u1e0l45?ZR>nZj)N&vhYwV5
z5Ft<E!r1BZGuV_wBX9b42md#t5wCB@2iZGwBQ(JPhy=%HjvTk#)1?9Sa&f#WU0sUY
zhOY4wAp9wm`FjY)Vg>(yo9F*7v*4;Ua_<9&C`g(E-@&w8M3R-<e`vPj=sqBH1^uw-
z;2Q5zH!hJ}l8*ph&qID&u@&&pH}kZQ%u3pP+=t3`lG-dDX54h*s+}=i7Hct5Tk)NK
zm(jabzDy7$0@3>GC@*5pzhgu|;CwGpX+``FBk=rmP0s9EHWXLxM0_bYu86IbdOq;%
z3%sb>w_L`-i}F5cIinjW(pXhxK}760@!Q?k*1}c|Y8u58g^vyfWZc9f;T9LBxgXG5
zYSIRi4O!!tlb{d(nG4Q;=i-q<&WQDdl9uAbXSD|9(!46HiFp;t)U)BP*_J@>u-VEX
z>nKvePp*}gS-+}h4_94G$pqtVF7pIWc#l2vGb*~c<4(WPly&!3iPjhH@ubM094Gg_
z{|qpfGU?(A#14f_ZM_+m_3tuf_#c-`8rKJSa?kqd!g$wO&am^3S28)nT(dQcrtPb9
zuY10r;fhpq14lmBxVUt`XO^H;E8A=9isXI+WPqZiMJ%iNiK}!yBVWhe#5)?xtMr}Y
z=Znu&y^<_TYTZ(8vqyegNHi$ydS+?r`=52ZeaRv7&$x7+jQA=Uf0Fx@FacDpTzozH
z@~`itd6Wn|R$@yJPzVpC94(LFiAwdxkJk$brBnAlbxueqKP*2=e0WPRXhlz2=QXKp
z<fsS#xNf&xmUPDHG;jOGP7ZQYhClRify(MX1i~T%1N`6H@T>p3hY1$P^i2p9U(gh<
zo~Hqx;@3ZCs=Vtr|DX2LYT@ie&Ol#6qwV@@`hW=pS^m?<mtXdOFXgUp`{z=pd`}B4
zM=iWA{cUCahV5S$Lhw%OZ7tCD8wghnioM89ja0IGJ;MJvMQG~D|6c8uzUM`jM(fTm
zB7tP3AApYnjO**Lzpc=JKM?ruE+HUDcJ#|@WGduUF>W%ccYbIXQ4i2^wkbt4ey5)@
z7qY{&rY`_jl=1z)1Mt6qiy-Cth4b0an%<5;x4B69WChfhBmcpV<%0inliapioqvuR
z_<NZ@p9TL%vu~mQ<46Ik|Kmu3f&b%3ft3IENP!8x0v*0TK6jW0@G`w?zINl|Wf|M3
zXUKP4>1kzS9`?y@`t!8r>^jvwTXB@32!(^=tJ<~t|4*CliB|<WgSpffAmGB8B76ic
z7|L&|>^HX>=%RG&#(Pl*5*_4)4sS`d2dxT-AyFrM5A<I59s#~vk<!yk^vCw4?r`*Q
zb9%vaWccEgi76<rw{m7dQ$mR+3m6;GSutKKUE>xq(E<Ztbi{@JH7lk3MZP%ADth$=
z{=k|fRh=x&a*e4mzFntZ09#t*mhD>ElOMiSRTadeU*Lbt*?rH;LTlj`5CKFpubzkk
z{kSeb@BZCO*%$7i{9(e~z%?T0`*De@M5*+Dr-QEdf3VjuSFTojUW*t-`-O=v(wLs(
zLpY`BN<+46MbdWVcv@#@s=SL1vWvY2)~~#-aGV$ClB_EouC~AMLdQ=pD;hOuH`P2H
zegatMKbLQ9L>gehliLNw{}Ws27n}WXa&H^AU%0*sFa~@UlY*<JJuQOqp3&q=FX@T`
ziqJ*nk5LR6KDpcwGyJk8D6G!}vH8?!?g#QY$>QJV{_m$gHSj=hw-_{vwe3WsM{Xwd
z8}QQ4sazqpdX2R_-h8TE*C2pm_%kFs2zyVQ0!d<TcCmWyc7H$W)r^S;U|CD0kh4EZ
zrF0h;8vR%Ufg`9&>Zn8q0ZuN0(Q-}be}XnTVc`HaE)j;Hrjk{TxJ0EaaBYg(u>+{l
zpebbwMnCWlz<^%34qSVr55vXi_%bNcYH~lakO6Q|mZ`t93d2!``=6lkzmI?M4b^#5
z2u$D+c;(fvp`UB#Y2o3HroFGnP7%}d)1TUaUN0p5lEx?ohFgq+1^_OEYxu9|+F4}x
zJ^TNk6Zf0`Cy4uG&FNV-oDIlK3&y`EiJJB5giJU7vlU$KUu*>+$ixEspxPbPybeIq
z+*&n5vt)@!**biu@gsQ_+eAbs>bZw;b+SwErKY6EA!Yf~%@5=YQkZuNv^>0I<p;3M
zwOi9rH(v$DMU|lA^dLk+utC;ap@#GYFLg&gA+SJE6Stt20M@M33Z|h5GaT>V9{!lN
z_*sC%DeQx~t2RgcG#qC%^X*EM2kD~qIDE}tCLMvj`yzKVb?|*CQs(dI=yUvQ*uLow
zPVUK*umb3>v(qV%O3spxHVAMywVHXlcB09VF^NnUOzevvzuIj@qUuMlFaHT!wM{#m
zHmkSS-0#KroCr_aaK{Wxb|MT0fRIPKYD!Nj_p2J=$~{(t<AaZ8Te4raKO(Fi$5JzP
z^=+fn#^<r-T9b7WHskdJs-Pr=IgfYB27VQ4t(HHUyax9$ks0k@C<l*+;}qK?qzeav
zwi!NZA!f&W=4XX&El7(Kk;<rSis##}XC+1Vs{p&+;}X*!7*sHh6T`nBa1LJ)*A8!|
zk^_=WwMJbkxb%O#L*a*i{h8(Xp$LYm<qL{|)toEx<W_$Iz^{Eeex$f;D`42DK~KFr
zyrwIovolDx<;a!X#NUviE1ljGwQ%_JtM;*4k?|ds``?z|SU!$5W9iyK7%5Z2xwwpF
zq<Zqy)HaUhC=T=OzIf1!t{gS`Ws>IQdDgq}7#Bv#GytY%p4mg6T20joHiB5am*f>?
zkof8>&{rPtJo1OWHNttn_>aIMdU>Q8Bq2I&Vvd#9!<h;Woq@H@NK_bM1QvhBp>OfB
zo9pjP@)Q1xTNss3C-)>hz1B~_$G!n!bU}nJ0`a`+M%_1QJKH!J?OlsxD!$rWegI@`
znl*D|ztw!}q0IXZ$ueW^5@;LG*uSQz!CmC)dmcx<yqw)lyM1NCW-4MFNVDb+JcLQG
z@hXBu<U5D7b1lO@l4k^z+a>P@slP_viq|RObzcM)j%zN5|NSNxJ3`&gBH?#g<tO}-
zDEeqBz4E-aTga#XtURLsd0s;}Vz<c(7@x-?()Y($doJLNV!3Y5^F}`Pcq5mop{GLF
zqQF;;(=a7z7)TkwI~`v~7&`e2-xp)iUL}?r;Jb-RRt1_cg<1pjlPrEMcO^tm^U1Lt
zc>0=g=Y*rEaxKmn2wq!d2Wv|!=|!oJvM2k_C${c9B~%nKC#&qyBaRjEWw;}rqv5-7
zR@>O8z4kR~V4*9`w~Agohkf<!58fm98ENgwiS5hc25U#xCYYH+!)aLmTHI3Ke>ai7
zI$;I{D-iQ>C9XcDkr&kZDNJO_kgvy4Vj2qMOW5P2taDt!cgca=O_~DHM^e{Hh3wB;
zycXYM5K2V!n-2jR_v8P^*jq<c)kWRIR}oM`Qo2JLq&tL5gGe_>OLsR&OG}reNO!k%
zN_Tg6*LSY^y!wv!_s4e~Lma~4oU`}ZYt1$1JZFoRML=;T9H;=$Pm@zjGk0;1zh+mo
z3nQvWFnfZqm2tAJu5h#@A&bA!(9y^r^yL9X0XWN`&7v$~YLd4I&c3&ZzsY=^_1Ks~
zD>3h3ynfCbmHmi9n4#m9(>DNN!W)jsrm*=@_zB**8`n60);l8loPxK%LH-B(e=1{+
zzp$2_9~Yhk#!YhuA!hM-S9MCg+?F||v$f~m<elpX3?Oh&YYu+mgVez1vcFsG2{Sg_
zfO!O#H}Bh8ALd!-YFKp`2m4`dNNUyerd|PFe`*NTK?A|GI4)KRjZGDP(Cd^1<0T2X
z&fkm_q+;}}OU$=#TEt%_cz=yfWgJV3W-=%JVHZaL@V;1hOl$bjbwAUzQEe(E7M1Re
zvHH4yY|)q<M-oz02LluVCvUjA=Xo$7o_@Og1FC8NS?&`P3Xz!jP_3`hU2OMIV~IIo
zeYH;yhcUmrp|*z5KDePG*>c_bOg(GEuYIp^U4OjH;m_6`VShB@kBG+7I@Rgwk;-|W
z3n4M;`N2BQ6Q)yABXzE~HT7Y=4Op0VOh4zxLJ@Y1O&yJF9y%k6V1^b-o<$Ke;`dcu
zb%&#=dw@iF`G;IR?eh)iVGw?;1up}@M)Z4CfLd^N!yB+}$sfHBx56Jg@6er}G{Qu|
z8=C77wdjfc6nO3n%e0*I$V}KR{&{E->-%_KT;H`FmwJlh+aHi2oJQdNJZIPcRvMd>
zxggSm_bo-Hh;8-uOY`qgE)~!@Q?$KhP#ing!cFO`E98=Y;H}DE>KfGFwG=lFEETkB
zE@WrM3&{s4*8a>-^rIU_pP8&!fINbg^O(wI%NG_M{}9g|iF<8VXDjyItW+fw)3*4l
zchRKD6z$KgRUCW*Mk^bUq<Deh`!KAORmObPMP<AhEfdSalkuKDq)f`{df>qZG`a7-
zeKTfkLN%_Z8=<q&d?-N1Iq4A5p*vCeJt7?Ye*^LjJ0uzM=8Ib_n)~&mON#DXpN*Gz
zuG{D&hHh?-xea&@yv+L5?iSEfnCM+k?3ooS+#62bl+WY$wJROWF>&W>(}X$4iPi#5
zlm#%?64?<wz+4L@NWG4xQlJhMsh_u)kf{A=#|$u$7Ega~uh_!$zxWn_aU7DWwafA-
z2RbTUwiq$85ugJYuN?i5X*9ZDzgza)NU&d!H|?E4K$ay~mR>n;JaJ#m!nzWl$t*U*
z{_<kjcTUUDKIW)}35vsIMcG9-TdiC+fnJsh!i^268IO^)*Ekh=G&}`x6Y%)}^EbW3
zf9u!stZHL<fd>p@kpZoBWxTW|b@&2J(<t7kk|U7-P`)wcqHofKcLWbf5=uJ|7$e|O
ze$!2(wCk$1E_|~|jVS9T1BSO-Yy$aQ!7(qBpmE$apkVDk`rAbfhXU#Q;{s>hRFduB
zyGo$B`mEs@pU;XrTFA>2)sK!dAiK;no}@(dC-Z&$=hBLaU*L6E>);T7df%+{`8GB9
zf`NceYkR2SdG%qKMF}@D5g?Ku3~Hbe!eqdUypjj-KET$cF~sejEA3ikM!$9{s9vCr
zK^82U$OT?M{Cr*|Ux!)tu?9!kEG;rAj5F6;mV=4H=xZbaFU`+b_rch7eouX<Fl*;0
zc$sag08D4Jb%;vn4I~d_ZuqnCmsI`<9!mg}HqH4WW6E#KV&g%e41P_B)PY30>Al3s
ziiftE!0}ArY%9VV|0iI6<a){znCiaTUpn9cI7ca@!;CFPg8U>U5f+u_-FTz<%_<I1
zue{Ag=+qO-&vDq`S$1k*EQZGH-$P|uDLn1!cIELyzM)@#ol?Gv!<ITZa8bquBxOe0
zWOC*FWPKtCYhr)&_yWRs3+n8T%KQ%ZDGgvm<7uy(ABXbkpK$DY)SmDc)jeC1y6t&9
z_f)fg(UNeH^mxQj4EewMhdEIt&8Jp5_p92@X2AYHR&G6Q=OdvCrH%#?8dBsIYZ4IY
ziHxv-*tl}1x=52w2{x_-oZZ(5uks5n&o%0>CM}x2oHNK#Y#N(96p*K2e=h~;Fs8QD
zapapPlzU`(-U2u<Z+_SGmoDRe3v_Y{jNsveg9m_u3jrwp+Rv5j02|%LcsP$Hh!s_c
zvS{r8b98A*SWQ94l-2Atgi7F`);nqsl^xi)oN>=@di^OF;d7hLj~YdSsIbv{mrJ+`
z>F%6(P3?^SgTWgHZ*?_Lmleg34mMWO?@=jX0!C%)f=0&E8MsTMnW5&R(&k7{;B6Wp
zqe=b(o`-yyhfGD~C&`+g``OJ8-3)UY_ODu|&u<tGHC;9)l{BHP95g)-NM-m>XY%qv
zgGHFB(NFAbgh)9Af75K%hq-p{OMDDMY6XV!)qrFeNBGR8VrRXTC0s{3MFT>~^<Nlb
zW{o(SOcIiDSE#L<#s$0Y8r6~FGsQ{DR$^W5Yu9<y{&c6cehyJMUyi>=Af{%$72?~l
zqU5fhPe;<Mo(ATCSl{sFJ=VuJ3V@vppa7+q0GGgcv1aefAPv>`zea1De{GsXG;i9D
zQnQxlfaj790T;^;A_Oiksre}Z+PIqiH+cY}PckNv(C45=^V<nN|FNj+9>j!;!enEi
zihu4GZP&kwo=p$-mFaX$q_DvHJ%%Fx6d7HB<uEkv#);~&sa#%$1%k#K29R`O00yt@
zo)HedU$$Gpq?Wi6?@LsoLT`j`P=1FZUSEquT1~u*6|%A={<qeCt;d8H*2ha@S#t8(
zmi?u_VO2_<>ZE$dc<Cmma{k$Tyzy2o6#?RE&tx7&3ZQ1-pR&KpvkaXV!CM2UH~+*g
zDv5)=!F1H$7s$c+7ZLUg0anfl8)iQ%2=f+&UFl{}^*$UBfob1}5(CViz7@dyA%2OX
zWijO76vBy*W~6ZsbN-RizfMIj_hpy&YG>dk15O=CAX?g6-TU>0-yK3Evc-j{fH_%Z
z7sMiHWh^nJXE*s;4YZIF-{QMM(<WqO+QpvNHWC|I?-#In{bq=MUm5VX*OMMVcrB2Y
zf(7h35aAFvbCjKcMdf1!#Wxpltz3EI5tg4mAa%Wa$JY(yTb*fG0O*X8efezsfu1QQ
zl2r;9G;AcdBK!qNrZJ5RZlSHu!Z<OifL13Na0ndWd_%&iRl&CH&D1cr%kBOYpA6ir
z!i7lVOv7kQbE&7bpOJ-59g?-4pokJ~i*<i^D1mz|1wO9L#qT1)q4`gy6a6hIB~{8%
zVG7F*Dc@)Q;SDX=sr_Co<BB3WR8P*TLAv?b(B3jR^e8`x%z3kk8_9&KSbf9xhd!Qo
z-!EUhZJpPrBH;k$4)~r+_|8@!Qx7sAZKrC&pmA)dPjMk8Mc$xt_$NmX*JI%9H<gcb
zmRw}1Bn=YnaSy}ZhH;bg=C?aID!HFv0A0e&`A2y{5!8JLS1SBUNDevhWTOs!>85I4
z4>DTdzovL3`*c*!e_L*Ab_$RLn?q@7dcMifmk0Ck#8{kozC(Cntwt}qHkbqSjuL(x
zU$rb_xh+|l0d%fEn#YEn2OajteNvAfz`nKUH=mrvDgR20fD7~%_gYyQk^%02^sYAF
ztnsV;$_<lRH@o3?Z3Dm3M)BjxLbj2IIV7P}U@h1E=LMKN-{d<vV=gYNgIn5jS+4DZ
zq+6wP`bM0lwY*dreODV|jD|P)&CMExnmx2d^{h+-<0?X%vE#BfMvmR(*2gnvt&Bg$
ziV=(W5Wcal{ziFT=Q@8W%n#;dfC?cRzJgn=mY=`w4ah<H^g6<7)fRSipF%DeEV`f8
z0II>KF~5&fqvK`|SCYtp!=olnsJ5{@^5%CLZrS+@P(UiU02rSe&V+sqY=^iF=K3T5
zXjxW@*o`F#q742NELk8O(xfsId+0<35Xh-+okDq#oBKujY~Ci?k2peqK+)XT_#OHm
zUjMD|pb4=Jo9&OXlPjnkKpO$%xf(S(jPhwS>KF;Kk@ZwzJ%H&*ZdUlxjKa2C5{KEs
z=mM<^$V1e;eI)Hb)BG+}#T`qq!qZy*<lk!d-x`|`1JbPH)d%Yrc!0O#QQjILSMkKN
z0k*;Nu2TeBHd##OK?I<h-Uc)R&p<A;&M-jS-(W6n`bqUn5o>^oj62;1lf>*3nrhl8
z!%h5ATT7h&y#sOJ4vv~6tdG^)-!>|kWKM=5zKThVL4?cT2)e0U8MdiGRDY;y99#I6
z);V1EKL7;+a1S^TfQN^>!@Bmy^veI|>i<DMz}WmM5=z`i5OZI@AjSs+>VX+QKrh#`
zXtbm39WPR5O8O#Xje4Fm{)2zh1UHt;#-esUwDE;X^HulQt(nThi{B%WH1L-$Y~kur
zFbT-718Q#}wEb^Apj?$^-9~nDK7k-GgR(b9qSCN_o`ZWEH(jyGCBC`T+c6EI{w7G!
zDGw-7SUTb1#(Dk;lbpXL095i0AOq`}M&Y|+0W~;AN~g%&qD()1cwwpB+TvOhcpgN~
z)1{9GoSn1&XFp%sl)t12Xv&TTVmem+*zby{3*<Doo?5m)*FXBd2yk;(;rrL{p&RcY
z9P3beY3^9X4e*9s`o-^e*VTSUDT;BeBn`}u48XzyZXdzm@5%tNP1TOur}RTGF_mM6
zy0<WpX3fw4K)ZaSe}8fu4%o^0IWYu4?;Fx#4m2PDz8hd*WJ|ApZzb(j`OlgIu?+`9
zad1Te?BcKmyJ>dNB~(ZTB&jC>YKUx)it8ULALn0|J;}gz9NJF&uX0}a(~J#COB{F`
z4_WjC?o-g4ddH2&jy%cuJ?<EO{<G^)N{<&JDTKgXsW|=<f**tm<Uyiw`b@^He{SWQ
z|33q(V)P%e``@~Nd@YFjzx)3F|0L-_9HVUHUYX-2g>`>6vyaF9TWCQgBC*IV!zKX1
zZbuR+78l+aSmHqktY_41yCw^AlHV6W^Y660OY(!UvQr|+k(<_KeM#KznAl~O3p7K(
z_=n|BE98~Gw6^%OJ9dlR->L*^*{;Y1`g0%;;1%%jSlu2R$08CG@be&d0VzV~)t^&I
z>isjX?@ITfE2i@{84O&N2>6f}YsRV3qspr$n;(w8R{pl2bnempiSDz%=PFV0EIAJW
z@@VRU^&=)w30nFG2kKjabK}$jTb(1Xa1TtY<CQ)fy$$IAo+wFcl>L35@xgy*0)xT2
z&I82uGa1q^D>(Gqv|8O-m~Oo8D&$8XhA0IpzJJ1m<Zt%<fIXN}MCgUcGAUE+;<2~$
z16j~*b>cG(vO2KyQ>h|r68%s^1<LP}b^eX67P*R{lsrN&X_hXTVxT()o_|f*kP5A`
zm=F9m^M+ol?Ah;Qyg2_krYxCTlOKfp(IQPto({OCxhC#Drq=J1gfRd0_3;#cCz@9J
zR%3<bzTn30DPC7HgvH92r{(`_lRa}D(t|xC4HR$(NM^~PstOC4++k(`UO))>kp6RX
z34d-bO{O@AedpH#o6kjO>QS8_bO00&!EDHCLW?6-Dz1C+2X#^y{?BxJ8JMJTEAoSq
zKGLUYott(r?jmJs6i(;r8)`dEF(iCvnHd<F5)K3!qW{`UD+uiW^HC{k-@bV=C2;zD
z>EnDO?@fn>IeiEc5;@>5q+`Far};x=a1j4<%hR=^=TIfy8%Ns!d8s#c-<%;=9>wX5
zbk7vSc=;sckY#g`^!Kkf3ivzmu>~m{zb^CT6TAWL%n8DcBJ>TX!(s~nCPzvl=d<sI
zY@`1=%!|48Z>sK#<$z4b*4h(@JxsJAYF52vzWF5?ys}M*>2Dp?i}@cVNYatsIu8<H
zJy0AXYFjrzs2Ezd!?n*ks>AR99z8GSf7RfTcgc_tu{!~UZKODDoeukr1<evL*c0R2
zxVe3+T}u6f1RMWZ2X;ZfM&KRrR6*~-I~yHyM>0nT%(*|{7VjTHzpa3WNeBdMwP9Wi
zpr8Ns@nSBdt9bfh{QqnpUhoC<(o2!saRoyTGb4Ryv1yjedUh4?eGuHZ*%4R%Kd;Da
zzx_Kfk6;R<U)r}jAHWm0(jX@}u5>v3K5Gm1Uxn8zTaxVEI8qA?mouoVTiuHeKos|e
z_E+%r+&#h<@V^dd+5R`iB$I!wzW`8}Y&A%I53XJ5dMEHa02r(`m*4Pf>FjSAY>BEw
z`?dr70nL0UJrv?9bLxh-K9;-I%`EY{wXW=q3;yAP08vEdA)8Kn8Mv-a@}mh}Jve$^
z9wiW7MSD_zI?D9ij#E-X0}KkllPCWToC;I=D^NaLI4JCw{})LUsV1I0P%oBH&GR{^
zgyq{|{>Hb%K(BQt;1`~1-!Bt-JPthOA6^e`if@*8PynUW4C`+sgRDCHNxk&u;MyVV
z#n0qboOUJ76r9?-pMWjL%K!@@Z)<9<I5Dk>^l9wg9N}G>j%j`L%;^5@;{E)g=cOZT
z8lbREr2Z4*c_}LP&;ahPt*wZn)#wTJU<F!CfX1?4u~0j=@tE5Y+C_Q-ip`0ZI)?H*
zkvYHrq{E1TyL^T}jNAN>!0vI+yy(7MjQ3!ds%o!kHPdjsn|2Kul}FbuPj(iA%v14`
z`@@GKN1*(f;r+c4TE!K+dyv`Jea=e2?-QJt83)rju3P0o+N)XX!5PidiP{RNuKOe7
zJAw6|@q0BB7E?MwAgrQO#nbdfW+Y;~*Gj)mHQanA*+*yYGg!S?Kg#klT-Q8&tDfO5
z=G_x!dtRq3{1hXu;-cAcZ^ZFyv1G%(^Sci~?fW<3o>V9U^64qu((_-QMRojYaXdhu
zFnLbgyMDp2cMm+9>}iHR*~tDdw1=fvvmQX4eKo^VMMD{Q#PTJnH*|I`BpYLS0eSsJ
zl^>?n`P5V?6jIS6<s$VQTsp=W!~G9?0y$h<w{ccg(%CH8uZ$!y$!srzgz4ip;Ge_Q
znND7?W)s8A;ixz`G<hb^PbWbB%ulCzG=KW}z+pR>Fn#RAYdh%LbWxnIG%S&_90)Y3
z8A<Gx58M0Iw?$w)N35Sm2c<TB_YjJ*`=d)Fn(PorX7u|8j5m6Q<BUFek6PxI9P^ai
z@^=!M0}XaD+$UaJ{i;<PJ^fOjo^HHM#qPP#76FcBYdfR@z471G{x5R>`x?Q47q>?M
z0`)soL0cim5V}^Uk2*cqZSfl%gdY><uP8QeI}!qZSnP{kLW%)0=i631#PLs7r!=UH
zO~laO5_h5HoVpl-Jz68nr<{)U3sH7Uvd0j%f3Oz4v0sq|jU;B13gzKrp%JYxkV-Ua
ze+~L_pJooW+w|p^EY8}9g9CX<`7O$WKD;dY9|zEXmnPs7*u4mxnGF|;Bd|Ao{%$=+
zk)Oq}nch|B3WBNHwPywfnF5oK!nCtW>>0~RwmHV<I3j!kIba!Fz73FRkeP<|e={tJ
zVU;)XVdgNCDym&|;cSSq6YqQy?|0^bts0Ec%d?leUh}982D@#oN%eYs2ZmYUiWeG`
zvvl)+%ob!Y5JyAQzy#=JM$7<VqiZO$J?<Oan1_?s{*X3AKrpLsk9M{D1-+~HRtm=b
zQ_L#*?^PsKZ?MI5n0!eIrog7mQaOfTjS$Yh2{KwKvZ5YVVAe1jM;$d7LKcjH!}iai
z1ZY2dVkd-S^eXJVS=T-rPrg1?EH+v=)CGB=e}W9vi~q+^kq`k?2GO+H%O#?W6Fv6D
zY{ej4UzMA_Btq6=DFsttwBDFz5AV@VC^kB(es~hs=+!W~9EnC%Nx~L&@)~?tahAbT
z?ClrTUWv5KiEd_^v^Z!e$4uK{LPqQRV_^+P(fRO7_%YKp*;nRXAJ0p*<0>vvW8;1#
zNbE-LSl{sV-MrMjCIToz?#LAml;QO&&c2SUSRBFm><Fh0dCX6zMojEBw02&nP-wB2
zOIeO;F&6ak4B4g>E6>jHg2b%xDIBg?!ak@<+w-zjk2Pj$mXjao>lF@2KM`0tM`BA$
zOO$>1e0plz>{#KJe;=yJ=e{)Is;4!MV(<`L`RHcVxV=$9v%G1vc`w=FEx}#)0wjL<
z@vMLA8{z%^MJTK{rV@jO)>Vv=b}vJJc{TCD!`gggwWviXUyVEI)y$4xAAECJnf^92
zu8nT-&bB3{`|31#h}i)fsC_n}Xoo7a**myulCZQmD38h>{^lwJCq%k=TDBu^Gi_&9
zOUGCiB-XT0WS(fGn|$7Q!7965ZiU`EV?1|uY;Lh{EBY<ZQFxRW64~`rge`9?%I@N?
z=-qE`Q!rQF*P@F2iJsLx+JT=>Vw+kCa7Ev>y9(>dWWEVs|NN<#)>Zfn9$o5rs6345
zRWH+C%y-*j=PriCVmh7g*i|Luwf^P;(q^U{HW&fldg6--i7*VKtk?WjXA-`U;5%gt
z;f3(RJ#tDH|4`c-IhKS53FHJ{<vU|Mq&@`+)jU87=G#WZdV7B?cppbc8T>SRlBP(E
zM#LTGOca*={b5x><Y#%a^FK~i1D8SZ_;rqW<GeoCVGc^Cs*TbYV7)!F!!gUCyl{sT
zRYVG0b7MFV>Y<J8eI<5bnzfGMjz-~a<FzSwSW$I;=_u^nW@p7>B6ht)(Z!4_&4dHq
z=by=M$=g0Vy8YDj<a4!E7RJjE0ujm9+!_@<7P+2O?X2S>Su8qPUEB?<Cn>pCQ+Y<>
zBMz_4U(WK?yO8ds{QmCk-2jsEru#SNbLPl$=FfR`F01AF^A&ERlg+ag?O_&B0&|E8
z^?Z}JxF6fcWet>5JD^M{{AhRbf=$ms$f8}<=UZt(+uONo<u@q~V70(1{MUK&_%NGc
z)UI4t1)cI7#)##y@q3ULc9Xt4k0rd_+Nej&w$4il%&l0CV;BAm@lSd^V`Ao(bm6f-
zs03zc-nx<YUaz>lHXQu=Db5u7M-gV5whz(+Lj*>w0hUiL4tr8|^E}FwS1MvujB56v
z^E{PygIrE{)Te>`kozE82W_IFYadlRE}>9M4Xi_xdr(ziO}^2Qj0=YY*(HBn7it(;
zhGXJ=@(wiW0CjAW0&oTUwD&s+e$C%NEui0dqIUOegq_T%*rJ--gZF3X<InW8CKQtA
zO?8b`^FYHPHy&itc!n6Xl+Wivvf&SzTgk-;_Hotu1$<j|P=u3jXywc{E~?Ikn*04%
zd~;1cm15jv6cdX*wfodQ80dG0K^iY!0MFkX_)gTpf~b6P%^$;~m$z}5s)rVNuA41B
zcg`V8hoyNa(l7}fbgl&yYq>nieabt~t|KVTo89E+zFrNwjlF#V3Yt*Nj%y-D;VZVW
z?^qaVn#%vVX-Dh<MI>#nxRY%u_DtI#98r`w=_zBF<?gWL;&T)?{FsB)3zp{*9yR?p
zu%HP4vK8g?-V;5-AbuRTb=ZqVxPpr1d2K?PuBjiY`ZebrLP8tC7?_xtuXwf9&Nak6
zJTz{zgmz4<rCrs9KWoNg5nNrh!4iIU;vv|sT!K^j(r$lmrioPOLPt6K_)b#Xt^O6^
z_P9DIuw~>E4UiBAx<MCkl261Y=J!}v+0L*E&ZD^}5nscUt2`TKh;URRupw5Yboo}h
z2ffT%qB>nA!-;Y4LuypIp{{Z9^yFQL&(Px*!R8Vob&$&l{w(&HQgGwL)leE(*;y^{
z(!FDvuhW3x=WL_(qD^4_+(3Y|;Q&zEU#r4Cw(H@3vOUVnC5>&>p?VCwU-)9F<DekE
z7#c(cR7?<P`{iVF3t!EH->&xYzLHy(vgMPNuhlx$J<iVs=NA^xio2FC++1DFWEP(z
zyB+T^gR(vy;EN6ryv&=ecTF<Mv@jNJj}Lz@-?ZbfV;)>b1rlrM8>p1e@#ysXe#0AA
zA668eUFv&2*(pu%-mr+3KxHvKA<eqQ-J;VhWVI6NTRlxxBHc^Sckg=74gSGY&kq@E
zYtw|D#$7|Fn8J$_^OxF<#>l6{r%O9B4C@ajKbV~p4{hwWZ+g<+CFX5+4n*&tI<V6*
z&WnS@D!fg~YhMYDCqcQ-(&!qS(psQkzf%ri(6WaT(7<0byhtv6zs<LBIQL3YJp*4f
z-5h0A5Qz2>>HBxMIM8CuBmrXp$RG>R1Hqz*XClvf9G4T$Fd>ve0^>2)jl7e<hd#V(
z_1MYZU&EMp%T(9t>(D;V$;SqA_0}>RqlPZxr!9L?&S)CZf##i`pry#7dcJRDXfNTp
zIkd7X7Q1#alyM*ogtln8*=*j7)lQ;Tty)jRw~Jnip$oT);%(wLQ+kH9@pW+*#VO*m
zZRPbCU&BnT$W=U>ErFTslA8iII0~!3PQX<_O|8*EX7m-`L+0OWp_I~cKB=Z1Amty(
zIYXqf7xUnVDOt_iN?|*leRG%eA{x$H2{O@dc9(B3$#5VmJ<C6*aYv`s%|XeCMzUbb
zd}<eM%@V%JA7jU?vKsz?OL8?yeGXWcCw%Pt=gzAOMj){l+_OJ`lB-IAtJ*7a8Rl?~
zVkQckV)i*NI}B|H3Tep5+Mc(dD(nEC&p_rA8u#}iio6^U(eb;(+MM1pN=j|luTj}%
z8hkaIGc|Yn38n-0yR+1p=6hc1F-@&QL-z~ACjwj|kL$jhua5EB;%|vKk#bG3&G6)#
zT8uEw)Tqv{&%D(SLwlN$orm^o8(6F6kU?UUVXwP9C+TDsRj$rtCOq_n21!Ur4j!_P
zANT5=&qa;~aM?W_4Yw!gIEO1)XC9^%ui^7go8%Q$$SPT#9oP0Fk%VN4koe4!C`Sfw
zYi^r>cWGbeC=NA!(Z&+n1>V&>vUTXeg)MQz@i=`;j>~Qlm^TSRw!B@q(>ND>uXG@P
zKu2q!*K)nF8O-nfX()f_{b6jbdbDWzu`uhoXhxB^PRsTU-Zn1CIB%xVdT>*LRdSl6
zX`qmAv%E0woiJe=?a<;<xx>Vzh^(k(+(d2W%|eU#?&|E=W>{SwRh~a+1owm;0FiVH
zPw>Ftin4}l73WjNI@YO+khF7qtr9-e(u#`ookfqcVl3z9N=_Qm17H00kSYAt!|_J&
zTys~jZSZk!z7`m2OU@LMg0k?WtSTcRU{kPaFl#02nbg#~)K(Ne_ZcEx-$vn#El-Yf
zAfs5l)M8%t*;@c7`-ZCspu2%df`TVi6MZX0&@c3KPmIV`$ZSOibsv8)uq>YNC(AVv
z1$&%?X|q_ctn@o%$$~xxlZW&+Z{^=DJ``O4h{33ptXi{5kiln4XpL<B9f=VytnT5X
zo=!FcHz)SWu+^w}*$j!d@W3r4w?mVHhRkUm?68njtWyo_n8cR5!hThNR@~R6q+H$n
z4h&!qA*e*?;`*SqU1(6k@+X^>r)ia<sm`*lWkI>C;L3uCCGNKhn8K3WV|1*PK0G+4
zGN`hyMwJcU9Wrt#6Q6pu<GRVD7PoF%9KdzzAfpuWTiNjfsPU!$hTmW^VOH<S){Gpq
z5&9Qyr0#Ma>Pf;}4VinWb6>6P${r{%Y|J=bgp3@yN5@Q<?}=3?@W!&DMTt11iM}!(
zLyMH0^Rn}&Ug+#M7X$C#L$5a;<&U>w!IPk3ns+bO=onkyQWy=vYSqp2a@85w>##`A
zpC`xh&Yqp)OH1l6<E&&fkPorY(NWiW$iYq49LGiW+!>vSiN9JaJ=#tyn6G)8`XZ|Q
z<By%fW~#f@aQ^EzSok*fS0SOevEWftP+)lx%KD-}@B7bimkt@snu=w`WNiwVvhH6B
zQzF@8OqFj$DkW0@i`sUkr2A_Tl;G0cFVVTx`Eu5S6#khJJTShS8SoubIS1C6B*<AV
zlT;)&jtaf?W~V_8UUozx^g!Ek91{Tw`?@3m&Hx?;n{T+D<!^M~$I+~{5zgPMv)}!=
zFUjC;K2H6a>sf-B<9T8{)`N<}E^YX3cZ;nt>oaTJueCkoxNa&M8W#=aJGD`FOI%V?
zQlDv*th>iqqQX{bN3UVb4Pz=APTxtEaX<O60My9yjYmD7d}m1WbD#50sEgHCZ?KGY
zWA&BFGm){vTpZnN(q(VV0xZczqA*ffGFkNdr_SG~%seWyACkyizL6Gw7X!C{j~1MU
z?d5b*@*Njnawrg+>OmW3fb)RVW!Bz#x2v#SgJJG{ApbVzV%sqpO(&j7dzVI1+YyN#
zU1Pf2?{vDEC7+?XI)Oi{D%2(+EuK1UEN%st;)=IU931Dm6cM?SyUkk?&gQBPohUCD
zAx)`;w4srp+Ba$DF2L~zm(hj;FpY<x5qW`~BvD-@@k}5;Fo}fZp*SvR%X(_p7G5%Q
z|4Y%7Yf@7oRL334656MJ(#K4pQ%uC3u+FOxEZ8Srfxw#41-kN@hZqsypN5r<!97X6
zXQ5w}HeU$C1pNFdm{NFAu~<>vHKP%TQEft0&g3p6H-GJ4PmjJ&{G1NVc8(2_6Rux*
zo+GvR+@xSvQQzu0H<wssicar~^~i5;>-}^53ohQN%FFt@g9I|D+GxvSwsGIatmBix
zcOd5^-hMA+YS4K0?UK@m8Lln`mZeWo$?T!=R^tr@`0DU2iSR<#c6}LzTM=;HvIl`l
zpd%a0xJ0_I7<Wl~Ld3b2js@SGDYh7~hR)``%bh6~-?8@12oAn_?Nc0c`|z63G$Z*T
zwD65TSOeI189wQ?uQMT%h5OAGwy-*$beJbK{E5>I1umP#Pc!QU=7D^RMx4)>oR5>k
zIpnhI{vZL3#_w0Ak_k;H71>QPk*&lEY3U1evaFW^3mU=-ij;e<<)&bPD;NPTTbiM3
z$`N*-yxrSZ*m!MKT3p%Ac^o9N3v)%ed0TR++$LiOLFJ5n^TfM02<gP2Z3*6DpzT4W
zO9vhXp;0c<s4c7Isnak8xnEzCzoH|5m0id{fCE&X@O;afjBK)=c_z~uMYfxg8nefy
zSk!am#yj8>92*s!x9m+C@YIjDC}-l;Vb*}!Q-R(DuFJ#)JJQa_qmZsl%qZ?w)@@p0
zL7+LCRZ$;fktYVAK7iVJV+pnx;|)6<U>dJ?3T!Yy#D3tw1!JK5Tj5lB2fG_?$mO3P
zt#8xY=vK&DPkjST62=kc?r_o;mfVrysD_@3yopTsj+fw`<2^;1FnAhzQJ=1p<3;ko
zo9c={%B%JO`8<(nm>1~JMFPZ3vu55?ouCc4tb5vp2OiXD<vfGI*xGLiyL~)gvs6Wa
z!Xy%I$4W+N3LH<yuVtCn`5exF;&}dgv3JqFUtRaXt5bq&<NM)o+X(v<JMyYQbWE#`
zq5Rf>xS=*wznMc<19_Fyby#-w_5#=Ye!?ewZ@bN2hj#|;yrZsjH(TZT;rIgf3Fwu^
zqYp>xIRnT`;pgmEteye@B~)--&-0eCglJ`Kbu!UZ)(DjnxM46{kqm_*e+eQ6`i{@c
zNr#a;)rCW#2<kdV$}m{9ZCoime9sDIyJu+iK_-Q)*%CJ&N9d-09FL<@H$;n7D+51C
z&mQiVdPGCbGQu9;8|-s>aJ&?L)qlt!pKZA3C#&um|AKc}8F=p&GRp9TXAj@Hnugnd
zUI5JA+pAk+66da>?u?^rPZka%`P8m}o4YGG6oCC1dFuGs->tYkos(e%`Q^?-3hVAy
zqv9897ov`bmOYcfLvL8Ft7p@zn#>~nWL2U>H+{`Rhmfh-_zpa8FjaZAC&d?uNFq;S
zi!|I%9f;W14liGe(`Y>dMHzopJ7$kix`8Rr=Ia|@hUY^gKmcQ=mC7v4v{6z=aE>bS
z`t2cl**Ur;ols^)Hnl2Dp7P?C+46N|h73idv2J9-!@ei?=dg#ruyF$(_>*OgJeYxq
z1?|41)A*pkfT-AegVrt~O#kTVSMdJoP?@iTa?Lzl)vvaM67Hgcs~o1)2uuj#11&E>
zL_(!6iVP*nUK9<pG6Kyk0A8~__hKmXB<ccZgZ-XQDXNdzuoJSS7SljWSxv*UNG9&7
zA8h;p(kMf|v#IKUx@~M>Lb(|ty!-n~Am$iK3^+QzR|yD$FP>sV=otH!GzJa?E1&bY
zX{VH%g?>3hS>TmD67L~0qbq}NqPjojucLtDo6sa-toXWw2A$@3N79F#j^!cJJu=wB
z8K}3oEKD-Y?IAl(nIGJzwbZTJKf?YL{{sF&^l}PAYzn_26jzrp_N-Q{F@_IZWdalP
zX|w86v+f)=mDSsuP3<$q$Bbe*9#6qmhvhWVThSbTQ3syLSUH{It`p|eUa2AljY(d6
z5^unjc@cOe8vP7tr$Iz95K2%$j>=-ctJy-KB07#21w3zqamLoY!u}3R*!UAAnh?i)
zJ2DB0$)-UH2Kb)p6ciEmMfA;&w4-Q7g4RSFmvX1=Kp!bYuYM+~zOt_kz9Pa%cbR)V
ztKxmq=ZjJ6minGUC3JZOTRRRgkK0{7DvXXP_>~|5k~v&YB3_TWb|OX1EJSS&H5Z3f
zOAL|g4nyZww{>_VQNE~~h9hnUY>L)Rp3~FNde0vZ=0C$RpLk^@L_Jo<RL<w-svl2I
zC@2gde)gMLRW=LC3;Uo5JnQ2oE3Jo8ERP40`*M;T?hy}-l2W!@aBi*Y%7dxtUtV6S
z?Qrls0S8nErq6r;dJNKMRs#OiDzbQP989qoq^GZ8=318nI#~DpeUW@mB-Lh2eRj94
z&$Ii)Btc@V4*}X!DpK^iU_qP#eKvmX5I>IgdVZABm*q~+ulX;EwlD$aQR@Z<t%Idp
z;7sU^$lEjc3~gj+!XOXS`{ijoYC|TDEFR3mtBx+u2e1SWBQsz>VJ@i&_{raj890u7
zPy3$KEgW&f!m6E<gHQ5C6zpvlHbp|G^4xEFBtr~6cp2G`Le6si#nU}Pc=i_sO}=q_
zII+p7_5p=VS1LlPZ(gSEN5tBo0avkwd584vF8tt{@LiCPj4i`Rfe9Ndy*{csU)M8z
z+3W&<&fz6!{E)M@It;&uafu>)>_^;y1mWBt@5zp>IbXB;X_W#!`arSfUcfGvYtqF3
z6n(>oxDO4kQDs(uM?p*&LbpZm(4A1Yed>8$Ow-B4-ahB_!$O<VV-@j-&R88usJ}%F
zf9t-?6DG-kJMpagXf~>_58N3ai%Jjl<%JGY{TF!cMLN)#!N}uJB3|A+(%;q&ghjl!
ze#36Zf|BzYiiJhv*q>lX8(T7`v#vJ1tIZQ>2NN&(OW{IPFIM6+FR+V~rGiVt8+Jjy
zj*%7VkyY^_^4-F9cjMZ#t2=GcbNH~A0T}_Y8=I7i$q98RZZ$VF!rfyaF-xO0nX9L(
zZL_aD&4ukfL?bV)$r5qwj%3cq$7WTDL1g|UO(Y|ngx^*uE9eq0(2Z;}h6Zzs_}9($
z7N@70%NOiS19yOQ<n^gf!T}b@caTo|zBRCG^+m&b4GoR6hVnd;l_UPDwf9!7D=SQ!
zeDc09=LnazJ}H#8O7TUM&26e1aC01HTWwa7;z3%D{7pyRZqnKI?nPw$i55uJY=k(v
zF|Q^osuK>alU0FNf_iVfjtkJw2VO;{L=}t3klr|oP~FtchuiTwN{bFc=mcHQ=<mxQ
zcosC?`V}Jz=*Q-u3Ns!C4g?{lVGc08lq1e<YXIgnTYM(b6x@pJO=d^`c}!7p&kswg
z0Qvr@u^*op%}3+TViwqk<23%^7c)y@XRrhxn^GEEI6nDEyC4;Oaa22lr+w!`*TF=K
zKpknsRl?NjQN{xw1||uMR>pFoy$Rx2;Xv&Wxbhp@`~VZ+c&7`{3gR?d-qH`sjg*Hl
ze(oYSH6PHQtJ5A$0eAvP#H`91mZB+Kmuk9CIo9}Dmw9yvu~GcGTE)^ZcTJo`ye`~s
zmng6vZ1ZNa3t{M&Qa8W%EDSNO6q5}QZra6hF$V*Xw&pJ8GW%Ps98X5zU7%k!XJ_7d
zSw4(p2t`4mf+;@bMmQ*qNS|$Q8AS&TSoEwKk8S4+8jnu{VV%jKhbl!W2adt%&}s_D
z7#REek@HGcMn($u;9vyeq+0|Km9PR{D~YfsWjG8zIMjN3>}3bO8e%3Fh)7rkzN0%y
zQl7b$Li-F&HK>Bao&79cyKbQ^N~eloz`dA_Ls2o_gvY9O^@Ecir2on(KdFV)HAV%Y
z_Nc{otoZnWk)~Gb(ZW9p0fun$$P>8KdArj+=L@?Nv)edF7CO}|`p)@gJ6;w4z>{Eh
z%7yHnfOI{L<%W`uCq#`r5kM-$rD`FG`!<vf-OdbFr)8rUx?Qpj`nZ@-R7#~7Mx7Ru
zw7VduQzyqqQ@vnbNdbx=`LgLoBCJ3;bbTiB(p)K;y~P%+Fh5IvGbT#R7s9RB*rz*=
zQ$;n1k1b*k(kF`yiO2(F_hWz1%kKB9dLvim8L!k<OI`!*`!ms>egLr;zWFmEzs9Wj
zz)7%>oHsRHNLmO;pXM|fyN5@9OZ&~a1q!an>0wcMi7UixPO`SoR@w!2QIw~Gquex2
zFDg}V^J5b;fz9*ejX`N{+e=1${uCy$M|e0k!sS%@>T;^Hum-oYY84L8R_6~nO$VUb
z6TWnRy)H6waM!KJKx2ENZu{~g*}$}66%%cd917qeFN~_GQ-H(^z{;#>-4Co-!o09r
z=s~M0Trem>J`Cs|-ev_xJSRQ;rsedEbcpMizYg;`#<2Ke#8aj<`!djTS!7icsNM1}
zu&qpd3PHGa^B8Um$00D&G$TVk)Gnko4Idj+vuwN{g|FqOXFa@ak6f@Br5ezlt#MMx
zg`mfF{9xllx%LfuvzB7W&u7NmD_NEazB=Pq@&$;}$6?;8$<oF|ceet9Re^+Q?3xH?
z-A6n+_ueUT2>k_A<pisH$sPJ4Pk;v#$Y52a(UL|t_ODkx;F{L|iB>$TU`poGvu6bo
zgXJ!KRZJdCztiQDvY&GDf<zSqsSFf>e)1jxZGvIBcmRAGy3)J)bG3HYC;SF@{rG+5
zF}85eHO<7sY3sSaG12;B(jgzFh6Gl<lD235htXcB7+U|;lWnN#1>DnKS1wO^=P&m#
zUeDzHRpU!29|^NSVO!`;6(-l$N!8QQr)ep81Gn!Kb%y<ug~8yt`<{AI%WyX*zOnX`
znKnViSJ9sJbdBn7_^*zJ?<E}{%MTfpEuY0{HXi+_MUi9&g9MoN%as~@GQ;9$qSxJz
zSQPhne%@4(U*$kJbNzKToGue9p0rBjVO-GmfUZ|3eef|QS#P-DN)^F=X2eovZ|Y0F
z{Npd<UPuWhjsYukk<2%Fpb?B;j1=M%yojLJFo{UmVVk2^;G8vVt}&Jj*3^D0@$mfb
zl?DASo<XlWMSPIx!uZL0JrAyR<3wGbWprLuy-6ijMk{m@wXH}}QVvj0@=<5^6Cs=_
z%~^#S2sq=@01mhB;8xW0Cwz?b-YhT+M{zuj5G`&@7aE{xt1dEIHMK8wrp=p2BBCD)
zLU<c4$^c+BH}!dEFxYPPE4cTADFxphi5u&=N?mqI;~i4-PuGxh{@^A*B@KXMEq0QL
z)hC9NL8v2EM73T-{t>Iljw90B{QQ$?$yVqG-M}wjQ398pILP}Np8IQEL6bzS2nPnP
zmV;ZrP3i)eHIEP$1D9B47igxi9l1YOFmhhR`!$zniBF@!tV9wa6EqT6cGLl^L$;&v
zVFl&Y(^A@AE2S#;tfkL^qESCBh=Nb%#UB99g~9kCw=Ddkv`hunQ-PQyI$$m^O>ZyY
zz`1u(spbXG?|7X;m~8d|LKYzC3vTXChs+BZ%D$z#P`A<vx=}(=6(Hh;5%*^RRDH#p
zw0n`W%+JziZ^0Wd0y%OIPLpbCN^k6&pPtq^R^?lM?ll^pex>>hWKE9Lu@qNSU9q=H
z1CH4U(S8tgNLG;(2ge~D03_raX1&S5ajn}Q!IkE)_VpF#xzQJ&m^j3a!jQD`2^(@$
zojhvG2LzR$5cTNmU?^}{M#czRh7TdHtC-1THTV=ZXN(rZFmqT)d0dV5AlXYW6yHxw
z(CxfFkh#-zX<*?G2(0%szH{Vn{7yjVyxKaBo2I#+vtFtFco0h-tAq$bHPSfllJ?6s
zfZlZ$IB7s49;qc*R1iH+=+v5b3k6uMK761f%;ecD?;9zTO+JlG!w7dmM|H!Wb;;~r
z;?UZE$?vGJXj+LQMpeTG7+j3ufIt<0{rYt$|ADP15$?F6hWsnGIkOs8mcRl?d!FY5
zwB{cWf4GaznK`qqNinLK>n@3Xyi&EEesEVL_a>+ChrO9AA|Cn~v2gT8C>=NURWW5r
zoz+7y2}#vm_tY~#3k|k89GvKHTXZPWQV5=Nn=cpMn0E8NSJc(r(g+UYt1yt}EfOL)
zKS4u-au*d0f$4<L<GcQeU&~X}v<z2H${EDtJ!6ur_-ao+35$S(B-9%rUz)K*X?UPz
zL6pF>esQ|oXy9^Gq)PcYW1>-MLyycGSHn)?tGke<8+{S$>!e9_5YZsP<24!;D``X%
zN((_Qlpqj%w3~tt{%L{m;mgR-X<D7RUs}^q8e202Sa5>{>MI580v&GvNfBthO0~?T
zvc?bwCIRV5?kmi!7Z{AU*{xbY4HzTV8A*-<;h%_kRa_0~otp}1^i-=PH>FLHTr`o}
z>_^p;%U|2}gACb}KLd??>*r75;y%1QBxWRm_!n@XTc8K}ool&Pu2pQ<pC(@1=pW+_
zUZ~s<)dV4&a$n@t&6-}r1cM@|I8wVe#8LrW$vsi=moU6<4iA#+sH@GwPS9yv)9V-o
zG^NSUrhpc#)ErP3hGSr5wohor$Iq%~<+f_e?igcPIiUy=V|Lb7BiJo$0r&PT-Xvvb
z8j0}lr;g>DT<<g(B`ITW7&q8kYVfV6LbRrL4!F{7&GYxuJ>g3GSI&G)93ZB*Ph?0L
z?<!d>tW}8xVPHZ1T#Tuwr8vuG<i4v{RRoQ;qOw^|ApK{umJV<r5E<N44x&bYcW0>U
z$7GmoW-u$5w#sjPUUzpF6<ZaRw`AU=kU8#uo3$JO6k2r&{l1Mjg^Djc)g*199cSGZ
zAf<}7oA@T5R~_CcfvKg|86Tj3VESYZS4>u9Wa~pWc31i9zHQX%J@ClU^4yhj|6*_}
zJ)h9u#Hz>|pQb%r7B2EMfB!gQf?{!`ieIJH<XP?u;5_<>kh4@qS0pH|irYIjPhIeN
zaahki7VA4c5s*fAYqw4xLkLTWUJ3-O62Qt*VhUBiMipMcTeie9(t5UY_KrpiCHon&
zBqLsi7L4eU`C$dnQpYt4d|{;+dMLq^#2Qw2zu@W{R#7^|K)GMPx&${v65k<32uhHA
zz;DYT2qXtsPmBQu=NBFn(`;C8zs_x=RhwIQe<z}3l)nx)<J|Nvwner&?p;IC><^eE
zidiUu;1xQg*@M@4p`tR+0o`QV?10O3-LF0_$ZnJo!N60Nkl3B{!-m?b_c=c7`9hQ1
z(WU=3E|^T}ez|465H@p{0z@`{;&VNLf#vR$MwzDo4~Rkaso{2`?15Vsq0O#*_qxZi
z<o@Rj{V6pT$xSF$s~bHR&h<&LD*KcmGW6!3JE&0LS4zE(*DPx8c^_RnB0u;+<epbY
z#8USAnKz`%@J|IgOFafRIfm(J*6UdwGFHmVlIvZ3Un$kQ=b<CEPAW>DH!i~``ko=i
z&jT-)1VxxVU-J5_s8Q~zEqtZ&#<)Oj+GGA!d0lxg|N3jsy6pTXDwh62v*wMh`kyx6
z1$Hc~rLW%fdqs^f>9U6gv)H@?h(s06kM!4O>Z4`%IJ+^AbEh5j;S>raZqx7A4Pk6V
zKE8o;J0n=M*v6NA#UE$+FLbvsbY3?cwZV5n52o@;f;&1ofHwoRuO3gS6SohVq}KPy
z91X&%g1<0pyDRRND(QGRt(IX1h96kFDG@~oo}ooD*$JWhQLOx+3LDB0p!Z*_kw6t5
zX85$igRsgBkU6OUYe|jbyKddBR)OAvlHPGtf<%HEBeM^Xe*ItmsGDLtT4V)3j{6Xw
z1m*+z18vq^)9m%4uJw(FZphSaY<H#PL9EpJ5aWEp-4E(cppE$8|6(oxk5C@@;wZ2r
zhzNs`z_v)M(Ztjd#DoY*Byk9UcB#+em_qA8dV8KLf?aKMXEBYB=$Gmro1*rX0FZ_%
z59yzSKA`P$XTdcM?qmWM6p)^lvx&&t>lvsRZ?Dgz_r8aS?|P6JeoKs>CWrZ_a%sOC
z8SVj&1$|{@XIfquFNDjF756jKrzrBw1RDDfdpn}e40mq!=^hD4*=M2S@M%SAx1orf
zA%TG$iEbP>7nL$Y{LeKDIsa1$O93utVlX%LRG2kSKQx$hHs~WD1ZrT66Mt?Hui!ms
zg!1bD_HG!E{_>(&9nQIewt&=fJ#E;!`PxRLK=W`;R@4AV7qE|kh{S;qO310HDcHIm
zp$<(C0Fx1Vs+h2RqG#EnWzmpk>o1hBRGnTb&7v`O-b{%V^Alu8Vvg#~MCxf`9iOH+
zQAUT38=E!lHy0^KG?_3oyo5*m^h`_#&>IwsOPp`}OcqAe$?!LWN&}xblL1Dw*ft!M
z#1;VaD@gLz+PVWm3qaF#P2J#6_9dPNgw!0<RZ<Q~d<cgkhQ|@d!0Yj)pp5rxwFa|X
zdb&yOq&5&#oeREwh1bRDbgeEY4#<QgC;WGUCNcy(tp*GLYC;JDYpD49Q34QbQn451
z)wJgc!pqfgh`R-7FSc!+Z8o`}t5y1IvGDRWWWy<=l<c5qUroO*qju|xJv2mdtOazG
z?zk9(xR1<JOkvKFx^q8NKL#3-Gcz-TS@`1#^(T>DxBIqk=IY*YN)R)3DvBGN?Y#Z@
z$9vw6ROAT_1k%$YPvK=?>!7kyHm@4vDSRBuek1PL%z~V{=)>>PNDdblWBVKB<th8#
zSYZn}jTy>-Z~!8UUfM0Nu-r12aNq|KRUSk9NE!^D2Brqr6B1nRfTMx-Y~?Do&(5yV
zX1L@Gk<alo=cF)ZuV;M})}oPpBgP<Bdi=0<F1zR%$SW|gX81y8FcpbcyNT;?dZVqa
z0q^4A57MXXu$_zNxm{%d%F+YLsv5hyyXYvYDt6?F4P4U0_oM9k!{(a=bQSkX%N1N)
z5=B4<HIsB-On9>CJLLWp<=CC|oNJwWj&;q>9zbuOirlk;0XC(isIIB`@``Pk;yJke
z@^FkBZl9|%y6W7n`=NU$FUQ5+E#Q?;apuVU6d`8^l-#SL0pYHl0nI0%76%}Ooylp8
z7ByK;wjT+k&9b875p2$VbG9w|uv{A1l9?{Vkcji09LQxAABW+5)p>+D(dp6w-1X7z
zZWSJr4-5MQytk{OZEJOkDjT`nA`Q>NCvmH)HoEc(nlY}VUBZ6DfuGC@l!d^?t}a3I
zArdx=u#p%D28c^y#3HYmTE#nSE?_B`#+vq{Z;XyN21M0xL+>B_y5s=g?#xq%x$u$5
zT?k#(97Wt337~d;zL91{qb`f%xiI8GaR+h?Gav`y@poPM>ZzgvEvNUlvMpqK*C)>3
zgo$~+w->yZ3p(>!6^r?lL2GF&10TNdVTva9DR_zmWt5#PL%W*|r5HISa2jzrAHju_
zR-DJG<AHO-f{)g~3QGv*x9J?z4qN0o{baJa!u1p2X*4#r45#(bGcqzoP;6|9UYyTt
zkRozw<h8G7?LFb?DdVt`d<!a()jT9vhsUSLd&MraKQ}yVRmNDF^3;okNjs&9c-rNs
zxa-`eq0f*JL|IlhfPOhaGs3<R$308E{;FWn#p+%uV>DQzi1o5q83XWiu569i3*U(B
zS65l|M7~1CT&Hu+G2Uio2u<@i-!Q+J2~BF|E>}>MGCrp^xDm-^>s7BiR?a{0&%5dC
z)_-7jW-NvUWtGysvap4ouXR)0l8kuc;QvvwfeRW`E_PldzX6xpyFw#m<ngSOedT)N
z<{_mQ)1{vJ)n*u1)j3;go49Gbrqn>x6Ob2&AP6?A5l{+Zqc)XF)D@ilU$TZ`&1IUn
z^`A_&EYY=O35N!R_E=!Xyh*PdIhbv4Kk9Oa#0jy?+OfY0!Max!q~k+<WAjFWvGDtH
z%ZNU6?QQmtD`hd>K5gg6i_*{y_#|5F)b0-&Q~Et;B+?^>r*))7IoYAHKH%oB5{l2)
zB(Abc5cpC2HVvMw;Bti79+}aUo*@n>alj5`aQUR8W?Y-e$Q~IRI&Zc6Cg#T#ep%i|
z#84rUnIj@W6^;@%1hk6w0PT1aSwCAqk1HQ|H8q%bGmiff91FOZ)q0mca2s7&+^r3s
zAB8WFQRxX+dx?03@FLeBL$BgtTv+&2$AqS3Hc$`T&q<MQcaKbZ)yTPb1&&^#skbjw
zgNx?p+C~7$uOBlkyU^<U9p@k_qsOzWXT;??Na2*V&0v82w7L(mx~F+I*$n8tOx`ip
zt#k)ms&}11lX~X$r=bP@^G$)9v)6km4FG#FdIE~*G+C81)zf6KKjffJ3SNfTAmXY(
zM9@c)m8;DCJqvDQACC(~Z5Q)nUdOvU{*+PA-uwLEZ9RRN1f(lx@<Ez&i|<GM*KhM^
z!_xr~4F+!0xI+ue>8h(qd|O-vP-Y=L9{=?$E<YlA({B$$x<d+l^JsGMBI0j`jXD12
zcxmEWcq(J}3CPP%9;K_UdhMiWNXXNQH&9O=6j9~SQF>q!c1W!HR1}?c>~i$jocF|5
zR1<uDVTVLS7bvy_v2g>^fnoPAEVGhMOhPjBGab|Tk=sQNd6h1a_`~%OulAlU>4rZ&
zLiHDwU6-wm_d0r8=+&Yj;?}Ct!x5%|fjmW|P(#^s$l%)v^bc%IoQIUb*rF)%6eC=g
zZ9J0s*tcv!jQ3=2VZ1bYZ}yX=H7>)n5j-WazasNM69u<hp9CV(yLw+)NyJnU2{TjL
zkOM4$O5L-~C4fUPk!5%xcm}UyjDI4JYB;!O?b3mFY7;l!U{FFfo0{w!rz8>lG>EIw
z$G#`BJl<!Px<uR;1s6<$LZL_?23n@`d}MzY0SI29j&~Vbm@3#+IeAENZ5J%U0i*@Q
zgpzCCt;8QEVCdW1KG-G%VFw3H@r!PM?Q#$sW+=%_pvHeKU!KvuG6{+%FZ6e5kCKc0
zf<0V%NvC7l<ss59m6ocmf)h@vMeP|@4u*7PfV?{OBb59;4tO*T83ibRgey17Ww%d2
zCO$#&aLfi3XKYRe$GDapnfb!s>rO+pRuNwExf&VKu<pk+qMm2NgT%Zp22Idlu{~sd
z^ps!@NI4;+p?Au=GVOvZ&2TziW*ohQ5-@Rb^An@!5wuw^l@1-yu4NSrI^p5_l0tkq
zyS?ZtlzpWE_(nGblr1vR*NCO?$+(aeJ)aQ0)^2$G|0sLwxGKA)e|V#UfrJ9mT~Y$l
zAfeJJ(jcXDBi-H7-O}CNC0)|p-QD#r#B=Mp&-2IoTOY+O`?^-nnwd58omq7pRTvpR
z2RO&dja|;0Cl=_q0~^UmZp_&_{EX?XPKM#uw5N-B-=Y!d%N{hQwyPMmB(L<ZH|E15
zJ<GWk6jyR}hZWjiwWaB7>SDdeB|N>E{(A4BxU@(T+a?bkKI(#(4@Zt{gT*&IZ4|t%
zayIhrn1it~OY%(+<+t*lyNbEi<+HxQNJe(9;8;bHKKC32d_YA&-GB6U@&0}#y8gw_
zgg@QkOTsYV-%{p!jp%15mbXXD*Q$o`;fdN_I`zQSyDLP)!?{rd--Iv^4-bW)P&~!U
z(^Gz<=rxL06LonhMDc>H?(iwgh}>1f$dFCK_-Fc04<;6fybptSSK+J#j&r9yLaWYp
zfu03~7eEBc61m2_Ie$y1ehYT-zqu4hua<u0LXhsfpq7F4dDdx_lt3J@*+Zh?!IQIn
zf(@5d)X{}U379c1Qt8%GN=cIaDiT6#DE)`_>V$6;7+yc)QT1LW{R~A+u#scyQnWnE
z2q3{}k&%gG4s=u-Qf<Sm)z8lu#fh=`h1v;)BsV)wRHp(^%^z_P5O7eq+(%7Br1fy~
zjGL_B03P0Cto|lWzB2`Dp$vFyr&ubqrne2Tbe|vfZ8nrfC^q*;orwfKu@MFKGL6*m
zG+-@);0?_J`s326zda9rY&xQLB$_)*r(?)=l`DSf&IMa#v))BXd^orjhB09^f3*92
zDAo|M<dE?Z$B_K)m|SemkdqD##eKZ@#&3Ao=bZdEXm(myueCI_UhuGqOT3A`?>L$r
zEwn%<{)){aHf7$*L|ahDTe*^B{@#A=V+(2|=of(=j=MLqI2jl`0j-Nri7u&CPay(9
zCn-zZed;=t)DedIP*#?$snn>v<gXBV%2_X)Ti}F77n-4Ay_|~<VOOZUM{{kl`EaL6
zy{jKVh=ifU#63RkJUZ_ON;*V<`>o_Itp1kZ18{ou(nEkeCp7ClOUEW1N9q1LbYv^W
zw>lXZ-4@oR<5;Y+3{wPlq6|=}o(3rw<So1WMEtRgGWX?a%7lqkON$q!?{^<naU?;(
zPAyNkgd*m9#^qud(lW~-c{@f~^dx!Z<(pQTLnhE?XY2L|d|0pKK9$x{(S#Hwt2@Y%
z!@n)o7LkZ*pQ6s3%Y$$uuazG6P3_3s!u)RJtg(^I^iYM)y&6_fXm=b%9B*mBGg3ty
z6${ndepgfoX<X#9P(~1&r>Dto;w&jaPV!S|@m)93x1QoLKEPZs6^xcR!)Cwdx3Q$o
zV3;xe_y!NSLp2h)PsbzIP&(-!v-c|Dam0K+BGX0nTHt?56ncB6gpXrqU}2d&JqY5!
zdl8fvnTnrN9x{lyUSuIp_tmW?WsH7etjl!5a%;3|u8;1TN4Yza6PJ(F>kKO_S<PtG
z&QHUcZdb^g_JF~03kInC%ajTSmWR$7?G<h*n<veajUR&#%zW#<{8+dU*re2d;u7!q
zOS*m}c)gnAtDhkpYt!9TDj1j_B#g2EAMoSqAs*Dn0}BzsNCjC|I=ZHtqr&N`$J+>y
zdP;zS{sn~At_cfZ8TV3?L#yE)a&NX-n2b#?mbSpC$=Ni~{Zw@i@^~nTS;Guz2;;?$
zpNR!>2oev$F&ywyZhT}-L;0Gpu#ZX>wxv$;fG8~B8HE2{;QlgI$#YR-hxwgSdOh-L
zLU-Wc<;Udn%zW2N;Qr|;ks$$9KNIA&1hu&Os@#tNp9SM>Kd1D5CQp95?U&<t2#&=e
zKO;a6{-2J(S@hBy-&5XY@twtZp=}aXun)ZQ&vK;D$HY<$&F`~Uo*62W(9zC@jZ;vB
zQX{SNYduhvT(m$DYRC61&Cw>OE4vW+=1Y_-6c9O2TjALqJQ7o(Pv#6o=|ynm8ai_E
zA^GP6m$QP+Bj7&d*`bI1{HVIYm@?$8psMG55<>5+iZ(nj3&@Q=AQd~`<-8=we&OLp
z;=+QB#p!YH8BTIwLjwzlsXB+=7cDxwnVay)p}vp{2qL*B<i<W+&m!v#qrd?0?yH)l
zl0Z!}Q{qt@!_tcm4%u#S>wr;u*6f+_gyakAy_$Kfs6B#B9!}WnfFVkRg#b+>{!p!9
zUh|cUbZ@ISI0PNkuNaFqv!Cygj;PTH#>mE2;$FO^K8QHOs4xB2AgOR>pfYmeJs5MT
zr!9eIS-v*oHC4qN_qjd8%k52VtEKy~-i5Ky8!`rx%Og*G-2!U(fb(0L@NNTOz_`|;
zQ(jc!<6DuLK6@T6%?Grw`@ou+j2=gdZ~02ulUebdWX!Rc4~@?Yt9UD=Iy^`+?wK%j
z7GFMZXU^6pGyeGlC(l#&X_8K#`&{et;UZtA>la#s1Me}D;C9AQW(-oLsb4+69{*-K
z>5X-qP)}X*0S(eWeK|SA&E)C1+A_1v^JKQ1>6+`-Xil2TK2QXlM4rx5G|w;+KQQ;c
zQ-$!my1fw*n2}E<2#o`<jb?o>v2f=W2}(FP$UH%q2}jGb0a~OZ|93+*iVY@wzE?e`
z=C>ifc-FUkCkexOl<89fRv^g~znA;ueJSF~`>XB>%4!y&<hhTh`6x|;U%B9`^xZmd
z@{GbN?pYM}*V@TNc+CxytQ{*iGf3^fa&u24Yb7S92!K<jC!_!HTFM=-MOQ@PTjZ3O
zITN#^WMKAKTt-WOvMHt}O_UgCV6Q*J3uF82Hy@omfwlQX-uLh{F7ACT7?;<H$|~B(
zpTBDaE*a$C1D-_^Y3EqwMf!evdT;3H<)W8}bokGo8Dj}GsEiTlF*K2W&n(R(cX5%@
z(r89uH{MO`BkSinNfzlYb+DvJoF#m|6f1ycWec;mbD(E~v7^^^BJfnG07IZ_mZrDG
zWqLEvs_N}vzl8*mu9)cFY1}u|FVl$g-Vn!zUS<f_llr5)S)20fl$W<>xW_|+#R}F0
z^oV8{$kP<KpSmzZh0#Tv-#EDFiYbXf@CsAim4ucsz5E2lT(P-H;hvOH_^I#`muxbR
zlPT0jrw^A`e|^pC+6W0y3%#jFu)fb((m41w@ji?jLfdlqeg!WvL-;jcCN3<87uNAx
z88vNOlM(H3D_b68!&T2C2-jE2vD0W@OsbIhvT_=<`nSPxLRqj(Xvno2EY$2h6pZBP
z4O@HYiTr2Sz$2Kmm&uP=dr2>%q~0s+>fW{?r$jR&&j(dsxaDgX)?tYMRY2?kY2d{7
z-DvN+(r<z~#+9tL2`QKtrI@I=0sgO-Q04WWLV}vdMh!D;8N~e$xkq_TJHF>D7$d(X
z&rnwo^VrO*_H00JT@te#yCOk3`e~e~*n&dVi_P=$X@FeuHi(^ocmc=aEJPc>?z6J5
z>B(%u`<$jsi&%9h4;&eZja^r16r`pcr$9%^%S)5~M^!=nvx)CLq8`$R#h__%AUDyJ
z0o+*wOxuEv&hVOIzW@Dhk%VDy-QzVlXe{)BTwYq+z&BqOmaWMqYH>!M<#Y$oGv;Y~
z2kzL*SGnuuLG!Z!W7UcQM|Jix)f%o1@z|AMu_v9-R;+ag<NmEp3yEZmA_~_j@~oh*
zgyI|;od3gX@Txh#6>o1kX^1||a`Kd&bPsWB*prKKu~?*B#;OJuv1@H}{}ge4DS1{Z
ze02xJRsmY9k1{L|dzTT4exTp?E<w6AjDBUO?_{$iQziNku^@~0$Sct{6!>6lePwwc
zugsRvaWMNad#s`on2$c~X+T<0xq9pne$*XUBr9)G8ob)9nLpw?z?3-J1SP4*vI^zs
zbO=Y$jwXHJzNrX!K8TkFuwU|Df(hca5)!Y6dW?ah|KXcj>1*!7A**@cSdDA4p38&6
zGNtgE>G|A=DWuC4gV#FcGJ$mst!LXFkN_<+Z^8)GA3c8Q{wmx*45~47+RF&v?I9JE
zz(5{L$B|y3$J0L@Vm_TIBsDppWGZ>Zb|@R6<3|InZ}iceUXDO$OqNMTy4wMPS4T?R
zUnv0C2_hPUe$p<l@1mYVvs50AWc3#fDZe&}CUvW(rV7W2zuC}Ao=<8D?*F20)B+e=
z|3bT1@0-`kz5SQ`8Mu%DqoaK`f{ST-@^#6}t6<%dlzYv}3X0DJ>&_}Jg?qiJq!sQv
znu_N!GotZc=oKBrh$SZlOr*8hgvA<4$FhyTNJxAY!{$6D)RpZ3VwfHQKcADD%Toq-
zyfDf3R#HZWGWe%r{7DrgB<>I8)$d<)#O$lu6+UyuK~?xDZ)Ewsq=Ibtf;zOgxU&Th
z9$Uw{KyHjP@$qmoYBanX2JBP{C6NP9V0G3$YcfV6R&Lnm_lOJSsTn+_EAZGKv)6w7
z$hXadb{uT_e6K5nUKGnSSDS8)w6xbhbo{C#_hk9n-;m|&iwzz_egIL7RNNIgqlglR
zr_=@LbF9`q69DQ35Zss|J&~6QgtlZHKF#m7mOdFJJd7bcm5l99Xlp~{AqZslAmm#R
z%-(w-S&A(ybH5!UTP9|X?5(((IJO?F5qw;y5~An5z`VycfmCy`(y6{LH-c+WF20TZ
zByBDf4~fLvR9s66+~L#<S24!icTi(Jg`9C9V!Pgh)Mr1x<9UFzJ_4f^I2Aqvf1!ww
z@~}Fyqc!^ytAkaZH&Mzji9JZ*DN`9xYPQs>l`12GB$F*ox+Ae2b(p0;r`5?}98A*k
zay>S1X=L*wSI{jU$AAz;5%hl6n=MwVxHT;Fo?JS}8CIlpgzi_Y)2;Xx(!mdlwQo1Y
zP$0mJol|?``SULPM6_ZLx#|r)=YhftefgD^_nJ9FW_u3zLd^?pMyaw*o{nINN|6k<
zhk3#P>)7~e(`s$4F|@bCg>D(e&L-h-(flOcKFlLCu6iDBm=zA<^}VBcG)-n(I9%-b
zo3gBGNdlp_=DQ<wmqJCUj(mT@N@6Iyt&Fd<reeG&nL(81a{B2qd%~l`SrUGp=~*Ag
zBC7TB)+Ylm%HQ%#ScjnCGh=9=GAa;#xrH<)WVi7^`o#n~uL~1qEwcf8OW7$7;Wu2C
zDc-`hp*_vjY%|kpOOX>Y7OB#s&6p2-{~l+G*hg2#`%~^JpYu|9aq&(Xp%aPcn>c)b
zr5jlnb%qQ1VuNd`k5UqEkkOJ?0geJWLyYhXl4PO@QbE3wH1HBR!j&DC`#pVw%6Lhe
zFkg{M##68oB_{#lCkE_Js0<MyKIvsmTwnuK*svmbyE^p;Xd`|W`0VFH1G^o#{z{Li
zI`>JK{kZE6qY4RQ=FQ~V>kBDOU7Yhw;CmV*B7i+TA~K8GEEI(izVp#BpBj~*zza(E
z_a(ysum}zasPdv&!OvzsnhmL)bK$*k^$eLK+vFkBeQ&LJ{zHHwl#tNj1)!tnlr7cx
zKTvMHf5|w%e13KSy%_LlEAPOpDrMlz*`lm!4NkKv^Mg-SPZk=|WMmf&EU6fKAx>5F
z%j_m@5gVHj#o?FtA#1s5UU~2-ijQiONX@r1HpEfZnZ>sxu;4{`13mlN_V~yoeP*)^
zcDD8VAMKBhO=pz9>o8frhH$NqS<|1?$djcs+T_{?RCbhQKDmwRv3@UhHd(3_L^)iR
zQ!dT<!j?pMXavCk-q2_D5U1?Trprz1*XD^(iWLN<=)g>L6Q&b3hKr<3kLA3lhQI|K
z$ZrMRYfy<UoME{;m&NqrM#oGwE^{;q81Ff%@!7j=%!T0P81q&qU^%Pz@OC#ZHPc{g
zOQe!;KxmrSjpU05g#Z%ieeLN}ip&mN%|rDA?M0las8LN9Skk@VxGBFw5BV__xMdqk
zaX2#UnWd{NQp37CFDx?@GM|?-O*O1AevQ0y{${XW(&IolNg*lDNhf^XKPP)Dtce4H
z=X=O{COF2TqAVtJ)MIVN%xfr<vLDl%@iK@qM$Xn>Z1`hq=`?SSwI789ou>tP6GtXX
zEuZmrdcR%uOvbZEAGfwY#In79as_F7GshGM4T6$Qh^p>G_VBCZQ2PQD%%U@c!Ysyh
zJG_Tr+mXYyC#8I-##5`b06L=78>P=w=2cyC%2w?0mHQOzv9W_q^z!O^0>)3><w5g>
zQCA5zQNF$kYii_E(3H8lQSyTF-;Te{M@7A^0jO~Xe)I>K4)n0wSO)e}OJY~IaM~%P
znt_j#Fg91f9dsKuS>3+X!eR-$NZ`-0@DWMiYiT;f+y=KN5Br&327r^sqAHS;o0?e4
z<8+N}G)~Pi3}Cc-i^U6Ik=QIgEF419-sG{|)47D2w6~T!zQh(VD(ZZZ*38d0d@;0`
z1&+vgjl(aiW~2K#kG-dz|4V&nE@$DV5qg;ZjT2mMFbYRBdY9oRK=_v+etk>b4|?>L
zy3eYsY%q=mUF>ZzkB?shPjjz1Nv7dAt1qjt2-UXx6#M$Cy14JPP+Xsf)X5M4Sp(@+
z<sw#PQEWke`*fyHX2N)ICTzTO-ubE*Hg^BPcI*}@2$0WI7<jS<aU7Jmt~o2zY#gkR
zw>WU>Y*+zM9pe!Ma*gnwnPlM~G&{5MRub7cE%(F49x5|WagNGva4->4hg{XTWXcbh
zk=?|_qMMG4AYg4G9LZ(jkTX|ubbtu==eKw>tJSf^F&XL_hh?7&r?kl{A*7v~(_oc%
z51;h`q*;0p<qsYO@3=5_fRbp(<gDFm74m<lFT#NV!nT=|YOEBbH`g*RdDb(-?32t@
zSI2zYi}I9Mr>>i?xW-}lFpelMdxV4%*BC(h4ZFboyqsq;d`Hpu8`AFgFtLTKVg6Vd
z1{Z|!<n8BC0g%s7wBkk2gZOxKquMcXV%Rl2g8BHbd_!oEWPj@;miRb7{q2Q@wAQi;
zHu51&)x=xxRmz(uk8Aw|Eo97AADr#!Z~JxNn0-kj7<%uVSD2=&uv&ezr(+?bF6L25
zkP`s!_^m~Km=oei9})^e^MCkUzEZ&`%jn3{(#}f9z*#kMpqbyZbfyL&->B?6*_kwz
z?<MVdgeFX+ZqY&gLCwwT4Jq#WYj*YmLKAFMNKN-~l~1|)-VQgd2*uNFe&!MoQm`2U
zZj5aN_mC{vLtAYlY>!SYEl*FaI~2002nKsLI8BWni?~%eHXDmb2-OZ}zCaCUIgtuo
zTtO(qm;`aqLB8rwfW^W$#27|T$3TM=K0?{TN&p0^^)v`MJ^P3*?eg}s+q0gxiTp7x
zp7+j>KW_8lsT1yoEELIBfjx6dDZX2PDiB%nmQ@bzD`4UN%uvytz$$b3ZfEK?2Cc$G
zaR~8N?Ug_F1_<KshW1X4`68$V6T-k=C$XYLXCh|Pea%JURZbc`CWbeq$`qg?5z%1p
z5J`<S(gWB#1furzmci$J1}I<t)Z>vN40DQ>t0y39fQ7%gHb+*G6tE(z6JL5U&4&%_
zor~IA_>aLO?HMvpEDWJQ_zUAwv$6Sx`1042V#ia?vV(o23{5F0=4?(}mrUsycKO&s
z!^xl<M&#@y%M5L$W4E0iLq2v5C+XgNun~!qrlm8_sXdM|Upv#CAZM5<pQOxl?cG{`
z(z_-*Gt53Fn-!5u{!Nih#%3!<bY?kRMi0*6=)1`Qj(V;8EH-YnG-vvNX5!gW&wcA1
zD6S{Z$?|5aN*q|muJBGfT(YJB9&MPc)_TaIY*DMGVT$trw@`3j>9l)2KgKTw)1hQN
z_YLX-NKC*Pd-<Wxq*J%9#Mp&59##J=4YkEXtpE?fApT}^@oYEfBfD;V`+y+~Bxs}U
zI4v`@0M&WdDm3eT75$Cm&Wmm)n2^w2I|fy{#WE(13Rc^FnFp$BnMr~t56EZAY8ulw
zAOX*;W925S2Ul##Z4Uk3$)#m7%rNI>yY9@On<IZ^+-Jx3HsNrdP4W}GI_SQN%0Qnx
z3cx#kx{lA0RV#BM;`lVhm>PtA^HYirR?;y7shlu3j5_xHXW`XO@ehuc{mro=?!!h{
z>K~sZ0IZ(um7k>z8w8s15vGmkrvg`KIwWPe0`3;K)n@a2nWyKMz~FW`N^AHy`u-*e
z(b@et%Vu=Sr0Ta&bQhpTESN8)5%?l=)t5`8A!w6qW3(`;EcrS*GMOE4Wh;-R<riUs
z&D=ZmKG5PRiu?-i1!=4<SHwI$p*>xPyBmab@s;0Nc`yi|W6e<z5WA?M+75h&Ig8>g
zyP4!D6}PSi81$s{1|lpI`pB8!0RqvNq7Q4n-NXf}R=uUvn?#PZ>!Lk9!cRM=g3*Y0
zGdpQqBfQ)wLSKjQqntT8s*7Wm>y!FrIF_denm@~CG0~Wv2FNhnAJ#om7c?)!U_D@u
z$6m2vuKPpTgVu8>AKx+QElyU$&YQi;w{lV@m8YArtX}v5%SvWL*YPrq1`Lmj2E5jj
zQvDuc6ktG+++zk?+=JESeO*B|Ol&?v8}iM`XsN@D{hO~W{Ddx<pPuTE7hrsgguoLJ
ztmEoo(f7URyZo9aItsXKgY`*yMbh)S#xdick^=mgHy7QC>NeA&=G1uG<$dN>ac(%>
zvc6n!zZyvr8WFoO-7jrX(CGw_0NsJ-S0}i>rOIqJ*G~3MD(odQ<C)kP_q0`4SA{sW
z*2O?2*Oi%{Qm@DRJ%=9AJgZL-?@G@}j5;Sw6?6yjkYBhnV;!EHVMor6Y=tt2+;888
z{-G?R?w^qqz?;qM=mmlI{miSczC|zCJ(}7xOOu{1PD=lv-zf{bPAKs(EhwI}3fBC5
zeAhpC3K|0G#A@FL%z>Mc_U9$RK&C@*g;&9|<L{O{Z$DQ8fCBM6`+E1xk7?fj7P4KR
zK0J;dU;`1iT=(2uR!%ABnF)@mJx!RGmkMQ{9CIqH16^Pr4c2QlRyxH{>o|Vsd&SY{
zMvH`>96H&`I~{pzc3T4zz`FEAyxPEl`n*-08rX@RRbh*Eu!~<MsU-60Y;&rFkDe?d
zyT%S@ew&0#bP-VvI-408{bc;SdB++P6#zv+@q9E=ho{IQRvp-ki8=>-Y=fTB>ph;>
zBperT$%c?F7_%B10^Ekf0VYc3iZuNB1P{;NaRBpjG?kq0t1+*g>t*J2$gw|nM*$JD
zQ@XGO*kVtf-U83K!c7jm+j3As2<`6U>3g|}sI%!Dgl&Y&)@l#IPaVR+!u7c`fD0;)
z(qu4E6+k9~TJ?z9x7;jlA~(egq`Y?UgTGq&Fe7synOs*I*Dp3hlb2YQTJUmSS>aN`
zZg_wRf)_h<vh&ThhmX^)sai43oRODfrAhM$Gu03=FvA*j)awuqjFCSmTg?Len~nAr
zK{8Zw(W8!**67EQDuR!+Ry6|TJGI4u>Ke{b%SU~C%IG=0+BX39XKN~U2n)QL1mwwO
z{4eP69dv<53P30lmGR^rxZ9GUbf0Be9~5_=#Pew6ONGfa>wK_#1;9MR3bfh1b5nk!
z52eolii5wxI5~B|^ad(RZ!A~>%6wc_t;+zBP45ZaBB#MzG><hlr<O6w-bK*f9!*R0
z9g0vG;0O8He))pRF{_<f>!MeRtTb&ph~?1CB35EOry&s{I%3=S>&yjBX~tY_^Xt}|
zK259II(X~TW3Swq)NMdg8l#Pg%3+Bcm4mYk^d*00PR<jXn~$|aH1JJ23J-6;S#qhX
zW8_G(_2_*yTD;G-WKCD1$G$o#XnI~>f15Q*+-s%ifKdu7zX<rmpUx*zWtAC_VX@Pl
zukkt5*)6=CT^fQ7N3~xX3lDp-JcjZG|KoI!FDLZ`v(<^eq5mX?ozaE+;c(+p^tGzG
zz#JW>a4GSt@qFtI!zVk+l^5}L4iz>F4f+{zLpS+_nfY2EiA6Lx#y<Ygr|{l1HOpFX
zAulJ<?TD{*bhtlo%a<D{Pie*1`5-Wl?+~Vbvv^}rG!bNz8eg$px7qS3SOs@;W6vzV
zfX+!mT=|ArlaQG<o@24y*28f>8J89WcS0bd$iY_hwPw1z%Z4-d7q`*-&9(9)R}*}#
zn{7QX5PozdV0dfTbCdj3>YywoU~d#M_Y0q4F!Et`Yd-3jG3S|gSS-6un5UBE95m+o
zlYPrK0Ak1}{iw*PbpfAXljU)M`N0=Ck0T9N4xAc1cBSi%i~a@RP>_m5-d}}tm!eRX
zD@7LC{;A;olquw`HMEWO#VBt#N}$E`n^I#SB>>sg)@B+^oVK^v+yolNua<fVFfXI|
zl#`}Co3H@TJa*Dxa~HJcS~86$T2fnz_U`dmQr=|*=9Zc8h+RVtb1HZ;bw(^BJW-9_
zaIAdsZ>?D7@<E#%bdf5zF?A|}uX7!UoAY!kZ#h@Q&>dqJC9Xn+(AZqfe3`hmyVz4Z
zxd+k6ySZA<Eo3(&{q+?bw%yu8kt@I_Azey!26@@q82skcAn@~Ri40K{?Am%u+4YQ?
zMT2M-=b1y$FQ!-+G`OA9AY;E<)B}r35dN*PvK-h+1GvvndR~tt;2-~>Gfmg#Pg4PK
z^EEb*206$8>@J|Tjl>Y>zS(<Cb{jbx!I`HC!eqs}D#OFPM+2rG3_j=;3}qA@#P)k<
zy~T+b`!rhFH<pnVy%h7uz39fLrJ(^6L6l%B2$tcR%K?tCKKs_14UztLo;t@MkfbWP
zQlSjS;R2)-QLI43NRW^p|8Qi;c|NiqN)bFp3u(h^e>7V@bi5_S`K|XS$SwIzs(-)7
zUWWs@x;H#*zVPZoUXbPz5t0~^dhi8qxtFgjWIFqzINC947Cvq5`hF3*xx6EHW3S}b
z<_IR&^4pxgMF7qMsb>Z1Blyps@?kb;%4ragWhky~9i6RZj~i?(YVFKfQ@^|N^|>tF
zJ7x%;`D29H`c=>iEnvE<kzk;`G>*S;dFIcD;RTXP+{OihtP-~0*dI8Kri@(5x||*O
z^O@CsZF#-Hu|@Koq`YaZ<OYIhE)R)-{@jK6?h{C&>nL|i5{Oj)?v5irfqzfM&$A{e
zncSL-zJh)3qsNffQqK38VU!zgyCA!`a*O7cB&x!MKr*;ZmRGjEU?$0SbP;INLosZz
z-d^Qn{Cr`@&Q>3M@&qs*2;{TN?Lz!rxVeeGwr{@`tQ<Oa`MDa??{496l?3MgzTp(c
zWivwUO#5`r&$BBTrRUEYpdbYTT9DS9(s(nH!e1HUAXO<-4+e`jXzr@yvc8R(|5tG=
zN56DKuIt;)C0G8k^eg2RbxosPyzlJVrQC#nN3;w{r4J-P0PC@ykokY#a=;H<AwabM
zcR^$-yEY}>O`lU#KX*^Q9Ib+L@<2Vex1qqFy65;Uv>xKuP-cAbU)^ZK`akMneCE}e
z>`<x~tWdWFJBgDvN4#|kN07K5p8UX1mq}k66{3-o?ONix=ka&@0Rp$w^RJ@!MNET-
zS8nVwa1^UzG_xdy8+CpL8`5U`7<RV&D2<{8T2KgcI7gxCI3>@HYPnQ8$GZY}EB-3K
zQ``OZ8wjNIS@4w8q1k~vo?VhD74H~?OZfSVdVK8iMv%sFN|`Afq)Qb0&l?!_L4N<i
z0{r^jXODj`H3$*MeKm&7FOzVwntg%6yLne(LfT<C6Q@o~!rk3ZzbkzG09a)IQ~B?I
z9o0V<vVjYv9QK;(?MyM0$i)m~Ab4BEERIQs;dL#5DE}&LX#dBY_ucNXV*1mL6SlFW
ze&y>&XP6MgBv)voY<UGi)3-;TwTJ3H|H_#8Pff{$|K1tktBLFjmH74`nK;hpBXMXn
zGwrUoEaJl@j6w8$NP&<nq_rb22Nu_{d|Q{<@9$x7$%F9Ff9n4IA;0@3yjm9Spvz+4
zp@ltjVR)7z%N`A_#^n?alJ7yNN|ht85S3s0U$4!>-R`IVF7+IrG_HKI6!d45j;=(H
z%R$~Qe?)-gDdqnJI^>Y^1k=fKNv!Po@b9}0@7{Kd{XZybli(}*5p|qFiJi?50(E|C
z3fzvpY4)4kZm4?7pmGSj{Me--TxR&I!>bpBzY|U0)rQ9oq6`1KTOivf!4!rqPgw0<
zXEgr|<n(5&>H@7IlTGY&rP4YEsyB*AgWi(nby($l?g%GULaPD~zCA4_bekZ#B2~S9
zUJ^e{i!P2JeXnq@m`msv_ut;zkb@pCrY){_sqqtjP7SX{?-y2Hl#TYH&~d&zrPPhx
z99ij`g+7c}_|8G=R&H))?#YP#neCupW#%^fii}j*lN0t8TWa6IMu;C`*vx_KG*3MY
zlf#V)1;y_vX`%gR-*dhH(N}z~?VYeCR3*KMCNeX+JKxt7%^E%Mj7Fk_-jUGMz?#J2
z$-EzmP+8TJ-#}}LpuK-qS*td|rAB;Yxo_8(*dN>XB1~uof{{}KlBt*)uwKr)PqKKr
zL|26waB4ljl@6dt1pM`O=3T@5_R58G%}E|ik7poJ2}9s**CjjXWxU)~@32=}Q}(;F
zKLav6@7@<CZuy5c(q-st>@ikV3`dq57Dqq_*eA<Aox}Vj!lv5N*E3fgVehTB+`^`4
zsa<>S;=%?avpE`5)fz-xA%Q~ib;<(agetFs|E}8fzq*BiZ(yX;Y)V^fx8{xAo+pIE
zW^Ea~3-{ei9=+Z;zf5;?Bj>Y=i^~#IhlS!%)!qACCExzp(p(2m_3OTD1~FqQh15dP
zPh^|_n`mf~ns4X*o6F2hOvPxoABwnD%|pyVIvTpivv>9V^Hb#``uKIdpZN%+ixPY5
z)jz+#`=G+>Goe_&XmS|aU>*+Xq5-K%{=1Bm>sUYq_w3g)Z|hn-tN->o>sEF7@504q
zR{VPy|9s(yP6aPsB?O6B`=31E#i5sQj^?`R5WW$1>e})`Vr;XL#NQ*-U|#q~2Ga9R
zZr_7wf#)<=_Ir9>9N(SAAl+)0f0%r#mo$vgKds(<2%z>GSA?qn#`J&owS?L6%j)dU
z54tuW(IzO4%YWz@0vUKKbe|di*gR)--8)fBd*^uTVK?{PgDN_7e-l1_lqlzxJqwgh
zUteXednsrBm&1_;?C6IUdec~qLByESb|-Nnooos+Am=h&*IF%gNtNrk$l5&Mcq-?A
z)s)nh4bdu-quS73t=qzmF<P0O_3NgnyluQ84CS{P1gG-Pm>qe&(nb*KB?tR5{hsu2
zVil<kv3Dv+0Xzx~gJR#s0a!;+#{Q5Kr2eq45aO4#gB<TrK35gZsvz}yCh493>?I6R
zu+R`XrOtcIg0f%rWx(1&RIDo!Pga_mntWbR=~&={YU{iHOupm#4?s_j?TGcpQ!Aft
z?$*R{t?Be1Y8~<(SA^6HGi6C(Qz=&@jJ`L#(nyz*y6qp}dtcE0D0z-+N$OBo#2R?;
zwI;>t|Hl*G#pku;uOPxaIk2;jIxmzcUqc(WdQZFLWZgnnK*4DE(=J39mFns2BUP)J
zzSb~gTeVj2l$wVg@^Zh)X0Y?mqj1l}@^Bq9(o^nOj_qMHoEr%~{I+Z#3XhOY87>ya
zYoy<|DP;WBf-j;7c#SMQX^GwZ)ef7j>C!Qqv>r~(>955z*TbnPP-_oDd{@3*xska!
zKW=l*Xf`ZEeAiXUmOt$uc~2-pAURKGnMbP*oysyt?Tw6RHf}QWRn?VgpnkVZ@`qdi
zg>qBeXV!@n0t<dyg2Vr;Atd63JZAmhA{9KqmCFp4;oj{s{Qus%27nYE&E8i3SJBE!
z2NiT#{x6RpyM4avmVW+Keu`pKZDwtZ!(lQOCno7u$@l$U0=k$#y~ZE{{e!9(9_fJi
z^@!NnQ2`75uM*V(=1F~$VK?WNH<{ij&**Qxf{N%HW*-^Ye(s3A_0vIa?AM3L)|=ER
zVhG)M{X$u3)dNW8i!#8F0K5H9k7!)pn70o6dYh5+pxe;F#;__2qLE7)Cwnxh&o-_I
zO=FT{W#6QHSw~%5^+aX)y!EDVUjg$*_t#@-I5*L3v8eaPSK3LXOF#qZf&ac?#nJrN
zN)`c%|92^t;H&yQ?ti|D=$lWULIJb+|L_7LXxTB+K>B|wgE*S6YpE)~pRlRB5v5?Q
zJdUdsiOpHPqY%f)u0P}k_q?PE;)tFX3WdctGgk10kAihPzmPh@@$Ol!VDx|X<agS6
zAb0BjSNK36TZ4UoGyiwdC}!5NfQHab|3AFI4HiTtHG1fODuaBF{U5Obp-BQ_#t}xY
zP)*>LIo%dV_dmRwCisVV6$K<>4=j71*_JyT=DPN5PTBMmRU@PPpI@@#SgJFM;#!~}
z5gAGo2S?`N^^5!0H+mM(BFD``oO%|w_Yn=a{~=3A1j%hp6&ZPXv{Z1|lh;=d8F$*E
zyqL=?bl2FZakdqvBYqR=cglZDtq3qCo&E}2iJ7@1Y6ICIyOr5n9SET$i|AL!A?cBS
zwWO7mfnNXY-J=8(tC=}ero~By@`|hzmFSK;@^!5I!zm*Q=3o91Jlpryi!S3@!!}3X
zM40vZJuICmvX#)BAioF97#H*FxTUE{BgeKiHPAEz$ZU(C<p|Ti<N_Dl?VdGTpFEtJ
z^l@}+=$^57_WQ7y`tr|+AU|Ow*c}O7b9KWPr-~Mww%+TL{ni22HU1GxD$93biT()O
zmBeYGc&+GSa}$m3c{f~Pkw01&<ogoqxFVOt(^?aH*B*13-+R<o^-o>aK1t>>5SoSV
z+x}smcPk+6*lsYrg1WuF00^OwKSBuX^4z2;=)sNI6D9=F>Jj4lUync{qW%oDi_6&a
zjsuq5XLc{8RxC)G;<<Z_%jK`BO;|L96yIuFACeM)S}GK7j#Bv_eO;L4?*aI-%8!mT
zOs61JOKD=dxS)uzd;%gf7W-}}T<CwE)%sQrnC4Hw+Pu>chV|Ps>(=!I`0g6-()#lp
zL`7apAHtXQ>4fC~;jWqpjz2>XHZ%siUWFCLh|+2^ZfPs`_oQJo|FgI2i&1~=Z9@zg
zBfZmYl9As#S!JB&AGL*MtNA%IWLzJx5UpZ3>f?(4TuAdb;kQ}jGkGVzTxTEMe&?vI
zCJ${yIkr15KJcGzz53IwF2udVbL~=>8VYmzI#iX^H7yG$-$Jz}pWZ{N9X$*5yK(+R
z{4>t)jzDE^HiUH30BJO+-|pV3^v0T=L+=>>?%4sSKl|e4XRV9=+-{ex+<OsmV8&(c
z^hH%)Gc&gyQ4{#PIncuVL#p(=SS3}?WGbsFJJ=NF;5ZeP0ytf4Y-cyxZyoNh$$;e0
zm>ON2;Eqz--q*_fztqTE4K|_}P|Vgi$j?~-#x+xO?1U2SW(@IV{9kFx^_Fj@T}>aX
z5IBomgjtjnVX2(2gDJTi+CM+(<ZpW^U;qYRYtx&$6uXssKceS<7#<CaS&r?VR{Sd5
z_p|@&%U8<NK=?(zh5Xj${!>2}+f~|4WkDmA!zq(LQ%TVCS|#H^h(w>{qggZzR?9=J
zWoa218BI-1%+9->{mozSWH7%lkjq#FCbBwaXgz;u^z=99L!&&M{YR_)M<w?DLt9+W
zj0d-$Ks2i5sMa;Lw6xN_a}bVS`Bq9JM}vF((6A;8taxOKO@UTRK7PkO99#Z+n~-N<
z<DtsBwg#m2XTuiQe7xewJmR3EGW!$#DK!&9q8M*RtBQ4b_A)$JU>;gzugZ^CO%ZY5
zjXJ95pM3!7{;krnKw-UqSmv+TsN2Zh*i};XMD+ceQA{b;Go$KeY;dSN#>p;`y?g!T
zUrH-PSXboHtvXh$+r0d>=+VQ6OnIQKXtu|DDm&iTC1$MUUz$ITK;7w`-yi?k4Ul@%
zkyVJ_dhRSk`4^s%)08Tmc>4(Y(uz*?`nz_t{uUI<Z))rzsfm8PjEu#0+D8(PEV5;-
zw{DNm9;KWgdJmhDbAB`Qnr@)Ry?vke2JId5NyBN1=8A18o$F7l1B-EKq=E1S;J#7$
z_}5$t{g#9nX!W$kSAN|n$`=$9v&(FyH$!x)h9Yb$Oh;EP(dS_!Ga!}RfLH;u%lNSf
z(Fu{iH_vx*EJGT0&psd~U;I5X;l6x`UsM{dJ7OcKXj``5@yr9?6`LAsQ^5cR9L>iB
zThY~Q4)`|c>NLXy5Nh^kD-hgyIan~?c_vCe!e_7%)o?MveHweAlD<!gkk0+pNc2rE
zN$(zsoFhz>y~lv0S6`%k-Wv#4(_b2h`uv^BlPeCE@+PMi))Ef>#f)9MJ6u+-;#ozn
z+sx&%o5k_=sCXqI9k8^#{LUTowqa`kwI`PSYPs2m@TU00aQ^i3BYbZSb<U`fWzZ*g
zC9C{vlum(ro=3fOa?ejcQhrV<XjRZp6)c1t>Aq!cqUVSqvHI3SZ@&Zu!MN2x|5Li~
z;GWE;TU@I?6T-3Cixlak)BPAmt&GEXbOeXJS@VTwPumk5a9&ax-Yk6w^GDRtf*y0W
zGE!)`oWm>H&-Bi~T}$UU=FrV#4R+;Y(Tdu(o(jo3jVqInRuVFGbQwg3*2D^>OfFx7
zNMXU9*r{~Fbxj%jXt9IN6Y>i3erqQk_aXFgf7qk^OTP}o`&<$XB#ud4mrcZ&I2YUO
z4kgt^bgne?^rnXw7H{&CexTH~99wU$&WX5@JpXl6IUG+>jD}^RJ?-g36y|yWiFiix
z>RE4R+DlB$54(M!ckR^T$Zm+ud(AI_6hLWRV;1ETCwC3tHcY0D5t1e&rA2RdUbo(V
z{aR7Ent8u3V}w+C^xM$RBALx628A)R9i-sT@~Qi3P^~MHBhIGo+E&a`C`m}M4(l_v
z+X~&Uv=g<@Zb#zsXC#maaSWt-<HOkbfyYeMXeU&-bz`^1!t#uL!bcQO7_ihx=OTpG
zsIkMYFE&2Kv(Ht)p}PKurz<x}{QTwFfodPXg+(&7v?oOd_%Mr#P1-F67m0^dS{f6%
zGMhMHDL=(0>ccdS5mjame_-TV@YD~nv5i!7fm!q;-E;8}R-5zg!A%y1MRjg*<W}QH
zd*C}dMGwT=Rr<(DZkSUr9w*lKM%_Q_=l3+ZO@?f8>7*lcEGc5x5XH)w(fisMxM4XX
zQKs%KIHJMbi~fBXYMmLUx1$XE_b9~&gIfJGdO5I0D%N&=cDO0bktd9sEQCbBoqVo6
zb@C-@CTAO-P!Jk@;tdbdS6?|!8h!$@O`N$|2q?keq_!p`^jp}B7VzYK6sh9(5~(S0
zGCsT^@P{f<x1XEo;teL7-Q48ZfMHL~{JCJ9WY9)nt<fvL1tJ$hKWcYGo<2+6c4|$r
zy0LS!KB~?%t)%`|sjCXo6y+u_k5wW`heEC|2+5H~vZWt=2GB`NKJe&gW|}WfR&{43
z7vEAKeSiw}_cwn@+_#_XkzTVhVf!(aR6SBT*X9TO+i;rvvOY-4LCO|mZa);W(T{)*
zG5phnROv`huOh1RV^0W|jcnkMQv3PtW^)r#6YM-}ofY66MJIwd;Jzi-g0^pZ0-HR@
zDf5;@t}ngn+W_N3zGWm>X+<Z$)-*sbzL>LMKrO}Eyad|U_ABGJX6GM!TK-<bosn(+
z?%DR#R68Gfs~eFHxG_S-`goE2v-H}p;H<iZ)NH|XZJmpfLpqoJm`seT$zug%bRS^*
z(KFd&F2l>jF;3sYMA$#DfvULWe<_nEpL*R)S$BFQb&d&j&oL$kw&SY!o*W?`aSkf#
zOVn7jx@B(jv-eM}7CJGGjm;>ovLz*@tyJ)y3{ENMr)iy>A3<BZ9PWv_wtV+KLy~-a
zN<xZ_tCRhvKg>n5XT{!DBHaF*bMv%=3T}JN_e#OT!Tj>wHB)m+(aKFT)l5~+yhfbE
z!nMPjuk*IoQyGJsXPsHv4<oubIIe>i@~}B9&fCWN*m5Innj1OK=R2TbSD4Dv{6%XV
zHr>9wZ|#lacx^tyV3H?$$`|XiqUf+de8hC*mLhXAhPt9S4EjuBGuLitD?7WBoHa6i
zW+Y^-d^9J!l3}=`s4!=6v5kGYE9;a=U8l2^lk*Z&>wB6{466Ld?!h}Z8)f#Ovlr`A
zMY%zW=H_B=sF2@%eXV=Kgei*!;R?lgUN0`(e&#`}iKlc@@no3KOtI?6Q*NO9XSeB8
z5fS5?n0!9uCGZ;%hyd%{V@DGL$dt)CslL6k?`j^gl~6!WOZvQJ^ouGzF1|KCpy_6x
zo2Z(%=3RVk9T%Pphk(R2I3<O#TE8TuL_P>$_8W*y3(z*1FEaS<<zc-~8I;CJyYt=Y
z3|T+W;#25{UK7gbJ>%oUXycf;>BwikaID&LdL3P~j|b3S2fl~4IZ;-EU<l=-et-hB
zO|!7IJ@)#eWe9$3ZN9pSV&<0-!Xe%-$ip@#T-SmFdvT4GRQEH+bT4o-i!uU4{gW}}
z8Gn+nRb6=7v-ut#<%pVXYJW&>j=Fx%?lxp+m*F=(g<>j|8EYy#H4-jp<~ipmu;xo;
zfyJV{dJxUwu$)%5b2bKik}9omAf)+bn4Obd=``W+43-prbB&$(f;zq;<<lZxpY!`i
zYlm(~504LV9>LaD%0GQ5td4uNIC8T)m@!~=g2Q41h~YjpZq0nB;*I@+b0<dsK*otf
z?UjrnZluvoU{RWX4Exo7pYkWVn_ecn>vO?EI&Y?<<3Qz_E6MWtPKWayrXF44F|OdP
zi<aCI?<+eKy%D6WkKsFtHo7LTr}?AVqlw$tmh-ZG4umx$L4{7dV`EdwmwF-&GwYJA
zab_ayt%XHZ{5L#h)Y3+e+S{{-uJDKYDBNAp$$KilQ9$%*v!3~4*`<|gJ1UO6BD>@w
zMulM3JV1kRZKzADEnPi8^)^h{3^k1G^4W}+!*)P?=ef|g>b+^Vz#JQ%9hmjweG;5|
zLAI|CDqjnZK#wsu;ON#^XN!9Kb%gOTHGtduQ7cuS<{8KIU{OKyh-ZF)aufTm_KrP@
z=;A|fNO<g*k9v7(?3a>~<~-``La!*=Pur@iLv164n~&m$U}B348Ays2eMTItpwh>F
zQXs$sJtmq{KK@K#yM2PhYxL!5UiPB>O}ZoctqJc#o5_ZEilJIE->`nhjk9x=O)0W(
z7uwExV|`_n1ygcH08YktGL`;yWkt$P0z&q}%DlR5KXdbw&@HlLzH{l59m*P!y+hw@
z^L|402`KsPlqv&i42?kkc+P&CR-vPlVu$yA;Om}OmxssfC*h;BXIC{#Pefyw^EEpU
zq(UAFleFUAq-m0-8mh5?Wz{r1GaK6p*%ysHol|rppew{?Ir6ExUR%iv*Ir8+aZoJH
z7@)pcAFE^E6EESpM|HB8`FRZ!eeFbg%ev4C#Wb7>Z>Dxzk!OIZC1pl|AwO?8(J^dI
zS&_Vpz!mYGwVIj3)ry-pd1NYW!vG307P!Hey?^A&G>4&K@WpR)XrwQ=)JJAzC0DnF
zxEf}L1*36L!E@L$kK=q~0zPy`uP>2jZaun6W~%tKgF=YDP-|;p!jLMzVKMN?HeCN|
zFRXd#gt@|Spv7s9+&to`cMYM$Qf%gKeplepv2bWwWQ6?Cz0i*AHu6&~4dPv}Lgv$#
zvBP1EUPQ3{g)-cvdBRCb!JA>dcl@NCGUBB-xNunxbInhz8xEf+5dBuQ5!5jHRu+iu
z@r+0OSgGB4HwGTuc>fWV+V6+)^R7ULuQYxO-Xq?8Rt(E#;Zf|GK$8e%kSh5MxXMRO
z5AH?%U`j!~k5VF1Qe#JNS_?DK#Y3Wuv=|B<1qjY|H4I1|N-Z~MtaD`&F{)!!f2JD-
z(tDm6n<wM5ZbB2$1+}46t3qo`R+RixNW#x}Qm@6h#lD{CItl&!5u&pl_@5*#b>p3<
zGQi$2UVn<(MFglylG%s3W_x?83TG|10%Rtl#a&SJ&8}kGv`$75feIe#*{&6=iv{#F
zIUsRyBwE$!p`liZAvcA4e28qHA9PD$>zU|U=}e`Q_Nj^fhUdaN?o8+AlOJqk6izM=
zZBr)V%I$8hH9<5^Jt6g!+3dR|yEx8#O(EH<RsGN%C}9}DW#ZwhKeZEON<6$iMliUk
zf({>=9=hPH+tBig@Tb(AslAEkIB;>Wxw<6que;vR+~JfG7vl5dxHu)gv0A8pBJ@Ds
zaE8fT>U^a;)5*?eWP1GcIZPZHS)I7ww&IRdrO<he)P2>RaUI=z@*0{J4$UiW=3A3R
zxsO+5&84h0$N>f3tm`zg_7mDoiJfd;Ym_iFAj`DBLxn_TW<X&vP%xPU@?zlNz&ABb
zz{9!nbe2w6sqdLa5SwXDBnhQ<eR~H<MWiU~$TZSdNUH2qs8{}k)oNU*9;6%aL;hCl
zOK2PO+GJ~`V-CV@{YX1MY3av|)WZ@s@XP~I<f+6<#RJzxH_L4ICuN3?&K}kcg|He%
zuG-HK`ksbXQ?~bC1&Z~Av6q5_S<#X5>!X#=fYE9+_y(^;)5nb!=qXeAh$Ntsv&6hs
zU597;4;EnNpxHUBuq;OMW{K+ANa)yi_2xxwy?VI%kbo1zZ<`kfwFS^NP$C+aS)o7|
z`T4ZQVx<>Ww0nEi;ibUX=#<F`KkLQZA%Cx3!c9Vyb>&XQW$+Eq5C)@alS?klPd>Hf
z5M6yTlpT=C&!~F%g#M^)v;{f}WA(9av?Ve3_s7G+<!V})NncLR)fk$4wy!S@)3#$(
zE%b$`%td8;0|m{v*?WD>ONwlWb|_LDL{_bGKd>Q<$f=D~r*f7hSnI%G=-jvyT1BG4
z)>?O;7zll2oGy203Pc)1i03R^I>Xd~19|Fm7b+~0HK=84Uk;31=QMG$+2e0y%{PiM
zRU3D@R{LM<wE4pdP*pE;Ube%PeUoM=h&8&dzrHF%>Dq^oWJr)+&t+^L`c^2PPM0{C
zFAxiKwD9G{pDNq<*V?IVHW2dN%#j|{UY7{zY^<Ly)LFA#95Wo;*bX&j(b>gX-;Bbx
z+SB4de?3N>5c{Oayt8l>zmjVBX>Tc)I`>n)5q-|Wg1-{yHNr_ehS<5~nGC0L^`;Qh
zur$L@wG;U2@RJ>1zur%^HtIoz1$j2JHIsD)(-$qUb2~DOrLLqhQfxk^pH=wC&5p*|
za7FDMRFzLWD)z9}NGqGlEJl>@e{$=Zqp(sRb{8DHt0IX(wPvMf?MieoW0-jdJ^$__
zHg@0nN^RuiXoZGMjDjh@KQc@61IW)e!E}x;$7meI`U?fxUsDJq5TUeiC<+C$Yystl
zD1`!6g7^+gS=c}$k%`=ZqX(|8O<)5}Xuhq2)Q|6-XOfmQ7Aa=3iN!Y{G&9m-`W*dr
zo)|<F8cKu1XUDw{B;j7Pw88pTXED9zaezp%(u7>2i;3+BtK6-Fpnq{HulUKkQ)<UD
zW!Hmdzy9IsH7Zffmp07A4~55Glk_@Ru^C>Ax_K#@a7;wX|18bY6j}B}bc7MC4fe5H
zAU+`_6LqeEq|yRiioVN)MTqXb&T-OMB<j?azpK%V$wncgpt9vFR|bZ|!Qh#J@fbXR
zkIIYpk1P<fJGa-j5pl>q98^xWcezD(P*fR7)F+3cY6PyYb}I$s+gpe|Vl&C5jKWz%
zNJw*!t0X-PLGmuJzuHE82b7Snic2L(AF*w$zhYSDlDg&>RT<i{!BVMwwOrF2JHyBj
zyIG=BZhw+lK4-&b@Tj-Wbi`?|%je*s@{=;Vl(5{=<p-L{FO&{k^vC8fm?jIi+!Y!z
zPiBl4e^_cZT{OY5G@$qHa0VXO(E**TTWI*@plJDSQO;bLeu*Spanl~EpWj0%*E8j-
zhLzUOl<ywR6=I^=uC-YC;N0LoRKlrQIh1Jici1|B)m>Y^nhP&7usahooMEPB`?OGa
zc91*6ytCe1XU~X~YrI(2DoRz(krme$W#as1*?!Ju#D1-1Q#e5ARn^{s@=V=scik0k
z9hwik)y2?Bo#{fgd?rlPX8V^42U5iv1IiX4vJIwc7+q20J{_qJ#|z+cpf|?CWq-Gk
zDcQq3bHc`sKRezA^@Wyg!=I`9KXuJ7UB+-)jx<tACq1eDS&0~D&}EhI`N;#FkX=)U
zYq^BDY9x-0cm-5mLiN@iVgJo1s)}1Y_aKrYkXGUR^ugrDU{bNT(Tm;E_f75Hb50&7
zQz0cs^HiMIdtomIM=ewFM86UvbZUQIx3_gn1QI-FzeL|x{=Jgw`Fn8CW3C0tTd*@N
zEpyUnZiN@76%*+p63<vWhB$JwF<{i~HTzPZ6jH-qZ}jVa80uqld(BMZjzJV;Lo$IR
zFg@nH!-I+RK2z{T#Hj$T^-=@SYg8$PN$uu71vgr!y7u#F?`!FPW|j}=`OrQWVoM`Y
zv?BSHUpVe&UxTxxpOzK)xVNbCijOs}AcJ}PvcAAn_4wZX2<LAOw2#qmoVB;NV|cz-
ztYLn^kuRbk?IM^mQEq!&o!(|rA(1-r!eB12q-X~RWXj_j<=dI6pkl8qtyA$p$RV3R
zrm<S0RA>{LK`W}&6cdHvtF+N<Ywt;~O$%73ntUf_$md}GwZ)YOo9$q{9Xv&CZDDw`
zUc2W$AB!)kNcZytO9EPv3DC$Uk3(@?uWBB?`Zg^bKelgcS$r;hepT058F+4LaL(HT
zH)FG0W-%3)vEJz5z%dy#G;-}w9Z+-Re0k`*Y2Fw%;jh-u(@(1OBq4PY@%vHXd)nn|
zD}(cv2%#4Smlh88Z;&*M*Gu9lKb@8ckr>FdWcbhE*3sT%fkC5(Po19S*`E!{PPdLF
zjBwO)<mE|_Ds?{DKW#j@-U>IM9GxK^F<i%})TLBfT$;h2Va~g4Bm?H_H9@Bc82Wxs
z4U6iFHN$GeYx|`iTRO%&^+<$LJ3$9ba;=qyJFm*8I@t1XEI+?dBAS{7Z_=7C#i4c}
z6g|Rs-R+KJ7h8z~ymRwf<IA~nQC40i;&%q)ZS4eSXb;53^C|WUbSxw>+J-~vsi5bO
zC>7w`_&$|ez2DKSP|8_rHa3dgbl*l<U-PZ62~Y(F79%0`$RCErOVxlmAX7e)v(z&;
zOaf6HS^d~IQSJ*<QM23xiNBvgaA;ECfEjYPccP?5QBaG8a3R`$AtQ-`%8!r$xs$jP
z34OM1KM(%yxL_jqMYE8}0bq{BC?v994SndLWVPFko6~zqbP_V1b+EkCb*8rh)W9e9
zi@_H@DTb1#RDo(nlBblZ@z43R;i!e+cE0Yvkm<L)T6?OFkVaIX?hR9@HeFI3|8p2E
zmb)8Q%`xNU*)aP-Zyw6!W4GEcA-$*dA;4ee=sqzOOcz}7;hc9qfnkjFU|_9L2u-kE
z=-e`S)!i;5#@J>0vehEZM4mjTR$Lr#(vl*!m0%^kzITr{Y`)gU&c{kwgoYoA+qO5E
zR3t^6d2tvku@S#xux2fk&rH$y9ymSv)%KECN40>)2hpHP{q3;F$t7p4_<K!5J3wO2
z&*jjRBjblgNN=jN2<Q5BYh^N<Fmbw-hT=Ht>YNFj1ZI=sv255(7CL2=mGSHdjoXj1
zT5#FyPOtoRRyPN4*x9E?aKqXfWy_hf@{m>zaIRKM{y);bGA_#QYZpZYK><ZT1VltY
zK|<-~K><NRTB(snx@#B)6+}Q<y1PL@8fK)MA*2~bx^spaCeF?C{NsIp=X^W&CzzSN
z?^=7sb*;S?%Blukpa9^d<ZJ^OlRj#zT@8W-+<bI4aS&9`*?B&2vt}!4)067z8OX^*
z0?6A{>H)ilk7a`O{E4Ib;6Ed!q(s*uo4_SJR5l+&5tNYXU9|jF>9n}0=*`0A&={YB
zAeD4!5L~iOA!}sdDZpY~N(|*brw#kC*VkDx!6zXUWpr(Os)?(8@q^0;PW?-y0rWgg
z*<$Z@XQN<!wX_AY5db0uynMF<kS(Kd_)^$aC^g%o5GKv<p-b*DHMhE5?kK!lBU9Wf
z&q6&;7M?J!0^jKlTKbstSokD#%WggUTG6x{cO=f{eC9AD1`})+#G$kbppJ-i|7y%@
z1H>hx@dMQFygg+?TZdT;I7aF9iHX^U|ArTQN!#r;>#J!c19P~Rf!kYkKZjU<9v{*E
z`+25M+Y8$!<jeM4*9K9&SMpU$x!j;RR3o>8wN3Vi@5D5V9Whh$zCn{QzAt_STT~&l
zeiTWV--Ak=ad`zXiB_qQ{GD-e^Sk7;dhf@I<XN(G6X77WAJ~tdlqPlo_cGmGA_Va`
z8KXG?nHEl<I0Tkw@Wgg;1AAtgD?Clbgv{NGyR>|5wy3}hF+|w)#>^J2heo)%>VJ8e
zH|xWvQbuJinY$$Hfua1pQ#b8GkFJZG!GI~{ojA*uVo@Sxxqcbg*lj~6zo;VsBnQpC
zv09T&Vu=9R$ZN`Jr-{3WpwAiT5j-^;;n33{XIsE~>mf?hwppr`*$>cB@(s2uEtv-k
z3M267I1W3i1S6tXzk!hOjl$mGX(-CQl(%MMSLdxTfnhp0d>w)k?K;=k0f3E@a<0BT
zjJ)SyQOWBGX1)1_y4B49y&h@)pXI;vkJ1^q-&=d>;kwd2wwliS(+7sG6J9BBXuW$u
z&i1m1h60fp%$WJB#{81alS|Cs8V~bR7;fI61{BkN6PJ8q8!Zh(nKAh2*RL#^Pfxy*
z4AaDwvX*nlWfq5r9hrmxm^0u(`2esKTL+s(#^)YMR?Q=yjz<vlr%&DseP*x2x0$8^
z=I`KD?|m*Ikr{JtfJiw6bEZFV)944_cpl`3oXdLmcxXA^$h#JHxHP`JEN|8COOHr!
zfD2cchS&#H(M`lZjcT!+Cj9`!u`C(MhS(<-byAEbqeOH{!54KS1GiFT4y6zItrUV{
zn?4CIN=6w~Sk+_EJbA9Mk8E->pWM22d00g^Pv$!J<wW_5Fy!%xCF)h&8RcTM&eU_6
zga)oaf^8GnMopn9*Nu~P4dFQCFGM$ffBBNCl}|stiql*y1!M`Oj%54Z+-JkK1?HaN
zI9a^R!7kXx`}1uEerF<aoJ}q~T{h!rIvrxX5ncz8@HkHz*%`CmrCtE*9(y7S+CtDd
z@B)qIJ8k-*_NSvorByOu!{ZmCQds*bxW_3xquyC2?c@zxopG{+bL%KyqM3|Ltn&GF
zH(5bArVC9E5pTx39-bF1Ov=LdySTwAL^AzAXNfcX-^R+xH2P*F+}E~wl83(+vX2W^
z>hf<{jV$NW78K}hiGr}#WMaDPv`zIP(CC%T`6HA3h6oQYy~f$;ZL~s$Pd)5jfCyOP
zT(JQyzonH0&qB%L-yJYXbG$I^E<+)!-3H6Og-}N+D<VXzn`;xMjn15RGIy>x3#|5)
zlvA0uar`oYU@GcZDA)oO>{wC$TT`VS11<fv1n>Su8g&?8A9wYsIM56)-u|?i9z>Qn
z^19JgS`?VKn}dTC!sv~STm9+kKlD#{Q9IZT-|6t(EQO##_&=Pq7;b`r`gK-qj=lb(
z{0&8y_H>rg{5(+nWT4(6ukDa|0E4-FHDj*s!_qJ39*GXhmg5rF<1$Qu(?s{$a<4x0
z-RXYfseQn;*YOYaVJ_NXAHqYnTxsdbxhjfRLU{aNpWFxFxi)Rq7{AcR9?BDj^a`3f
zV$<36fW9+L5DOHro7fAg+z(T|rt~;IHG0-U0ja1-uE`Agmi^Af&qb&PU?d3rO1<rD
zfdD$A1`^Y^zWwsqRqCVVA=P80u8zAwoVvR3s*Jt{eIqEaL;%)|!iv_S>V~bqXPPLq
zpe-uM3jT$ojp2C_nI-#u+matsA9bAjtBh6C6NEIafld7oF2^^$Xf=~~-}x2#0>L|W
zKMsvz1dIv&vu3gAt$Udt2}htIo%-q0cX0i)t|Pjua(0YFTlusuxWUjVkXOeUwAuF5
zm4CnsL{2T7SD2o<#Z=x4{wcbq=cUa}GQ&4?=E=?IaqAGC<{2LeFI4jys?#pIdUqxM
zOJMx?jqfLL{%=J8jcmg{d%RKxg>~0up*Au^@EZ~0!%1FA8GLmqd1O;9?BKJ5sOr>a
z`1zigAb~C#iwR7OWOAJLe9@rQz5M7KiS|nmQ~C@O0nxzf-o3pHoY<VV^2QE&5VYUP
z&nhiIzu@xcK(-mO)ZEW$e>AcMzdy?vC;-(o8R!OlI0O#19~06<s;zP{iE_|E&$rVc
z8oh~;BftY&GKMxC>S9h-=NfAwC<2r-)6w!8`cY8LwsKf0ywkU&_wX%ot}(O@(irLq
zu%W6XWajEHu>JNV;Dnhij+R6^&Uo60t(7k<yZ}&_mMZ(L<A=0Fbr^QN{Y)}QKnggT
zcZPpuTN-Yq*)h;NO*fVjo_sy}{q7tVbj*DJBE-%5Tjq6dB623oDE8vFMg43t+4giu
z)8<ar+evBr2mQ0{_K;e>dL~wvLC5pv6_DrBm4E0jK%U>v*DA1)8XN6P`J0Vowz{6T
zJXT%@6$D}@W3B6AVa~gYozXOHN(Ckwe9zKCoF3oZ)XY8s0^184jR9;emr0ofpih16
zpW8DRSPyQJvvnMFR(A8KECN)MkjNU(b%1E_Hv6g=(;6*RF}kB<KG^9}<PcL3YvFij
zjx%Bjm<zDt2B{-aENKGqQZdz##L^8WVLxulT`l4vVf)utD+@oNHyn)2HNBEihxY*)
zXXxH|tfD)0^vFjU@DiN?X3dSbpZqCO@3yz*xTb;cQr`=nK3IS3!kZg{d+=rwAr)O^
zMh1R@-3n6%z$F*+lLd%8BaBZ;ghOk99Lgn->b;f~(cTG`HaRy6#z}ynIL2&FM=m=>
z#WKb)Dm6HHY4|=$YibmV+XP4|V?Nz0iQf)$y!zg<%8$uz*<sJC;QkWa%_O##qkK5>
zoob?&u7iKOwYJhGC7>Rl_3j9b?&ngY-I192XU*@=viFej9nQ^_(5-z$1gK|sVRQvl
zQ~<{l-|DuJ3p}w1ey8)Nif7H@vY2hPv*jcmAfRb;0Iq@niaxEC(3?P(Y$Cdp3g%^x
z+bTiLQuCd3D1&+d_$m=BfnNXgy4xRrnlp(ag22%8<LF9AMbi1dxhg&2(rpUB{Z27E
z6OW}vGSf4R;%fFGdo$d}Qqg#_*||!T$<!p|xFs5gR!9KV98b6wJ(QEI*=v47+$?p<
zl8z&@)%l?2ah+YDMeG*Eu;*p5yB>!L6&8FqJjOq%oJiDy$FdeK3&P~dO;3kw)%58S
zqjle3o(yCWy<c*URq)XGtF&Nf=7hIlZU0x7QUWiJwpV+o&(TPDP6eolrG28YykbJ&
zs@0*IfS=8K^ZH`%7OrE58NKbed-tbK*L7u0jvRaI))pY5x^7bw`v*^s@^4Vv&tV+L
zgUR0HZClRgsl^<jJyg=|D}@nprCb^o*`igu{gIVDKjSwlW99+JQ1II|-EUt&k2dH4
zsay4{MFYfll7FlMfq}S^Q#_>5;j%FOpck)vMR9fLU9|)N<#cz>Aben&cj#ta%@TV6
zH79CXqBW)ylfb9^tmkJ^GpF#Qd#c}GI$3DlmY*0s!J?nc@cC#6l3A=0meUm><*`>n
zIW7Hk-pWniOir_j5|84P5jxO*qX!0!Ub|+ZGb(?$0pxt-FnIHLI;(&2YVa4_0c7^?
zB#u679DrgU!@A&^6b&JCNRMqpI1*q>8C;pv22BJ_H{Ev?$XF@Zws#vOWqsMB0T!+9
z1}0Ym%UZNsbjcw>GPMC#3}Iris&rF>rk?P|ZmvIR5z78_FZA0KKwV65Ex`*2`(KKW
z@ypO)XTx#n`6_ddDFL9G=_rFbv)Cy3&#TsOe*EU9Qpw)K{@1dSvi?h>@WS2emv>_5
ztZvaq<iFQh$AqBSjI=x=(F}fgvV~x;a7$-qfHOJQz-RZ=>icb>Ab7hoV?TFSzwCi#
z9{%fWei&>Bw4S)W(YzCp3OVnonD^&4Iw+IHpMX;I$F)ff)?N-}@YLzmI5i$iWkOI>
zaT{VDM~#vtGot5_KWr??0f<n}A0R%{{P(8HxnR+axV=felivaHAI5yL(vbKilst>{
z%|VmF5!sknxF(Vo$TdzNp#ox|)VUL9D}(Xu6sKZuit+t@XL8e#sv5u@P%FRUaX1d4
zw&rxXNxcm@nwZ-#sgRM7`Diy%Rd!Nt;Jgx-gG`e;+nf*f;u5Ti*)Ta8v@fxvFq8I}
zfH@tXM{PHHyCY&Nc4wQwe}?z7UgWJ^5!nt&ovS;4V^vpkQr`Ha+EC2ra2#Yczx<-<
zq|CqwWi)cl`)K4~I#|5I@KqnO3A{b}zH^J^{**{6oY)(Cek>ye+gkT)ZT7Olg&JL@
zI@K!1p8_*}@H;kDPs&kJx?6s{PhuP0a$Bsr)@N}%5`jb8RqUMkmW0_wR84vg`B%k(
z$R#59yAC>O+B%`)v32GSHzS0Zx(@&z*p+koplyn-qNMTo=;gEv$<>@oF~7e&QhDHG
zLi_r4A0^w%k<vwR7Smn2$zYlxVBv&jM`#$w|MWRnd}Z|+SQ^=XM`2v$%NciDp3Qir
z-alL31_))?S<Vq%xhB7NVes5~?4wEcbHayeY)hqv1rXG~$P$ZeKK-l*`$Femm3*M9
zcNW~JY0M>EyDHfbW37+XEyEr4`jydvKRNg%q+}Hn@2ZLK5fPsHGyo15Rsx;c1?oy<
zj(mfz2ofxDJa>t{do3sTW=wn%K>j{qU<=n*dG|7!Mym7v*nW*lZmA_^g?O#;zT<BM
zD@sULU9dkE&cA-r=c#bvCxM6r6zx<1@XA4Yip#&FcB!+LCBg~KwO)En=~rwLwq)qk
zQ~BZVKGbc$h4iEO9mKs?{X+K2e-{uBlPc!4KuQ3p$ypLSP1G=ccmLP2Fp;{PoK1E!
zdsC6#4?=8Sk`*V+XcvSEdvpq;;q@D6$~6Y>^?{I7j*|-}C>>J#`yp;j1HU=lZd^x1
zUS?z#V0ro~$lQad;JNn;yhMqTbz5s6-@%(Y63=0kTP?0@Ve<f^=u~loDv|!4tBK8E
zdGz*a_~=N<(fW}1lkcIh>7VH~BIa%+s+yO@tm-eY#M6~|WQB!su+2NKxa!_`2V}ai
z@@8LWog-Arf-ly68yRbQ;0*>OUF4(fNC%I)vCqmP?<_;by(SY~Qmfi=sjHRg5?fOz
z<1I-IpZaTMU>sXPl=^%ykN(A!a&9-?iZ1Es1D>`O7(D0PvZc4Z0(ZDvCfuyiPRc~|
z&74#0=Wc>cm3&U>jk%h!T4qQOq9aK7Jo$o>z9m26ik1KXCd0ZprigIZW_x?quw=F6
z`9x(Gn)>vx`1JIStdy;kMB7JFGVW>n`YFpk+wuKtNo<_s(QX9X+4>358?gM=Q2o=Y
zd#aQ`q9YKPrOdt35~$J^iAkA0iQI_~T@nkKeVqSFzPoVyO1LVh+G>T^Bt30!c1@|u
z$x}%27@u^F=li3V7C_$@AWccNx+9!3>26DO&z!dDeF`~B)ncflX8cLmjF{^^8^xAH
z!Oh-Lxd`T&zEgeEF8DWS_^xDGH$0O=>*}Y**Rw40)mprvxlYSYbbWbF-&7+1VHbC+
z0z6+SjPH(g2!sKUDqG-M_V3NoD-q*(Q;iYwh&1HQ2+B+Mmb9T-$m0!E%dNJX!X8Im
zfH&|Q!X6kUgGSZCT>5KgW~kW$@l3}SAQ)k^VqVKSm28+w|7rctMwSRDPLsSQ&j8S?
z=k38^`A&f}>a60TxS8_|!c23_4gx>W3rbK&-St>ruZ{#AW{3`z)T(FZX}Ogk%|Ybj
z1qRLg%ie)Xn}=}FI}g%#S}qNka7}>Gz8TnJx3xiO(@3Aj;M)kWPUa_~C61Z$cAiHo
z9Zo`m8>E~?^_v>PBsV91lSI_{yY0X4z3GqE3zv+aPm`h>#Hc(qIQNdzM;g?6EK=#u
z&$QTQ4%=4@NL5-IqOC(Lcl`Ed^hKy{=tIQwKG~*xdaIG<u`MYR?l?Wc8)Q`(zNeqX
z3Hzfz--z#y<>(ed3o(?n)J=^UHCw!U#TTh-7V$nSN8HF~Y%uKFm*DQY#QHt&rMv=B
zz2Bylks1x{OMq-SK~5^V`EOkl8iSwB)yb0+*F-gfXcUK9W_Mj}V#Fd|cp52qrYpoa
z-|BiU_B^H2MIwSy(dlI#yZGUwLHZk~uO6kXB4sgspMi8QM){uQFUvJa+ulGR<&(rR
z%^NuP^Xnr^_K=01rfTKNTb;Ir)lJO>Z#<WLv3;=^$;!b;)47!pAVGNeb{CxU`V#0u
zXaFdjk(00f;X{b&dt`joGv<xd-o9!{EbG0SS2kMq8Yltlu1<{!tigNGQw}^(fTbw{
zX%2KD3vXHTuvO)C$-7(g?*z)s0rh3M@YgT*Biw5*7BTJKlV4{`O1^ui4k$rTQC>nB
z*xy%XYc+3D%GZ#0kVLSvyPFld_N(7gTW|=5|6MUJBSkEMo}2}B)}1MELY9&nKyN*E
zGLYWAHt&I;ClbCtWXGSuf;jyzqo*q&a8s&bv;81wPrsrJZSOgBmi;%iw_@}81h&6i
zNsS(2kRD-ywKTy90Yw#14^LcIqt!~|7vqM4z~m&%rP!n^Q~@G(wj|wuR_S;Qn%Ixt
z{<Pm|F)VJ08LQ&Bx>mwlZ>vd8yV@JB5zNy(S$J?Aqzh6n3b>d5U_#rP<=jdw$57WE
zRoeT3&ZdXb-)1)dM=f|<Ev%vxoo18XY3Pqz;$0Ep_ETMbeE8hhXMEfoOb=$`v1tya
zqO8^HU3eyQtLixjzp4M)$@$N(k2G!B3Srz%)ujsy3tb{Ts!>LK9?$QKhwUP8*qQW-
zx=NaR4nj}aLlqJ49*xi?Wd;J2#zmr>o5t-5fy3Tupy`RWNYx9`!%<O{6&E8StJj6~
zV+#zq#l)B-6s~$n8LP<2m>UQrQx5z|ybF%@sp{niz^duoj&KUQSNF@Rzf^r>hzoYt
zF$0X_YfPfJf9OgU$<HYZ+#0>YX?y5B)p^;{<PAJ=x}(c_C~QMA_}=fcH1zSTljzS>
zD*=LV&0V7HOCZ(;(i|oWZM)&`^Ho51-+G<Q7)O4e<r}Zuir(Fp#NIccm0xcGKadP!
zhn(z2H8ILL04V?^=in2}GP1?rtJujy!mAplqojmz%b^tYEZ>M!8f_DO|0N{~2E@`J
zK4Kql&c<D*ZIlGAdg&c8P&s1*xvz+dU@G1;|2x$xeN3T-7uceC&`s?An9%KH;xdTi
zcQmMXCUEo){wtFX-qYd6Ebj5giTLLjZs`1OLlKTEsB=n9s2;>!5GZw&`Y<B8$v;jU
zL^z(Og?gGd<Ogh&fhSMl5Uq>`^nzkdN~52XV|;fz&)yBTIm!Kw%2MbgP)+&Yxd7qe
zH?}E{6#YrG4ZWZfb#H*%8Qic`Gr=mp+WBBIN~UV1YlNo@ShcYwA=7RJ-hDX?yUHOv
z#Bb<6{;5w?=4`ie1V!a=z7=Yvtm`A5_pmtaUrGtKHLf4$J*2DZQL+|dfpBM*5&E$+
zEJx1v)Zpj#YD&VKyIo7Q{0qRHoaqiTi7Pc#PnTj;Pd#1v^M|LK>Sh9XypMHImt}87
zcg)B#IGFM4unfou-|m_4CgBn10}D&lz`uR0Z&#%s#ncA_kfzVa%e}BkVt1OU{1^G#
zI?*A4Uzgm;vjE_lYa&WW2Ee~#Q&`h{r30O(aR*a`+(96Yc1ev`5{r!haC7p76Mte=
zGxi0!HPp*um(?V;&gt;YoxB)SEV5eT%|P<F3*g3SS+}kSif8PoX7f>4@_gg6`#Xi>
zv?x-Gky*LOGDk?B9LNnbX&gXv$4^IJurw`@+K_I!RYFw*UHEA8JqduoUKQgn)6`_A
z^JJL$-g>YhOE8;uh?#7*H_dZX<P_0;`hg?*r{KzM<(T1Az+|6}zUJp`u{>s2ec5R;
z04yiV+dj`xC0Av)r&Bm)E$ZTt>c_`Evp*j=2Y*XyxPKqmn$&tad8G^f>7c@(uCQ;J
zmDq?o>}Q(1`*R}MIC6E~1z^Hm%>=Lc?i$V#-yB@y%yqKBNLPt)eP(Al`7*4pU-N5V
zq`E-#BqD8E|HVx<<YY^=m)E>}Ncy}E2sD56T?x>x4A@QOysC<R`_8KacTP~zu`rhn
zeSKM9!{eoY2yFQ&xG6#d5sCC7^u(_@^&m^Xr$T_YeGHO8$0bu<v-&g%O^bycltE99
zDvV?Br<Uk7W4O!w5swCLzdWZPwkd)cADl7JKl`*>*!T>f1JP?hsT<?HNYSHN^8$J!
zSvahKMAej3<C}%Q2A;T$gxDLcLHJLtg#w%78xIYuYbx&aTzMt=X3k?mIP*)R$x9?S
zRx84gQLjj+uTS%7tf%9pkboVbH?ALBB1URh<Q_(KajGd)I(ED*zFAVl%h}*nVrKk|
zF3M_2XC%C{weM4zUEN?<S@_P|Tf*nvY{6((D=|Ulp@pan>T6vCNo%R-6vn}i;s~@~
z#yPB}@~>cCWmJ!k{Ve1vKor9L9tVEO^C#i8hR*YgZ@{7v1HH;BW1=&jcgUJXeKCWV
zhO@6YKfR{LxS<;~2`cdk$yI!MiQbmwlURy`b)}QzWMBK}VSG+<@AmwD4Zdj-j_s-Z
z^deQ<ZF?ZeoQx`vW@Zx6ZP_}MCV7BwH|KA`ZZ1pJChIVtCjI0l`uuLV!}!ByCBJoT
z@qS+?>q~A6(dF+c?^y>=(p^saT$>8Cikh86L+nKih4Q4<(m)oDge6GB<3O8=*9m-z
z;7-N&FI$<5bc$Uw<?gV@fKRd?L>3K)8wJs7Ut$%HZD0bvS}sx^X=0!l(=f|R;ET*(
zREpC{n!Y9fb@as{;w_sT+l<U4+by+1HrdbYB7YtT@j=+FyN07y^u2VF#zx+bzqHbE
zOO<d{>@(&t5b|aBWs5zNVet1L_;F)b^}NA6Qo6gK5AH`{1SZHJj}dG3mMNdrW3EOM
zx|(m&@(;5vHW1c~5%Y*86K0<X6+VhYhA?Hn?M{2?Mi1!S=@D7!2R@5!tu(dqAJR+=
zI3euMPvbDhgbj8zKbb%ZAQ-v)Ow^SAWLxacu|sbP-!m5%vB1UN+w2er!_a)44&}O|
zysFzDGgtCt_oHV#y;sjB(R?%<8%$%n%rkz6yDu8L_9Fa6{fdie!V@v!PV+w24!W1i
z9FA5FDj6(WP7W7DU8#ZBoo`&XSq{Jh{H*ErxYbESddt#_lBJaL(>h6gUGAku-8z3i
zKklO?zg4n5q!C*`sf{o&!9j%7c&z9Fe!o_6JPHO=#VRgWI&7rl$Dp$<o{s_*uR=+B
zOIEzHoNhP`HBqq4%%9rICt<=rySZ}PT(z9^y_)&%it`V#-;ESgQWMQ=R(9uP>FYD-
zsaM-G?_&Eu)l{IQ1^DOt-MG@FrCk3cR(g?)KS{=>r{uDzbw$)Bgw1|Mz7@)l5_LFV
zfy20n(`YA{k!6sAc6W}?(nuiMX!+$(42Oee&C2tu+2T7+@Pl@HktPdeLvps17M<PS
z)7yD8T6DJHHHIY2=t{$@uaA!(rr_0pZ*Pt`u}s^Ovr!zr)$%UiURLfSK6E7vCh!|$
z)i5g7sUKJmWnfw(!EvZecXfZ3OBCZm<;sF~UoZ2UX^5J?RaCPL9e?%cv$^7Ucl7UQ
z!77qA78SvlRns9l#k@p$1I&=l*<T3+o-~SUDoT}-f{HsYmFiCkXQMoxBS(+Cgn4G_
zk|i#^Cp3AW-DJhwCz*xI>mHNXJjR1OhKW~&esz39cb3?VksE$VxIgD_>ie)Gg64}k
zQ@*y~APALAMRs^w?4CC3o9czuL0!K~V-N^C2NPHE-B@Fob!Fu);|HesREE|{j=-2t
zdyE^i`r@D2krSe>8dJ$$#;^4^yRRm%;h_{+st*;1dQY}w31}$T6#Eyp&vk<-nqL~S
zvL=I{_+6h+cb{tEwIDM=i5i9P+q$aHy1^vigv0Ia;tn*dY%vS`WyC4oP9~He9!X<G
zQQ*C#?=)dN!z7DnjsZI(%i=x@IGzo2$r8GBoV%E;{ZF&TIW967!r^Xs55ZK=qnWL?
zll&^FVxRC|rOe({1);LZGqorNG{d`tUlVsLCeIkAtD)Hb4e8lhlp#vfEVshf{tNv{
zZ4ei~x^&^H@a5_%cq1yvQ|Qbq3u1Wa4_DrMtp3g+5MuZx%%435)|#Dg^!l;f=oROv
zfS2))O3cX@?Ke}KCf{~9njA4yyE+hm%@Ax%%c2v^6d$IGANLGVbHdx^LZnYL{A;hy
z%+9tSO@fH6<>is(21odApmd_&!(Tsfy$hOrib6c##Wu-?!mxGI6-;pQ>W2i%qpxu)
zc{y_7k0I=}YMSvcgKk6!BuTp*jOK;uE`PM*v37phdSwK2Zli+B1HoD3U*%~v*o4=A
z3)d0W%$?qTS#F8BoRa|B{roj86m+j8TGCtR?TDd)Qh+?^ip-+X%hwn)rgTc0Aery0
zoG-ZyNTO-R`&*LoSRsyJVW*zUJ&Z}=xRZTAh4Pu4>TU-Kp}slUH$d}7@~qCrS#&uW
z6?m(sL*{g!5e+k%ioc)l@;6gc1}K#Dy$yf-CMUm(<c$h((te#7nnk2sO>l1{sD@&0
zL^oEW*C*VeCeB@?Y<?V#DJn|u7Gi4$ymQ+o#k2<b<Q79=o5D%60%2g!Tr#qP3iWoj
zj)9?i>s<ao$a$>=zxo|Sr43VqwJY0%RnQj_!TOD=eQcA*E_ixxVnRl?nZ0F`R#2GO
z;j!_eVWZS(n`$BY&>`b&cg@7R<WFM^?2tNx5Qw>HqdAG6sA2n^P~@^xY+17t#X5g*
zKwX+%<m2z$rJ;_)nb#T7BH^y%p)~ncKPq6jP)0-w@g4$p;z=&JcUfq#*xkEk7v7_%
zgT(akqcvPzh_TY++rkGH2*O<Nu4FTBqnEA(?&+RQij|b=f}LM&vT;EEhtWNG)Z<eJ
zD>H9rH({{a4aXtN;{0<ehYDvL?#^T{3s4J_q?h*BU+CrdatZXz7Ml>X!=RU@<2hVv
zAb;rH_Uvh0fs2w+gX>N{p(X`z@z#dYyz%GeXP<8&n8p0ew6jYB(xzgWAXD=&LM^W}
zt8tC1<-wM;=J~I|vs@q{C}N@Lm)Uop4Sw!ion@=Yf&{*b=lnTseuq;Xfv)ILHzw4X
zLjKRcf=)x$i(JQ7S1YZnX$3j$mMy9ihTvz1p>|F~sJl$6QIbyFfN+j#HY;$)QH(J%
z$9}j6?9B=64UFJ;J;Mhmy$1a|Z7+Y?amNi=aRF+K;tA6VsO{y~u(cj5%7?FFZoc^{
zlNPp@K^>H*$j>XnW_B$6v^X=IBD=zR`&``o0c9S;C}2@jiIe(=ORk6PA8kX0uX7Mp
zK!h!@{M?!rpDh;`)<D4;L5;b-!e8vBf&=z&{&=uKOfE5A!+piVZx;~5CZ6AwpC?~n
zY=5G{@F4ZYC&Bm6B!N#-gCT5l+LdQ8alQh%g-20`HApv&&M+&}!8j1BK*BCg?sRCn
z_nel9&7qOesyvJIAIZd|+O|m+-z4){J%cgu-LZe?x%B7F3d7|AyGdhzHQt3gP8lVh
zD}{Z<6nlvZV&ucm?^CkvF`NOXPEYf}iv$=9tf7L6b(3Myj*>r-)N&-hq<%cwxcv5|
zp8r0!7u?;Y{z$8JhJf5E*kuJ<AWoJH2QS{Y;YvAImziFpt@krJsz%qH88tj9Da!xh
zUW@8FI#mcCUkVt!VfY7!jrH~kePpoDZjFA5=8fT*Y~`i#p95;_8TI?>_QNlPJmAS^
z%MF5j@J4peKCg{~1Q5~%&AJzM`Es`VmsA@WAHMkPbzn<spL%mbi5sAav|a8t$b9>~
ze4uqeeC@X}+6kCS=AHEZCp{S?tSlmZv>~@eLvyu?a68&p2cE{Mta+*e;&4N0yC4co
z77s#3Q~pQ<&W%~^=<>|fisctizMfIP9v*J;ioiSYU7*e0eeO6YVixuTw;7SG8f^0#
zW3N>+2&Lo8ZGbk8Oztn9ujTZ=N;EE2w#O*Chh}5K*Cl9;tFkM!HVTX!&r|!FF$^Q;
zdYB0X!))smy&EK9gJmfh7xqtA#!%aw<m)=owxb*iS6L_ajSxGVqleBrA!c2ub4qAE
z_~&Z%5rpt3f<QWZPHhj(T!j?e>2&^`eEnf3sliKIdi{c#wdeLo(Pm3$XY~8|+xcz|
zApb$(r$ZD%^Lu=0Fd{(~@f(|gLV%r*3RW}I4dQa0a`j)M0w(IVi<t^`VTUDkR<`(w
zt+AZp@^?OG%;P%KbUf&;#x%&@PO_&7#AhY&R_2D@GEw2UKgodjoXB7GyGy*#+A4*D
zSnOLnfn?IZK1^PHX1p$Vvm)PfR5bIG{VU#F!%-(LIq8`gzJSh0bze2bL=No{Giob}
zA!$=+rIAqeo!Ruek7tYdu9gI_hz*|t?)KhH`O(5uC5^n8h-Ws`0T*K+{7p7h?b%v7
z(uep_kS*|s|4H4F{)s_Dt+nBKC223en6UJ^BSe4Nqi0B3QmdUwlvNjvSn62t+4mvg
z&31q8sX6)=As%-l;_(EV?3<j!5DCX~`4QEl$J*=(yV20D^`b&G$3x;_iR8ThQt;UU
zEnta++WmTCRERuGEZi#HiBg9_Uh?D#e8%6XXyrvHavmc^;|yX{^7xG#6J@KVeJjt3
zt~8{E(a3*~zI9>L=st;5P4}Yd2b1_?z{G+G@g(uvEvZoqk+&&igKXNMLF`I<CYa(J
z<vqH4h?;d&7jSMsS-EahH|XE1`QYnO9d~|WWVs~Ho6r1&sag|TRKrXBQ)Suj%8)T<
znT6TZ%vf)t0Yko8RMU>sNqUv_z*Awk$&-$Fq|nNzSFSGV%6G2*dZ?xC{N?*i65AYN
z$s!C3-^?NMC=nUZ-8pQSwfm~u<XNM0XtG1eOQSVITBs4|11DlMW3>P$MI&=Q6LmrR
z@tn3GHJndOU4nCg$tIKq5^QS=)gb@eKt^lG-Mw>z!6?m^YL$_BU>4e?t81Nh4Ae6*
zwKFAjwPGVK3w0F4=Yi@4&eBr7w-#t%I#~z1b@}j?4IZAF)ZTOAoi^1o7A{mq)?Xj6
zT~!*6^Mb5qe+4|8^xHl=VlBzSujrRkEyS%8{_&dXbZl^FK~z5yUf!bB&)9s1ZDJev
zjymY7nE!_K?<bRCo<CJ~C&qzV=-O^1-Sg>g@Eor)=)1D{V+YnQT!XsXMwt7YM0hAE
zPy+9{XZeW5Hy_HM{7mRkb2?FUI#m40^~BniE>w_AEK%}ERX+=lUgQ=kdjz>~8V{2K
zLxGLv908v>4H`c5V@XF|)d`aAqR_2e*vWXR5J@jH#4<_iu>Y(Q6Ff$nK+_>!=q7T$
zdh;3BVvWqa%OoA_X}!|NbS!_NIT?fb($5L6V{?!*vh~GZidTn9D$*#QRS2X?R<02?
zqJ6AQz|w*AMh~A7GrH5uM-k8WK3^%!m47Jb*yc%yr<<Hlo#m08&cdE8pD{8woQ*fU
ziNsxB&U35ISt9ONP;^P<YHjFn1A?J29uG^Uk=KFtye_TBY}LK!x%qt1syYkU7fe?(
z0E5chQ3plFffV`l*cn`J9X=F4tQO}EbYG^rn+Lv-E7NKqJfKx*ecaGBa_$F22Cvta
z8I<$191_Y3pWHX&yTehfGu+6?Y|8yC%k^_Uogjn3L+jiHZul}Tfw3kPf>G2=)cl@H
z%Ow9Da0>g|I!_P`#$>l~KU_aZ(hbomT)rLzV9f?Ggo?iNO=2<!9)yJ_%GTziSCv(z
zXxjAqM39;SF%7GhEu$OFPrrqP2?jy}KVP+e%^=ewIUIoqUQjBLRxkX0J5k>=x80qu
ztHYUG5*YNnsYIKbMSgz#!xv>1g%7M}*0Tw+@Q28Wx*WgKE?67xCQqh(m5hrq_KoAU
zRi>sF-nsT;!5~00PVfEp%}?Q@tbj}UKo^hpzU+sD4#LkB=W7<LP`o}cAW3hWOBAY%
zz4nx*WYUBIcB1G`<l^!tR2r5)nZ*PXv9WV_Cg@h50m2>WWX#lNLEm*kINHWfmX=T9
zwnIc+PbfbN#v6Mqx!d$m6ReIlCAlqqmRZ-5;{*YGLe0y`(ED)l_vbJ&`6<2H4+|g*
z{c-1*@Dn{o$|_5u`!hBOYGKMuFSnh)fcW!#jD5k7iHG^HD9ndiE6(6@$YY$b4U~Mg
zU!;uam4yI0-_l$3JDYghHSWDRGs|`QHDDE?OZGlO_)$O(*x3))s>odUmjr!Jb9vBX
z=t!Y?<Q#Dm>tN_B)97g_6tg+kK#P(ELei<<uPToW?=@ld%fBWCKNz|&@+tb$Nu(!x
zeh#zo>1wg<1At2>0)-cY>{>ci_r!FS9qVFP);2#Ek7C~UUM!WfncZMt*Eo83+eG|{
z>3#4rL4Q&MKI*on&g{P(ef9A9M~j!$Rc^ghl{edTzRN+Xb8@shhVVmbxA+-0Z>e$e
zCpV<=)C+xKKhfbfR?9H2HKJF%<FDz0Xa-#e0FT3$T1D>Uv@4NGDtR;6T?u^n{JChi
zj90&(dzIUZq^PLD*%UHf2`LcEh=j<IJzfR3y&T4KfHtx*cMrTmdMBH7)x@f3w7`I)
zBkS?AtWS+_J#_ac1Y;7lD>2&){<L8^t*`ebTL^Ut_9qJ?(1UG277H(oC-CDzHw%cb
z0hOv{n%db4`{5x_i(Cu4{jB!4Z+5G@?>Q!^t%H<(fya`%DhIxAf7Z$vnLmp6iCy7m
zzF}$?*Ijh??9u>#P{ZAJ!xu?XVGxCoyTZZmm{s#cgPTjn_F+GAw*RWV5idI{&87;H
z<ElEBf9#f6*3Piv^s8AbD$M2hY^^=btE(2gdWMS|O~AJ**-DS3L4RN*8+yKl)XCII
z`jyE~bt9ct(i}F32C1b7NCaK3(^GuAQ~VPq8?~zqcPn7c2gxZFVGbwr{lVzl&nArB
z6b88(^Ow2~L<>F%ns_gg5*G**rmV7%1#klykA&kOYlr9A(Uu+#&l5{&+H!VOZM}E;
zX8LL5rzN9F|JP*-q;5{S63r5LhG}i;ATVo~gIDf@-iIw_V@3>m6I~^v=TH&jL;ehk
zg?U$px^4c`{&F*N_NZ^RPzG$5J(R(=5OAH`XCo*MN`KglGiTnA`wWOIdmxoLp>U{A
zjkF1Lo>MdewT@u%UY`0m^#yPt)<FP-Gaz-vMZ{ys3iUTcCm{|r^Fkau%$CbKXJUiW
zH~Fkq@`Gzzgc=<6f{^lW;ox1F!p-H#Dk-<^=yPjduRh*8B71axiQ~;$0S3#~&K-&d
zf$uu5yaZdCV8fzBYb0CTeHuE#r)E^@ltylxxs4PY@^fa`8WLJZTeOOP>qsbai`*58
z<326RZfics4va9mrbKvm7#TlkmGJZfkf6|eFTo$L*D%SnC=<O-0OFE}V(x@sz>BDw
zjso=ZbXcASU7qVQe1H?t%P;*NQ`;dDVFK<C5(6&gJFKM|XD~<`#%*mO^VZ*^`%Dx>
z#RfRD07UquIZSnTmIVTu;WO0y=~Tr!@ohw0IOx(P5;^obF<S)-=iVzJmQZ2g?`7`c
zRD8iulT-y_D@+<lQp=F7$IF|>bmnAP1+_AoE-oIEDIRxQI%A}rGo9mp>6l6zei<XH
zU1PBU6I=80gu9sw?A6Hxs9ENs=Tqw@Hac4u${-xE^v84>_IHcS(>tY$Fz;K(KcM$v
zFI*Uz4TH(T%GHC8?vxMtKJ$0&nOJCTiy7<)SVoGwuRzb-`F0lDymkqW-dNgBQ+mXm
zYbLoBbK0q2HG?#2_`wTtMWvAfF2`?(f+^;aXurKI+>gi~r7eZ&+1^weOh5LG>NY1L
zBQ^4LQi<mXJf^P#V&ma*ts88eaT*MVTKZ-f?iE_SD|mRdm2m+q>+d+ehoS?gikd)o
z7dC!GIxp=Rf!-qMyo`to6~t36?K03QdcQX<PdI^jZHe`9%Gjkx>lawx#`n+={S(=v
z;p3Pa5$qaAxT-&h6ekCJb&pCB2oK&uV?M)E`aGi%_H&iyzQn#o>)qe()2UPs{vxX@
z5T#xt_C0T|5gJ%z;(${^@}G3ZRa(BqUejUomwro<Y@<AS^~Y)H4)K_FcA`$-W(4Lx
zOH1n3RQUTMw6eC>f~HkZc{vc}rj2JO_wG;K5sa;`@(p{}@ZnF0L!2jN76iX5&dg;{
z6PI$$Q{V3+@P93`K?65nS0S2J*iczZ458q!fy~w>5#EswJYc}kTvY~)yXNqVLbF}W
zw+m0W+W}O67J0JbI1<n8#h;wm0>n2!Ae)=OhfolC*~`#i6zQr-rL6@-%W(#`<!$3P
zGS99*Ftw$^>2xWHq_lh=H0et_1R$R2?GDu%wUQ&+jrtu*z)N^rmFv(OQ;XY<)ZwlH
zAjw%v9z7$l+eidU9VDxd+H3dno$Scx)Av@@F~^FmMg(4=MLS{Em&8C<{^D0bKknMb
zHCZW6Cap`#c+@26e9zIrWIZ}G9luk;_8F+eWMc{l_AVgUo>(~N2GPuay7A|O72id9
z;zt(h<p~1w^ue!>ZJ6*V69Z68XZ%98*iJ|o5IpUk`t2UPLXncyov=QZG5qkVTYrWk
zQphEX_zMtzZ2tWdh|IB=hwjZsR!R%{kj2G6f8rM)+SDfwjQ5|WTHYFUkld6Qk!I)6
zFBtsXZs=M-ab8PB?R1~{skoubtL|5MpV}uN?S^JlTKc_mpCOAV{(i8~!GxC#q!4-%
znSY<SzA*m+!r!b*^K}Ft(x{G!mJS*lKtUGXppXbz+yJH_9lF|-agNE_Yc?FVA0Jkt
zxG#(L^Ox%824r5KY<|jT2a_XeaylnVK%#KCouY0h;{Fl*91TrWSqJuK*vU3E!XAb3
z?A8<7eW+15tSoLNgNRvSg^5DSMdu~YJ7ftfYh9=TyW@p)UHv<(8_C3#YM=h%=EE+^
zjPvfb^wYE1hM(byOQolEOcg~Xw(p0KV8{ZiSj8^%P=i`Z@@RjWmn{9MDvErgBu2ez
z>UdTG&d}0G>`f;{$Os*sW&w+28+th2R$PIa$lVOM7l-2$lB&0a&hN>eGaWU(`Rtb?
z;Qs=9?R%{8ThRc$-|_C3o~rPO9#PwGiXUt*L?@U?yR8oVUOhfO-g259tuA^}A~@Jg
zF}1y!+6lY(o!`&3j)iVTS8#1zMaIn5OI`koXPu>ULi;<B=>Q3njpyt?Tc3$TTd|bY
z;7K(wsiy>B)(AQnY|CjY+4m#YM6h``kfb#=6+)=#jZ7&;^wlTK%)|WqUr5{OmrGR^
z%&9n#XlEBquB20F$7bX_nHyo=vjKmKU6TZs%xWI|E3c$1`~h_Sv>$3L^@Rs84!a5~
zx1aTen0OP?-Zq%8I$dW|HS*?)qX3r0@8;9(<yNhY+Wt_lqD%~JFUn+rwl^<I=LwO8
z5BaPPqE5W|lK4EViqtiGIa=?A%U6ma04>N8c=zC-LbqM_tH;~B;_PSKMz747M31*e
z7;@*1vM|$AG4zOFps?Do4i=7q3MXDTBm*x#gFeiTfCLPD%m8&xcurIqb7+t3l6j`(
z9fcuCbBMEmrsH>I-v~UFP*pV&`_Tfoci6}?Br_LrVw`t#217@p`ApVjf3(ytk)UT<
zSsV<ng;l7hkoEhej5)A(rWHj)g2<FvH+hmGD|7gz#oqdPGy(-OaY5`zdeO-2WwAI4
zDmJ9`^J@bPUR20aC=?rq`yJ+mfQ_DP=r0=bH@P4R%rZ_jp@hFi^{HfxW5y18R%N$&
z2POgZ1xs0hX?e41#P;-jd6l2e(Z?GjPiUW%g%tAZeAr-B*$S9&K^8+#rlACY!O@VG
zA;w;)ZV><2Y1a_ct9zvq(To=*CVE9}$(-SWRa_;?&!#3Npy?T52YpOX;yx1;1p8?x
zdw$N^#9?(#=Y^Z+0?EcRPy)=y9{w~174|{gBUM-wq)!l%tGYjXJPU0kS_eV@;7YBD
zu?J;iaC}*KrG}It@GKX0+!DqrHw<SpHm5<DIwZmbNwm;NVj0y(o=>Ifk`9cv?l>5J
zdF{_|kB`_iY;Be?r4J0tVsVQZtIQL)X|Y8*(f99MfD3rFrYV0m*Mo#)@H@8wKSMyg
zhZ%xZA;)%dkKjo<VFg|YVvr}#z%6rGLvlY@52b!^s%Wlaz7uv$?mS`aF`X;@p)i+Q
z7N;=B%a@m!i--%Tr-rppCpkrx@S}OCKpL$g4z|D|B4auMw5VgN?7yq%yK$I8^!9DY
zn~Le{R9tpK#3l5boeox@LZ!Fkh^rG*JGz{2Uh3)!hRF~E@~+=HTYbp;5c@6kp^m)i
zcezI?Y2<81E=b;+6{eMEtyxX3Cr(J5QF|o(uB+n16+Wds*tz3wGVSWNya5*Zgp2#Z
zF~!6HbqU*5{`UmgU@VX?B2oY6jp!uvra-@)I;Psf2E!jq^0V~X36#TZ4~cr!c;#O4
zBg8`!qtdQi?#p68LhSbAp+AqIRe7_0KqyEfe~N3}VJMauz0sYeu;YsrJVAPiZEz=o
z+f&F&8YMbAY*mXN0*)h1Tm0gUh(j*BEgsrqP2WZ=l{}C4VQ>29jC(4L6!$hSl&zs4
z6l!iR%&B|pqAKgcLKAh136hSh#TPw!zo=vgsDMl#RffP2S~FOb8!1-jr@0Q%Py0f`
z#}joFNR$Ar%SymB1f#qKkBj6d{c%v(7QwDhmI(Y#y=A`zy~3o*vQ`a;>L(4;7__4N
zJ(jHZh$Xot?|0*t-*uGGl#B$BMDU(xG>bt3&Z(SArm#KQ((uyVlkAdR9@=~IJf~P%
zsA~l(i2n1!VEZ_pPe(*A=h7u+!!Kz6=PD;Gq1Qi_R}H<rEM_pKZ8?^f*cwX&H|Vs+
zyQ824`%w;(-=(}8mq`+*8?3$Np?_4#>ALCD!8_yY>gQzo!bN<UM;F-U-0y?FRb8y2
z1wfYU;U3Z&;=XA&<v_Ujnqy5GrOpXw_QzTc@kKfkfU^5tmCUZh3w))W%zR6}M^@+2
zN%?94+dyG*WN9Hpn2|?8SJ+M69}{xN;kwH+<`J&vD1xw-?SY>nFN7_808tEZ#sQrG
zQ(R`dMUYmwevf0+;T&FNd2D@ezPfDhBdMT*b#B7$w1Jgy`;$mVBq$Yo3;bY%X?WC=
zc{An7=tHEZjS^}0Q$;McmI`q<J=$@rMC!?6dNd<t`q=}yhcX8>PJ}|@uKqkP@v1t1
z(}CR!)+Brq41-x+^J2vgO*1B!VF48bDYn=Wq@%b(pFVz4LBtei5GE@U!n^8^m(qy4
z1YP|Bf@+=LdU*xp&)zkET&zB~kih}K`khVj#%T-kfgr>^i0_UD(JoIc(l&y0fb!)2
zyvg@W&4>8e;5r9egTDl4z4sROx8w1$O>Zl8XLV1)U!el%e=H40WM2uty;1vOHa0>G
zuh_O0tQV8kaGk{MH3^yAYw>?zQl8ftVkg_?BQ%{f<;3>`XcFzif`|kwq2*dc-NrPq
z(7%6?3~UAp+sTs`0q=$!C=$UP*JnlPpZPd+pzw_rMtxO1ft4w`fA8zP0Rq1C@IiIR
zH5)>($-O&Tm>+YpI<NFTF^o=a83uLT<j>UO)xTi!qe$?epQ*g<;EUzO!oV3yCR8x?
zdd!fFiNXBYX`BPwH;i-?ue5IXTBz;)>1Qs_%GZDI=iGdC<5G+#;uVPe{7YR2WMo!L
zl1juern8j=!nQptkqQjpT`SgcmZ&j*wDFEj&WCjxjZtb7MD4IbN*=E~Y}xLE63H^^
ztA(g?OuS^O&W32xQFt6w+%WZZygGkN`T0=ddU%hlvCZ3Ca(4BL5!6VJ?#*hb+c8-;
z$_E{Xjx}^La@pBLHFzuumsHppee3i9utcH}u~O8Za2TyGXpMl2r*%{=&P4L>y=}s-
z7PAYM*RQ<rl7?e<SP3|yI1V8TUBT(x?gm0Lqxq3)EY^_ENzROX_-P8PMmmsP1HZv<
zPWaW}-`Ns5*nKxljeYio*x`0OT9zlNCd095vp6~&{p8t|#8zo`N%>R7vkN%!rKUoa
z+3hUfqFhRnw-}(N>(78V)|N8-uOtA;c3%C3mjun?-ymk<v~zAC{>~p^CyP*{lC?)o
zXj_E{1W1j1Tc1)MLHiKg)T*cXVQ7Q5p5`N60GCh8FyCvHGrVh0o(0gYs+QJ)+^YzL
z*mdaf29w9YvQx;nxaliFzoazX+0qgdz1Jo;=Cs{>H%IrAN3XdV$dwtC0|gaHGOuHx
zBb)b;5=Lw3hF?y8Px$Zic%&Ci<rr?3G1{=4Y~L;NLPW!6W?1Jm*ODlxwY^&9MrDpd
zq&==V_*o!CWwpkrykuP0(rVI|qX$3N>|TCKl28EyGAs59tr0xCr!B-cV-b_aZfN}E
zoDrfZ)X2Y+w?xLXzxW#S^;h#<V7TsDI`DiUwOp*z{vhclhq>j*=WoSi1jA54R(MH7
zPehWx!OXmuO)b|)g&gltN)k1$0Mbml(z&0O_>W2v3e-c)0G?!Sd9NIWe=~=DPjWR|
z?DN2D=q-vhDW&Qh2Dz!f?(8}fK%gfx_gj8>`9~`UJ0IZD7l{$Tvca)FF4jQx$l8;R
zsj5V3qf`lJ8I_~A|6+TiuEj7KPg(Jy;S+#T2CR+{|5X4R;KA5|3lP4D*>-=Rql-LX
zi$5!Xa}y2PsXyBCZu<-Y@y$Xv3a@2f|Khu9v`XGApW!WsGr+KN&$M8_hKi>Jf=>Q2
zcx-M-APbVadWOmuv>y6dS0}obTSjLx=6dy_&a7I_zuGI1JOFvFa*_=2UScNUH3ta$
zI`B~1-|j5nBy)=1u54++V`pnIe(pOe=G2!$b;!Nmy6g-#oI3&SAeZ-U+eQ}J{AoS_
zEIlZPN|<O}UH9e<`G6=d!@tio2(aD234!B*1~*_O=p`@CWVzA@(6y6}DvdV+^!8=v
z_oo48+_sIy-+8@!4w6MZxF#KJP-En+)=ziQs7NS|@UIRp0l;IO2Bcl?qDV1ik`MBi
z{J866T|CdfdcpjCrT_(n|Eq>^OjS`X53oq$AN}9;^dA*n*suS6(AU_HUL|=DaPhSM
z$MHZJ=vHtWkdgoK8Sv+WcYiOz?f>7|HUG1c|M&clo}{?{KRPJ9iXov2=-^du{NKL@
z{(Gfm2weXEKAzf6N2xSk3FveLV1fVr6|luB2B6V0>5~T5Pfs!JLqod}5zEE@D(5`?
zU-2X%LBDu?#o6L{P9&<u)aL1KL-~A;hOEs4pk(jG!TXZ<p8~g6|MSRwk5#RHIw2c@
zNZs*)cJ1J=#-X9-In*!6&0jD{NKPJw$0J$)8K121-|<I1e5FisJStBE*ac_m_{OIj
zOl?Z%hTCYeM}nKUg};8g>9B4zDLPZX`Onx)5$u0D@sMPCaNPbC&A4NzUtc&mRdkT(
znFAUBCX*`BG5m$%q8{qS@hkPeclN8)f??ujCY`YC#p(3LP6Ythu6K;XoJ^1ubia6K
z*t|)Cm-L^4@p5h(|7mORzP5_3CdyG6o<?$H<4K}q^=J3y_)Q*WNHcaEo`2Ip;L5+6
zZv2l$4v4Ot8>!fHQm+I?T`^DF=0eM|eyAEF_E4_Z%<P}L5n#alr-2X2uTR6kt^$=X
z-!!Q-S$hHW;(|mxG4vaY{|NcN3s%eh_rcwaR^u%Mo}4XE`+7avmf<M*1J3lK^0mHx
zgLODiqW9v+Wc;7=UmocoiJ(}bZff*}dxMpBO*;>QTa>!1yCYd;{c+x&I2$K>+J85n
zN(BG+LEA)aw>4h+nZ0_|*+M<z8#Mc5{q$hzGjeQ+y8X1M3YhU&MhsbX!3j9c$-g%*
z=-Mi${pSvg|Dzg60u+^{p1%N`6ZB=VMI}1V9Txt(;JflGd2POq{PuG$V2wHQj@Sa;
ze;4de9sQ5Olv~QB2j1I}5N`mx|D_)>mx3$}6&)M1>V1-K{i|7kK*4`+by#3J$RmX|
zY2+i!V@peXYvSnUH-m|5HmUhnojdpd^+ee3Y<GDb6htGxj+`>CDqH%JqQVrBcCVuH
zUyZy9gZ{0IfF?0qH9s70#sgxOusQd`&gQApqoJv(;qselMII0>v4Ads$KTGQ^o?UI
zy1QX8CSI`e&#nI|UH^{&W!?uAS|UR6Ttf6l9K_S!=7qIbGRsrr>tvZCuYol=wH{r?
z0Ixr4`)BBTld=EYJIN8#PKr8eeFp3<#l*-eJ2NX>1F?C5vUxs!iL%#^@=s5y%)8@;
z|BjXU@7P}hRg_y3Vm=2<ElSnhUY;7^^&9K!_oZa~e|jA*3@4C150~Qqad)tqBw`>e
zxVuk!vZ+$iGohX>%-BwXN&lY)^(4QQ|0{O7|A(@#fQl;m-W|kNK#-7Dx&#Rs8l<~B
zMH-}28kJ5#x+RD16i|@{$)QuayM}sa)Zb74@4dC&dUq*nfP1+o_WAa=zrF80;yf${
zDK~9GNaDk09x)JG4)*hM;CyASTHoylZ{|}E8BcKh^y9hC6Ko7do;O?%>-J*~^Vlv+
zy5qh{g56aFLXqe%qKoXN{PT;y+cwDqzp<3@`8@Qb#Y%9+)5dzFuc-9Fx4sYkU$<Ba
zA$cVf{pI1rp3G6lB|lG^Qzp1yv4`_D(p$wQS_N4A+2_t)@^3qlxT5tASZmT}HFhuZ
z_hE!q37?(!9I~~^z3G4%86dtz7r(szs%IX96?L8cehV98wI=W{Up_glFD+YvFmwR9
zp%BBPe^x=<cXHMxBS_l`tYZ#}<KOF0v>LP(l1yMd3iD~nHA5jvcQdf=9xl~M_CSr7
zri#K}Sr5_arhor{_3~L15~ZyCS@Mw?OK=Fx|1}P8&X1fPc6-w@P~}H_tPPPrItJtS
zdLYK&4Qfh`MXPt87=NGbEBcYH6YczzNP%NKHc0P;zUnyyqKs^7B%{yi&u%0c1nxES
z9{BTVB7gsOnu4MP-ZPBQ)C8iW!OHA%#rjE(lv1VnCy&ffhb$!aW#M4msd>2JhSF?k
zEA4W$lKys*efR(A_@CupMU|3ou#D&CtfE~{ue*)sJpAlaN!9SL{y^k`a`gKXaOJI6
z%;&3gT9<9aGf9JzbSH8ucz61O5|I0q9KUm`_j3w`SICV_->Sb??J#MhblV7Bah%}y
z)k{JDth_n?Ee>_Zi6Nq41%*SA%bdAFd(AsvwABYD2OT$fJS(bWi0h8{bn4R7mJ^By
zG=DqnFWP?rUjX$F;B$f~zK@rqvAX-@D^`P;#UuIAO6T*MIse~xWsCoNS7Z9!_WOXq
z^kB^1PB&QI-OXS{7u*=<P9fe1BbFbvZI)p!4CVU`u#YkS4pZ73*N8NF2R`?6UD(tK
ztaL?oB;iMG(H-^n$lJ5GpEx*d^P~{-*Hu^38xy@!+n&ytACsKFV?WZVZJ+UHn+)-Y
zzs)c$ST(YgP5JTBI_8gcJhDga7SwvXytOjTIoe?V?0rOswC*k!mw8>b;~|L$;_KG<
z(I0PK`w#3Q43#9*o{=IgNuS~UGbeMN_jd2hTjfGUoa=@i*z;LlzK~?`HqZ>6M;B__
zlnbBHU6=hmDt#1xdzUN14pwLuSVuwEIDpTL?*o0IvPUc|G$dL*ySy0}N-2RSpNjsU
zuBZ$A#bX9Ln6INn$W~}We{!i#cz7!m-@|=G-J#goz&Yc`JqFpQ1YdqP{67{1DG>k@
zSEz4yhe#iZb=fDY-KH~s=^bZ#Q13CU|NZQUOytrugNXe?j)kdoG_c3FAO}l^$2wQe
zJTwfwa|UAke<6U6McaPNi^*kCy<>0nV|PGrVZ&-dVH1=BevTz$etr;xo4cNyD>DB#
zBY+r#+L!;dushSu=G?T#t+`Kqqg<y!EEV_HmA`3IY(1G1az<XDV*c)z>65=BNfRMc
z=Y3k3;kM&)^tIoot;yZR^tJhrjGCHtFCOXSm}Mqf-(x8(UN<Vhh^%x#$^;rgJ7N4%
z$Hb(T=hFjZL!FA&37^*U|FPZRzecl$WCasRE5Jfr&Ry7-Pj+9d5(<!4Qab+ZHr1;Y
zY1q?3D_Ea=u9RVl{j*~_qwQ8bp%O?(Bab0*(7P-r-i!E5DF|*ob?Syh-Sw9W69{Q*
zp#KDq1I@|c*N2hRlPKgVl{x2(*Zba<>4z!OLvmHgP>8bLVYJDH`paW6v5l+stTe-f
z?pvx?8=s%ZJ`wqSlV1z%=Y_SG6U8c)x5(xdVy%^bo%^E$&X=Eu@UK0`lUHgIYA6i4
zXDKI9DZRY0$X_^~SU4XZ^F<*)SDLISdE#BeIshk2#6lpBnga){#=x;k#eyeoxs(QM
zX1jzciCtf-nar#(NdATd?}Y!s<?gND1ZYGIsAG+rkYn>*G16ks!(uZb!uv1&@TNC^
zjrW;bpqMe(*UZ%F7w==qxC^?7S6SeAT1+dXM?#|^{m1<*z5P265cLNM1YZ=)6pEvA
z_WjX}b$PVg_VhlKHZu|U-2OEZ0uTQI=_kNVBb$W(h1oT8T%^I6z-ttaJXE$c{zJ%2
zZ~g^Xusqn9y)(X@G#3?3Gfhu_a8tKE*Wd2n-PoN+vN?b=<Af96|HB)XQvSA#fFe=y
zKSgy3cwatndazsu0xW6^$8eGh=f6Ln(;3FUF@84d&=t1zb+^Ixo=+$sM)iLnA=mGp
zKSfzzVg)J8eS<=@n{l`PC=4%z_kQ}GqCRgm9{W~J1GUxlpB;lSkl%CgQugm0q$yy9
zcR%=o(pLcqJGt>i@+NkSC4TLy-_Io@(Z|0X(3gMcAE44=gV#7-;X#6ly-t44)nt7G
zQtpA(2IJT;fa5GQFe4-Cq`xg}Nbv6>Mf;N-0=H5Be=szNcMLL#)FK*wKiZAXs^QO`
zR&gW$SV9IeTzocjje{jeF2;lL?@7J?oK*f=oD%Gu_o^C{y52R?<>67?s+|s<Dy{#H
zrm>+pI9BuNcRSN$|3U&J4F_<N40{>1)z+vS)*IpUq5KT2*(C#vzt4T9_V)&W+z!1(
zvhgb6UOuOj7En4vtb{=iCat!C+<1RwzW%2Z7|NPf;D@JE(Y9pjuCE=#8+DM21q@`!
z-}wVz0fNdHwsm1*=p=Tz<Hk<#poKS_=kghAiB+@w>-7`(A3mC<@IUVF>xgP9A_MkR
zoFa5Hs0!O$+aPQ=WQp$<&@<V1rTgT4+^&B-fa^Lt?vENz|Evyzf3OqKD$5>j72vZU
zyb`^SzPUy;uJUH&8>VlD2kgYR>6O+WN1g<p<M=%Q5P^4wQDEOls+%?rlhcEh76lt*
zdN|mjGz`#4=0-q;WcJ~W_*5|rf?n)3P$##y+bXq@;L?8=ewsd@LXeWtr)EEExtu&p
zt5Ysqn`+AaFnud4e*?U?>pv#S{NFYCo3xlu<ydULoATA-vZ{gk{b4=-V`6Uq1&Wo?
zH|M4@`QDD39(4{cihn#e@?&nT`g51sPydliAl^3tpGL++QK%q0G4*a<g$Z@ot-=q*
zSI4a_oY3<RRz&@67YIdRH<081|4UEel79zr6v#XX@f3HsNn7~V=4KGS|0ulvc*Uh^
zz~vht?jk+^<u)(=HPGlvN|&v$t>kx8+_!{o$4eXADT&F<->!EJ=ev06Y<c#3p1w@}
zGi43G$D0zMT4AfLfBO7UPMdYHBlYi7O|kys*$^30`sfY^y9MF9|L*e_!M~&QQsLnj
zl!zK!wAtU6{`af;W1xpp$o`)~5c0=J1*F>t@Nfs@E`8wWd-k_OxOKlr!x#I@hmH{R
zgF>k+E>cptx<4UL+&{5Nuqa(+nib?}E)b9esiN-8&pV>y^LQZIVM<SucsSYEheykP
z-yn_g?+tQQ0AGTL(|MCe;d`8KF%h!=>DRZkzq60HGYb*d#zG;A?{BP!-p6<Sznn9M
ziR}B*Et`-dN2^EX2#w#2;-%GJ{{gUTWCIe`&-Y*Q#qoE&K668AX=$}6M1Fs`eM!XM
zOVcosLsb+Hn&89qAA9@q_8<D#+U}<e+&+VoRDyFJGD=LaO(`mRp4`KM@w?Z-iGLvl
zqE0iO;FQft0{mBH^>T3a%=A=0vp@A8T1xwOd`v$g!L;u)By3k{v5nNp@o;*m+;|=h
z`hHN1eS?7RP2KSd9#nRk-(#U2X0x}@r{3S++KPVcu*HS|8&HA6o9Z8)p8-6yxHfPS
z0GeC*``sMeOSqvnCEN#F1au@m9Jqqc&c}xDR(`xRb^zK<{hrtpBy80`;U1l8|097t
z{?|)VB}&#^^A@p|__n;wn*c=sp{Nde6jOzy7@uJoZXhgZ(D5boXD=Y6G5cd*e+5^F
z_qfq-p$SBq=R@k<!_%RZ_ncZn1&FxMUrpA03rm*6!+1e7;aL47;P)tY^Zs29Y|=kJ
z-~$aZ0rv}LzJ-j*YU?OwKu)UxMqlZEa7$>L6EEMm_rUM#(v1I(X|5Q1Chow~)#`_7
zBpo+xCLM&tx`)A<rks#XVJt6F>$Yi0XVSX<8<D*!{#pSrppUO1VxPk%+$2RG$7o$+
zaJv3;bPn%7I?~I=f0nfo5LW>H>@$O%x<T;PQI$@f0Vhp6xA{ZyOu7E{MERqOp&!sM
zf$-u`M5_U(e;<`IueKIsp<QHIkU;Po=LLHHT5I34KOm|J5i0<6sTCbNMC)3C_C!Cq
z5c91nu<OUKqbKL@B68K%{8a)nT=38l8B7|qXdy?a&c38R)A6`<`Wc(D$VB&-SOa8i
zEAaeg)BfgzZ)09>SgVZ=uN*9Ir*C6i(Uc|_=QOHS>`eQgeyP6;9;><icXZP9%Yfq$
z0LCQ<Tvi-rfWB#Soz21O*R3!*O$^JG367KzHrA3Nd$rXkDBf60J|jI^)$(Yb%M&J7
zi#^UvpZRUgxD;6%3_HM12ODKsu_CEw?k6Afukh5zS+Fc=kAv1Gf;&IwgPsNZ|6NN2
zl!=nRn8XNJBw%LhoI+df;}H>#_i+xvE00&_xw(zac`MTGbMvc@vma1Ylg9v&8X_V5
zLit-qditILMm-4cGEE_-9>1I9EwV~@HB&ru?@4AX!<{F>%tUz9RD!pcXWWice_S@v
zGwQLQBwJITV2<oEJ#(d#T%j5FJkpzHm?Al~E>3+qUiNy`sFt{EQ^e@d12@=5nH&On
z>GSUz{N<v7NcfS|=j5u~`KnYQ=}qHy)8^izw~s*xf1UYd66%W%stg;l*1Pf7<MRsy
zM=4)a&%Lx64XGVn!n!^T|3ov*KawACflB@}c!DUYV18D6K#d|r|1Eo+`ogUb)Sb3$
z9S@in`X23g+x=XGd_XCl@VIiBTwDW3-rCIOv6R9Pe~N2L&U)->(emlXfR@6qF?@S9
zaI|LOg@UQ}W{GhAGBypM)@EZY`M9;_(W}yV2+`BVtFRk5Y@}S)^)3esldXN-taom#
zCvwfGw+BBZ5&y8530Zba3&NJDkjoqK_xwykN{f}+_=L<nIakiI7W=i!G_FWEazG%E
zmcaG@3L(F26@neiaMPxJfcI*&<hVuFKWn6M@^P~4y+j*SKr*p^HDjO3^Q}kBL;56_
zjCv)x8TsP()KtR4^tAn5A^y#4UkkZ2vJ$eJQYtec9vI3RgsmSRz%37B#!*%Dso{Ad
zA;{B1@&%z>PI7W2`fAQUHml;SETaE3DByoKNIYD;2<#@&ZVd5nANMlVg4d4hUFz%}
zZ<Ns{2{-d^&E0$nfn@5FxV(5zYxs6w<@f@g4y`lDmX`Kc!EQ2G`Ki~f`N~(wN6CY(
zTTnzQBy39Q4t8MY`YNF-n*eU)wILD?E~j4XhV)#ugKMs7q63VuZuG%md*f*n4Rt;K
z-7zBZ&WS1-t4so<SD-Q0{;x`k94-Ma5Eno=W2YeK8g<Kh6Jbo{k5o(@dyBHx*$c<r
zh7WIlSFyY*8P?hc*NRRr`A|!$w^u*&4Jp%eG*<!RAbsb?jfS8|(V|gdYnlv$V-g*8
zGgsz&fjC7gfrNA(&1K%OlAt<y(kJ#3D2O#Uw-eZ}s<gBH5pliA<gT^|XHAbs-#(xA
z)902(4qDhnL#XM<W<74Qz4(*kLVgQ@m?5)M3)(TL9KtbMr$68FJcQn4zP$PJ6{~hX
zdNt_MOU0o5FL3(Z^gE}x*%5cp)spX0MbuO(2jsWESqKX9k5ge6x;hbUWx&!|!lLJS
zgxGbZz`7$RJ=y3+jTL{cBbcR=s~ooT{rzQK<VrIYLRi8;@@G75|Jm&RDMeil{z#!~
zlCNoqa4PZ67p=#?cnkK;qQB_`(8k9`xYhADAcktgpp~I7+!tnNnpR#K&JFPDonIzx
zx5(zG6r0ySm^}gpn@MAn*UQ2$*b}T;zS~q!G=4UN&RbWa_^I~(V35ac5#2=@j#xH3
zN}IgxvwWe}6@QoN?dj4xq#dVRcaqkWc*BLs=q3ae%<$;^gmzEevmc=kid9M9-FeWE
z&RTS}&FHU8S`it!^Ih2Gl^{K>0`W5^#Y64QAQDZ7A44>=szE;T$WfJlRD;}iR+w*d
zC)~d|8}GU0(pvU3TQznr@55Tck&mVcOBQ$eVl}5yonAi*A1>JBD&?vihZt(iuBhfe
zR4^MnN@1%Ws;q1-qnvRIJ)~8c?61{O$&YTF`KYM^DcjD+(7kXnWIK9ZSa80*#%|Ri
zx~`MhNSye(;xp`+(mxoD=xjx<i5Q<XdB8Zj_#`nlue{$9Vs@~<7`5EQ=eCbH{m_9I
z6Wgozc?EXZh07%FWdJSKAzVAV^wK7&_2feg?CrIw@*)<zG@e1TmOL*Fy4v(|xl*qV
zf3)6rAQs3zaDZPJ?()LyoG%j1tY}U;LkZ_?o9+jiW?>99X)S;LG)wRQbG=`ZwB~Ax
z%S)2?dc?yU$My1Q78VqZBr5QuV*;hsC-SpoKc#eULn0(<Z_hsXkY#b~j;V3H`_=0-
zry<!y#1L-Kg@vO-ShM4HdWOdx6{e~U-#PO_D_Gn@PPAy2b(}7nsP)RXhY#t(UihM8
z%MS<#DI@%vBgZ-t1A0B=InmJ*WwC2t6|8}&UV|YvTEF{iIlmH>1#7*$pogWD>`Xc4
zeP@+?(+G?rOq}AABByjM;Sm-wgX`_!C`qhBufj~>3w|DE{SEVT>Q2F%zc-r|4?_Zx
zv<b?US-u7a85a5s?^NRpdfrkl-g#e(-inz@D+vlBbg47H!SU|rDdhauF0ZGe3QVt$
zkEiSivrSLea(cYMbtFy>)QarcOx84eH^PY}2DQh!b{GC7U*}zhs`dLeQNE3}6U(un
z>AJOUr^gA6bCadvTzo{Oo)!*<JIzI3XJU$1JZLkfZbmRd<vcwQo54$Sn#*EEsid5z
z9nk$$FFTp}vf5SijqyrWYY)#Y_sR8Tu^-7-=aUmGSEqW#X4<obwTaYMJ0k8qkUia!
zqD|3H+Y&;SV&;`^8tMnJ%s$jS!ThjuA}3}(k$K=lEMdT*%kN-F^CG;>t4?9UN!abB
zSa>i42BcQSrs*E#){db{a^eJ~Z)%+PwdGf`_aKj1CGAF!1}fdI=?RnJdCm-1nI-pp
z`I=>cs;6DEbLOXVN2<Ww^z6cy>5<26h`{)F6O@(GduvUctlNW<l0!*XnY5??o54S8
z2j{gbm>SE_=|(^=Hvkw!U-(E<6IbRXA#r`qCmNP<v`Q+K!QEurvXrOfY7C+N&2xcw
zDp5zVmG0mLE<AfD<e^xY`JJWp+w#6fwskJ{{pCY7J`O#R(A>0`OpT{8>oZi>fs<ms
z%(+Wg9bm?Pez_K5b}as};#Kd2;$TERPKzcoe)s<L{g4TG+>9`^zyYYk+%8%Ux1x0i
zqepY!=+xlmbhZOU2qsFzc9>OFV#Q-Jvbo)!O)({_Tnh`gPa3$b*<4U(hc~$%7TnZN
z(gZx@6CZmB;gJwd4jsN&axG!fxb^e?#=2x&WBR@D`&MiCJCpQ~F*{Bhgt|Mz@9Uvl
zBj?f;!j5aaj`QR&a7uNRkh#ahvtaR?3`Qk}r^m*oVGb4KGzYxCA6rOR_#4!G;<LL{
zQ2%_|6HH9}t5tEG@Sd};YU5a0*3x&b95RpO`$Uc*3)1NqjMz4PbEYa<s6Xm&VSf;}
zy|@MO&1D-QexOn3)SFT%aqp=P;RzX51S#L;$wDeWkF#t8V$^_%zs$-qtd$s@NPT8J
z^|QK)N;Jb8r>^Xre0sXqbg*%!leC#n0+-L+rxRitjc@r(H-F{jt`}m&ZNC**C5~bV
zqish0%GB$5G?pb#li8%IqEcztQJ#;r5%K27zG^#siN1a6rif@3goCMo<U7+RyXy?B
zc%p4vNGOj6KyBf!-vINL^eQP(@`8H8lFIcQv&DTij98WDPI7&BzJCYIhvkA<Wqp_R
zRAu9BGybJS@K~GqBN4I>6VBTl3`x&jf3y^o;pe#>rHY4td4%Dg?4EQ&gvMLGG@QIZ
z&Lh1(^XLXvN?d?2S<KvHSAF>Kv6{bebWgwx2An<+eL+~P`p0RuKjUh8Y%|5Ll#f7+
z&4pCv15c7=gj)J*7_Ih|<eRV9I15&&hT31e#^)S%cl=m=9_B06=%;2NAGpflHXmU-
ztSd+j+!oMZvrE##oSy8D?!1t-+?FvnH+Q%9|2(shw<r5UP(Yu+rPAX2*e-V&%w*O@
z=Vqg<*N|<;BGy^m&16OvcJ(s_$89M_LP!KJR4#?@G%%_!D41GnYi6MCwG~|`nx{v>
z;irtq<t9$txQ23DpTzOvWLtGkZ+EN$3wuOp!I6B4b9Hr}GUjUDMMsNs_Kuv`nyb^n
zkjCdj$%#<goXW{uUW$hLWaz!eXLq-p(=!sMG6nqGxT|jxOiRs~IHo<u)I@TfTPQ?D
z3_A5rz2lk9IQtSb*pc$Nl}%cC=9)*()$p1hWnK=W=HE3Io;zUUv3cob;nn^#S)K;_
zt}o)OJ}b|7qS5$f3(BA%l5b>s_@ikU+hkIbkWqtbx*FqyiW^EkfJH=w<EJWr!YvKo
zluyS9q*E)V9%rmOxQSv)NjX1vPK=%PNu?qZ3x`fW$~=Al)Sc`T*)y!o?p8z#dwHw4
zC0^oTF_+nfJ_S~(b_1O=Ie_+B(xCR0QQkw<#drIJdg@o2X+4eBRDk8h`oG{5oeoQr
zHg(tca3=_+tbAJ5m*qbD^Z|&guq)0_@dzMDrm~0%{__IH(ymr%(*4f2a$sQA-kihc
zix<^V?8Q#amC_x)<w95(+EVcK1l6^{i993+m>8X6MX~*GjH$=({Y`#E-<Hg^SfM9-
zWqe$p57QDYpiSbQPloDp`X+DGa(YqWd@1i8(!!`sC#X)!z~ZaBx4e1m3rzYP3?eDz
z3Pzh0Jz7XcM_66Vx_72WXY=(`n#x%kbWYNSJHESR)O<nzLlYF_8VN>^Ev0$HTrv$(
z**1@-g&_D2o8MNeL#ciC+B7{67lOv<jOLeSeXZZlByp2^?4Lt(<QPXwtg4n<rBU#K
zU)e(v206di=kE49=c~iPN+WHcz`VPdy-0Vi;;74~xA)B^zIW1_i$1hL=lFqJOAT5%
zitKNTiRv)-RjA!Fio_U%m?K>&`Idwd)i`^vmJBVwj$lj?1QF{KT*B5JVP1n3%5t%s
z9zuh3%W%2j!H4dPIioINB`TDVJ<+chjE0zapQ2+BYt-u>it5n!T~2pu_JP-Vh1?5%
zL0h7T`&sCFgqHKx9rWcoLbLuE*O!k+JOs#)08}VO+IsuY89eR8;6~wnoj^~0+lIh?
z;LWT=#&rh5f%zv4Gb+XTF1M+ZyQ!A*Wef46t6tW|0DC$M$LE0OsD3nas&f@-NHl%n
z|8RV;EHdh*iVI^BF`_Qd-R2UYiKaxLs^TYv+2Dkt3~*u;1kISFq8Z!Nt`}CU77Q~3
zl9I~7b=`%EIMe!L%l)NpfO;yN+4yP7B`=fH*bm*IXq0LVlgU3o(-}}wb=~&{VMNiw
zTvHzeg5Pq;WO!?&39?WeR?kfCQyFjJB=V41IFq@?btF)+N1cpkdae}4S~?$DIq=XT
zA;j{q$cn49bx1vP*!J*pu-pCC-J$-`4m8cxBEJ=edzdh<V%VI=e&9-Obxc))mbpt4
zEmxPIFv0rF1bnA4)}939+g_Z_R=2vN(R$T3(k6f?5q`Fw5XP@y#RCoTJbU5lR#I&7
zeUbA!(P-&HV74x=hmf$U>UOa;*QIcLX?Jw^=i>(gR61K{tu_*CbB_MQ1!39N0qv7G
zU8>bZI>+bAf$P?(O&hg^ao!=o_J*xYhX$d-hDVb)a+bG?Jd973)9t-#mMV_E7Rz~p
zd>O<p$Aad_O>uFWxc>kR6Oem1v2D~+y(rJMy@4>!=R*0D^%>69DWIX3Q0I~3tz<m6
zsHL60C-ySM=hzeTx10{Pic<4(eq3!$6E~FVgcV;Gn5L<i@$IH1L5|9C;#$ty2SiHt
z=Js~bro&zSd{ke$pY(ZGkhBNUR7`Iv@~VjW;&%8X!G2k4;c`>Kv_lR2Cl%=qj`n(b
z_PB^R&&eAMdT2zWW5mSnz#;B$#K_1o73B6gb5@FxtYg4f(zlXt;g&vO&9cF;!H#al
z8(wEd+f`aQ9I;&T;SRxirkd*#>>tp?yEnt&_3h2Afp$;0FEy3&iXVrr$eMv>k$ECo
zCe_-6av~f4iyJ5$tNKn0LLc90QUfyrp$a=bQTFZog_#sQ%(m-4R-9vn=%SlN2f7n7
zgYz#sKhESIOFF7=U#*NbfanDPpv0iOwMMq*C_|Rl0JM(#5O05Tco9bQBHkeC>jw<6
z7cuy~xko|V2@911iHqyj&$kXg7obtYjm7Cwu_0m)f|>yfv!242IZ&Ouv~Qol?6ghB
zEf9T-S99s>{_U`sN3rG>M+7Z%`XfMsWg2#kQY^T3&D#86{T=#OIauyL9F<BH2*v~9
zG1auM?-#9(KEuhvybPBUA*xVuqq?km2e~UO)*YDZ5DVWFAgpd8<)|C8A=`5(9~`}d
zrTNYgO$IHCIN2{dEa-JLtfFY%^jt&%;=4d@a3V6+b?Eo#-c)C3!$9S#7H_V>0=#%3
zIU4LPtS;9^ITHN+x)TG;dfZyW%IKnNnVT=<h>K^>r(F~Sgee*$f=cN)i@ijyoe%ue
z6B}x5H(6u(FMob^b=_T$>cT5d+-!v1(<8jbLSms&?&@n)JJ#gB*>P3MO>l3oLyr;4
zsJ@Z5_plHT-hA5CRh&v%%7p&dnkeX0RJFGHvgd@w@vwncr&6gDZ?2AH6&B27vV2I$
zJOh%Mgjh(MhN_Pk(-3cA!JyayXnTE5>c@u1VVy!^*<~W#(U5jR-=lhY9f~S~F7)7l
zp{8nj_LApga?%2&@$HTvI+4gmCsV<}!GZVO!l)h+*1PRU3f{Y6VvASe&PK#D%zs*9
zJ~8+1R^kQRuiqbQiafO5@QQis+D&OxbEl|;yH@9N1xr5l33sE}jJ`<lyuW)x+`(-D
z;}&?UME6)KQI|^2G>fMbZ`1726gR?)?%@j@sS1*zkb}-!%UZbd#adKU;;+>u6k#Sd
zaRW3w*gv9(9#mpl&oq$Vv4c9$*$R<v;-ti#W0GS90)gWoC%bU|Nx3NTwHZC)fh#5M
z@agvo6j;pJ)+g?K@Lq8W#XPfUj322g?9C1@^VH}_lde|2bqeza>PdPS9=c9X*YQM?
z(_{b;pE>!4I0=KP4b^LGDgJ=rMACc|mXo3#e!dT!In@jDXlGt%!V`Q1!xy%%nD2i+
z;D8@*_O%GF5|Y`0g;FhqXMf2gQT>@XooK}DtNHak?2)Dj6(HwCv|YQ5o9h8wyVG8#
z_3Y4>90eovLhQ9x;Jt)#JNd&LV&}uo=YxeeP!?*73Yn?>g`;qV398#th$d(3#@I5*
zci-f;;n|Z(a1=7wI9Rw*15EWHbb&t;(WWo17N3z<9cE(YA>>WXww}>*Tt&fGta=qF
z(?N@R#m;^xOfMG!po~l_CKf+sif^gR8N#x?)Un;WDwpDQ5Q_^MwV{U~9ZuEeciRNd
zb6=iC!<d<Ruee?y`~m_P)OS*2<(?il6;T$RE%P_xplzsME)ZO4A2*;z5Pq$T+m;9v
zWENgq3q6)apSfLUX~vPA{$_hT`tbPsWzBxwZXhkQq4PIQ@a{OJuJyo9l=iG|XF4=k
zT0L$~GQ&-HZtV8)+Sjpb0#1Tfo3x8%Av>Sg-e6`HeR3VF%De11W6`N&VSop2JyMmg
zxu-sjf4NuQQSSw>plD=D4p3Hxl1@8GOwlt8dANV+b-ErDpU_=k48_mk%{QA1%;utQ
zptT+Nh@xh84?tTy>we=2E)Xj{FP1`2?Uy-==|Zszxp1nVkOA|!x?Gb7QcMeF9IhIw
zuX5E}k_CN+v@+DB17XVBhNil830SE`;|#3iA(o3aP7UL@<`ZU=lvtaEM)&=DmzA{o
z;_TC>6^nc0-=t=`pKSrBT;6G%tJ8j}NH-(AXa#5`^83@!wRwuoa5H;CXYZ5y=y0Jf
z1J>ng#@;CMcF|H}iOg?QV85z*P%A1Pe0~LZ;o}~Ff*Z34Ja2bgZcUd;o&Gc`!tAZH
zWeOZ}-in0`0@O>`*==)bRVYHP+N44XGC=o}ElF&FSB5>uj?CQLg&)`pMFAEo){hmN
zJd1ozHzE3|N}UjAi5v0|MnW-e2G@?Fl}o#^%pAuM9bAbE?ujq?Giw?Ii*4(9$Bt-J
zdMo+-_+*(~mp`g{%gQcKBzvb&BW^&8Wm_P;NBQq}3CYOlNnE8FWwRZVnjDqB8+Eqy
z3Yi>Mb*5@qPsDUz8m+hLaTOt!I{4jfb$Q)gbk@si-|5w_6L?S78l58)6F9mBg{R?}
zPHTi?Ph-(;gzC*zyt$-kSvK2KdTXs(?|g9N&|pVi(D;T@`aX*MBV^bRH;d7umYP{3
z_i)acOj>QMKI}R~96asC0CPc%<!#A}3sX2O=82<V2o#K0SxPup*Y#ivJ=pJt(?`u?
z<l{b-R;~Khl$Bjysh1gdv*lXJoV%{l{JnKz*@UvBnC)TUj#L7?jZ}b9kP;I&c<zg$
z=NM~LS!jyTod{r<K5L9q9&n7C_6`w_^rHpyD!av-Sw^SK;ah6gA&Kyl=?jsRQsBgk
zVSbohgN8{9uh_Lw_abO|ai|^r1;*E#C=ovW?68vsF7j!WFpnR|?b}uU!<4S=Qj@Mq
za-=|=t^K%LzU(w^m_qcTrOonn)Whj(=|SGjT4hbRrf&JR#nnbrs)<iv?urni7c>NP
zgp4{Zd+)2sKhASKRSy-ooJZi5`2b~UZ#1WsdYozFUh(k4!4SlxDPTe?`BE?nCiQ#{
z{e$p4%wS7|rk4^n>^1(NskGH8gx#A5`x&1EZtY;E#M#gr!~B3v9nnHIE9VOF?hOnR
zC6<Z-z-@fp;;(@hbx-)AHf~6qePfEY&>`>NBR>;9${9XsEP8Q~vqX7Fn9O(-TMBD4
z@@k;0<b$-Are2L0UD%#}rg#^CtIW%^%%>2}<aaYw^Yb?q@RIhVa6awE+yOqX=IF1*
z;SId_o4e{=O2LC_%--F>r&Sr0r@XG!RF+CJ?Yfw6Mnt48QyP$5zm5RiU&LLai31)u
zs%=u3?>ux68fWb!><Br{-SNcB^RiY`(-AXdY0F?X3nqInVlYky9%8U#Y(pX9Ki}E)
zjq<wOndCg(s^1<^U7x=?5qW7jZY~uSBqAmJ(+Y*%=4Ar`ifZNN&JbGSftrOTA|YSI
z@7CM4aF1cqps=U;c|fS3@`^+;Fi)3CUr|&9+Qav+cIZ4)lXh86&t-%jD6Qrk8_lb&
z#(4C=3h~2%PAC2Jcq1=oup}!|y(U+3^W)NsQV>t9Exkl|npJ*dS5bQu+wrb}t+HpW
ziWQ-mXSQllL_)$*ay4o=sG_1Gl`=K5QQSU+!pFn2lXXZPI^*Ci)_Ok-$Q1|P5Rvd_
z0jbAZ?BND0ZR^>cmXb^EDNP?5gOtsMVA=RH@O=(tj&C4B?K-_WN{6iJRC)rzb4CXp
zrm=3OcjpJW+EAPHM}2+ls}D-O=B0T2S$Yb^*JTeqyd2x()b&~NU$+7+MA=USmJOS7
z1GfN>${al;DfyG!;J``e25)XaXd9jyC;tOsVT|#<C=NxuCMDS6sxW%EjG9H_ee~r>
z5}C-bM`5eYDsTO&HR`SMdm@ziXd%^NeOHIjBd7tY-O@yQFg{(9(DY^NZY(pfW*SrN
ztDgwGk*1nU5V#b1=_fo3qZoE?kdSPfF5Dhm=zm1jkY9=BcRL)-<QlbVTev*>iCo*{
z;WLTuws;ux$fr%a#5>m>AHR|l?)bzAzM|aPC*m0+uqPQ$DX7Vgw(??P<wc3G-TKCW
z+Z%z&%QD|qEUcQ{cOXU_`w63Rk7`+iq|Hi>cX;VR(;aLM=k!PRcrn4!qB`eidY&;j
z<vMJRo6BA!q)aKiBQ22<YrKO54aP_3#cK6k*HI$uTLbQHUK(QJ99#I^Amg<{y?tT5
z<)l5Vfc^@r_`7>~zMN1fWT5D=7hc}BZOS!baoy`QnkE>`(%29?1~p)$?vp3SS3_@(
zUW?Ls@sL+$o1S!Ltu#{xYu~>|z)_M@$&bCSNHaugzIj!Uz4-F2k5@D5t)oOdtUJDf
z>GBp{&MGoec?O%Ab+uVkhfe}v2ZLv2s#1P}7lqts`5*|XX>VTM8OL958L~^iv|`dS
zHy77G-u20eda6CgKe{ZZ3**b~8Esp+tJM7PsnmPTv-*b^?+DRJMrV>6kw%)Z_k`8v
z<?_Prv)VMC^@5o#44X4KL`AtVN{%wbr}K7=$8i{lW1DbD^&>hLoD6zmmU@$q`Mnw!
zYdzg|m3mZ4-saQ(q_b^qKS)n{ueuV_Pq<sWUOofAX*Vi#c7D&`rkK0LAtK`*y~qMk
z-sz$!6<)sYT~0(g^6c_wi__UxfbwD8a!U}Ne+o96K;@QB)v>%?*$yClp49?N;RVSn
zj^XdSvY(tC9X6&xVC_mL=1LhV!@T)kHY~e>$hMs=F+*Zz;#WC_YR%;$U~B^o3?oer
z+E-;-3F~?Sb?`d+6vPxF!PtD$Ifsm+6`HiwC+YEZc}QSYU^Bjb%BFkjMy29eX*V(9
z{ar$8)tq5&NbRguUCdChgK;*w(i#_6XcYg(521wY(|~En-9+Z@Z$4|cjCK@VH80bg
zguS$<*Nr7Rp?=*F_!uIj9?wg*&u6nr<08vD)y{jQ;2nx0kcCKQ>}H=@j2<{DhjzWE
z6zLY5*Id#{bE-AE?irQK;^e7mWmgxe*5OLjmC@HpL$k-7i(i(PsYv-g7xj;^K=#_!
zw}Q?*>?l}?ncb#zSOQeg{KG;B^Ku<hE>Io^-#v4FLS0!a^#}}uc)s>^n{%6;cq2Df
z)`MqWXkMrBo?7W5_mUYMS4T-MNcfG@FNb+S2G?05<VLWT&2rYqZS;?O556lSRy{N(
zD21lDF@}aHR-MPW9&#H#?SDJ))4WHO=i;?W^|%p`Q*`spn5VM48uGvL76$U^=DIEW
zh<GFxjTQ}Fe9f7!e%DgMJ!n^SG%&<&(dplvW#HCLn!PrbLwqpVjAB(i@xIP^tM;es
ztyYJ{HQcf~+h&&rEoae|HP4nAbR9O8t@D&t38_}KlqY%AMdLx?{POuyg)~zP2QoY$
z{r9UD0OXE$;aqwmt?;6ec^lhh3nm3$#CkS{(ca0=JZWm$@6;RtNFfgg6HCrl7-S#C
zDBn?Z*2)w1FYm8fb9VR3`+0u8?42OEmV;<);=)deldpXw5z7+8OqB?t5wZ^fz02i#
z?mR?Cr|z!&$rK?0;y@7$)fzKy6ZhD*vtxV1@$)C1GZ{B^#H&~K?BXMx3%HBOTwl;#
zB+OC%0ld4jh!Hvdi;u&INVi#^TV<1EWh?xzzA;YrJjQSf8s5@r$Uh2s@=oa6ob8y+
zIM)c^bvW)nvdg@$Ds(@v+sl{>rK2o}Q!EKDY=wQFmK6?kggXki_Am=fJd_K~^$LH8
zA%7C0k2_A}Jd?hC(PCeYTI>`mEC)4W8nJe$+z9uh#wf!6;N!u&w1Q^zJ5}UGuebX{
zN*9X@#n9&owh(S_%DQ=fezk}(H<Q;s6ZEjn>#*-NW-X`7oWnYF>C7aVl+*cg-`=|G
zeCFJ>r|!m)PTiZmgATc&<~<X1-pbK8){uq3xi}^P^vHCyZ_t79=tnw2t@!Kv+Kvn-
zh;%KXY-=ZsA-8}XNW=}_^ZJN~Vr>R{&`A%01<axz6v$m=kjCY+a$josvC58+_j`AC
zh$8z#l!`VTFu*+XD$B72HRxjklG|G&ZSQ>R!ZA(^{BzKV;s_*FR0hke2fqcaoRrVa
z`OKe%!**rO7vgdzoK_VH-o{DtH!xXDqMP@xtbCsHncX#*x=6qDJY)7TJA2&x)!+j0
zYU0S!RZ~ye&;M5A5Y+CfbT-=@&p?F!{H)3YcT+L9$q|q(WfeHj8F2QgPW-scLsJ9;
z{NoEho14GRBO2-v8A2=4`Sxp6Dlu*_{R4fFSU_z<lj53?Qkp0*<1nNOAGCaWHl)wR
z!}B;ywQ>U0oukIY{6N2(J^nlm%dQ&rYRkFFUXkIW>Bmepv*HI|4lp_3(1ZCy5O#o+
z)%-eLzMzPmG9L%O=L;N#bX?%2TbuQHMTLE}0bz9=_l=rFOSB0U!&6<uCpwBo{>&PV
zdWdN?(~vPB@bjI{PZ2UqoaQHF*>J4hkDAvOw0AG^A0NtnOm~d*&*#FZ|0q;|=JLe~
z)q2mex1;(p)3+KUdO4r0w?2pdqo*wCpjEKVMew>F>?mqokNubLPBT3hj2gvk*F7kj
zrc8A@Y1_0IFfKA%t1Qep>K<A#DrA$;hln)i1ZKLE1;HyTJtoT}3oGV$oyUYBK{HWC
zukan`I!jYtT>M}zPsyG6X|>u_YQT9@caxx^9wW4UrF}M=gi}}S<}LaAOEdg>-zQFB
ziE3i6Qc`$8PV0lPugC&oB9)Z;Y$0aY=ngAKkq!rOq5h?{=V8~4l)4*fCCa8OAD65R
zadpjDtJ0qq*q6zJRHok{v>wQP9C;`LC^Q8#zExFiNj>i8ouDedth-;QY;bwLb@e&I
z#>K_uvCGTw-AlpeTH4cZ@_{qnUznx9gykNQ@_a4?Fa6J=Z*DC^z4fU%3aEcu+@sn$
z`j2j^C@4(1+nCQ?8hnj2;?1?P>bHgt0iF5F?pxRatFO_mu=R=b?8Tk0#?3CDgUr3k
zZCi{H@1&<Ni@*)aXKLb<zrSASGtsM?kj1&I!N9tE=<Hm*HCgL<fjC@0-r#&d^vFiZ
zL`Evr)<{M9B=LCl>3hLfm0XD!vYUi=jul}nDuu%usr6HISv?6=LfP4iX`XSa%F5jQ
z?K(uWf&8$x3p@zRr(NGeGAs@_9#A_(opUZzT`8jzR1m+&V1VCcsxDH*IeNZwitRe=
z2hv=a(fV@~n}U%o_H2zMjo{*N#)2f^u|dwogUnAR$R8GUB>7y1_gPV>FwkO(tA0N!
zDo$GDWqUyn_q4m9zk~dEj|P%3n0n}lQY>@h{6i)u$=nr(fz}n!O3-!*zb=11x*@=4
z>qWNKDGbFPfO?ymwhC=^`df2fbz1JIjQG307aI$5R&|95>0P!Nt`bJI-n8JC%z6UR
zc;+=n<a|&2@-(|GyL0(s(c<hNR05gPs)?CxNWD54hDIqo?n>1zx2kGugtU{6^HfV+
zXYCIi!63$m^1w-Qy;yz6B_(Dccf~${*M^Mr<3DUtsB56EWCZ@RMm3?Q6@XGHYvQJn
z1n_PMD-#jGwVd9J6eh%++@L$xGOvS|!1|)|$q?!}KIIqvS?E**OB{~=G{M}gwM(0f
zFMrMbp-5JC)|mNw_3o+ir1sgn#XF=M``lXrE#KrYhytcR)2aR<e%mpl+;Wu0mG>lD
z276T<{RNdMf6~KrRmd-#++fO<6DAE@d5IX-^=!3I5WVPZY7>POAK5MHs$RmYWTwL?
zptoGg<TL@<vopFFdiss%WTNRp>+D+7+0?dhW*%L#5TQ`Iyes|n2<z#Utpbpo=nG%s
zzLBe@eAB-gayz;)J^Evu9sR*1T4AOkU|4{6(UT@wYrG>V1;2Te!9281w<G9ld;h}4
z0~C(cIVYJb-+zM-ln@GYK{JqgMP#li&D2G?*Z9%F26&adh;F{ky<DMx0e?8dd*L_$
zjZh2P(7R!F>MZLwEJ0GV#<QU)D?N=HSBY{wYB{UwnPRh7$Zs~cDYUqsN0UZse>W?2
z@I}ya692`joY|%<`!5utJq9R*j<<w}omK<swZV}bV!{Gz8dF^KaBGe!+gCDC@WfWh
ziun9^GK|MKm|f+HE}!k~xtWO|16%LLt#f`tb&I4n&zeH;{p}gvylT0Oj=~{1TYER-
z$^D9_$ybc)NMi7EXdG3ZLbS1O5E>K9zjOtUqBrhUkt@D%uieJb^I~Cn?ywcr+k0+4
zzlO6!$9RhmsRXNEc3X5U8P|0|vV(=h{9aKN^x6}p+IH<FPW`uDVjHU&<%_1VC*Il$
zQF!Rg>@g*Hp$MwR{n?YbS$>@P9icYh>NYK$-W)T*G>qQ(yqx<)_n`(wMd=-^AU1*R
zxzol#w7#uvKA9AW-{b6oUhEh+UL`(*@gu~dfTpU8*E-2qmC@iJ9EaBGj(tARd-D^K
zL9SlP#Rh9vdceh0;YBLEVxyd>8cO@sBwxLPqV66Q;y~g^aUyXJpItZ-Gl14|xPq8P
z3ir)~s};Hpb*h8;5OqYiR;XL~a168GmjRF|0G?)i=ouFm9og0XvX)sl`t2l8Bed2B
zC^_A>fWq$G?1gIv3Ga55a`6u3QF(cvfP|{Fz5GFqKPcm1^#}sOiNQl8*B9!0_sMzS
zq|Q}A&ynugH+4~XeCzWVnz0wcjl^}P8*!S|G58PIS05b#d##Y3k$2}yZH#=%$}!~B
z4-KM&y?QQfG-Lpe0g7F5?jGHOpyl@yzBpZF3*F7D)<XlKnYSRSwM}URE|XkvXxjA6
zcS_Cu80zbTDP}^O6-Q#FwpSI!O3&Ban;Zw^svVb>CzT&23b90&M1#!7)Af)X&C_K|
zG^<1;pYtXIWgbOdS<`YwUYsC$?Hb2o=sqmIPU14_!!~@=Pz%&!u$NcBjyVrG<lU(3
zphT<+4$bjj0&CpSu2x|znr2GYqIGv6Waf1uV$EkU8Um2n>Bb_Me29kJkhO$TrRHhS
zbnL6fsMi{`wq25(M+$BJ<VsAsYxoh$<opoq=w+{lEZY~-POWdpf(b_$biFpaw|aX;
z^tkcBet$uu0ino*7|W^6Z!f5^Lho%h>!^hAGi|8nrGC}+2vmD)rW1oA9z}d_v+onz
zf{xh-X#LaHd2%Mz5gyu!s1WSjjT<0Z*gF|3cpk%_mzNg<iZ`O$c8yemak-dy+!qHH
zGA@S?!&)D=OviW7S0g16le=BAPtCd2yIqHpvXuD(Y{j*6o3y4rNaf|3r+7HQ2PLdK
z7ez%XXY@9Yp~3AtJ73^Ty(GcJA&=Iqr)ZzvfoksQ^ucA#)y*Vg62~7@%^NC>7npGy
zP8OnCcX6I&QrJ8m`B~0>a6drUI<Q>cZyjoFnh?|K>FMQ2gS4ud2-9;PGue6)YW+On
zD+Y5~6<tcqxKB`FVM>pcx!a&l2=U#e(hf@|6MER^M{+H)H4dvwH8KX2B94kF#|?Ww
z1gB)XV4k7J=`wG*d=LkDGX0?GlAK6xWinA1+|x9QDT8mx!iW8g3PcJ;Wc)y8u^NO@
z)crWbwG0zhgG;`K$b7mcJDk%<%{`W7Ic&mnMox3@mWb?VTbK(UNYa91`WJG6h~WLY
zUPtUaBK8i**;a>q#V7j|y>*w*#67*#xiUa}GybUF^((SAAduE6d+I@xJ}BPj8NIwL
z8pXRN1|?jUcn_*8K2L9BLqho0YOp<WqoEDuEHdzz_(7<HmachZN*G4lr{1<p2kE@Q
zXaq`H-&3!Fq{eiul9z|+)K?zI&?tknAA0z8RoxL~wwuxx2UfDZSFX)<WB6M%-m6de
zJTmNx^H0|f8?w1;4;|o_hUdo4nMEdIa>>Ue*Ju>YE_G<YISwO7L&%3YW>t=^aWE#)
z5AXV`OrVoGxS147*SNNf)o{LZn7k%EiT*-5E0vZZpUIqmrOU=7-%wF&4&L+hd93D{
zLEdd&gi!kBmK7!}PvXk*D;-@P{W%<dY5d_Ai9Z~+WKS5;(xQybH1;PX`$cJ}>@CcK
zn&sABU4I&(yZi7G=P<|>+RtO-vG=**KsH?sA8pqR9n{1sD~h`7chz3F8vCbs0neqb
zYN6T3#;p6asO-M<qKb>Zk+sYLpW)m5geX4-mw=l#k$Y*^7_b}#F#5$M!{rrl#)Yj{
zc(zl`nG_4Zi}`b8y5{m75>6>qPaY=KKPml+d&F>j$}kxIxU0HsCN;#aM(6vkk1>zW
z42Uj&+66>>I@?ZkT>p`~|E)DgPMPl<nN{*lcKft9H$*S7R@mAgvF{h+%Q}vqHe18{
zRuH%P^yZ$R!X@X<(Fjkp7plPWTB^|4DHP)+sI`u7J!G0Nss(yk@%ww+ZHG*cO|)f2
zuPT~3b~qS43?#SLbS?)s{ceY$e^`;Z$-g=JGGK<IbX*4yf}Pww$yZ@}b~b>d-!^1i
zbl(i`K#a$U&T_ld^&Bk;`cpq1H-0SA{n(&wG7|dKaj|KWATXt`^>UMbEV5`8)bLfg
zR5`l)kuUasuIjehz0sB8dYKrNnoCW7HQ#h~-qn#7ntC<5cUs#ahymUx^|FD(d-ZE-
zQ4qr*bXvM}pbnvXXSi}I4q<$(q?X+%a;-h|^O;UNi$@0a6fxl3Hgpz9<@S_&;#jTF
z+~wT%&BkIZsYFAeQI}#dYEezYJ#~_<X-Lu9(mQ`O;!&@H8T0Y(1<S9Xa0uP_l@6e`
zP|}SF?8)+SZVIviFNPWz7pY&r@!k17Zw7HTeXX9L>iBlcq-E<`7Y?EnerMgf0jRQ8
z;3p#GurELn;8hY@<WoP=YdE=pKT*;Z{gj|`dJ=d4uz!f?1Qhg!1n2Gn7iD*e!W#Q^
zPJSTnKE{hDM~jD(Xq0?L$ZGmExQol480W={Pj_2NXhAyjVW2Pw%W%4FLCLj8?C`Nx
z?^(|Hn_mjTpjvVYVlHVdb5(@;4hMTANn+&OD4u_aSVSo+Y$1H1DRcFaEW6CPq>@+^
z$y39b2X}~na3*;X^6xtVX|`x-i>rX$o=EnCksZU!4*l0J(y!yb<f89P9-Gj!@4dR|
zjhT<>>%9(27uWSjY%b`7L7B1PxA-G8r`QKnW)@C2=5_CLpBP>76Dg;NTT)PYkt-nl
zu(UHQIe4Fc8)L%4#=LWaHXmF0Tn1O_T6=BSj~-cB;d+AUV7LB`K$^&Jf$gD)W698C
zZZvD0!0l(=?gt9~PnOBjDGz1N=5;?G$1dFSWL-*^BV*ZFt5%X?gw3oJ;@(4M3nT(V
zgEQw94+5ilZ$G9S`EhdS9`v5XVywp1_q1hQ%{sVSwrt^0YZZT`!d=0rY$c?n$yJ>%
zo=7nLc=K15$+PT>!Be@XvQj3H%)xZ>O7upy=&h+;tI6Ow!VWwl*(+$-EWIVP)j(H5
z4ADyA)r+cA$uHwk?j}NU`3_Y)XZ5UJUfS3{Afge_c{?Uxae>y5;hnR4K9y4C@bJ~g
z)YTOG&#p5^dm3-Cy4KKtJpO^Ss{*SS07FdSq-Wi)C|&^xpqDJC5`##+xXODFUFD>D
z$RS@uP^~Bkdlw>n)|0^Gyo-7_CgD@6DOFX9638T*us+o7!KqGS<n!A4eBcC^=m4=i
zx&afY6_oc?Q5JLOeI-8A*WxUMfmzZ^&-{EdQ{7Y2&HUXW0*WJ#5oxTIdmapwFey+X
z@pJza*ihpv@}Uh--AbM5vQ<v|9K5-99+b<wMYExZ2?KV-#%RPPYZ=z_xl@m3I4#tr
zC@eYCSB92Qh+lW)p)V+rq~z*Y;Bm|<sp*7u^yx<@%PgITzaH?y0sK?#^TqRCe=k&Q
z6r6i26m#^^FNGf)X<v1c0J`%QSQUUF0g@A6s;c-IC$`l3J?1%L&4G;&Dw*+|n;fZ_
z8c+Akx=+@2R*%qR3DB#pISQ-m*T0)ncR|`oQmG;fAAaBtov1we0eh(6f4h`lfZ4T7
zn(<Bd5Ku@t84>W*5ZZ>=IxcojbsC<5QkcbFsp|C5S)^-?i(kk}rI~$Xn-BrN!O4-g
z_}6UvYb~ai7yBZgLj&WQY{K&Qkovtac@Pk<5<jB|%`jfMD;XA@^Ef!sEv#s--;&4>
z6&40{Xgk#XZNSf&7?dqwK*p5QJ0g~?4PoJ-Tg=%8(!&Xjpl}Eimd|9#{ppr_2r1TB
za_cxsW}dhw-rYxoTp%6=6;~wUcENig7AGF>`~A(zY!sqhAJo}B$J<ncf{F^#<7$#e
z;>wKCp4la@5o&S>^Mc&mBPc1)=*Y;3^NEdAS5-n@9t<ip)%LKcOPCn3(WM_7Y&ere
zU*$p6XxAa}l-BaSHYrOH$3}raN`*k)<d3RK_k0j;hcXouyW2FH4SaB2T*QGh6^k!d
zu7kpg2{E-tGjK`CjeGfg`7CpRfxy06uf8iuiw>yd^+xyybkim3lQw*KwFqh!p(aZn
zSqZ9nW_+n~pUm*guaw=a*U&`ne};pnEdZxBn~_guBlZXk6UM0k(o+sz2|xRxbJURx
z|J@`abZNeBbl$DDns0}L^GJrTL({(+a2jeg<uy}4O-Cwwu=Ow+V`8%WBCylnGNy9w
z%~$17XhIYt%8mErw?*cGze83D9a&B9{qb+eQ1K#2kv`rlnT!;u1Ri!527Dt;Jo5-O
zo+7V!n~QrQY0xE&#~QntH<IjF*b#V|b(*LVBd>DE94KiPKMQc3zLFKzup|%|e$6bc
zhcP3I5V4+ZlXklKg{|V{k=|-J)0I1`#%`jQgywa~WI$-$Z7O<_0}iesMI=cthBkm2
z4>Co}{T{|ulj{*GEj#t{;1{Y0zEwh?E0ca)syg+H9a*YEZOMGqkM$uUK>TXVXK9Tn
z^L66ZwJOs@BbYKG{#zM39x=y6s_&0cWlaU0x#FTDC<X@6D85cCdVcDf{7N~K|HYd(
zQuIQkmfM`{=Zhx&J`Pg2W9|_kWZ%NT$4hb?o~x00j2bB$=WMq<BV2AIxC<3O=p6C~
zbIEQCnV&1^UAB+hNTUm6g6TR(Q>ATfty2sds*Z)cUc#=ab1V1n{*bu=eJ^?Uww!eC
z!Cj?WAV5zQZ20+PL9lo*F8`-tklLq;dzk%_k_UV_zMa7e>36cHgeWc4okrCajcnp#
z9QT;#V_5(f=73R{;b;2xyb&edaE%N$b5@l4&^Q;g``wD)e%iAoA?9MsuA5ZMk%|^h
z#iVZdlPBq;ZH{CD=9gV`3KpPL0>*=Lr3T-#W{X(9%WTvfS>3d>ihx3}6Sp&y2tQQU
zTt<u5UNNTv5fQN8Z--2*MPBKT9&JB+B#UtKqA(`fwZ~t@^fsf8eyEVKKD>5Kj2v`7
zz!YtJVzL(d%fYUEd`O54A4xsKt1HTd|Har>$3?lcf3Koq5rTAy(k(5igfuAKpmZ+X
zrP4@<(j_2W($Xc}(z%3mH!QL5ta{GF`JLzaeBSv3?%lm-?zv~Kx#l}p7VRWH#K`Yb
zEm&6+q<I&WBG*`s8h-G^my^1#&fv;GU%YdtXit!T+ldC3=;C-}t+Ty|=a~vdx5Lnn
zNx#KF5xlxH+aG1G84@>ASiFQ-<*@m$KXRj2ThVBN>ep7@A1{@Ljcz!CoG|=aDxUJN
zbzmYWn@j|;$?x;TE0Ts8XUmBV0aQ=hsSar^YE=%1=cs8rZYoj#mPyAWw4!RIVBff7
zgn~Mp^T7g%i>0?{N~6jBr>0i6z;Q{rAp66G$GwRwnoKQZ=|5s-XA%>pfKvt1BEB<U
zQ4u<>JOtY??Kv3~yKBFNA&}rUk4_xQ`J$Lk3eWd$T!Ud6K?@nxikf5`2L?L9QoP%b
zEVSGDA98M$oUly{vo}rp?#181uqKLp4=iRa=;WW(<YWj^n9`&wgxYXy6z?{i09X^Q
zA43bwQsFP<L~l^!DfiBYjf$h6TRXjOOc#w4zmwn|FX=|*1J2cI<fQUX!2a`?^82^K
z_qIdpRbOzecl+&R<@NR%mnm}n+Wpk@il6Kcgz(&cKM{2jO!_@(G}!#L%z9Jk;EK=!
zHqun_TBtwLw?<IoYMb1u$SVYtPsKC!%9js-ND>if5bV-G#c!|91ZC9Tw4s1~r<(5=
zNhv)%O$W@g@F)#Ugb=BE&6;=4Zf`*RaFX}=Vfht)Jz|O6G@GWrpTb$97mwnr?|Qin
z4sO@x>;^yFfPdUp{%~Y|+wpt6|GK35*!Z-GYONGSyPf~28cA0I<o0(dnO3a`ohM?a
z4iqlmLf5g!pbE#S_nHEdC-e2GHkhMi;CIh{CX?@Wj}FcgU>IbpK&Tj1NhZIIi=K=j
zOsv_xgqeQZJEU)y_EE#<{c^x_)OPyxY<A*Z)*q}3=fi#!m~`oIw;Jd{Yz&S<v{e??
zM8@~2XH`U|OKII`7KviCBwVdmh{jQPU5HzqtA9A?yRE9dppY<cwUSnlWfOonvFCZ<
z1xmsCB#EYN$?kx_Az`#?cV<-KZSy-Qm>-mJE*Lqc@;=V@I*_W`eXHQ{$Q<0Rf75>D
zt^C3I0Eh){K>RlPjYoP>?kkAOTB0Z}O1tkn*LBy#GK|;X;uIR+Ko#`1vvC>iMjthO
zZhmpp!~dhG7uug@jN%&_SGwa@m`dSDs;bX%VKY<TW9c`K0hAfe)@|H|XJ1g2n~#S2
zUP40f*|_J7Xmj+h0Y{n>McY%8A3Ht?gXwg{Bvp8c(MW}HtbxTx)Y-5081dc=H0Y0{
zCVOVN(!)maXolG5>E!z{Oie5CS()Fj5Jj%v`8Gaj=?v`OfZl=)>{tQA5wS&UUndgT
zxUpmy-0WVZ;ZqT)N50xk#x{8}M1dyV--*=c37SU3q94u&@Lw>-U7x4^nIj2WYjHfl
zM|XerTEXjV=@7aPEN34MXkEqZ{ezcpAuvTr!FIE_&>VWAiLm)Wknb5_b>7JcxYA8_
zlOOq^_t2B>N$@Fjk+D@jX*URTs_UcR8<O|)KPS1t%b@O>-~OuFAF)5^hh;Hxi)#P|
z(VudV;UP&RElYQJ$U3d4)Hp9eW#866sE+c2hLb5gM5>=!MF7j0_Eo?8BVTHRAT9P^
z_E%!pkL+dapVLEEOV_)E4OmnBOy5_-BdYaOPEb|u%8}X)qe0Y4j#V%^squ#9f#GwR
zOY>WAQB;JDtY`Jv1;q{l?qZjQ0M6mgLHk0`NOu{n`w0dfQg;_d_vADxkH~Z7OIAYx
zM)`9*&x;XWB5Y3@NQJTvoq{v&yq|~zH3M9-%kj%8KPP@v?Mrxlh$Be`)L-0(N>~B5
zZb@+q?E{mrx}Wi$Ci}Gq^tTTk$$vqi3;K3-DKi~Qf7gvFK}1;TVTRAD%tUrd=qa$T
zN2)1ELuy(0_%Qjt8Z8Y*7c<6GM7*@Om6Ew}3N}sOM4!r`aQIu41{7C^%39a(r5<TU
z{EdLx=-SU7woG_L*^PD3Y_a6c;e<7%1>bS%99Z!*OItxRoJl|C)#Nj{#>mza%&H#r
zbu~eb_dOE5^5ehA(qP%#b4_+WkcgC&#o@je5uCwbz%!9vn#cA>`(OvKs^$Xv3&8wL
zd4`de1N**#<zFi}To%(QvY04ipDf1vH7gCQ&Qp&+rAz;LacWq4PF96*)_8ZJ1B}TI
za5BFImb1v1sqA}&mY7?9{L=maeQdo&c2|e5Br}6E`hrb)`ODq-w(qq@DY6`YAQ!y<
z`xmiZX=;ZA7s~!hrJ!i-!GYB8Rw2OPe{MboU~*uLL&Hx7d8(|x3}yE!**+VYY?6oE
zMjxpwdcf?_wFHlP*8|*#01qSrJn-JJkD(LRW~Pkdrh>vl-?ilLHMP-^obiS9>w+W$
z<IIV+6NFwuSgtPIu~u_R_3V4(?e3EE2j5F5W<!|b8Z{;q_%^Q4;LAnI@S}&?kw__S
z=HyHH8Wl$x?a;@ychn^sIB~)^P&o~b?1S-TjTTPXhNvOa-F2R;54)&gh#mD58$rZl
z?g;<Ko<moUt`HZ~t3<5H<=44i6v>Psg^wwM&fiRZ)><I<-v>rR^|cD9=|=P<i?&@&
zyFb+?D{b0G0n1frPtb`Wmj{ip!f`}d^dfGh`JJb@_ZPQ>fHX4C4(}HNLwqtU$SP&#
z5;rpHk}?WSI(yZdm^QX1+@G^eW?qpJ3InRK@<T@r+}v@ZjQh@i*#bCl+v@d%!>%&`
z7+z!!D6Ttg_~Jz)Wloue#*e8DH{z(J9>wD$i&JwO<M{5^o2aPD#~7rzO`mxuI`Uj1
zuMc(mm1_k$JRXEq4!HR6ZmQA6aR{Ol_`T+db85cwHF@{ZdpSxTnW<>y&+?>mMywMf
zlgEuNe&tM#_#a=<=_yns>V)}JXqdFmZ1*XZx*Wv(Ot-tmaoxSA>lCY7a_FNn11vc7
z=k|%4x`c!q{5jn$gC#)Uern`&#lZHNf^o~Vse&$-(|ATP$>dq7==-<{34&4a)2tg~
zKVJpmmw(&2j|B2Sz}3<;3-oUWa9_9vmD(h>w5R?@+#X{iuebAX=>9G>>j(jav-(_V
zLC4_f4&ZXQKU<MpPI4M>IqUOI;FOfSeULxQDV?U}y_L9`sWmWYMz;@)@JDyPKf?&X
z?>#9QP$<X$WL1S2(26`5Of(6tvH#pVNd{?7%aE9uwETu*a@t?1HQ_4l(^ZZ$RNZB2
zyl=A6g57_-7?OV!$3SI1Xm1xQE3$Y2p?RpW+TqC?A^-UB)d;G5opn%Fp9<Na2d|-%
z_C-$zgMF{$0q?@Qxaa#BkK!r?oSxM7K5TX2VIoL0T_9qEaY6DKRSyUEU?Kxd&ud4|
z4?KDH2j{M*3km9=yPY*VpRY8@<;86zXH~3`0?lx5uZ%#IQ<3O!4K6IQt@IAHdPbwb
z&l5jfSL3@+lbvkKl|fJLk(#c#xI@q5(uMdJXWo;_N#Le)zwbEqjnAyoF{^rYZ0MN!
z%lA$6LED;N2)?keNEJ-V7Q&i~>50eMZ2PIj5ya^L#{>%Ia85HsvPHy7TCpo)VV>bo
zS3T1bWD3qp??@L>>Mm<SWpkx}MD2J*DjA4DHKyEHW>b_n7*WrI)_E+N&5Y>IP%)sz
z8q!~~)0;a#m)}M-G%+4({ooxAU(sq%_YvGN#JqRVueHwr#GgP~qyPw11M8^NbfQcj
zl{CyFpQ>-M-CSef?798YEK>WAkYFz$ul?$7vD(uTuuD6*mL=^yj4dpUkSH6rNU7qA
z<B6%@*j{dwbRR%k+9QAqQ$+16ojt%1)!Tp8c|5v-)%~?^rI~?VVq`o7+8q5EsS$LE
zKn#-Z0uz~i#N5`(#+SS625`l9eENI{An>H-dz{?JF{t)XfKglou(Yg^;%e$|@G4}`
znEQ9~Jq<*)&!T$|fTP7M((sQD7(d_E9_Sk7MNIKeZ3<%h3!okMmb3yFAm#C&S;;NB
z(!yAzP?%L^6JkYCqZBy8b5D72rE~Nv+pazRyllZ7%zrPq+|e?j63t{}7)YuXvHKn5
z(1w8R`NQz-<Hf^Njf1@~yCHkxsmW;q$8lSC6(z`ZgLzm4K~jj@=+Mrp7)pNz%Bc$v
zr#sO1J??UOI3j_|oEMF==k~KHKV&@gMu=z<8PErg=Z9bf$B@e(Ix-t%IYKnmt1l#)
zbn}C#9g_rXcgMSz_u}5+^^~{~%PeEwi$TOKXk$Od<vWSn%`#eIPMmayi_DC7oAs;Y
zpSuBf-R({2-A9jr9W@zQ|CP^~gDNc^kBCfp<sE78-3i?m`aX;TWen-|K2BCfC%)gy
zRYu@8^&O^n0cf!-Z!IEgWeb?d+7FeH5Cm@6rzwd8Prnz&b$70Zw0wvDl5OoOn{X0o
z{EjIx;W~V*z7*8U=P-L`^J*kOii__qWKbIBTLr4_-xh^lr~GYxTGXzS5d225ed}tZ
zx}bt;StHDCdqoDN^@waBDmObjfQci_siDL%z0vdQ)(Q}Z$#lv%EUm2G?$yXn8HZ_Z
zKuJk&dN>~z_c8?(9vaulX9=R^s>B|JCjKz&_7iT&3Pg%-xzOC4&qj^TTMo($VCk!Y
zfp1V8%3V=d#bpl5n8QkYMv;z`faQJc$*fq*zfF`x6^oxFY4y(Zp;flr>yeCXwz$2V
z`3V=sK0~+IB(tYUwH`pWZg3^3*Ql$G<)!V^Ky>W~|7kzK5oJ6n)coeNHp1nX!sO(i
z^l5I~bq-HpPoU&E?WT)+4C05J>m;N~n-_VrHtfM<0EvFS3B;vRyR)rszgQzFpip3-
zdRuF2RUWYt{j7O*``CA2V>qo8IB!Nmuf3-4m(306Tv4qtTO?4Zea_9tbZY{uU()e9
zb`9!h|9$odOAD?ff6j78^{8P<?8(Iv`ZBDIb1mz=$FYPZhoe=|lH7d2(SO!iyhMKS
zC8QZDH5KT&M#zrdf=@y)ImltGRcYSLOV0?ycj8D7-<>~*>0HV5$hfevP?2U3clzby
zkAcm=vK9bmBv?c9y3eRuw#cOtQp1nsE?~0>7+_7p=j~nUk=!_olb-I68E6rF)lA2V
zZnCYD_GZIc=k=INk8RFB_J+D+sW6!$b%jzO=4i^%;-<5`1=s7k4^x>Iysx#?<3<#W
zPM))rVdXhOyFMMLyjI|AnFtx=+4!t%BDpqNlUlOT<|G~SD-wX7@6}f29j~Qo3<!aW
zYTX@+;O?cHV2#Rtv(q#{ChYMtx8bs2CK0gQuQ9$*nauVr^E~t8HzV-53CY6e{;C~(
znrE`IXLZC~3_IU+e7IZgd0I|8!cKYrwvrNMogm5WpjeAIcCV&Aa}5p_?qJst0TAo9
z7ry4AqJ_iBNnR1s-bSSdX$3lJ%HClw6Fd_5PNjeXKVVIGz~2%G<s(dI_fAAa`1_2$
z_R__X`|>xmYv;wv-aoNMxEBMi73@dtDW(`K$^K4wOZe>VDkF&SfU(lo#JyPujr7la
za?+9?NN$^qJUgRn4Fbpg-PvO*#dgo>VcorzK&A2CCo%IUPm7-v6O6?#=+g2cdYnO|
zo7d$D<MJ(c(d}#Hi`*W#o>o0c!0CH}N`T{OOr3(BM$#+LYl%|MJzLM_(e(zEHqmyl
zH#~e(hl(u}*!zc(o77%JU?V2X0t|PO{q4-TNo2J<Se5m{-$Bst-P=jBbo>$S?_UNa
zt$w1|T(hNEKR+P0@}hLi&OB|V?&weVcZbMgC=XCmJ&86(It+E&GGcKlkB8~XiwW72
z_!6a=rfqRKS8}lLc4)joRJt#rR%~8p-9k}F6urt>NN93%W^`o8nij4K4kzyIW&6lo
zd+kdv>smniwPbRdYhoDC$$ig@x?i<y>Z5}`_}l0D&QW-QKm#7N#K%qTa?f*N?9qt0
zQX(vGcZ$cy2eZS=%lSZ$UHsv0=v#*~#04ylXu8*)Im}tB)FPzzL(FTAyXz?2#KX|3
z%qPfpv>zK;QD|ShA-N?a{Ng6Mo#moqvPtp_4?LUE-fIW<>R<X#-##1BbOf^fP|4!y
zx$%0RvMuTIE4&jGN6kGoMcK)(27U!47-PM6Pp$IPy#NS^!0<chz{4`VJAMV4@)+O@
z+$CQJ`7p{mjn`qW5h0=*Uux+_AJHhO?LHdc*2IBHhOHO>ndu=wUcl5TC5yP_=N*N8
z-|x>!zq|&*I1o8t*&^#}C}2R!^1!g(m$WyTC!Lgt?b&+*Vl1>p7TLy;EWH<kiV0S%
zsQ48NCr@o_*?vxp5k+E=<gM^1R4mMHQ9<~rJ73MdcK|^Yoq%-##0sc#Xy|09hoiCB
z=Zw>+m8;@?Eh)blsvvd5?ayh=(yHF}rcD!iT`f94*GiAU0-jMuJJb2T(b4Zkf*;df
zjn?Ko9FD_c?<y(6sj)w~pIvUFx7G2+UF%(xxfG*wiisd_JG-Jf`OUhsu|TfQ{c*qj
z$VsnYMY{FaBs)8YtMz;=V;vlBXVzFI(m`f+GPH`52u^?Z)8W%K>DPj}$E&BOJVJ+z
zW;%zeTpg=DmfKzwK^-DzNg%s&Z_b$Ri`Td#7k;}3h{FUT@@R$`N$X*Xiu5=*Bu2=N
z#hId@)rDRgPFKGL`I!bR=46I00*|?X)#nhV=m*LmhG|sRTrb0k?Dn5PKaa)xMD1XJ
zeP~kY)lX`LVXtu&Mg=;cN@1zZNu<gimSA~f3J1Ohhs|~_vCo0q*FnS59ti2MV>P<8
zm*s|rt#HF{=biZ5Yw=(9Tx~f_$i@`?$0T0tx<v|ja=lUfzU0%4*QRa+x{HU5*3=7}
zwC4KoH6QJo_~~3XB{CV8)=_}M=U7}+wy-{ovynK?$71!D{GdwV@><EhsL^`(CRHO(
zRwS)<gWp8_O6=gk;DaAZBliNJbQ=~|Xo!cl230y$$g1HR{<OZr^z?|T1)}nwdhbOK
zF5sporn<0`1}VMt@{UNposm<OMTOTly(kv*T}_)R7oy}65%BeH5Tk`kM@w74b0;NZ
zm|3;@L4q~DtqNZ$`M0yhnNsLmm5cK;@qHo7ld}a|M6svyLg3hnI2`gq_l)0F)TMIO
zqv-g#@{@v#?#lbCWip4`>)lhlB5w=GDmnBIx|xOqP}<6bj<smQKNJ%HO~><}550tc
z`rTQw0RdX?X9i{@<*lOoh=>SH@g8ZKjFo3qKIG{DMv^m~^R1Pq)P7{j78Bp?1@$$^
zd&=yGa?Z|fTJ?P5(3Kps!YWU;C6ev~ewZUyEdr)VKwOrHPyXUyhD1&LfnhaZ&PatA
z8aM}wpYvsD1F=y?VAgdRq~WSI*NqU>d{kEoW@+r}pAxQp<V#BHy>-p1a|5>gg65Nt
zHdyl*$Z$&eNPWRdMf&GfN}_jWlZ<^;mF;>8?bk-uJ6?1{iju8QT}h*IJ$tA6bzmRg
z?7fgv5QfuV50jzis-1sqoC;5XD<X#K>ZS(*Obj@l5k1|Bu5p@D2pVx#v)QcV;GJUR
zcHIm#$WdX~`w?A(Be!u^AmYOGioURxduS`=IX|bwCU9$W$>nsef^POG4nb@l^nx6{
zrcSVPApkHcKD>j3gc2n=7qV?4;}a46mfD@KN+9iuJ>S0NU_8y;#JHi4Qf%>I>Ae~X
zogD~2fm9(Q#r9xpS}E36I`*N6Zb;16y~=+q%&){>fv8nWhZ{w70|A>X1=h<7qU>^D
zCw#z;6(G)D+*Fj5j-mjv^*Jo37(`9B1bzOD^V$K+n}X!CQ_OCo^wusHA;jx#{U0U+
z`uKNgV;e&UmS_(^pw0TA)x}!ev4L@@ZR&h>Ebw_6x+{1Mawk3a3XUs4Qc&fwy9b_1
zUHJTA-MVjq!h-HuQ#4A70(8Zi9IJZ@X{qJUSNZja228p;jb8hj&)a}WAAdFY(|c@-
zKFxIz=NI6qOUU^HwlV(*J?2bVb%mIWi)o^BSs1dNS$U!hOaHBkU6-f%5-I|KsXE9i
zB_pSz!g7cSS+9r<NnEK|b6NEhnP>YX@_`Hl1`71=n0b**1mrN(4-wUQ4N6r&kLJnP
zhl_AooD+rUUoGB$GxzcQCJT_Ma)LGQb^rf#;^a{ruuUL7%jjX?M1?eGru@RYz4ky8
zNNY<JKU?NznR7ktW!DeXKzu*Y%5`KEx-i!N6QE2`|FPHnyyq-(;z9Xxy?7r`tyeX2
z?Eg^$MJ0yS;m5+3e@6s{c+EBZ>#pa+SGhfo_Z-JeL9CHKrH%Wa4+T-i_#}&m`h0AS
zu<I(alPv~1L-~r#$BIfyO6^S4|3m{tR>lHI_perd|8IVT)Q{(y9-r6U7-oHq`S-8Q
zNDst+l>3R{lbm(g!To-Je@#RB&scW=U1J)62%%$oIdeLir6weQ#cu^QUikltg!Xq`
zz;<@pfHvia7-^{I&W;$q=kHjt;0CX^zv3%hx)l;A1_mM*`X`kN@|XKG{~87MBE0%S
zyQ%bV_xxe!Cw{r57=JhM1sj-O{iDzjj@_9FOLM&_Y}o%XHcw6PCyqI|$*b|-&j-SE
zZjs>|9#F!!P7YgjZt>lu|7mUos4f2U33(MLtc=2)>-)f=_PFaA@T>pp*Z<b^v&dgf
zGnM28IxnnG8`K&R3L4&j_*WIkZOni5-7RelYO|8a%+$u#d6>vi?NGMYY>@CUN121~
z;#iNl?#$o6t;nFWy6mf+$?DYa81TsUh{2Kk?{<G(_%I>K9_dVZXLa@wVsv<0&Xu<}
zRx+M{!uc@Q&LMsDMbj1tGeSEI==e{e_&1%3aH4QCoplB=r}LEkhltvK?XG`SL;n9P
z{^@7s?<lF4zy5RmlnKNJ{x9Dnzw$B(hM@nC-yqUvAnNb`D7xw(IobdC4KhFo5@Y<2
z5<Pr#=70PIsV`BEZ7v(W@$kb_+=>6R&ie0InE~ONj{mC_NE;=Hiu<2}5Oh5vd!%B&
zeCt9{Wnu@S?eoUZpNF&`dm^X3-gH`=H!|*7!`pjitU+&E1?yYNs75Va3G|69b}^oH
z1}=GG-O{6Em@0c_gMF8iL1ej^wAYKrnQwd->{Q|ByX<9SWkTw+J6&upXdW$7L#3XR
zo2wdiP$d;ddwWj-A&H?p+ukQf%0u=eHA`-P({p__)xU<%%UK4**!e$YFZSQ_vsij)
z==iz5`6IRGNs!JNaBZ+9zkB!Yd*+SR)!XEg&P#Bp=T;R=tH#cR&GpavZ?8Z`O#D+A
zbtA7?X;wS!QfnaAbuGWYdSgs8S6}}WxYNkpOg3x1BM7pD4BdDLRAr#B@LsN)_vPdY
zF`k?Q)O}kw=7XQXUE;~TnFfAPE8tO$FZoqS7a=0+$j8tz*<e%!S=;Y&ZN`kem|z{g
zF?UB(wY*deYBXXuE49^46g;AF+FuWYWCE-7-&f#oUo1#IV7*g6rGW3sZK50c41_^`
zk&|S<b)t5ZXn2I|-y-g%b^O_YgVt+a|7fc7P}+Ta*82^A=lj2tHC*DVUdqGM-ki*&
z->UAFUvzMNHBvj;HQ!pjp#sv$;Rvox53DnN;^gKY9UB{4|CNPk@7?8I&5avYH-DT8
zu0B$kY(Sm5@h4Pr%;F_JP*NR(KEpkBKGU$0<F4DQER6Yk4Msn^#9`}+`vg++wD`7-
zhWtQq>y2-==SQ{Wd?2B&xm&-_9Xug!M34@#prh62-KnM<hvxj-Qy7lS7Eg_SxAhrV
z!u~y0y0X#zkdTjpV+ya1bw@(*JoevB<wymSXja;p+{{ram?m)%{^qn>FLz<y+hD_i
zMvPc!p86|~0?Ya3(WywbxG}+V5>!-<Saq5G?t|ef6)M0`Vt24Pjlmr6l$_JGhT`gB
zQ&fxon?jzVU9SD)^$iqmkRD>vdalv0y@ga(z0&2#4^MWT=6tR}2mjBGP;X~@39l>$
z#@w0t=5(RizQ8NHKg@k>)%atm#=!SMBCls*{;v?QKvYb7`E`4u3n3vvb$kLZUhjlp
z1Lgb+vlw!JlvYoZYk#=>K>MW)B%0Td!7H_d1Qyq^-GnKmdS3H`>&kB);}x@XdE!zC
zKv%Y_u(%*=@WUcnmFo2gt>!zZ9*(a%aDG>K>U?<)Cm-uwA*P(Fa9jDbrdoN{b7TL{
z){x$F>G{SI-Dop^Pu8CcFDf%jNNdF&UWR6C^WIV*o`bvV*eCiw82||Xv2sjWu5w4Z
zSLHa(*6--Z+FzS%)*&T}ehm!IkmUv2<+0gBj!jzU7DJ((Zm*oIY}PoEE=<p66FQ@0
zv($fA$?Uq+z2C#L&sZ<SeXQ$vE{rMXAn5dUaF+z8q2-wHyFwVSy8e9y{<0XA^r}ip
z)m7D+^*#-luqSb2!I9Nm*3z2iuVPAE*H(5pezgq7kB8?E2jT$G&?|cyF?JL?bzQgi
zRBg-G(D6$2h_;7`R&f=i%d8bqfaqIi0?EKH-i5MP-Kno$tOmQ@;H$-}+#R7$@6OJ=
z>+IR!wx-y5@I8pSs$9V7tG*Jscm4q^&zj}Y?~V;3-zob<=VE}jd-Ezw!TC}Wf3?|C
zA>b9!b=s?AGri=W11|P@NiXU~Y6ijD-W&0BvK@iJ&OAA*@aoSF<O<HnLbqIs$7&#p
z%vrgRlKe6PbF-CAvJINkx%vjws5E&;M>1}!-MMB~114=g=b5KRQ(1#WyC8?*=uYa?
zUzxG~xuBk;xIfp>IDdy%iM~lI=)eM)CYzO2Z{o59NW^F9yu#}j(Ra{@p)8mD2L^`4
zUmj}eE)aHIwenPq_mIztMQ!$HDmrxq@Iw1l*xD{n<DGHC%7&pDY6ocz<Vvdi`>xAt
z!PuRRwEkY>>hv9EI~m~`EoM!Yv)}l_i{Kwip_1yFmG@V>BfntMccZ2A-~FnhI7=jd
zKUW-BRR6vLe~m3<Zq3~6$N!GaOUQn+aLwlqS>?%}4$Azzyt=*`UbphbjLhFW-);=R
zOXs{@2=*JiqRsQ)+tbs%%S4}-9dL!19P|Lmd5QrGS<ZIGf2}eSf{j;1c1ZYF5{TT(
z0IVd13e`&F*8`f&55Y6Cr!jJKa==rlnKauB1<$@ZmKOdQ<}RU23;C&;3Dg1gb!7f@
zjqAbYR1LXz5FWGd9XuxKY9#K*jAY`N{xeFrmmz<N|JeQdd&PDu@+jzkcs$!>emIt!
zH%9XJNd39+GW=J1D#(8a1$i~d+Mf?54<-Rk3ki7+t)q8k7WeW^g0*tFJe4vvbNCm4
z_#8usRc~Fyt0aHaF49bx|5@vIoy^OiilvU7;$L_KqORmx3_^fEHRC<)YU<u>EU8+n
zaFk?9LYvUehhBa%<f)BC_vSyoL4IKW_+PLK$G@Rrm)Sq?+OKw^NB)gHSqeZRjQ>+~
zwauU+lsT@<n#a!q7Nnx6=!UJXZ~WPoWG+v`Ng6qzB?cMJMdnwJf45@HbeU-i01NN`
zGJOBP{Q${F{!icj`+WX4kf#m?I;_4qh?-fmjUR0OV?WZQ;zs`vO*Gc!UPi_=Bk>pM
zaeLCn6Ydwg&H4ExYo5rxK3L@UzVli1@|mz$Y2t3F?ao|t5XUD0B<>n}xA2_I7xTsq
z^i5|Z_NuCq8-DR;CA0e?*-q4XDNsB9jtlVTm00|wjUcVfA5kl}YI!3h)8w-ZIwKi_
zNT=K#kKP2nyCLTQt66W@6~@@Rdlx{uv8gH4yF$OaqdR_iMcWR7<kXsc&bo8B+M+18
zY>zhSp4{EhccfdHX=*3L*i>_HJZg90{Ugaxy8I@#<LuY0;@4|k$KJVS2R|gnAKjZN
zs;b&UB~M)U*w%FqMkjXLwmkZ?sp}H{qpnse3b$$1yAdP{mzl22xtR5PY2X@Jf%Yd=
zMDW5TCe?+4h6d2c%;vZsrtD<p+T0sKBx=aa7qrD=`8^_pm$_dw86iFrrVjH5Y@Oo(
zIJx#F2+z~GIzQr04m=>v>o;NaVP0Z3^=|!NAt9B1U?v*t6Z15P9$&C{TW#b04an>-
z-+b1e2?@!Xt~b;De735^>rf4NCpNV#_!mN=HDt<q^|!aMSJpo;)rZ3tMYf?GKLfiO
ztGTw#At~y_eC-;Z_05eMry|$2cc`6Zj+<xAEERuFfP<IuO0?WKp}bVYj-M8<Hn;Vx
z5NBMsVKE+)CU;?K&Z>L0&`q#w99ubNeejl2Ir4Ct^(j)id$uX`cKr9teg@$}{$38c
z+9uMFRv%adndi58k;A6xshp^gPzr!dS5~Jn=+Pd$2R+nx^qFvCoO1tSj<(dJldH+~
zySn1#-J^>>_|^<f4%l95a-?dI8xysIqN;UkC>@&MU2^X<m{!tHf!Dt)_FWd})N2lg
z0kpW(8{v^`cJA1T@ksN$X7>?hRc%eKX^vDBOs(`ch==@h5Oe_Zd?^bu;Nplqn69Yq
zIIk1B`D(0u>gpfJPw<c4$!r8?{%qX3v8oOoE3XowDK!lz@(PAs&8>~R0x{N-BjqI#
z;^wuZEdQAw5+{69aO!K3i93>4EfQWHEe{iMh}e^%&3O))M>hfPB=^4Nh=~Gu;2qgf
zOtWc9p^jC<<61c-#ET26ivI(}#iet2rCJU>XSxdMyIH!Z_K5-Uvv7bdhUQ)=DW%$d
z20>oCG^dmMqH&rhLv!8GywAuDP51fbrBXb;es>0Z=}Xa7OcO<Ok1;+*B!D_JAs@5w
z9%Craz-Z#9$<;}z1-OFP3-KM<kidXd&_=Y>8T>yd_UHs8*X&XNJ@}4I>0y8I(P{Sf
z0ee}wgd`^mh~*Y^yNqiIj6qOb&Lf&Hnzij?$@EPl&o|vKzHgU)Xb!^tZpOk4c);fc
z#p8o0*P&j1@QB~V%MdBeU?Cg&WMsXyN8Kp0DO?yHfHm;Nv`g%he}7onYS39GFAYT^
z6ibISoZ$47x$ZW(W`zsCy1;gSMng4&G3F5|DtJ8{1iX_m37C@q)iMM^hfFHL;z`+F
z>_10}%dGyQw9KS?CX-e<m7IXs$S?n{4$&>_H+T!5-k)XC(ol4)(`XwUkGCzWa(ahJ
zrWK8u-%PGy#HnHDi6l8*9-x5Uei7mtiUMJ2O5}6kfzxuiSx_8FvB3=-Ra(2*Vl0ui
z9ly&{WG-<?Y^u|yBm0=4VfxtM{g;n<BJ*fprVNQfyCZT&gc<fyiYCiO2=Ci~w@&#S
z{I*?%+*=d=+*QNWj^VpbaQHLdf?yG?1m*7Y&y6KETh<VWq4eg6&i*?lY8OJpfa0I<
z`p5@9ExtJeFq>Db(CxrGVkwmUImJflGWStF^u@Q&_;Y>%FL7Pl?QlT_4f>F43A9fX
zdg5=WZ%_>0@Xmlmf8$}9;#si2i~Yjtfu?z@opBxp-Qdr5y$pI0^XQ|teOm4kL)Msu
z@@cCk&(rVJXpqHs%Og5W_K%IgXn_efKGBpc|Bw94_zCju)4$Y)y*;k6xD%-ojoov%
zO{Ozx*gQ!5Q`Sp9#OhYv86A;Ow)9E-Tg8x-nRgxW_uacce)`n#3?<_9-R#*y&dR=u
zd@w}!Fo)K;*lCJhr65Wgb`zT*2#}RcU9mor94)qva*4Y2Z{wy91Z~2cOBxstEZvx0
z^D?u-WHS_aVbui<!B>i#XPe?AN%Y6dp{LI8G+dkqlEWF(v?mUo@|5@;Xf!JGH$u58
zC+ZojE=Cs{I&~JArOILOK6fHcu5W1r1?cgvj>j$yJCswkR_);%z06Q6z54ir3B-}-
z1Q)*0@e*Rk^8{Ass!;5T(uEE?nX9Oyy|{-hB%ZOQ3Bo*QC?46CyJ@G)dH>kx^-%Bp
zu*%y6d~%E}W=E3F^`_8JiiKgxVp55(#wu(2sgtD2Yd!2zcJ_PbHL&v?oaR5UF8P#M
z#syfSKt3S>oo=@2ZQ8a4-cPx7Tmw|Gqu8qn6KtnLJhoZsaS)Z8zW#>@KwDs1KWTlq
zO&%dR<<`^~6TfKZuq<h}Hs8@q7o}Eh+yHJNnkuz!7tFMsM``C%<apLn(`xc<P7oh#
zCxk-B%N@Ulo->EpnXT5PQIH|^ZU~(BHl%)iq-t}}c!l78OjoG_bG}4*ox7)_u=@D%
zbkImg_YlYiINQvT@gs=T*y$%~pC&VOmDV+kU}MFdV(OyhP(Agrr0e<%=t*SbBmSC|
z!_AR0nZoR__V>QN4n)z{WuXEGam-RRVFWd2Jz01#3&aR?dL;Vs%>wl5;~uM(DAv>w
zX0+OGhE}sp80JvBpQtQHZZ>u>-QrV77o85Dqf`aiV{&il#mS}KX{;ZP9`1pUiP`3Y
zmU@Pg4z6^u`jNS(osGq+xD01ZghFK;iqbdH6}fpL3wOjTKB|#_BtT;WNk6<;k(|zU
zKygz#uDS=a5m0wNbV|9%&073fqZ${Wz8eB?cInCMT-39Zi4>->s@fdXDO3tE6^iyA
zHa@|0A=Cc)!fl4do`?k%PwQD|Dr4|;AK-w1*GQB|6pjr_IIIKi__BK;w#}Y7t~(@z
z!wEtC!ObQ5F1q?l8CyabuP@{vNzT(8B;|j}u3I;S-@6TYDz&O+n>~I!-?bd8zz-Gt
zY8%L?Bnk89aM&0OS|fFu>!#(r(S_?rwvy!$=KN)8n}|1q+a%eoOH5m#CX_gvF#bbk
zb{K*C&%RDpZe3aKvdzJT4j8;^89E#fKcB!bkA5vxP>Y!3O;wW8@?07Ws&k(SSjrg1
zXURCX-_p4-K#uZ57=Z_;+G|nd5W5?d8CfYxV_i}2Og<=36l5;Bh!AeAt8<TadBJ&+
zn!6i7ZXMHqAco!vSW_%Wn`1i>bB?jGR-m0*T+B9U-P>|NuVQDXnEA!FQwnuBl7>YJ
zEx~zhzaEuLlaM(6z4qq%di8vWYh{P|3WLfKQGc%ES0%wGEU3~uJtggH#I>BPJdtP%
zOS9kOLgcdYc-V|~Qm(7yYeS(@G6t4%SQ5JorjnlGqlG+%Oy_n@9o|`(j`ow9o9868
zoLD~4k~I094MLL6=g(QkuCJ86Rl5E9g`}m_PGd*aYdZ!F`_^>%th3Cp80CYhT-uiF
zy3^sN*C31*tIyOq&MmXykM2M4o`FZL=_qJe2dW=@MCl%>$~T3#eCvuma)H&>Rb8Yb
zb4Rc{WE;(+@P-b<@L8m)tFwOHaleTHOVsc^7YDp8dHeZ^2?<*6ZR6`1Tq;tqXe;OL
z(+pTG^tMXZjqoqf_f>YQeK{EBFilUrdd<U;nserh5ktqT(W2q$P?6a1h7T8bT$GIP
z;NCQ7y{7eXfA$K|JnI0Cp*6)yOjbPMRbJNAwNUSQ4YLhwNd2I+8B#b!EvF2aon`{P
zvNnk+2RAqV4b-E9BcC26IfT_#pfVX9r=1}p^ymO1(#gOIS-#fQG27*j?>5zdgafw2
zgIm&ZUwVp#RZ`*xloLs@F=)kE*Qm0~G<k%>EBGZSdW-!z86_0O!W}YLZE2^M!l9Nl
zNwqrwx1f-@^YBE<EQmox#wX;-zdUa((i>YBH)L3SwckQnfv~FXr{LwxwgodO5sTkx
zDAtOEslH*i$t4YMVQ+P^XPF?Ei5CTOebAfgh7He&-U&aef6q?m@X-WB1kD2;R9<6x
z^R7Pc0tPlVpwoD1=KJ}sP<Cw3Cv25KM6u|8z%f*uZZ&GdGnLp;TDnuSUV7G&Uo%3K
z9)5JC>uM+oopqSAg~q3%o!xX{AJiBbzb3N-nNn`IZ-b8wCwf!I(d(*B*z2fcnnHan
z1ew)_lBcnhERtw6Q?jRJl21!3UC_F2fQ1ytycadue^QX7Lz*DVdp>wtS8+eI3;(=(
zagQFDEa)h-4x`~Qs5yx(B~+AYlW-qxZXb-^TZHp&B~9~1RAcB`!siNH5lMK6HKXp$
z8!SP9zx6Bb_-vFj*O@FieQ&|zhqZaw@FN_%DrOVEIvE9{@$Zd|U`L(LRcO2xOu(h;
zJ;nDpbP1liYM2u&CK-?C;@)}O9xE1lKM~KRgt@vltx$`*q;~i?c&<LguF`$&a3^!Q
zwkZ@h5pUI*eci3*g7hk6&{7?R`^F*P@l03#O$Cn8I_OB+aIMxrpGKAAK?I>&H1um7
zR0V;WSCjRW-f4GqB}`CIFjkn=z1$%Qo?!BEBqr#kKdDMcrnrC}^{jV%i=mZbPq8@n
ze5U5HsI0Idljz<@sbK|6gxCtvqnJWWAvz_VF$!v;MwsWxP`y~vb2%|qy(IS&rHV#C
zJC#oTZ{K<-PZg%>jUf?B+S|KPiZ{ntNGtCN;S(VcJ3fzM%%hKwudtj-a9s6L9;ttj
z*KsS6+`GTf9F0HKL=%y60FZH?KL5z>jXN25{9pH$e-3PNZDGuwCvp_tYmVtp#57Gv
zxF>XTw&u*bbLFsnuaToNbec?PS<8`bTGc`CrWi_JCHhjYxOn;2dn0YWi}6EDbTSW$
ztp~Q;-Y!_&oH%8J33xvXH+DvZZu+^91#sEZcyCa`&%|eEvlW%DPcdoX^kIe65cQ7S
zEEhPd_wD{Dry}49>j3DZvZ1qL>n!SBYvV530Xv7)6Av#5u-u^z&8>F1lar-3w%q4~
zSuo8GMk4+rQ&oaBjC#CtIm8kd!nN4-N*5iBfa&<2?WxiAc0#U=Gd&vV?>A%QYVS3Y
zW#lNFSfnM%Ouh(Ux%XL#gpN;R?|h^1W6gv%j8F)ce2@wfM+|QCtC;eCxqrjq%Y@^i
z!CUjOLl>IW!w;Bc!&71)v!#p~1htDaA>#Qs8XDU9?D0>=;v1(PXTu$;D>$w&cgL-}
z`-S8^>q*4!qhl4TiNR|s9>uAT${G99NjJ#ecicVP7_a*|RYM=!Kj8d1P2dR&T3{;I
z!~XE)A>URj_U)b$`$091caJKvkD6r!$}_SloN?Pzd9cc&zerK)1jc4Ek5UFs8$;=d
z>ief7D-+v&F^kAW2nZrh)fh$!mp6z2D}P3MeIRF!LySmc+aW_f3xMhA-2l`vOsvHK
zNV6%AcjXZ`q~2{9P(^s%MMp$@?WDYJ5Ad^{gO5m?X33#<5jrR1Utb}@-kdmb(7qvN
zICVPPeQIiHsyAJ1=K+5%dlZMUpZkPkY`E=|d~h0<MT5A5x9w*JRC=^*Be<F+s4Kl`
zwlww;H~Ny?@-mZF)8Xc!a7|&lq_;bt15I6e!c;+oxz$8jw$hb6a_0MbuRP!_-U~$2
zcm0O%E?PEwIPaZoukOaixg3zq_gZ$#!P?YL&QR)f9O*`K)({JBqnfT7_3`^FQ0Bvp
z(;P&>*c8MtF0rm1fujISeQgKWk`|1tG|A9k?Ud}V5GAz^a9JW4Oz8Q`-I+X47DFDL
zEcxvI;G!}g^WbCddbnq-({`si`fae73z&CvzHm>G^NXGlsmO_<nsafv+>vm#D_${_
zUMND#pp3fCq~)Ji9Uk~$3gNHwdKg;JqjwUytQjMYn|=s0rOId(UPxPNoQBqxdiYd%
zJ=U25;!Vn^B=^Nxg3WeKCGvvgm(Or0Ilx4>ujLZPe0U)fA<h3THJvYr;w4E0hv}kx
z9@~c=bS(msPQ&7{*g3kCBlZYMMq??h-uDp<0|$8%DJ)TM1;he~D)2*=8vx@^AyN8j
zvZJQjA6P*X7O(hh__+fgjmX)JK}LAgvdWTRmS(Q$l}w2kxk16ns<XqG&5n2$chBCa
zoC@*?l;StoSCtmFQtCSRV|%M;+I4()&LbyO`jZkWh6GP{b8K?T)8x5eUthp)^s(f7
z0+Cl2R?XV3@6#v%x+crn+*h01XTOh4V@kFdDYAycluzFB4v<eP=WRfWTRY=+Ymd`@
za6OzTOLKM6(>aS0cB(E*za1{nA?JMTOm{9G&2XlA^Q6Y*Nt}v8%zm(l(iO(p{l(dK
z>XqgRsV`8GNNAnjcC!?AZ|&K!ovQX;RuL?uWn%nA?TjzwtD-!!C}D84J=^(s;fPr5
zM!?m~`xz03FrT_EjjLS>o>2(b<rql=sEgt3Qm=NB+cdV2k1e}ByvM3KU_W8u>NCmQ
zV+l)*9~&!OnJnnW27YASygU)$#mvVNJ<_Nhv@@TPXSn-4Hxu|m+#|@WNag;>8?&)r
z(Zzkex~~PZ!}0;-Y16G4uUjPa!_bv=vO4$Sv{1jsz(}G9E<5<Uohr{fm&!A=UXXE`
z{;;-WLEf(5Yb1%qEd}{kLRy<cCwyu8tKbM-DfIn0)4LgP&LrZ^`SQjwJ+GiAdxwC-
zz*9u+F583lsuxVzkGj@toUsjk294|(pl&Mwj=!Z`xPO#!JanvGd31vP8<3ZL%hog$
zbDQnTFdy81mQc$=5={l9F}tAYdD{o{dwy5YUY_jX*0^*A>%?VkmQ=5ty4KosE94SK
zqs3!pb}-Q?>(FQnc`5`38x-TrJ%OwOIndLA1q|xiI{ALCfmbqY@{fP))+YNbkm=J+
zqwbg5IqB1(#%xSMi3O@@<J$Fwq9rk(PxA#yJ|hMC4T}jPd#-C;xOH{7u0>yR<olQK
z{5UE~Li%BTEW)WTai@6L0G*8CoeQ3J0`_g*fH*EloGNEOD?%{G4O8kK9zVRGIA^xZ
zmUnaJiI1Ds3|iV<l@r`s^Imhmr}6lFJ%fwiv2!VS1D~aMxA%@kp5p$2ohoeRW3L}s
za5Iz`RtdPdOLoCvl&5y)>!ha04G<fb1X6z=Oqv7<{npAl5iyXjHH0ECg*wrmo;9`3
z&tvdLxfQFAg_IJh9EdtzG3)M+tY^SYgHq5wAt4|>$YLIT1#p=XHrH*s_ln}w0Tcyg
z*)3GcKZgT*AH~Kve$_t|vy@=kOWR~5K`vp+9;&i(B?3DaARBr4>G?>xCB4c6X}9rI
zN&wPB2|JY%B-7?BBGx6`UX^UxVff|2pXJdeQdO7)t0j+~P0k{PX(1<JFUYc<0|AFE
zSb%++Tdhf7a`X#eN$43=sLu-S%ytuMXoR81wBM$i)Geik;<mYSaH4dP`pwH_H?Nq>
zwbq`;$dFeno7qm3QL0ly55-<8ih)T|Jby;|G>033QCM)^D43`DZEbizs}p;HSrDgj
zkE84`V`KKmGB|2&7h-aLqIFuIAN?q(rme?TPwd!PcOHt^YuTE<FGOSgj#5{kawVp>
zcKPJ_;a0v8E4g>J30@ugY~;#%4;Ki6Z-W?{^Iqe3lrqnEn=ADi%VECwGVYw5x3b;w
z(>Ym8*vZ{ASP(9`9=3blaV6(5V9+75c^DALfH(y8yB$NtBkvRquY@tg3-%WU;F$FB
zXsClx>tMgiRmGddZZU6OOmLYj9{=j-uMKxZnP3Dj&(6f^bza}tn6ud11H!pDDW23k
zMq-O6`@@^di@FtPOw-?j1vTqZ9@X|RjGoEYOvZi7R_FaVA(fNp;kXr;%y#POW>>w^
z)6>Jdc``d)N-}WJPC&?GTe%;JTW9|%(yDPotG9*?X1P4Bs+1ABNR`Uzxal~g>Njil
zC<sr95KXQ0Y4vWBrD_bzXMI2~3h{3>><5H`q`!t%@&JJyfNq#Vt{QaXvlZuT;PcCC
z&l_4=VBax?IK9m|UHJuzxp0K&B04B_z6kkihe)Oi;b%qexB>yjU|}}c9@<57zQ|{V
zbw95ldX>`LU)VPtxe6b)K%TUPIf-zBY`EsoMEZ&D{IDSR_Ldo;J3&1Fa!I0N+E_NS
zab={0QRpThz+T3H+B<u%{6;`D2Mm<-hqHO>Q#%65fGr4|G&DTEqW%eN4`ZU&H&?%%
zfUP8E<@r+g=5<4m%JO9a@EGhP7^&0c8RSgljzE89<8x#d(!%ZH=_13>sMqaU1BJn~
z6tf+tvZ_r8OIzF^#o`vPa8FQUn#j%LB4sT)cw6q5?Wj%{w`QcF*5}$d?*`f#*V)=e
z{6Lt${C2kR{bL<3nNV;)e%w~o!IY|ZYvr2(r~{79zQFmQehg+~9<oi_Q9*<w!t_8`
z3bDWyJB6V;Tw$_A2H<}W`{5&R=ER_LXB=FQ^BC{7W<B2Ug6my|3QtZZ@fss(+1)8{
zcjBjG_&W6fu334)2J*3EUgdsZxI8a~jj^RF-?W^^cIvpbM%Jal&OIvltpM>(Nser^
zD#cF2=n(PEDh=}nDk$L0XrJ#=>04-vP33q0kVLS!XdF&rhEZKr=8{(N*kVWfaBE^n
z@Cjd`en5l!x%+}&16oZfqSZZk;!NJvwkS$!RkbR;Cw{Dso=}0i@bx!pwK_AzPsQmU
z6;A}qW=&(Q%o>H=!(sF8K*$Kr0=O9}-<LcD51^J%^P@n~Zk!CQc#=waXN{D*A2h!R
zEreG{tnV1V3%^GMXe<t@8cpGzY8sR28idE_@5t)n;;`ItA4Om_OiD{al^d^>#mz$B
zjv+3n-Jg@X93z!6FooteC*2X%^T)W&DT*N<q&bO%6kR_x+@sW?z$kA!$ZK2_A$Yi9
zkwqg+F|uSl)nB-icm_YC{<s0)3g-Ck08h$pKOQw{0;Uj+cz3iUIObxQS~@ok5YTy&
z8C?kTs)*|3zr}c^Ssa4JmKw4jJQ;Yw@)@0@dpzNUwwBQdevw#kFf$VkVQjtiCVM^Q
zbvHc28Q5LuE`GaE%1-?SLD&rR4jx)hkymq<-f(1+taGwpeXD2Jot0?ky4ro3f*&#+
zyk$^eFvm784E$CG)vv{_6t0TSkzyn}y6~{B6bjeuw3BU!ay{)%B4XWH-tLSmN<C@K
z2$L<(C==9XRJs<up?%&O8(47maXXGV2QNprNqB>~mW}6vg6%AhBco@k-+5`TGZluL
zq4WF5dc3pP(u1a8iWyCBxdsCKoKg1^Jw@3+lbP;KY)N!psjy(zKHZSed*+*zKx`iC
z!RZ`*lO<@#Y1difCO?|#NgrJGtYs&UBUcSx{Q@=%B%yfYm3ivfH-YUS&<lL@mGum(
zIvogyQfaOI=%1?4c#K214l8$+2zcy7>AQ}T_iatSo6CAt)mhF+XE5r*y^0Ayj~Py_
z($gl5s-uXtWnEbQxzw~4!rDiMVEuGQgFc8W<@?vFw+E)<XMwJ^U!fxxN;yCWCt<`y
zGNtwHal%oqbl)D2z>||lBzKfoCQ47|B{gmrE4BA|zu3}0<=G{nE5TwU1Nwm5_lT|&
z5MO`w5J%{C3XaraSq-BS$6)HJp@8$c9n<8O&t;PXZ$bs!!LisEmF6eUi@x5;DusiD
z;DU&WnwQ3+mL~ws&}dMQvdR+^VDeuC-SH#}{n3_?*SRRzhKobPq{bj|RjPy>9E{Cx
z0xIreG>1wBhZu;pwiF8lY7lIxI<P)Wo?%nQ#M;HfyWrGjNHkD@5oyi78{anerGT{p
z5l#A*FfU9bfCV7gue#2goj5h84nQ^!Dd8EZ3j~<};CNkMK2$$zn&QZdlOVlMYP)|y
z%~^}rIde0n%UIPm+b&<M@;rcRD<=a-H?nP!M4z<4WD1VXx=?dAn%tO$wQ^`7x<kOc
z#6N*6<OD?PjkLHQqz&2k_bnyM>&j}%WKawvU?H*svWSo4OaeMp#ErYL8k24??Ai17
zyCInQI)rE4u??_K3~qW$*agGW*I4kT_L=LdO2*tFi@vjRa|4}-W|;vMXVq`qYL%^c
zy;I>DjP9pKK}6$}pG}a^C&y%pHXe`XY?=O}->1VZJc9ydO|Gr6NX*7rvCvO~%=?nP
zsVj#g<shLh1QJZ{g+g_==3su3yF?$Q?cuw3@5sG-lr%W=b%>^@O_#eg1bfgXh<=@K
zE^6YD$Jw3Xiq#dee1u`^cb10aWVTjMH+I5ET(oU^Ta>~dJi`}Oov0wr6Oz{fF~Pc~
zeJ{Tjy~?lvuJu<aOgI1Id=3U$jXj^^R{li*x!n$yGglT}#Ythm-w8c>!V^*rwz0)l
zSb3vIwf&i~4no({v*NKUnlp2YAVeea;?@2I23o?c00;ZLr1<1=2e}Y|E?2o0F7~_1
zqw>)>9aEf##msedM?VUBJXv>s=*ZM0+cbxNl2bKmw>TL(ku$AHcG=Zd%0B^jWuM8r
zyXS0WO@VRvZ41B<_!1uoL(_7Q(m3R)pd*d(=R?`sZ>du<z#fTFs2>>4kpy^bnEXAu
z8>T;b_;_$_B6ND+D{a^s0N`CzC=$w2c<W`77^|q(-DJjy+lBtYK-MxCqv?}uEZkBw
zwiaXF=DC*IE8hfLHpwfqdE?8aE{WC}AE|8h>Dxj}!Gb_c#dl%?q;YXyqb7oGnth~h
z@#vtSGm4q3q;}P^=3tn>@&j)90}{y)&iwZN)1;d~=R?nL`GmS=`Axzm+08u}I+Dgu
z7_2AS(}Qg4MICmoM`o0ZVn2J)XzN^Lsn~#wH3}BiaAmJ0smxC*btp)w?4*{OGfA+_
zd3!esXh`7OC`6ocrL+()zQW8POd+_O$5dN~t7T`-*`XknrrEBM3cSZVhNPZ&Fl<kV
zDD*oz?Z$)|Uz4Vl_VRY>Z^3ul>^v*CdaRi(q$Jb#npN&*jnyn_aAxMS$d05mTw!5h
zPj2_KWz@pB$9dzb6A9&w*%=t$p9k9Ks)AE>SXIH`uhUaI9MBuS=&+ibMpj4&jw?p4
zs-U$(_C+c*Pv@nA4$p2ZI=MFs*pazMVIz2@zoaaH)N;i&SF$tqZcN+T<xZV8S(#M(
zG&Ep>N9c?b6%^<-Pov2B`L%EruP^%boK11o9&9;mB&n8q!Y=Yng)w`Hdo9eY1wp2T
zLK_JmA~%lp8r6gSH>LDcBRQk<jW{1YY&xJO2WX5Ny@<2kkN1ff9OH^yII0_0hHXd^
z(O6@oZ_cg`<#cb}l)+4cez$!*m(t}A)Ey$&r>UQUza8%`TW8{qe&}9YP``h_$DRra
z?g%>#h2QQ6ivt*#NG`ATOhCLaGg~p+DoLb*JWG)d2PyQmXQjpon#CvSY$+BGcf`Qz
zkbd9NV_1yX4jqANMx&E!*0>DS`96to^dS)yK7PP4T(a3qRjVdIJ=jW}LFJ&L<j`Ld
zi&Eu?ZVn>#m2i*LWn09p>DcZAhG)6b7<5hPz_(9JMWJS`NXeFu_0iUW@9A0&S#+?e
zU|ThDqg|T%CkYr015S1t^nK@h0UMg?mjfh9Z`4d)2D9n1BAj@<;hVVD@T^e!45&9I
z6C&!(KwQ;i@KuO$bmD|SuT-0{)p+wZ_0BjRRn8f2VJXo5v9n`di{>{KWVYk>9#>$5
zqc_iUX~X1guRmUS!RW51X=r!;!OCq-ar`PIQ%m?Nw~(r77g_Nn*uVVX;W!(UyYe!Q
zQ@39Y4K&k(iT;+&OgP1PyhA1wP5hj*Pq0vV*1vqT1Dk7?&Ke~zn`fBR$#Eg(9Gk14
zgX<TFt7HBNbIz`opTOl&o*o$EdU=waaTJ-UFFKOlr(?1wyc9_mIsFyKRcD;_aeblZ
zOxH9{1gB?F?T_=eI$HILh>T-xJ4-d42@Uq-NIBV3#tlHbE1_hu)uQ<9L%SMKZcjCM
z3m?!kmz1XkqR`r`5LMg^(ER>flP@BbJ2_PVOl({??lp~1jKqy^?KGGs{vUha85PyG
zt%;(7WEBMk<Z#F!L2@n+h(tj$l8PiCIp?BOl4L1zmYhLy&N)k#EICOoC}LGFJnz1H
zZ}<DrJx2e$HHJcMHoG=!%{|wg^ZVwvu3?r$y8AJlw`sMrjzRhW!Zt+lD>us9yurq$
z!2fVG+~gSy!5@X{QA3J%z5r57KePz`e5MA?J9u2I%vy)OeKdI?YH>eQdoZ|2aQx^5
zd)EU`Ub1rK9wYbkg|y;0d1!skS4WQvi<wUBzn%(vx|aWZ6TLY{S_@4uH2vtdbXnr$
z9TRaWid^0QZy4+Z=-VH)@QY#SFj5NMqwJ008W%fo`O?AtvUP^_-UamLYr-88&IL*>
z>ac|$S{`0f$*o7?#Ws(NZH*RE(~iVN!~Bwp6$?c{&6J1f1NPcyu)cUzS9R9Gebsd7
zrBYxHGNwTOZ=Rv6K1$DGc7>$r6NMT5Sc&1@$A=yg<U&Uin;(hAm;^~nnrxsLoP4kf
z^w@c%8gd9n&pt1)TCGg)UDK%V)3PMci@Nu@=A0$?M)0`0$!cyRnai`>Gs3(v>mJ=0
zV>w1C7{k#X=BlY!=jZjaD}kGolmu#T<^a$sc{s6JY|bKzFEs0JeQJ}EF$#q_uI8s9
zec)ya{;?(~F*w*@nKc90cABpefz&m>W%)6s5<SSux$?N(a9CdI13)vNOB3K|GHI;$
zP0U}~A6!!^UTTv;Mcti1<~McOdOPj&7uVTWEIyXFxtbqKEcvvUJiRv~6tW${RB}u5
zX^n>YT?Vf#G-s%>?H)RJ{5lo(pXymi3E_oZl&yJ7l+8cGq1@sG2o<FP<(K@jpullV
z_A+S&ayMz|`Fb(tyqIE|>JJbew;vbRcwXV-c5G5?O74K2G08-C=T7xbwC0=F{tJ-v
z=sd^iXN?=4mm+d%DrwNFalD|Ip$_9XJ*xznVi(bIC-M6@2@d~DL##~FY@6A0+jYNz
zzVxnt1iG3Z`z+=mS0^x5({KJ*U@<rf0ZqYAp<lczTi%mFMaYWbi%WC?c|&ttXF%t2
z@MS($<k1FeiG=LMNqR^g|D2Zyq;Gj6dWXgb6kHDn@u>tUv!0yYz<Ps!gOWtt)O4N7
zWp=c}Yc()b=nle({mkUBRe~D&1U}ZZm-<qo<Iv`-T7Kw5i^(F*I%gYP1G9~OY6W7-
z0qS4@fXZmNp3m8%$nzVmc&%WTNBIT(f_I9Su=;z>Swe3-6D2P-*psxn_}X}ku*$vn
zx&O&OYGW;U;T|yy$sQtqXXYSl(j*;G+9>vd3;OK05Thq&5jlswtgi|+x?OM|+#^=o
ze*7DEwX%_Of^ApkRJwMHm*1@sqM%>oi?%q2yYGVZM`VDe2ry2phBq{wpCu5|%-Szc
z;yEk$hDyGM0}jgVqyT4Cz69NRGlNxGaa<5b)HXV4x1qEJl{Q|Pd(HsxJiQq##CDOA
zeDaSpzgXixJQA>{{>J1+$MTrCo16-t=8>?nj@kiBH~?<L`!&l&Q0krdtMhZGt=iU?
zTJ?zEr3kj}j&vUquj4WA)ym8ogeNuaXah~udcbPq%^Lc<O~PucL<t?jG*$I9&^R*<
zsgxeXTR{+pG=t_Z0P7aPqwKc}tN4nzVtpr)dK0YTt>H@p7fblR(I8xk1&;%*B7GCb
zxeJ;m@)!0q^Gva=PN(~{ZlkpB{8YPMg{_yL(~u#FDK>k9eLuU6sqq)rRkqkTQKC*P
z9I)liK}J2@r*PQnT2WMjgzNDY*}~%_O}_Jlfpv{FGHk4zzlN+*;0U<v+R4x({_2oD
z<7KuCOssyCji*>mZ)-8A1l~d6#zI4p?MwX5jg6Uqje|<v3guiaH-|TE#2mSJn=&L;
z2J9d7J?)k)r=la^q`xS1If*BNpH`o$>!^*n;D0luqGA=ADAFte_9P);8j5j+Vq7)O
zQ9FAq+_y^RwYAkWGS_~vb+YvA4E7qG$xcpuLm8f*jAZy|Yq{sp^GJ$#r>__vGGYDM
zV|;OwK3$Y0w)fc;ld2(j_-l|p_79dRSkL&9scH`a^h>|KvVO>b7aeBY2@rSAdj$82
zS>%22Rg+hsvsGqQ1fIxEO}-g*hGSaBDoDg$?Wy*c4BHmOC*vU3G07U85T^T}1PyyL
zGow7oo!R$_hNI@0H<K@2l&qwO7~Z6rjU1a^`5%GLn7yQAWqrQ^pb#*IOuLN40}ln$
z40Bamf`%FS^cc>H(o2^%LZlD(6jpm#jdxf6RPBW=0EXj-v)*Hl$)?_-L_pN%%y*hp
zzr0GWk0bX5TdIW4W7=y~I6D6na2K~FR*m(S?b4GUue7r9iQ(^9mkNGLl&W)6s<uJ?
z{ME|(%xD<0-9$?@=WcA!2qvSGb!DAcH&B#Sb|JPL1fE_VTOB%!95`XVs}$zDK;jG(
z|NT--H_sC4%fyCMbr%6WNH66X#uCv?r0<oxN+1W|xsub6Wo?Rae@72+cK2}M=5cZR
zApWuM_Z%Qb{vq#{ZL`SO@b&uPGTBt&p|PNkOM|3_SzHB9af&(AlE+BRHuTwGKstaR
zOP(0)cT?i~t0&Rd=lQLN<_aT7)!r55hktgI5FH@sO*_aGC&Ox!Y|)-`GvpQ8T%hm|
zm~d(G{yELk3<<KC_M;c9FR^?sgFhe^(gF5rSt)>>rWq-(6oD9jmGFi2jluEQOC3o%
zZH5uU(&m>XBzbW4x<;oNP`^Q6#XwhOrG%Z+{EfisZX*W>cGuYd<d7L-aEA~gUV5~<
zkV4#?PQ~4nO}ffAO;k`Q+6*%!F4I_c&^kIDXV2n;r-t5Do%qt9q%|ICF<L++1{)MR
z+6M(mm9D_5+0zO?7r(QwSz)tNrcbZ`(i@{y5R3E7Hx{eSmOnq}MCqy)>_Vur=Xd-0
z@xGq}GPy~&2S-tlDo2#p&j9|QN6|7DgWQO(hxUAH3wJ93ez-D#Gh(6eI#gGWOx(`@
zBZI)$g*3vue{gp)@DYPxw{f=Zm)wr!;tGHOKWZsq_E{nV@R-o?5sR9Yod5g}`JhfD
zaV+8^*(8Amhf^1}q6oBne|nXqm(FnI(z(3R+Lxp?pBfq_TKD_I&Ty=CUA7L}@2wrG
z(2xQ#t*>bu$#w5|Fs2jbU@6}3Sm=nCS*WQtKU*K{SnGwWSBP*5Q*ORZS*ibya%t2n
zw-p+~r}E-)&XI$|Jzy1)Ulgbcl?^?FkOCaHuz38jh|EGe7ZAVt(f%jLlgdtA(;@li
zX=-;MXbD^gS1iNT1lU0IOxtr6P-m?H*%=IYYc;uf;Jy(<cd|8kB<po>C0km1qZiyI
zcI{r<$kyB1DgO50EQ$%mLQ)OJ<}Z3G?d8%mSmjfTJt!p5@@S9_{&lAdFe(!U0Hfv$
z;2ckdUyg#k<U;M;{=6$Z@rB^?X=un(#=H2M^=5yy)YX7TcWKXKwag~>@@7ZfbM~zt
zfGt%M8$4U96xk_lBGN|d9#6RF4(#TQL6_kalfj=S|Havi8NKIJ%oY`>*I<9_uEfh$
zU>+=4M$~tJj5CaC628kzr@>?(QRA+~Q<O=w6)VjQ^Ujq%4d))`yJo{Gpi!Q$rPe^<
z`V4q0uyt1b+wXx{dstkiKP2e~Jn#6!-+hQ)13<C;6~0#cC&%tF;W<%Vqhy$#t5ofl
zPFT|B{?f{iRg`#jQ{~!+sXstnIE#p@VycpD?g9Rw0BP&#ui#JY5WR}5%ZGXjrFfkA
zm+w8QRc*35)3hyEsfqWbIp*IKp8C-*0@P1Y>?v+)e&rh02e+MTz98$VDBtT%3K~yF
zF$=LP#-^fN&+gg7voaa28uVFNmAr`F>9Hv$wjTb%q5r#EYU8CuFRSQwL_yT%Ta0W`
zNm5tUt1!IU`a*NwGgeCQ;aD$>%6aRsIh#B|EQ-|MUxvtQk5eZ;&9jS&bJ<)Dgi)TK
zDkHSA%7wEJ8qDf_k_tse?#C}@Q>dc>lL{(&tiRi2CWAB|?^z*?vqFOb8(uj>R7h^#
zrH@&EO4cQ$x?@^?9W&yiS>IMKqBldJa+!FFL6S5?QzU7a0fm>6q}g&W1#i+@6vf|r
zGrOzW&MlC}8M#Crs>%{I5Gv)Rx~u0)8PX~wVV3aY3kXYo0Q{GQG*l>BHglQ*E$90s
zaMdD%;r#74!%8MRIy6sJrjnSr>QSrX49G3CS)yj5mqMOUu~Kj<AMaSCUBeOIJOKSm
zq;2#Pkv*h(x~r;Z5s#;#F6d5+x|?`LC6+L-#&PdENvJDmy`!FN>a0PlQA{?)G9K>z
zoYXFK9lt3<{}|Hn$6U_DhpaQV`LdY&hs9c7!<F4|;vQQ)iB;pnMD%pU!3*3auaFD8
zIdgQ2RPe2*epoMHfusB&`%_Rke}yXwe=xKl{aS_fMPCo(%(GhSAP;m&eFlY2?mVL&
zs^zeAoI!bw@gAu^2Fo?6Swqt)gsxiJU(^amt5)-Cw9dJid~HHUnfe<iy%w;JAQ8t!
zu{T}KuU1W>SO~*CORqp0b~@aEsh<FwVCjNL%^vW1JyIagL3H<<irBMq+m3wr^2q)%
zWx7#-Ar;<BhhqmjirZ<c#{(mZe^${bvA2OoHP8s;-6mz34vzsr>%b9VPmy3e4QDq>
z6TkCPVYNZ)@BzZ3$sPkrkzhH#Mp_`;^u-!yuzbpU2o@Gw@S%9C>DDWs2tZj$37&>Y
zs}Ck=)#MXaH^Ib^Hh{j0@Qs}yXUjvnaL$rZlRr-0ZKd%y=%_1{*9RDosD)t{-XHjK
zVXxL*arXgrC179j`LCy46A65(9yPg~?OZ$h&0S&#z11+CLt?sN$b-9><l<#$a~|}}
zcasnD=&R&w)D2^zZ$Z7ed-B7rj@r9TKKZ8+fU#c$4}sWU<@13w657uA)ih6!wYx4x
zzq<Hf?#5btE~T+~lhKBeKe5f$CM+Wx6)wSBEWCJdnO$+mK}4?smvd&e*7opVV=q4n
zeN}M{3W^b4-N1@XTSA8T0)DUiXn6$+cxi1pMX1D!{J37XRJM>kg`2lc)6EBN<z5*-
z!g?{+&fnI5zJ?eixg&v_irLWK)+k`(uigFal)>t;9vwpz$rfdLotH0Ai{pC3AYK;l
zjw=Q!_b8~>s`4x(5s$L0OT(i0AbjX=vO|W)go#rCo0%F}CeA(vPb;0a(3u}1Ho#%v
zc$APey&&+7`I5W!nF=PWvWi*jd7w$R=;xdG_J2{-tfft%fi--pkr>{p(f9SS*N)6c
z0Zk2}ueoQ^cWQwb1r?e=dnU4`@Sq~5A0t90%GXg=ZyunxMoF;S?@7PAo$_Tvm(>*x
z3_*9HmaX?46WQ#yKjt1ieA*MNwxPM}dr~>Jp(!03qvWLh%IM+vmu96rP!baHy~YMX
zP5bm$U$JbF-j}@Z?Hhsds;;aFQMO8>ld1G3G;57I!`V51c+;X+&Go2F2f>;?c$-m;
zeToJ8r%BBx=8C_Iw}V-82{q%d7Z#3`?nqtklV3eZyzn+YM}NUK@w=vYI1i8KYuHpT
z`mH&e@PTHH)o6;Q=!<l_pXrJcxWg%<!`=u}K6^~0D!QAO--_hxb0B+smCs%TA5Pna
zFYUzlE|9h6&PMdtlFS#pIc=!lB#n|)SAkVK;nK!!z55VX*>DjO-)n=P%7~&XWA>S3
zAr*^s`I4(z8u?nZq2w4^6z>}j$u}z9x7gg5-p#xvm2Wwpl$`$vCkv_(0w-d_P>*=W
zFgEtaq~I-{;rZgQcQqob)LGsRRP@oP_U)bSY>++y4JFwin?NQ(;cJtl)5i0~`AF*Q
z3bwDgmAfK(Dramo^#~7y8;{Q<E2$U)Q~^}&g&LBHwv)(iVrhNCB8q)WWx-#xHN!6^
zZ##*m=h)Fz5=pb#=tskrjn;(r-p0AM&jcQ83tgnsn0dqMoc*(bpz_Pr2UhhJ9%&(Q
z6!2JxtUdU}y<IlzvwpNL>~0tJT1el_H4by|@nGi}XVRYJaz~XTT*D4(M3ot6X6|wb
zLbk`^I}tk(_1?pm(K_~jd1&YTQTDaE^;6EL*{x&P9Yp_9(YJ?$XOWLAH7=2==1cEc
zqB-wNevv=mre1}qvPUE&MB17t^(abv-mxr728%S&*#XXN?(|n(;&GG+nHrza4NNIb
z@VhbqG2k(?DIt4|S7`t`Y^J2mqLz3NZo$KQaF>#*PY-m9XvkPAX!H<U)ihRoS1rnk
zSe@^U{^9dT*)}Ht?q<_yU=XV)EMoA-*yR)rXNsD@6H-t!=7g&bTJ$B^GHGO)TN}w`
zoZe6c(!<-(W4o&$H$zjQdI7<-(*}}V4sA!>d6JEIsRta={{6;sqBOq>H0w>x3vZOQ
zS31V_cXU(Y8F);kC1jmcC}v8;TD~ur{UIfwQ(18t$VWv~tX2zquP18lh*$W&4Y9wJ
z8Nf(Q?DJi%azmx9gb{3dJrRvA7uAXH9_uEc#y61YSn3iB6@BYyQCJ8{Rf-Qr_{C3O
z0h2IA6dZ1M3^C+i)M$&P6u>}>LDcRIHWLpa=XibxI-kOKL1J1c;F4w@bbb;6tb;@c
z_w@rrQV#%0%%H9?IbW8e5o3u=1i(oqCVmk$*OwVyMbFNKT%dNQ8!(035)u#a{rm>&
zE8wW2p45*N7B$PNUDMPG6QVz=Y$hd?b|gx;l<@1Q0|^g3YTKEb-<KKgjpj2vMeTW*
zq0pCw3r!cPJqC_4?vL_qzHv+^)NK8IEm{7K_)b*Xfnm}P1~XNMA<r$shVEzA;z7m;
z!bMN$!f`E7Lf(fNt}YC9te=h(@0Ko{GjkjGv<%o#;&5{Iceh?3g1uiwD7CvH>MP#8
z!zG;iahKfENLbp{nYTW~*X7w?-x-403MyiFXIv?fCL*Vc-^09~&Pwi3`!~*>>+0&x
zjlYJ8iRWxj)T}37q+7epc?_Yf7ZUaC)(4ZMT_QWjsAlx{+g45&OK(GTDu0Fw@<#<L
z1*k05Jv1$-{AkGXm&_mHhxRA^j}I2h>B|<LuV=+t>~GORWzchO7mOCSSR<v+Az7l?
z&Wq8%GVSi+`HI^|KjS`!pShJC@^UlS@vhZ!JvT}-6lwo1MIENYJUa$!%MbI+3>WGu
z_ivBc42dQEoPf^Sr7$tBLwpW-$>38o)=n<LNjt+!tL+f1YQ%EKE2BzWgCU|knY2-v
z$uqI(&<;=>#TR8+V6?e3MLV&~Dq}jSB;wVao@cH5tB#9y99<cCCCtTgM$<0}r{gz?
zH4bmo<X-r3T%BY$B4<(p^Bd?1U(?FezV2z_B!s+d7T~>~#V06w<LeH>GOIQ($;!-F
z&(5a#7|7gAHpPPmYAVB5e8+z-RF!JgXvp4w&KFvdiXE30y?gPtuoXqkY;M&#XJkbf
z%Mu7V$Ij1>qF|C~*W{n@OJ+7_Ub#SxG1cd-e2FF+)P&BihJWnb&V77#LteKj!Ch6~
zD+dm(u+P32%F-K%7hZEPu~?h`;_ie-8&%*z>{mRuNF3TQA~pr!c}U8lgVA1W5;KUw
zBz#!js2s>dqw^S3P$wD?74Dje=tQ!uVkPGe70CuK=TEh_u-+&YdENV3vGxpRKjYD3
z<?JnmvWC=F8MY)>2zR+#{7lSL+$G3O*fk%UccfV6g>Mg5m`c?NxH?{}_3{t$&X%2N
zSFwrIU72|vzRg_Sd0PiyU=P+xuJD<()j%O_Scj6vIWwx_@g?W`&hlgn%ND@YtPei8
z^^i69)Ogz5Y2vBA;JvqO%TscDu?{L+M!(KVxc9<z?i%pK)A)R$dy|pUqHW|9HK|kA
z4{t<rlmsB~jJ|cep5qU8>faK81oWKCPhZtxP$@;Ec)D@&r3%2dXdcSmpX~n?$IVUv
zO&*SNI}R_NJ(!cI(C%@1w{$a${y0XOucd-ooXO{VRL@E+SIm(H0#OpdU8<pSB&)95
zQK-&D3V<CQAI<qxOD(g>;jC3B1r<BB%bueyiG#conorC*;Y~6lzFn$r_-$h!n9anA
z|Cq$d&xt?uWjhm6-w{5ha<CS1Ub}j;N{pu49iUEZb^^@&zs)3Ru~~HLq9{{$1tK(3
zbvXYTM;5PP#AQKxc%YqwC;YRW3xmhX^eDE3SJzdsWv=E>Z^{mnmzG6VIRxx?((m`{
zaI8#0>imwnjTS~;CiUaT)efAvtJlG0nkhXenZ#VQeHSQ^s#NhWU0Y~fdYrf=biqu@
zT!o0*@L}Ue9k&$LvC+mBn&#7jnv3YR$q{yTYN12VYL}^agm`vJ6MWx(>zM5#Ub;ON
zb@tYqkt3wBmv~3h!CkJy{iLCE+49_J_b3wteM~2F334O+CN}Yr&X}5<sfE}`zK`#h
zo3W?nvz6C<leeH}$Ap=T9;g%4qZBfC=baT~QQGm4$%VVhy03f6)=(5m&?gGA$tuQC
zZm@K52kXno_%6{4sraVj{lN-+5A(>4x7E(;5wAt<%m+spU99tU-A39g8D_|{1?5r-
z8cs#=yT*X^<xftQ+0I1BtxqJnEmcbP4J_0;XdgG86zJA{oAw~R04~c;TBf4eSAePY
zr$2N4$s+<90vPfDm`7g7ZS(zuvXZD*vI_0<T!Rhe#zi_BYrxG|cRM@1<KiyW-JA`%
z-{YEU-lymCYy$YE>;&=}WodMWb@^(WTBDc|2G|Kh1hWsxD8He#WDzU3$69c009}A%
z$ss}L{mt&Mz}qH6iFOi|+kTdldjl89>whs`7b;DH+t7hz!R5gj$Xn&ZYljL^aC=qB
zI(Q69Rf`-=SN_1!Hm6!%cJh4t7NGQbM|AqjBM4Wq*kRO43h_Ppw@w&_@X#Bo%UH@W
zYLV(526Byj(eR+)vMvtm1L61Q(2Z}D55Ce~Qyer@0P&yJGCs~L2igsHH14{B%<!Ju
zPl7K>iH-M4Yt~-bT6(7Dq<*2EvRepmynxnBO!R7;?+TYLr>9MbpCpI8=Eae}x9tE7
zExb`T{~Q(EI9rwnpFAA#mUobOw0%NGZ7l9(?6%OH_i^PodHf2y@4`8%QvK;j35|w(
z_0~*9J$z_ucrqV{Wnxsd_#?ByR9ag0?PLhJnRufvDd<!9N)#1}Fj*14VS4<`A}{s!
zg$Y?hBTw~mBkfgKV&lX_<LJ-=+_)<>Jk79bZl{zJ>&@sM;X^{blmcW330m%4vNFJj
z>!(mvDzg5&3XIhYky!9p^B|&eKG)fA^<(LgxZ?3K+E{;eV>`;@!Xssib>|rSp)St;
zKezx9W5QS;^6ig<upSr;jua}g?}#2OwO|(~QBhLD3^g`3H*0o^q8jAZ!`8j31dn2@
zlbRg2YaVbsg`7-Vm`db<&Czckt(PA~*Y*Z2jFx@tDH?cNd)#xofsx1y^Fmp-z;g3a
zkS5VAi@R{@Q6ZlxdV<mABw4^~zu7`+blYl`-(!6!&S5D)E6gKZHN+9yn>x$)kj>TL
z)E<{6i*YQeA>7iJnq7OJafrH@kIv-&Q`MPvz4%=R$>+GVs!*lzwL)!&$WSpEvd{-J
z!XP?IsVVC3IulQQ1DX<nri?s8QN-ax#CC5aBQ%oqGo-yawDtz20ngg<vgZ}z{rE|z
z@Y*bLYZET7hd!Kv3uKwU*Xh7fk0CXyJ*=d*!paK2gJHNEn)u=|&X?MPw9XEW*pNi+
z>Z23I&W99P^OMFO?*7IN2Ec}<3bsF!XRo9Jt;q-1rGOEo+izO#`qr^FmgH9>ekty;
zQEb}{MaMx{dnUWrrNqZ8#k<!k>)+(>HS_HE`_1Mxj(GPUHtI&nnL1psuG{=#E~2}b
z%__C0;4`^!erS5SaztAfjZ)ju=Bq--;YCdDEbx4*FV4sc2e#SQrB2Lg1za0G=AgSr
zE|)pRPhs9=#dvk3hT~VhNHy{_K26s8Y_s8=AeD>t+e=xQ#?I9;Z^k`}f)RDxGWSXm
zz5G2ati?K;M=4e{XKi~tLKY{ZneVAEt9ATju(R_)oH7YD-!wnW&@1iRLR}wm>{*jp
zL^;MTRHGtzr)JA6SR0?e(UCd%ImyVLZFE%z-P?oZ0oJ$*tU=ABdyS7W_S-xbmTGkF
z$vzm>ga?YxroDK4NFG@W*E^xB^Dh#b&JsvksV`fogX^*^0MXFHeG8IeJwkrT$<|Kc
z&AKZ{ue}cf7;=_y0x0BqiD6cw3H>62!`cpHiC0)F5}1@4+uBoQLyW824V_#n*G@)|
z(rBbC`T53j*pY@I*2{|V`x^Qc=MnLcY9vQbJIBz4V=Tkt8+Nd#!bj-tDDg|~>ZNJv
z&NuWH%b|Nu^E4L+%Ft@kg|&3vw09hI2B85RTr}-XH4)#!uCKRU^5WNP4F(4l6rlt)
zOefjs5X61CmHb&%9k~$g*?xDL%%$hq;bG3+)LZ7A?&St}@#gS#PPQ=sc1|-eRc-Wz
zVfG$wKhxML?&?|mAn}0G_Wq}#Eta>EZEbCHb8`oUL^_fca9+B{lm{33(n;nU%}!B#
zgGLG>Y>%~ICig$Q&+nLQnqFGCvRmcucT8-&2wmBZ$s^CVQ^UU2+9BD6l%|JAi{sMN
zt!ATmG$Y?{+hE)e)?N=))sX=u9Fd@2mJV~B@fTQ9xHLX$nU5*QDf+Am?w^1zC<A@{
zGieFl@mzmqYtscTIFa4>5sF%Hp}DJ3xu8>x5Qqtph`X+Ihn9sLGYfn4;k)puN50>>
zzP=o4&(eh>ws9V_5bCwQNxU?K+408gE@03$<y8=Qh@9S!%M_dQAZOGkJBstcfskYi
zX6{C;&;{_O;#b&VUgvflKisvDk$?6y2{Se_9f#+2)S6N}^USLXJ971{syeND`t2K(
ztjV%TiM`fBr!NLZfMcT&<tWdUHNzckNlrpl)!}1(&(Nt8235FqJx|x8`4$rMd2QF-
zhOz_y<eHSY46oN7NrJbQ+?7jj_epioVe3CpvJ-f1X;{Ea9zrl?^G)ORUgm*8efFU4
zXE%}__s;>uc7_20CA%A6K|~;XSQ<W@w-fbjIz<G2`6CV$A)4U{39#YiQfFCv2i90S
zM57WDsn{S{;o|9^4}dSirBNwK?l-fl5QYk8eY}hCx_^3$aLjVo4EuWzqRE-1kvCb0
zf@4;P^~KEV_8Tig5qgqfBb0=z7^L>j_N1!nA?({_m>bsDJCjWt0tsI3HZq@H#aZrV
zqbw=mavZaD?^cJLJyrL1S1ql<T!c08>IZ6KX0s5tLb#@89VjM3nli5v_=<KH)NDS#
z<g0pr#a$6p0LAFM-;(MG%&=*8GMd6iMbnu@_O3PVx6ZLu1D(tU?GvgQJF^^xk{RF~
z#B44>kUd$qU;!P?4r=7A;DcdThgu4p9{>FzEjU>tkE1Inh@!K+RR4fsW0_F^eW20a
zVQh|*L8wsnWrzcZ7i2z-yw%EqwVs&j*&6K_(;JuAcgqqBqv=SCgzwouv!o8Gg^#@M
zSKaHnO`q=FYg!M|k~<b&$4E$4wqjlI2V^wMU2_>ob@lH34ukhS`f~!y-BH6;_`cO!
zbzE^e8cIkRhKR>7Zsv!EaZ}u;kp!c%P0#j5s1}aDeU;(5i_M!o?2r4kfq|zg$JfBT
znUwCdfIXk7(3wv~hq0vDcd_q4Sa)O0={D9f$ZC>|mhn-3`?3pHv}YrAfn`&1aj$W-
z$o%boCbQ9i0B-vtry!T4bZVAin8H?6s|v1yi;bPnLS<|pcRW;X{iTDOg3s@psS`Z4
zKKFU$!;5-THQ)HsDIF8c`6D_%ER~*5Fcdp(b_$f+wUN$`Hlgm_@a#Ja@Bt-jgKQS!
zKU@kh``=d)*z$a{oK>dY8-!S2m9n=p6K|F;oSQa1ZJkFyj8ZJI6ySVl^vo~?M_pBr
z1Nmz@H<J7&mgBW-=V)`{ZI|O;RKNQ@8WO0zY-X<4+Q$uj`?EC<x`?0t@IY?ufL#YX
z@<sk6C~{d%H(ZJglzBiOPRSPBCe0`uNjxW?rPwNvX#8E(&2;4by#O7Jtd`I#+o?~E
z8^5zuHCVE7kC;j+BvMLWRmH^tO5Ui$L~5@a_v7WE>PpTVCS$I+mDTR3R*<$t+AwF`
zOeeXK^);H>D_c7D3Flo;pKV$7;h#VC6%~Q1>Zo%?S=^FF->KDL(QPcF>%4M%_caHP
z;Cd(7F4YVH&hCM}>W0Yn<erWFpwmT{bH664-;Y6?rNT?M40b0g+pX_}_pw)r{z8M1
zFm~12AfZLP1>PT^e~A$eUkvkVnoGi(-i6GaCaJ5VDBvf5)y1<Pn^MmTgpXrKSnv3m
zCL}t5WsBNodC03FcXZfdl7Ge&>q7vLU(gqdEpuFLZI}t`B=eT4vJunD{Ml-?P=NI0
z-1vh&-EibFrOZw7THDyzgNTwj(B4QP?`#$oxGc^3O44y!I`z>S^VpAB==qK!)$f(2
z6$I2>qMVaaGF$?9&%FV_EzS^eor4EPxr{l-Em`7!*q$G|tAef1@@3!_)`Px(3bbXB
zUr%k@WubqSuYXmT*m^!9*>|R~#;I|JcvFY>RY!lapwUKpm=UrQtY}5750sY~8Pgb3
zC3}bpx5Df)#)ILx+t$0)ha3dZMov$`T@QMoH2P&HUew{0`?lk$^NkND@j?dXSmjB3
zV_2N2{}kTRmj!SIr&a&L`O4zF17j$)$o}+HOUsHvdvbE}%&yi{oj#fp^;En=>G^@b
z$~ybCACzB+1=o!|6$sAdJLwb{xvOU6z*zTg7CT}S<QOuP@~WjymSk9FLAj5z?li0Z
z6z0W}r~IdaPOK2nWRPcdSnG$@I-y6LU9ht%&C?A*VliCO#t=elRXn{7Db>Z&mpup~
z|KM|d!;AlN!{*m2FghB+fw`$|WKBB-Z~*sj_PGcFw#VPE58KH=*zBi}vFd+m2L0m)
zA=~CI|EGf3Z!MgEobvbA@59%s5&!=B&oMyi{m;1kKa7F>&b8*==nU0@=-4iM`D;cF
zpw{;<FCE3&>AAn<xc=p+S@v~=gM}45c^z-=vj5AE0G)|v0#p_L?GSkLLU{G}g|bpq
z;FZ;Lx;Zw06PTTyJ=gH>XJ&mYa(4OmL*U)C%>OBCO!Ysg;Qi+q78VuT<UfaBCk1^{
zOa`LT7uhpYRRwdpSVHj-{|_fxhplVQE;lyJn%owLX69~xP2ZVvgV)u8u^tWkZ^^|E
zckLFUQ!2r4OI>V+UZBlJk9Kx%f3NWxtNOKYb&%?ey1AhPznBcd6nmg2f4{?w4xwYL
zC{946|6+Z5zG-Va0cz&L@nP^dlHd=ltiu1yE|{VJTg&}>COI3d-j=4#>3&Ueu18w_
zo}+(xK$ge<$EWS?&wn}R_qSiu|EFVrKb1w)!g}tNbGl|D(e$h=3Mq|pJgIqw|NdOZ
zX8n(@fDuFm|J%!v=Re2(GX_{#|KG=;{P(i+J@x$Q#b3CKb1s8)D%MxO0KfTlm{+>h
zZyCgYn&O7^wIm#awj1&lX?D9yhpqf4bBZOQKj!}_Vt;*d=S#4(yr5_|_K6L6!*vBN
z8mX9FA;jWNdw=h54cLENlJ{EiE>;y5cJ=jD`>RCsuCOV%krt;2#G9vg?~#=?;A<j2
z+Ha9D&isg*_}_oo2&h_MedT=;5cheLf|n&=k4Tr&{b*&Im;*W^n(u)qW6`3BR?^cZ
z&8B-x7m+xiB#iw|i$bx!Z>(SD-`3v`J)ofsD>e}E?1QHmwK(8Ks0iYJweEk(wCmR4
z>7(g-I%11UlQ7VEyI4u4FT{!Y38^Uubfs!u;CO-}yH`7LW3@2hBNvRf(Tx5v<%9Cg
zf4jQ=6A)@+$+L~};D6QXZ?D@+bVDU^_aLXwaG%4{J=9$e9TH125CNk<m3f%f{!&fI
z_fSXeXD-HaD=VAak9N7QZc;a<Tdoc9@TeHl$mjc~=;HM#$_s-j`M}JWbQ|nROewyx
zyGvx$f0K#$@H2O%NPW6<|4YQzVbl*|bDk?OxPBajT^k`;tdGN@pGms;K3oXI7v|zZ
z^cds$wst_Fo?!H*EctY^z#VAYB*uSUElm2zRT%Pq7}3hMfc(d-{r<{Hb?y7IzC;_}
zhy+SdcCwyTYDzbG4H<g1!(MB6T7WUDs1nCLnDZ3Fto2M%>sD<}49(9^Gx?OTOD-42
zSoZ$t{oau(o@#4lp8AG~^c9}3Ji`O=&oY#4F@8s9m*Df=!Tu&)CrI8OI$$@Vt3PrD
z6#`Nj3A1W-;0)HIZKV9~X<lPFk~RTtTr5{NjXh~GDA)6Zq#Cnj^aOvU$M|Ba)I+Vl
z0ZrtawgI&QI-rzF#O|JR8+MBUifPW1#Yvl$*iLzGC+DHQQ}jdG1movKsP-N5%`?Lh
zy;WK>zy`hb^|b<|$z=aBbLsumEL+g7US(DX8`Xl#uYop+7&**D&m5=Au3y)BMfLNq
z1EEb<d(>iN;?+Chx8e2mheqF?AWw{n2`J6&=i>{BQy}P;eC$2UaxU=@!|Go3!6@b=
zA_c?m071nuXy3|;7FC4oD?V&woI5YXsZyM))SEtO2-;K6*~}X9w2;}ZBb3j9U9qwx
z7rw1qY|bd3Pe|PAu@<VX6c2x>h~o9)%2DNWvE4b(TEBxueqF9<Rvm7A_CZt8Ew2^W
zT)++LLaHYk{EdrX7W7<(4e!7FGiAtscq4=fkmiog7UkMxqNR4`)wMfwG9X|IFvoCs
z*J(_e*43t;I4;14?pUoHosoBjXV}nf(=fc_0khIGP5$`tU7B``1TGmI_B^78&W$oD
zmtL5Y&GgpZt+>x$vIHhEC`@9gRHeSr4aejjf*U|zQa^Bd&LgK+;LsDDE&<U(3^a_@
zN(ZU=!;wl+K}=_t(v2l+c^ZhJ^)`eAq}Rm3P(WX6s?6MU9DBHXP`9tFX8o!8vnT1v
z<8>}-Hq6f!YFHCys^Jn1T6k=APYC3xyCx|G-Ev1L&ccuULw^Y*UbN!7aG1%+U4-rk
z@Ti^@UUq|8x$rJUy}>~Pg%;J>jq*Mm4l&Q7Iato><v~@7$LZ`;>87b7%=Ki<9|FHb
zS_Q4M!5w$mA{My6zSDF6$-$v|k&N7h5)};2Bm1kO<tr*HVKzVK^0H&u%Yg``&lAl9
z$Kx_yMbxo^t9wUt;PYFBH?id1uAkK0cZqMYo_C~c(c8_(R=&CKIz0Qm`g!L4_Ry#y
zgu)b4%&w_&X(yN8RvTvI`GZY!e|93@tz3Hgcp^hr;5NX-)R{VUt%)NVpFe51m)*`)
z;t!62v&M<E-|M&^7|lzmGLa%gLIIqsLf%=69o8iI#NgK6%Y2HFF;_LdM_2)m&|bhn
z>7P$kth{OP2s)ybBieC)f6!@Hr6#(Wo>9|QXJKqR_IbZrU8+I$L$0y@+2A=&+Nv;4
z@{Gt}$)j;2vB3E03Qw1$`mpikZ5&QQQ8*_2sDk*9v+xKM)tP_Fe4S+?+m6S(uGDk>
zPdZ`%9$Q$bTRpJjyCZ+Y26A!^qGSIjgAi0-llx`c^t-{QLRT7lI$BNZ4k-zVVmH`M
zkXua!7a_k|x0*c9<7^-T@21^1E#PjcXsG-detr`QK|mCHc<zx>1awhQn`m$12|{@*
z<?&O%qkk`p)rGWQ3Xq%$3x!nz_&-HQDEoLW&bH@JA%tGQWp_4F&<fC;29sP5PD>{B
zgoOnJV&%85fw9WE-Cw$#a3QuHwby&ZJJaD=KZpoMTTaWc49{Q90a>T~br;!JEw87)
zSJxI2=u)zf;??LB(b3Zd9yMMJ#y#mFedv#A0N;^&J->KRHfqyw+?g959)7BhP`;O7
zJw6SVsZC~$QU2ZxZUKZ_S6s_Imtc!2>RdnWIC+BQoF70HqOEU|L0FzIVJCSUwry%-
zlfk`AJV?IXjGmt2q0bs5`F{5`d0Zo+IUb3_*vRwAMP$p5$G3#q0kOQZzi;c}GF;Nv
zd2@s0!NJ9CETQ0Ss1*QC6M*W<r0-mV(<pyYZy01KEfpD(Ul{UCkWx=ebusDH2)}Yv
z_s<WxHbZ%}n-m(WF?;hW!er@{;0Lb~p1f@o7ST99Ey>o`yyF<XY_SD;DTUw*tX*(-
z)Tgxsp$X~pPBqgo<X)XiTl^XY_VnIPmzSVz&RkOURwsZ8BIkBpE_bdkoFv5m&3Ynb
zH_k(@@8Y(6LCn*fgwJ+rUb6;|*evp|Z;DqJYyh&~U-je}p?u@Vhre$awM!*X?VLDO
z5cO*(!Dm!ts{a;}66-4{ihlp>z?giBK1nyIoqhGFPFQm>spa=8T?*xw>}%0f<D~VB
zusL>|e222D+_Y@UQw-v0mt**&-G?Y-2D!n%mE}VKeZo$P7bl+urdX?^NSy*r76UtB
zH0feMaP2iZFx-u0BE`F|-VbzgrJg&~ZwHn!S_tsnvry<W=Gt5ds(wFmyVvw_ezkbf
zD;qjI-!CiH!DyY~?r!N{raFPuGa|LtDk0>dwJneBR{8G>-t$NvFfrm!(_xPpTF)mu
zED-()mzUvxU|w*SmMU$pgt{J>%;^C~K%tCY<8dc5t!u(}eQ6c5Q?ieX6IecXPp@ub
zaRz{@m|1|Dzg$<EsY8u7G=z501@-hPbM0X`^(IV^^3uF4gbJ;+3G1Kw)h|hb>Pmsm
z&j~Bf#$ATFw^36)i(6gQT?M8nRxBiPsDyt(AYVUT$4dG6#O5<I1y~H(MM(;`h%XX6
zX5V98oGCHIyv4{D6vBt{M$*4GZ+%0@^*Uu9@q-m|Jn+V30IWE%UR7kZGMPg}j3Wy`
z7%Nc`rlZ&|q!fw-8}t|I*W%sL=u`wV|LnxkSw<nOYhFaeQP2su)^j6j*`hN7b)$TA
ze<rMf*EXlbHOLRPJ6wlpOAm4Wxi>)t@jND~+3#S)_;bLrGMcZwH2c+gWyjI(TZf}{
zu?IE;lw$BpZw(v;jAyx@?H1BuXzsGBZ+n&xSDtyM(Is^yNO@O<|H@TYzR`8lVxjI0
zl9w62_)&C`alr4-%7Vf@!pJNe<FYER?UtQ`bx{u3PD<di;2|~?e;Yn)T>LSh!rsRB
zNPMU$;O#?f)|2RDGOZF0k(nieR`k5zGJi76R;6v6sK8aI!ffhw;VvK6i`O>=z7|;w
z*_pd2KC$2J@aNwvbUZRBX-W6sr-oV8&MM#i{&kvmVT74r^Sl2WO1^tf$e+aA!fD5g
z7(NWSr3O4hZyz!A$>91CSaKQOjlDtq{Nh-!Bnx~8Jt6&a=km0)Ka|cJj0oZ6zzt@w
zpD#_%p>%xr`jJe%v^&R@it|6X01u1<k2qwx1X`=PX_rd7oQ>QiFVD0V0#W|ufb20~
z5)nnh|Hl2Dx_tQO?LyBBU~akOvdYShBKE%x(e3cql24n7bYzplfaihn>BxQpBUEuO
zC#1dZCGUs?!`lkt>$oB}fzDo6wo-q#s}}-0P<b5ulghvLLaig+La?py;{*#PC{0U@
zOhqN%3kc84{|U;$S{Zcxe+oMsm{GpMr4ezoRl?=i(?ZlL3P4I(eL@B*Dl$&-Wcg51
zk~QQ|*oHon#jY)GP~h4Y6?5TgrX+Eu9;6(uglpikmuj(kNL#1wi?qL?();3|tCp_c
za1QM+zCo?2=lI7H+|)QJzakc^_eM%zy8N6jByj6o*cACIgnf|D8HQ7em}p_xMJ_|s
z?6rG~(H$suxsXJy{nI)8*3I=jYLOI52s@z|^%|Q+R<6<dsDIwCUmW$erb0SKE%Cx+
z1%Ns$Xn$zQWwkN@Gx!*;!%6ti7h2vi6P79T?bIbY+W(9av!0Tz2RrDwEk~9C#OE8|
z{e>F2%@0SHEMxl@DF^|DCR;Bz6x)MQ1O1yKSJnrp4DBY-Cq@9UmGinRBsmdjI8RZA
zMxZx0NuF9-Xe$=8iwQ<vjGa?wI|xm%o!K!Jny+}!c;@HoVL!-kLRt=ZQxNHR)S#y-
zXnt+)u?caEQHwkD6hsxe)-Ilgl&B@>mHuj}crc8M^}^|<KvwQkREW}403!%{evFpw
z16ByW7`LfHX>p>k%dQ4a6a43f{xj}sm!k&`#5PF_p~=o4&a{cC=cjiSu*Tc>eUU#;
z=q%GtzE>kk!ez`Pg<8|6H|mhcAF|4dGFcLY^w@{2&b;rL9`~fYRG>>gk01h89U4cP
zD&S3<VJI-7gd1Y~=e4){u>_o@TlFh<Ss-#D*|BQ%J6cWZSuH|uQjnG9Y|3Y8-I02h
zcq^Q4d!>+HK@@n@v-$gz&hk&0!@hjB5Gz_iRE>l`=xIjBR)ix(o^x^CpR@w+rVrSA
zkQtY;cRp3}5z(HMo%m;N{`mA>)UQSh`6`mAD{jnMm#mZLS4~T+-K5ZQP{hN9hGqEW
zr^N}!QZ_kt)x^f-OjE`hMirgQ3N<gf0F{w59Ud!jR43A|vD3!quodeCeGy#A7pMn=
z?J?hM*$!h3fer#C=-jEY64R?j)ao13#Nk*q|De4Ly-O(^*653^&NzR#Bikc7Xd>$1
zun2Kc5;V$xpO{Lg-l8YfqQdji9@l7(oAa?V8P#XdaPrDZ6Qcw8@Lp>0ACR@wN7x-W
zUqYfRT+NKEQAMgMYOP{;o4xU{b8b#uQeVT_H;IkU&F4$W#4#}M7%6B9)IsQ10{>dp
z$<-t>UmINSKK|1t6!k6yf!sQTi%?AjH?VyT8Kj2q)%Yu>SlKX!HYH`4_z16EZKm_b
zkO^XTGc7nCVHE7nU!`3*grNt?h#S0k^OnahFsR}+uTAd49ac_*Mh#?0fxtbZv09uI
zX1ua)$T!BzjDX_5*vu`ol|*S!JI9csAG5VMHdph_72RQ;qct?mFQ>IV^>UbCMq}}W
zI6+(jhpZ>QrrMEzRomMv__W)QswE7F{^)Hd!TR9mSW$=l0zb*)pm{hSa#ApNwi2No
zGpINRBNvb5HXXbyR1~Xmm}wnmmLjfqwYC?&`(`}W55*sTn-eVc0JAKcA`g9&2C{4d
zsn6a^<^&;f$^_R1j=8d0BdeaA(<fzge4^(epxQw@@kPFW{iuX^-2F$l>qSOt?An*$
z5ye)9!!ys|2;g8P$oxl~06x4)5Sp>+s3T1nRy2bI1px8u21+LP^f%J?*{^w<$6OIT
zJPq|B<c&N!)%Ka5?hJm`?bc&`H<%Tr16VoPEoUQpCW3=h<}uZhOAHDp<u4Pjt_{21
za|a~7wWo{4_!>n3Y(?CPV2bHmwXYFRz5Zh^=ja;JR-ieDExC2?w8fNpWbWp&n%D;C
zu4%f-T!H-C#S)gJ^SnMQ903BBU9WPv*)D~vCF6>ciA@*6s;p5<Q1aB8|7*@g_>H&J
z{dn&89_c%{Se)S?oXGk%*C1KNuK~m(8*87#nCYND=3Z)3lU1nE{}zoz%Mbg;sBMx-
z($1H01F?*k6sB2u@AHP5>3)>uZ5Zh?!)T|!HJ-jGH|>8NMF&%bc!{WZl4|MZ%DWp@
zw%%6y>2B}&dTY?{*Dpe~F1LXhTkWm!OjIcU@Q;PCj2SKUhM%Q1xn$HyHfhXXUwa+5
zEQp@;HL=JpR38Kd*(4&Z>rs=$jO!`ZI!H1Hn_Shc%b)%|L(ofmU!PDVez=$2;gSDd
z(`4Y&n-PndIDw6fYO?L0Cq!et2(;0cayu5=nb3#upi8@BhN=Z+<f^symBP&kOMS^<
z##-g`SLG%W2X3h!C(4_OF<J#ud>RG<FUrPG9(CBLLfHXkta8q2rT;|O<+37Q5u;UI
zxJJ9Q)(`lZJ5#3x-&z`Q6rUZFaO&e-S9L1dc;hKLFF~aR?^yg%T5JLdHm_{^$E(fV
zx$~(RqdalxcebiJKG(IJp7g#^j#1u8a@7UV{JMjcaK~apc?kl)&yOiLg^U)>+t5br
zG8LIjHasaqAa)%w=Z+(<zxRM1FZV}~I4&(MEe(UAs@8lofeuHmNKs(8Dv>lHutVcj
z$DyrIH};2B(xpckE`G+4yB+yFfqk(BMRd>UigJKNo4m|oZxp{`V{-C+{x`vK(#5qu
zLJpI@wLA;1O=!I9<g&I_=~a;v9@{ef(s_tZx^faYXeQD_UgDPqj>L&$>^C`|Kauy_
z^7w_rhLHqV>nPZss99A#(WE3teb{^I+sZWU9g12LwUt#z?Ws8489eBmrWLuAouPD=
zxuG<mCtbz;LU0O6G=5|8nO^<q%p^uKjyrzW0M7S}dg$XuYRJ$raRG}?gQFJjW;%_B
z^3|Y;NIPkf17=@ot7BH!;vvk_xtgQqk=P+I&9m5VdJp5?+Rch^J|x7<r8Id7z+C32
zj?j)nv-`6gaIu;zR}U!54dVkV=(#-pcH=>W!#3H?EWe;l8{-<AahCICOHJ4g#Z*US
zMFpd-;>kq(v5ymR=w3cG=WyT<eqo^+a>{aR#rDCvFAKS5c;Es^My>+0FSL}bq}3?I
zyT!D5+wWoN|LJt?r?xM*8uC%pC;f#U3@(EsQR#{|gfEg!Kip7h%aUpu{`uq7{4JJ~
zsbV&#YA|;W->hh`iLt6B&l)pbINoz-8gydtS+S@I7rVzZ9kg{p{Q)v0_V_!nj-@5w
z$9_j+h)B>}hAP-4=T>^2MVB%}tV?N5_lYKDXb9c@%dw7ryT*1^7EEto37KysVwFnK
zOuRxrJRk6TH8isF(a?@fcTfW)!}A53jIWqQ))RCHOEAs1unY==j5UHcdUGq#!?NO6
ztWJ@6{5-hllFeZok;-h2GqP;eXK#%bENpR}=0EQ+=g||mTbf(l*1id{q^y?I!H9H_
zy|i0Hb^C{{sJ!h*aH8#<POKQ_JBZ<Pt%kO9KT3@y@q=_1uFJc*&_v6O74Uzea#K@Q
zkjxCl6;KC_)dYJhsr=~MB%69%fzQF>-mbr=Im6Pskn<pM_**>fu>>ykbyGKdHvZMe
z=q<&j?r5?7uk8|&IB9CHd8!Pq(Efa4Rc^};36qm7>OAM^hN@CZcd=`K3b_c0b|3jI
z6zZN1B8?`7AY3OCUi?n9l3p8UTenl-9LGylYatEvU$3k$muT_PZOJuOpT7&DPnuI#
zIueh==5-2xvWfCDKrol%#CemaEeVq>iN`(aoo-kU-YWvi?ki76To8wCP)touk58N-
znX?ovsV(?FFu#MIOEw&dTLgs?EH}B?bd2VnhTv$&6xCK0GkK-^@?v8J%Zxww(CJbT
zDKWbkCK2CwTUWMZls4FKYU|0Ts#(2byDBA-B>bDgOGQ=h-lV;;`POx^cRusT0v-Wz
zS?XBs((~B)rOyS$6%`e&nyV;vX}J`18obFffl3pXLG_skyDSClQA^iSdq5|AEw&f)
zS{f3_b_gph*tOTDS-Fe4kxw>%vx4h(b>7&#3tcc>JMrNp7bo}J)1b-Fk~>d)rrL9|
zyZN^IFrh(mJHP&=k?>2PQ8q^ruX7^XFgtClX0#(;8vUF4>E#}__4{Hr*1AI3hx1Cc
z@KD%z61WG~ReH7sN6srtcO-?(pyV0Jxv~4svkw9{n~V+JNj&D1wt4YX<vvN9+O``U
z$(X-bUcC35>W|#i4D_7~6{Bje8U#IPItO0Qw^S$K4SjJ&(T^QaveZZymHkA=6*t0s
z@Qxwk*<LK-2}>_HFE=NAtsOTwHArGoPyKF(bdZPSWlg8E%uN-R#sp(qP{<6_PUYOq
z#81D_tN0vuFbTPvC%F<C^q8_Rg!FM3^4{JU2SzN|Yu50Nlmm4`?Gn>2@<|X?tWf^D
zKk2%D+`*q@=gj(R?3K=_U?-{Zw|6~(R2-thc1dyPPvjTO(l`z7qh2)ux3Fb<)puM>
z0>rD9e}QTYr@9e-N4K#=HD6mF{lOzA(y)kPG7uxWyBPc1oji^u=)g&HmwD7dy@e)W
zFu`rhOS8Kll_oS&Lo+3hYv6==VUV%H1>;NT`?c`R`lb-Be9=)%@m1o5XM+?OhM&9b
zOfH=ReddU36x^xlfMW9xqg$2*g!5mg_%7^BtWKBVw3{W_RU%@QuRfgASsm1SO!Qyr
zczj~<!&v{++TSKixtv|kBWrkfW(D=UJWlB%^VfqP^x;RdGc`$Bp6+zL>H{2KjCu&(
z$J}?829Hv@{qTDTd_`%3H@6jnrl-|aEqgTGvE)TN)_sY_m~}n%%%Hf(jmhJP(4KB<
z^SRd_uvr5YSS!AnV10OgzqutAn07m=3X~+01wb#3?Zt3|D1Mdb+tttxvXoOakPux;
z^!B!!s92cQLWSm@xv&#Rg@HZtpu%3(?^l-?L#aX_&+5x$HDq^L<&V7xQ28O+9S_}{
zd<mY;&B(hH%_&!-eg@n=Hc2i+_O!`SY?w}+<9Yw)e7WzOv<K4PsiL0KzT&Iy6>hI+
zFbM9`d;d;vdv%s!?Wu|>cqZL0hLORGMV;rcr2v0cM-Om=lX<X6bB_0Cnc}fUBcPB=
z^XKDFdLAzwg__cJ2)I6j&88~qz0C>7$t$VwyKc(dks@FIaH9F_1irEKhHGE!q^t!!
z3nGC9_hnTk{9gcRqJK*Ps!wkyY3`Xmt`$(_`U%OB1}nI=@n0-{w=;|c#0i1{_dQO&
z|AJ{#CM%!yZS;!&f#1zYHp}IrqN-uzFV|83;X#*8vgdW(%707PbVoWewsBZm(cIWJ
zSfzB&H|zf6EY(Pfx$i#DB{}8GPu&zzf>(CZh5~u3W$ibObTO-|=>qhP9<VN3Rf6{{
z%$)C&uIt0VcJn2#+h_q{O{h2TTyo?o-lEj&s$;w1!m)HnLw#M<$cN*{K~M8y?u-O`
zMoembxLfXS*)i$+&B9Q~eqd{^ZbYg;r=$Dp*y$Hbrh@4jR%q%QG06Ssz1gS%tJ%G`
z*_=r`7sW;_jX$eed*qIT#=luAS%c<?e2zm2#y#P~H7n5+xToHwwVmfe2|qEht^&2Q
z$?-D^4e#ca^OdHgDm2IR!v>Hrucvry@8=Gax%6MyCvlVU!_UGg;OJBTgN`9;F?3ka
z`dr!C+7OIk9*@TpeToCNgF?G08&um3$0uknyw^9h%!krh$uJO<HxxQ-a#&Ucqs!Mb
zi<6o-mE8mXhF=;oh`%POhZdfmlbmaf*~4wN#s*s|{ug^+9Tru)y^Zp!C?W=-<cmm1
zs&xC(DIlGr(m8a)pn#Nsv~+jJfaK6fcXxL)#K17$gYWO0>zwa*{`s!!`}53&T-(je
z-g}<C_ImEM*8SXzIhVsb(pBQxrVFoa`bk6Mxbn{tCX4;`6W_I{IUCglmb`mqyE9~~
zs?PJ{9Amq0nVd~W;BBUJFsDoZT-;aJE6~cqk@M+`TBEu4bjq%OXc^cNO^=ju2@-9t
zpY#|JTQWM!j*X)?TYW3lbRe!=La$G)_FOp}{2S}$VAg+~-z%sQ`DURsJk5CL)Y9YA
zLX+W|WjdP?KEHPLD*;Dl<K#84?Y2YZvJ#|vMXxevaT~nttWHzr-jL|bnsU1Yj7vQ~
zHIz|p5d%*WbO*Io#)`o+umH#7Z0okGmfW`Kwh`PR4_UT2*m`g1UgW5VREhfhqA$YI
z<*(FTjTsOX-77g1rqX}&dXhS+k8Xl~LYw6sSv~tD2#u<Z>_ddymrQGxB#R_(KDuYX
z_$uHo34jcS`Eo4)-3Y(2vtGv-?dj?*|9s%um}CgYj{C#pnIRcAbs_inKYtOA@|M`X
z&V2Zc%587cqxfesH-+M2v+7yYeVk*TmHx|zOhggSdCb0ewywP@_kB(yC;3R`T*MCg
z?QZJZV|~5W;IMH<Nj~N58=vC75I)xbBTuW2(zH4gh5nJa@$6*pj_`VWj#Ab;mFUcx
z-gC8VqR7X2_hqFJYA3?3{!cAVZC4t3j`~buRzvD&xAKvV8v&i1YqK!KvPsa9%B~mw
zW_;%4RKv4GI8i{UAmZp$NJ}*tyGjVuBbn$6VfF#n3RFFqEl4ygWY^d|h#68}YfVLz
zdn)76U&Py5n@QBn6dc80!{u#dx9;YA+@{=fu6;G#m_pB2o%vUItg%SSB#EbU`mxgF
z;7mAQVy>;*(!il?GJOp)##9Uh(key4nJcyuoMhpRH*m*=mNO|OC{36@?$vcsP0t>Z
zaJe6Z?l*&Q^`VEY_2S3{ZEgsy`XJ_4$Eey3yS_!f(XpDX(QVhYqCgW8D(^n1&cMDF
z<`9{5u>`b4m|5%(rnj2cOe5<|dT^62gilMwJXPhQ+~PFG220P0V42(f)xDVr9K#Cl
zhrJ%ij>Gk!oQy++$JW6@ktpTG6_=J4`70x781dxfU@6;~j`7*&CmjK4+alzb9jodc
z8}yS7)wU`_azNgy`0d^KWWsQx#8`@)5rp%eo|n?8nDk6!w#~1Jv#X6pL$rb3A}w7z
z>$gOf*}JCO$1y3NZ|n7jJ@V5l`nE@rO%|^PyIg3Iti7?}ARC+%V<}vH-pZ#Rv<G>D
zkIHoy%2Sm~<k2A$A4Rwv>51f_!USXD$q@U>*8LDRfv{1ov+>pFz~&9wwv!zaT&m;K
zqWX*78r1?*cc#UbU-g~4vHWbP;bV$85@EGcy&<lg*r(6PDVfMcV+8-~B#I}alw+<k
zgDpMES&=X>;ae2o<5YFcUUM>iYl_*d3&ER--wx>LI2g!!*FL3I^!kZ4{);4Eu%lmD
zO`iFu%aNCeQc`-y0b56X(ROtQTgbZ*!F(4S`QeE{29JF&N_Hjl)KU5l2m4y#54!%o
z60M&JHd5XXl@P?p(h~^dfP}-lIh<DSK#<h!TMIWxqYu!dHe8RIvXrx7>Ad-7Q8_TS
zk%-Eh+b-?-!Y54z;w2P(%V2}_;5N}cpCt4#T$_bBK%mU2WAP`Ab>3iL%Sx824!yp#
zn$GBOXQOfC0BOlt<e5&%O@&nya_hhgR&(rtvk=Z`S#N|+7W(`{rq0dj_^_gqYDTBJ
zaA`GPrlw|(jHkRkRK%5xhv~(&;x(v28L%>s1N?(<r@t4J88h~t$lKa(U3(<v(a_=f
zboG%$pQ^r9L^&1gg*Q3Ulm}7~mR7L(4eLmb_FhVJ2X8wUS+0k8Kq-e3o2J}pXyXlm
zmDAFNPleVcC{AGGaz#Q2lf^^0OGZ>t-P)BL8|coX#UkW*Akh)d2gNk6e)5Sb!b~Tj
z8~I?ihl}U7d!zI=L!EHOvAy%3{$H&0wgQ{9Y$NS8^>oyvA$%8+?Xa13x=aYv?Qomz
z4fp)l{?Iue3bKxjMK0c$(1UqNsY`JO+qK-=SjJvh!M995>5s*Rt_C<*_G>?-l>2I0
zz*5JiGS{$+?W-y~Bbj49t^DSBv8ECVb^Ep>em;&8^`+_TC}YMIPRs36Oswf;$_)09
z*#_y;X@TP0ag-_bDzD{XQ`eK$3Nm^vHOO4U{AZDr@fcRf4>EHD3^Ic#@y+YkoZ#m%
z<tsEgZ_i`DPz4)XZcL-T=;F4c5D^QJtJ!m9JL7fIarX~U>!h!N(X=9)GY9C4%y!)<
z3wLnDUF@|zh6iW&^)624MvQSjbtr1g28Rh~SuCW#9>MROxpVt#02fXk-%~r}EVdW+
zP}his_4$_^Jf?}$6&*EK8B!>H!?aYpX?lhi=i>cfZp~DtlIM>QvlH2V<^Z0qWMZ>!
zu3AP&Mt^(qI!q0e#X4F{NCKmo>Rnb$=`+o}2ib$Tx2bo1^VAC4Lel7Sw|z?`gKxiy
zF`(3m@}TOS%<g7!-U{LZFsUb4Kehpnws*$Q-Nxs_54cvl&@D(KvS<?s_*XU~S$|}u
zPJW!_C?_|3<RiX8s&rUozt%j#K~t2_DY}(;M`Mi<H(iUHUCmb5GuRj{YRB)*Ts9D<
zzy<PJ3k*il^l>xwD>0v%IVpM~6E+U37J@!*u|?wE#O?Ieiq}ZQgy2%KbJf&M`xHTP
zlj<J{|GX5^tt#FqJlVYCGQ<Z%Cv5x?bnQ8hHFi*am=iUOIvY6h{%$Lm2a=Uo@X$vk
z&}jH=vTJJ86T9B8{f4<>zN3VMB%rHY2gYS7d?5x}QT_hIJNQ*izZH&_mTpgn-o{yw
z5nqh4Q~q*ga(4It1s2xM!-l_(G9?}t&T1JbfCeF6c@m2Y+8fqN9q-X`lzAL3jL($R
z1=R~YHW;WM#yy4oblKZwFk2_(FFqX{{#i+KN)#v48W#eu+s{2fx5lYO#{PI3H_pZR
zSbC3NMGNU(x4C;2>hHhtI5JXpZ)i|4W8)w@M1@j|yZEAibrGxm>qY8uc7Ew=8ZNOs
z`2J=chs4G^R*0*yDP}e0Bski%E)Ml)q@7M0vm9lU9BMGSUK!DV4}8}{VHgaC8_5am
zbztJ&XXh-1Uv(G{!-3RN{h%EB0Oc^Tcy&ca{O9jN*ZmS)>#fiHq3wii9mLF{(BbHg
zmJ$-FPH}?k#Z&J_{mM823a-uA&<<<5L+LGRxKWy69X_Z6s4KFxN;p&m_gWjm9{Sov
z5aLo11n|(7II)@8n_Qx_pta86u#TJ)e-Krlxe>y|4flo?<p0c#t!B2ur=+N>-La6Y
z?jp{ccx9wBS>V3rEix5vYutRlj|Zr&Vx|XH-->`qoO}$Sjk#9_L*u##*B=XdP7Jta
zec|V<by}`Fyt{SBx2ZX8Lv6Q5bnSg5Pd6o*%fZ5cE901rU3SM0{Nyph+(gF1D(Wr#
z5x)g))f4{CT?U!a@x$(uhfN1~zJqkBZiPF`&y&Y$uP&qMSh$X&yWx37v!{cIdaFKk
z+3bt?Hme0SA!&*^=L<^CXY)rVd`4HxbV4o#iX~&TeL}qFA!e{sGgrY`2kaY~iib~>
zP8<bK>X?<N$X!E{oVJJ!gkIypILy(r@f>(RvV!q4mbf@86Xz7$B{%wJ<M=eV0GIp^
ztN!xgvn&#?i`x!KLWr6i;;xz6R3Yz7s4hUIpOy}M?Y{=d;m5=^gIV{20xyn|Op~L3
zET*QM5Q(=f_^O!C8_Z$Ey3Xe-xgw9+#ZQ9jYF+m9RN8msat1=(8C8iK)e8y>>*bXs
z2E5=<^C%Lpu@8{+w#w1zlAK9XChau`Ybt^1($TZ=NXE%2SB_CF_*U%1;D*RkqS$76
zH$DPv;;5P90R>*%YOkY~!Gk{zHVYv&iM2IO9m+13%#|~m_BS0KFDm4HRgeEx43v~K
zk#?fhLt%!#%Hu89e>Ngr8e59@TaJ`>U+pc0znF^aEza(CblTa26buc&U`Smak)^d1
zDd6Q}G=UKSxk?uOHw6Y(EmXul!acWeeZWL1Ecu1grb=7GRQys^n`z?mm_n-MGr8Io
zx^uYt{S8Vg@9B%Z_CS!X)n=&#R{+-4JjhupO9?iz(ib{TS+=ef+aM>Q4C-;q>?C<+
z3VSL74|$9FUT7{p+2?`M)Xp^hyfs$j`;drLce)slt=y=q_v~r53yHs@iyJ{}2*?3y
zbGw(C{*1E2>`j!GfhZK`Amy5S$pxgb4ZXG?DtV4*GY@TyW_|6-lt@!8p+R5a%NT!;
z%~^6PZ@e}9`sl=Bk1J#Pmlf{pygv)|#R1oHEfwF$F+Mh6fxCLaZK1AqA1@jE+E_z&
zoeG&a^rlAFg30F;eyto)d9PRLzzxmJdEkuZ+;%S_pRw1zju5lQQqGl~Q?IYUT*+0o
zhp@)Gtt9nA&~mvAD;L)t$XPjgf+*Gv#xRCXZ=4U$3m9WBd=?VePSN_j71Yz^Hanfj
zIKleM)f`?{h|^k;FGPNkhb8f%!|_7R-NHTcwE<Qc4QiZiPCMG;<6M-lFpg^ynFk}1
zOKG>GET#NuzZnJhYxnnlrBm-5OLrIUNFQ?XYVpjhPLg5aMu6El7j~2q>&O&K6CTY%
zq*lMU+B?k6^K91RqKUDn)~#gn>?iN-ubihLL9)u<-kE*6T$sr-b<{HBdLHM0{QakY
zN$6Ryb_xQ9oN=6k=Pt-qX=)BN%+G_L-G@!f%V$%>x;#3(Ppv<h&lMJ*x*@f!Im(K(
z_();}12Ws^M`nnJkR#;S>h7<WkDK%^M%g*ci;fKq{T+lJ6)=rWM@i3rxRDz#Oc@CA
zpQ|%+pQ;pPBF=i%Vhg-56^=daRF(++r-SnI()&K0RmTY<#Y2mCq=qG#4I1x2>kbS@
z&c2pPlgQ7ouHH%?5(uOJ{6g}1ct?xkO}H@qU?+im^gp-&o9RL92<7-*g|m6D9QL57
zH{4-i9bU?y>RKaAmayM=Cq@hu6#$e}>Yk*3sZFm>#N4@dG*+Y&l1kr8-;wpT6v#1h
zywu2(0#8L;Pn6gT_UXT*-gGhfSC${H=guV>2G>bPnn|ImdaWb#^S0axWs<}f+m$v*
zeS~p<UJY%ye@N<2*?7{!vZVUaeXsBzUIoBxJs((K=f`NrmQiX}XxwbcImvj<BP)<L
zsV}DTR@!)MMiJag&Que$;acUj1r)gDBkIXbksLScEVgWGq#@j#wSv*X=RBng<H1T(
z(p~0LHVo`5%AC?8QLJRRAGPWZu^_OV7w&pe7-~AGJOW`lq?5<#V|57Xon3^GuDa?C
zq)J9B7DFw%SLj^`-}Z%_ZIyr!EERm@cg)ivPZt<nq{;u&w3SlubV&%^TXP^oj5sPo
zoO_ji!Ya#r?oM5@7pElyYM5##UyGr=B)7Y{sC|_VNcoaxh6g(3iGV$FI+ojDd7j|R
zC)@AL?i$<JTsEyUvO;Q~(rS`I)n9@Va|ccft>yZ#{*>8$_O0a4Gp=foX)e`4{3lQJ
zzz?YPcwdbqAehx5b8qM78ybM^foMn~NqSVU0*cY$`1)r>O%bV?g6$@e9<tWmBv%$)
z;x=L?Lu9Fo9Qa-~CNmT}Jvh3Xapy>Mf&g{Ag-88DGtygWSq3_3kL(<IC`=%RM^{ew
zjLbfjYjNY?fR@Z-aSAE&d~!%Oi-T=NP>R&QjA79);H=;$*P!(ax!wi&EJfKh90Dbt
zuus;U5|4Qm>>+bDQw2Su<;o{k1sQ{QAZg|EaIvCw<F*gJU^ZNe=H*jBXbJb95`uqj
zog6Y)R1Y$*SPt3YwgpuT;(#50zFGzPWt%F6Z6ME5ChZ!4Cxw`P|4Z2D$ukBa9QYxt
zsX<PJm=61Wb$_9+ksGfLjLNR8Rp34L-KH*Lf=Q5Sa}!B1XiJgNn-~Sz4=>Y$2h)vC
z%qMtU3sR-zliW{2?eW<mPO7BXN4_EAGInG6v1Pk(${X`-3n;Zazx73AfWF9D58$St
ztb&e%#&b9Zs%?jcsAplXy<KJQeJkvFkkB<!M*-4JiYE*Bgbz5GZZoNC8D@-HXeHr-
z2t~TpaQUmq_@t#p%({XkjaBzg=Y}tj-P5b??)r~~=*SA>NzVPK*a!@{Y?hIy^C^<!
zr?f;MR!(lzEg#NZ9ez60-+|5zbiaW@P<b24&7H?VCS7AN%dHlC->iEG*Q;ed$oSDg
zMch56ha-Bds}4E?%p8MUpDya3S-82Oa3`kPjka80^y3v=s@fioD=>KIl@Hx`h99*G
zeX4s4s~h(w_Yd6gB_nb9kZ<-@+RE@pesbT0%s;%7fP(4UYP}FGpHS#>T9z~Q1z9xf
znmpQ5Wtdt7{c!sENW_R@hT{$R;(A$q_wl+9m{(RKtesy8sU!#|Nxpy4i$4D~DYUs<
z#NN?z;ne0jAsyG<<+s+MHmbmxhX8`}at@PV{*8S>qTI8A(;6gq=lQU-tZ@!nk)T!M
zvGq+cDJC(d@VkW1=nXXPZv65(d9KM2jhc{$;8FoUbg%U+1kmL)%CjHdAsepPPU{k8
z$jJC)mWeH~xI}gL(9esILF){-m8ib8rC`9x1ckC0(ha+iGiRI}yAWAg?}uwHZkue5
zy}G|-+z&n~$&6eEqBIy%BIO}r=}Ip<=KYdX1qwTx1glr}pUeeMIfEx6M_#TrI#7dj
zh>Dfc?V+%GwfP2EWbIhisywv(&CIr7<E3xX*-2@fpv%e(461H3AxF8>iD_LWH44+P
zU)zJ7z!5Nb%qKSOP&5vf0B~^(+L%~?BOA@RVOBa%Or8u(`bseBnj5}5yGRHq^?Q>=
z5;6YB$IE^#4@#-yTJCXMBEsG2c(!=K%>~X}@x-`!fO;#A=h4ojqBDi@s;KghG}#2^
ztkUK!rAK6>!p3%sxf&|8ypKh1G}az=4xQ>)Xq&DGgt%ob0%DV|0`(*61N^F!SVCbO
zI#x6-(%iWKSh(Y^#??wR3$gsGXat3@f!B?ILQEINt@7+FlOrxxx}CmfTU~?Eo!X&2
zaSnZ1m2^q(VY90yG<S56`9{#Q*Tu$3@5!8C>D(l$E`azbnv|7$BBh9vTuZIo9OCsD
z-=Nv*a7=@3>z1MZ^1cEa0#~5C%;x@}`4tQM<=ai)<@FScossuhK#MljynX8qST~II
zz;N#DO6Q`D0kXTdhtH^1sBJx_lV@3f-lU`?vJk7nbJKct2|uo`>wWw2-<Y$WV{#)p
z@0)1+)2Y+-0FgVfq$J$+FuCbIbX=TPI&{Brg!<rl8LX>pHy2h`WXN}aJ)2i)z3Xx+
zqp}Z}?HHy{XyX@&*$PKXtUsHd2&2cz&=?V`M*M?$Kao3wNJqCs*ik~%1Cs1&$?Ak1
z8tviG-gL>qKRzjP#`Udr-DK@_c{9^wI4IufpfN&@9wmj+mi+lQ3M}HXg{5ze6{-c%
z-rex?UL>MXBmt4kY<BJpshDEvND`_$fy(>u1Wfe1Imyj#{WmFRs70DUMlOqBBi1pY
z2TMT#Q?Qk}qRXK?A*uH=ysqgZFO!ZfYKaF6ES8)SRjfhA5G8&P{&){{rVcuvsnNHD
zGP2DLFegN;YC%Bnc}>)k2HAbA{s&_SQ){$YA!wU?)Zm-?9qPC!8u*bYhCV$j2@3uB
z)?_MJRZIV3ee$6$Kd<tyNDHxup5R$Nm3qAts)lWPKE85&2POLoi=YSOG4JNHT_{X_
zNm-hkj$9|KN@iq(@L8b+ZoIYBnl7-k#nU1jc?gqW)v+{=BfNq)+4p*Gtygt(Tcf*D
zvPG2YBuc*FYg)q;e>7}w7pmjDT-UyII!t1jwr<Z=)w<gaucxmaIpHcL!t`$RS4W90
zjOCb)!hg<|^)TJ+j^RP+tm{izNnEuoq!>nZQd01S=9{k>Crp(@4O1F$a?HHK^3<U5
z^cvNvWq)=%mZFfOUg(+o_#6VYd&gIGW_nDm{@7{e;G;@p$Xj#E%a1b^^19XIri{GT
zuFmT8dS?9}R40Wm9)F$YGv8$S(4kXEt6TqyXMEVRnq~HBUZdjha{Z6lRZek*UgL$T
zQGLCwkaF+y?1ew}v@C0om<t#r={&&j&BZS8C=Xh<J=OUu5gij7cw<@`Fm2m{XP!NF
znq)=~jeLf4YCY(s3cLeRi`!mqf(Vd?W5jsSI16-y;DGKpHgHg<zon(zrcz$gHOR;7
z@%@k2EjR7}JL>Slv91t%y_X!N&EkljMA8Qw+PC!^zUin~6FmE~qbs{wM}-UU0@{Q8
z$KHcH;dHKWN`A#L%a+rxE?HSedmO6B=EsP~Mx)9*pH)N3QGj^wRH$1OTTbV`{5!VT
ziKU^@Ql$4@98h;l`F^Z>=u#bM_&8%gmOwF_<B6~nYP*ceTVhl|VNGZK<pIxf?;OS)
zXW`oKnZ}&WXukd^@GC}NmBX36*G=q%8fzLy8?*Up48Z^-ncI-8mRMQRt7Po)+8%aM
zaYIqAl_8_9HRV00y{1Z@4*XA=P#Z>g%`b!dY;2_?-rd|#w)%REr|!h*{vy7H7s-vG
zBo@vVO1W$*;w_&wY~nL~{NUzsL>KI$E3mbl5wq2<70*E_{Bz?8BMGtQEYR}U<7Bzh
zO?n|b&W+Ch$lW?&xciKYa`YAxL6(T5!N>c)#c<1m1r5u0aTH8FbKEm^QCL_JAm>7;
za_2R&K%Iv}wcrFN`#Lo{^UDl5dvuUorLX-#z3%j_($adY3h&K>)kRM@>J*d9=VFGs
z{lT$lbKI(YRNNbh#5=}&aR2^pVthSR(<Y=$yzAXwN+wW-h8qT_J{0~@yKhr6rCOL>
zMg`)8os5mN;vk<j-&K;Sg4gB!QBky2XyB>%N(X7v7{jt6Hwa6dq+w`DAH(5XArBf4
zt^|qx_gSqzTP@GZNu&n{y)hYGQL_Mt30-WMr97=xr0pf}qb3#JVX?4}&X|f<rU0^&
z!T^u6sFCr7cJkrazu|rlja+d|=b{Tp<#v8?U1`i@(t44=`FPDPV&tKi;vq^|TT=Td
zRqQ}ZN=1c>vzsY2g~lGN#cNLw1jM)$1;23a&DQ92)7{~owV~H!@?=8V=eyI#D&z+P
zr62^p&hYWoNI#TK`A(LhGBs@)Ax_pg)LqejR}e2S)Gtd#URL(_^whD2`Bg~Ez-@M^
zw<7HsnCP*6<85WKZ(#O30<9cC0_HJLM*@BBxV{`OAG$R=<I$YnHGkdx$rb00l04~G
zpffugMJ&@M2+Qu#^qgQ1e=<-JHCAyjZbd~5c`|*VU6U21N=O1R{$sUr1wmMNWw>d*
z3n9H_b-Cr5Lx*s`$cNuczBPtffxTP%<aQXvkhb7S0q4JmkgN~Y$<bo>wxU6=&<>ea
ze=r(_vnaj6MtuA(5fgO}y9`Q;z_o~tB!XN1EZ0D?i2{+|?^wZ!s(Sa;kpr3^=L<(Z
znlpJ%QYU9h1R;xlI&Ws^zg157{5jm{otlmwHUR!0J4?qZS5SR**8H&TUD7j>Y+g>*
zk$`HwSyST-w!HhpOk-TH%Bx`J=^0Zza(gNMp;^88ydYb+cxG&UE2z3a@4bX#kAUKv
ztKz@9NGKyyRv#r)-_WdU^L_<pLjI@aAOLQw<CVG2rI6oKHm0tAj{a#LaB-b+5KNMn
zVCXLX7^$?r<l59X*o`SyeZoRgk!va}8l&m|*t+MS93!CRz?n}slUsVn8s4D^oP?%d
z+5N$1cYK6@fpvqz*=~8tEYX3~oSpAq49YLP^2v>(<hm|0Y|P43zS79eZ#<ChdQgTE
zm^OVblGS4USY=3Rk$A}|l()z1B|#c>^@tjk_jQ8DeP2J|<u$t!B=np`Y2{4$q^avZ
zjIhbCJO{C2@jPe|ltu_lf#h#fIO>n}?W<6}ntj+yBY!AD^=<T}o1l9rRjyiT>;Qef
z41+O$6mKoWW~_loO|?Abc-AXxGXJ)5>j`-uT-iel$2W}b9FJ=FBAbGl=Mfy0GHbOi
zX?ngy#r}gumDl}KwUv}WkQaY&R#LgUFgEW-p@gn+&gD*5s()82{?$s6?|uTW52Pxd
z*O{qEtuz$CtbTgL5~U5^JuILAa~1-rkK*6GutSrjY|f%rG9Fr5Dy$#2)Zj^DmF>i7
zC7I-0Mm?mEPh3WjG@Wh|c$6-aYSe|5RTb`ueOx-0m&NnqZ_-D#r>x)wVzZSw8<`<U
zq4yR<4_{#3;vg-_+27k3{@~(z#&CY)<>WLRkp3gIhIelF-{LBwpN(<`6|#;)$M(~`
zyw@aFzy@1aPIf4Z+=MvdcN#ktHHleO@)>}Z&N4t-Lur|`EOk0rD+ValnHeBe{)83s
zvw!)l*~wftGD<elId?qZsz;<Jq|Gdk?Lsv_zpP)`i$TJclN}i4c&vY@PDtXjB1OmA
zFKi*SL2*EnYa`A0lfD|fncn;O(Lk}mM%v-Prt8Uori2~HT+;;QOzMOW@s)%-el^qC
zSSu<m2!{(*zja;Y*{qy)FdKP^vStz{qklz;Y%f9}+u{Jwc5p><4V%(;^a{)w(>?_8
zBCBrc&vzX!gr3i~UFY(y{{hFxqElhFmI&>3o5Uc^#+J+@4xJ*nm;O|G^W2QVJuMVi
z6+OXXi#G@+x>kVex?~SXeZfvuH=FzVnbgfz0KB?N`Ex-kD?6=&oWPJucVRO!FES1k
z)t2su9tYwHy|E2%9B8eVmRXAwRU{Bo<fQS?G&*o%Tr5uPbPRu(l)DqcCBzbbMI4K>
z0tghI&VV>Eh5+0jpsoAEmT$0Q`1}YN@F4OVwWa%Q=S+PS5!O*w_w;4Q3s<EO2!HAB
zbmJm|nsigerkuSst&zYvS*9W?o;-MzvHGJSUFzVGLgDV+7v8V5%`7sI>3enBJp<Yu
zImTLEiw(7Qn1JDMnWAAv6f!S|ZUU1r!hzP^WN^LOJPHaBg1tm8H`7MkglnITgoL%5
z!7hR`qqV@qcz7lOL%&N~+`AmIYg&>^Qdi7+@e0NtL9rC=TB<ux6y+wNISRLDW<8fa
zi45y*mZlqYHi)_X`dh>Fl<Hw9MUL9>=L%5aEw?yMN$9y`tP0VKaLId@-^L$w>ju`t
zhJm<DUxW|vSSRxMN^xX!nDtm*y&vzFrrJ&lq(Y!^WJ2MZZ#az2U38vM*HFv|3GxV7
znu7dD_m#xhc5GvEj{YBuza-0+@B^7&T3lX9D2dGGeoJIxIY0C{w1({1uiD8Ik%s3w
zmX`lzKdF=g-7nE62B&kc8EuKo(jr#_=Nl&4EASvLop@dB3bp=I=jl{s=5q26rY*Rk
zW>jubV`4$Y{xKDK+eze3Ko0q9ckRkzIzy3;wW8%Ab{?KYOZxu1*1X~cjzcjxc-Xgc
zdku>qu*rNwtMc3{Z;U#DSO@Lcy^!qK{((>v49J&3ADSz2SW+wPNwjpQidicNU(RtW
zK7TOdr^K_U7FPU?8<JL<cpP|b>AY?!HjInw0jofCl}$?jN6Jr;!<Q-JYW}KRSy53B
zA~k)sPDc<rF0Lmi2Jd|kT8#2wbrRh3;nf)vj}VWyQ;NBIf*1@XBTJ69UuMDsRif);
z8nC!bA_p|?>aef$ZSJw&GN%K{js4ZpW~N={xO0#6cM~bLG-ixdfTHt_dU>vhe7R)7
ztt0&ue@3h9MZb=LGZCp-B%f-O-@su@j*;{z!J))f&(Ne*ww*=J{I@|;=`}90Jf)s}
z)S#(m%`=Wy(1_JfWPBQV(_8H<nTgCACAlMn7McfIIY!Y4ZK3U@jIJEpZYUaBSy|1Q
ztfuX;rFrONfRTuWSe;Y4eKI&)o|f;`&Q?2HlfTsGU0v#xzsxdhHv7!7#dt36@CjOA
z(=HRnl)AE2T~Wx~C7pZT>d$L-xw%9G>8KwQrT)yLimK@^xdSfQIsce0s-SEvJt~6z
zXVA1Mmgj`hUS?;v9<Z#@VX)@C%w_L<ShDbQN)K93skxN44U{N?A_!+mPgD@LBlCK3
z5S9n)u}>!|^Q1{pT$IxIB#b&^^*TXql7Gh%lvb6$%VheV8lV`0Mx&oRc`{PrvWXrX
zZMW1NpllMw$?-_q?LXmVn+fM2eZo37Ip0v<i0vzE+OIRyDUw~%T4wg)G^K~wk%ON;
z)^eg@w=R!$mo#rC1|7#$MD{GFItskqm6b+>V(VCy7?mR|r5AV3-grqQ?+z{9r;Dp6
z{$w<gMBoD8eya=-1mRGp`uQvAxYeME1}l8nq>L*K?s2>Q+L6&>+Q|upa7G5XV&}4H
zRWb>I!;%mt=)B3DyK1e>1WG200GNT=8E~6l7r(X)kSY>@g$AH@J6rfOFl<L|Eb2u?
z%gR#Xx2Y(>D8Op-dJ!7Y`P4^~FPmCIY-FjJ$&6YmI{Wh5@cFOu7Y}}ALy|wzs9Z4Y
zo>#>@nJwn(3>klK{a!qNK*^y{a<%}$)WEdyOM}GFIPvn}K34fJWKqs-5C|xiTw}Y&
zT6<Hua!cNw=a|_lYSPhNZWYz_R_ZOA4F9J)-WukVK@SS!3TY9$)1mlkg^bsAj416-
zehv}!2<m%7w?M6eR2V5e%V=T~`KE#Yp!>uaoif~x>fdkE(dzPI2*xX76?l{|Nu*i9
z{Dk~o&XIy8dSI!~qUiB*$-#Q);1fj?tF{w!J9;&n=e!kZny$gonNHr>lhvNLJ`NE_
zCndeBIqRv4PQ*W*M&#u8Z}q{P7rAoQy5uW>%JoH7v}|pxY(yW#-DPMiM7dyX89%B1
zhesmX&e;_!fJMGGj{j<zxnTP9`f3QK$nk4nk(|KAK$_51%QdG4p7K39>Ed9&Y9s49
zY>BQMdY_DY;?6e7o$x*swsr8t6@l!pFBNW=H@_6?;40m#M6P6tY-ygNjh*y1^Jd!G
z<)`0k$X}O4-raHbUQcl#!Rsy^ic3p*k^|9rN9MYpVv*H#FM*#^89XOtN2?d|Da(~f
z=Z|r3Vu8jpfz>lu)v3_^hS?Fc%ADz|-dwPqq8ZpGLrGsF^FG0zkXW<MdGA;*9w3Aw
zb1~P<-5W&0#=BP#M<LZRz&$usEw8tsDG7`YyAXv>#JGFid0<gE#hL;`&N|$#L*h|z
zmpj^`1(R}Nwn5B^7T13tvG~~It~k@MJBhr`Gk&lVoAK!zgUYT$bey#vI~sOTNr;|2
zv${J+D$UQV6Jp~hHhadmicYTyEm;i1F~YDei%QjmmDxduNW#RjcX_5@&flIs3IdTg
zTs4{&O49?@>Ht?<6e}L8o134eIb+a6Y^f3-@kmUO7qqntYfZY^-t$0}AzbIHLILAI
zHHP7F04XWd^^z6v*A^Ze(6-};#fEnhvucouDUOIR&iF=NRF6K89_1!2!L?pv(y3rS
zS2L4v6s`T-6OoU+A-kC~sg|T|eMF}=%`f8~>33=)=@U5&WHkmSSR!vNW(rzgS{*_w
zNbwaUvU&&mU(TSHo?-p4Erh8+eL*{0(D>4!>Z<ii<x(>Z;%Ie_&(UiAa$s>NO$1na
zCe)i=3%KKQj)}pfh@-s(zMs3719+`#q`CkyXq#Ax`w7-&*2WIF;5`7l-VBS<G`N+*
zkGH?~k@J==tRE~6Es+LOXh1xkn?ePfqm&HPDh{iVzF)b>WD+Uooa~&?Dby=wjQG9u
zMRZAgrC00wZRv@6AfYW#dZpmZ5$5@Uss<+4F>{kO$#r8xu1K3+ux=koDd;}evT~f|
zA%#I(6c~?mWYF}1l(zevOndzm>bUw{J*Z_vTu|B_c7w%-fq9df<o&pXn8Tpu1G%K^
zQ!m+!R>*kl6>??f=oGzQBd?7*pXx|1-in^{SnH0WyxMS_l_T7`VBIpZ-X`of!Jm1L
z3feAc_eC5oPb?0VLELSAvOl60MGQHF0Y(L~CnVCH&J|G4g5(otwA|`X&-{Zz@DVDS
zL(FsLl_L$?ld6DW;opq^?LKkR1EE(oPT7}qQJkMNj5@4PQV}^S_py>mOM?Qf^hn0t
z=llE5)|Il_1GxG(_aNwby_!2HVI2E*tdr|Zxrgy(_xeV3xPb@^vspt)Mw0KbSLmce
zEKP(1CIv>dS$8$-n1kP8-@`WDXLADmm)SML0u8axVB8~DB8~~q1O(o2)~&!vDTVoU
zEnBaYAwE_cI}gJSEgTHz=06|h(OwhXJkA$e>_(J!G(eI4<a=fP=f)XquNu1cK5eI!
zm1x$lb~`Bu1Z?I0)}Cj5BOvT7HPgcncAzeE^Sju2U)v;-coUSPCySPL*fYnK?1R6t
ztmmCr|NV7Du-)C4F0`uH{<Iqdn40^tr7|vN=cI#>s|cOeL#|<R2iivAYu6HbFY;(@
z$%R+2p#1!unc)aq6^)a|N=UwI@^-;>!1f!Ju;`X@Z)9{v1bihQnPsW**m2Ro5*_@u
z2`9K`ulk7V&%%BGCEu-xd~F6xU0w<5fYjp?+L=f@vG;ET8#*iUHUk)3p{i?a?d=O@
zFpOz)*a1b+^mpFa__Bm6CS{hv!8^sJvK@1BWiNN*!IT9!Yn^OTZa)p*#GKjNdVRi?
z?qy{1!@u_mOe;)e|3fxIw}C`w*}`z_c7IhERrk4@<xG{=h>QdvZRFKkmzRgFZ%=l{
z2k)erh2-<8xl~anx$aFj+H*`288|A(6xFXTtWP%@l9CCRS7?HV3Gg(wzlyBTr6PTU
zm08IR<W-W4tZTths&oi>$fq1deS;$szW&YcFxQHhLCLP<MIx7vsuZ#k*Q*W*@IQQ>
zIBxB`3W?UOi*i|+98ueI&1viR+BTA()kOy3#z`}1Tr~-RV>t?xp7<J^lOU?N4>sMO
zaQN~tZNQfKkdqIloeD8MQMhXmMnSEM8d$~J5*!!P&Ke%}v4hUi(pn{N_iU*MBC}?u
z&3}s8uj0j>E?@QG+cmbPdF!}yvD#jLcvd8X2t~UrEj5FYpu7O=BrYQvM*qZf)yVB{
z4+_j?P7(H>mX-|Wc0ACdOU<NtY6Pis&c=|SO3y;RI-ZLL|HC<-ukVJR9#I%1%_t*-
z76zqL2Ygrm%t}+Rq0OBIiC2~mg3Wi_-p)`rOsu;em3J{fT(Om1()U!V#GXxa$D+8|
z_wERt0i7us2x{<~Su7y24a|X*zuj@LA+6DDna3;dS4Y%zx@yhrzD_4IiOhU?6}hE8
zYDL<3RKV+&Z_hYQXg)cQ%@L}LvfCBOGAtj>$r|Aux<Lj?d=)lgDGqLL6M{`mqFd!p
z%_p822}InoYPlwfbggW*TPRlphAwIDj$e!HiDBwy8|?t$!D>t6^9|z=mUl#Qd&k?p
zf&}l~6U%u_Ob?|m=arJG%1WD5E3<c&a#Nh%H058-F-**Kn517z;jT3H0Lb(kK(*2I
z;3Z-%6GTW?=i`S6%N(O<SAkH4^(%i;q<1r!Rc{I&8~<91kj&S2m^XKEgSH^w*Y}Y{
zg+KQ~iK}P+FocM)h})~^aeV6eTZ#MNSG6|V);&;VB>6}gnP*n?n8=Sfv3_qMq0g~~
zu)BT%=^KkuSMwkFF(~?5?i+9A85vwHYDSTIJI%)4K5s&Wxz)>Oaem<fGc~p;$DZlS
zqwKcu4=w;d1{8g`4jPMy@?K~7V+8{o0D*061opsRYNCEYrbo0DiLw!wyOWqT#3f|L
z@>bc-p{>w6VHQ;n_mQztwP$=$gFs|uD!SfvZi{%S-EsD+y<Lq7TIVbTTWoJ$o3+qp
znLa+pP9oJFLEZb&cXb3opVmz+Yo2%p@CR<@wwroCjJ-d<-d_#W%*ljG4v2fuEysLX
zW`8&rKiuZ4SO{@Ii&bhhZu4>4rRK1S?5GDl_`{0dYH{BVZ=iSJYvl(4XLDnh1jQ48
zZM1Y~j^>gmQ8txvSXdUCAO+IU{SEPkNXYo`3R3(^F45Ue&K$OL5hJHt{ZwcjGGN-q
zX{Y!`46zL{W*JC!<~wzxPHfS&XEDBXkty4=tlAj`ID!KdQvw(ub5(1kWn7^})Cok6
zlHqxA9ThD)j~rz*;7GpQI-LptqPf?pdR?+#KH$st<Ihf0b%fg`k1}YR=`AS-UMD^k
zSyO&46cjpE07LoN$B3<Pef;C0+zEQAj@6Lzijcf>YC%sZ*a>hr=&323@?D>rmM)4u
z02nQcJGQ!&rDqvCR~XMl+YPLKDbUS~R2#GSwRCBF(p(uK8Bvf>KM>#=r>?d%|7|BN
zh|yoLRuy%d=})bssgW~LWj^<`TUib}#%?6{6dlwOK#Pf{sLW0O(mts<PttZY%(~Ag
zdYJi<fsoeqJHfXNg|D0r=6o_h9OEU2%8aqlhCsEZ#X-x%g<t`|3deS5+#;?ngKpst
za(!5ES&xCb=MQj|hH!v3D@*nk&euDbri=|VHb{$KnUlI_AM6QR=$m#I3QHOL9phnr
z@Gv=Zzq?x2!IpI$LB;ug6H4d$mLSN+uCOEM0O;Q&F@BVV+6#h?OauqhU7fD!S$lyL
zbL?=X%|Ce4jpkZgV{@zkezTDM2CctRc+FfK^UiD#%wRvtw7yj8NQKuh>2P~1FnUq{
zmbBtt{j<zi5eJBi$+lV%l;JaQl$9;|d82NTsJ=JTV-(SCx?gP<1=fPm2Q0b;b#9Da
zQ)<Gf-W$?NOuB7((`E>{P}xsMk(WqC$CBP>QoucC>cPwOg7ITxbyETDc7Z~ewcfa3
zHK#Xm$o-YWl}eItfJ)^Mzo{Q4R$G`R;$zC#WhlNkGiygPLP}=f;5X}e%eX5?hl-4J
zRslU-jEw1(D|Irw2!0q0<=0$cA0Dh;bm1gpUd~Gls0f_(B<;%$Kh3}QKjh{?uGH;A
z4#wR%JJ<Hna*&R^=CMTTtrcBmZWBE5nbRCV?C>RM5x*7bQkyw|BWjioePmz)mt8Q|
zoteN6Qpefj^C`iD4LQc$hy3?&XRG&MjNVIh^@o8B5U8cbRe$}vHg!8wovr%Qf``FL
z=*<<#8UsV;Rd4B@g~r(Dwwt#4TV?p*oa5LdAmiiUqNJvI3#<6U-T4m-v1D@)GUy2G
zlwqmhX9}{Qx(4>*0o9ME6Ug-pK!C?PCJaFV!aHtSmPA|TS4xtQt|m64p;^rY_pwk(
zA5>ta4(twJL56oA&*+JFN3Q>EDbAMb^fZ~peoo)?ImmtJA{^YoomZPhpYk&2Ji%8a
z%Iw#rNjYir0G0i(=K_4k4Yt(Q_`^aU!k&h9)^DOhIZrzmp^Frf4<?;FKf3vSJ+C{x
z@!_w1AtkI2e}T?#6hM-kE8W%cTSe9K$pE~d(9E@FOImHv!Rzz6>C~u_1xwY+aN%gD
z(2(q;JlocetUyDP^#z)4>OMl<{xDI6nHi9PEevD+VpXrdvV~RDpDzX?tS$o~#--c2
zd?3KsHO^C1VS>OvS&am|8@b&rp+$l6h?6P2xT;(=*D*lmU+79~-dOdLn@Y5>?I`FI
zWFLwYy6Ep6B^s|CDjLFSxCCX)TSmX9QktS;s7L-J`ujZ4WICJQ3|&kSdR+k>QxE!*
zcZ%YFi`~;*zoncmT)7F&q4_vF_qg`lFv5pq_+=HSz<i%v2EJ&N;n|w>FjNGMq%God
z-sWIp!{->nwvH10(JnDLoks!-f6?kqvHJ{U+OG|JApBpXw-X;VG=8x?3IK}|Gm}U?
zBtxE#yL&*L?B>Qrcb<bFP_V+aVeJNQnpfyOogBnSTX&=8s(856L+g^|NcsR9N;d^~
z6Y4Jt0~*F=0?h*bzs$z5xduZyZE>Gp-3I6T!3fTm_Alqt1?Ouj`(yz7@dK<Dzq%U1
zC$SX6e+Gd0B}h#B3=FYC;(thrKR-D=Evmg}Oi{sh7rftZT}GbsXX(c%_MsAk8+Wcz
zs;P?H>Yq4D{n$GLY~Va^yOFFeQCrt?fxlnflsCE6tx&feD7g_ea05Vlh1x&*k@YuC
z|Gq=bvwl)TK83d*8ksgewNIM4Tz@E;8C>hszg(-wwJf<ppZT48I7~lic!~Z#^#A-e
zo=vsn`nNsT|Ec~D%F6=(LBICD`>?S7I|ly`W57JM1D+?VB`5tE>={d8-x((@Ah-B0
zFXd~1^e_Maz4?~Qpk@YaefY}-)bwX8|6QTguI)dX@XOz)DlQgqKbf%PN~!ku_rJQj
zvaI~~536Gl+1G$l$iKh(fkxK9?*Tr?<iGp=^B92P{l{R<w;eYPo!ZyjeOtumbarLQ
zbz|KGc0L0K_pE>DH_ZOM>mt}<-QkI6;>jB5i@8G!<l~#DJ+c4iARknxG5g?{;G=z%
z+Ld~1#+6z(jti-MeXpj*dhKfn-|nHh_6pI6G4gE(M$cI!BFX&>Gkz%h`)xuz^#_bK
zSE*A4ST19y23R+WKQM1P{NvdFxK6C(xc|^dVd%g6`iDwqEWs$@HXl{8)C!Sf3a9Xl
zo6*o!5C;Bp@UA!3d9G8l|MzR1_J50jAO8no@Z)mxO494wZi|Vl0#7GfA_9(@YqqWZ
z43GWyr?ERo?Dqxz`<d**|0!4h&p)`t{&(Mh#{diK|K}K7!A^FCI17*JPcKiczD8(r
zeQA;ovTUy&NUobCVt&r9{*lGUhS+<->GXLd(u<aAFr8c~z18Iga}e%~kiH%qelCIb
zh2I>d5(}`A)u8^K>7hQRmH*u7&Mn-#wN1PV-?0kzZl@_5=8E1ckZI51j$fD%N{nG!
zJB>g8tM~r7$LCkvMu|uKedn*+7vtB?{}$c%v}e}dd2ns_aN|r^l3zFIK*VG+3)d3W
zlqgaukpJfs@h=-646acimcnaM&f(awkYBPG!6VH8&D71~+{Ua{&xCV&qL&qw|MG!<
zZYInA_KM1$Y+8TaeLUeA`Ec3qOrcfz%E#jTa)z&W<%&6zhyq3sm`+5&WihF(Y{U)w
z=ZF7gL!JEZ<A~o4#oJ%#E>pO8<@CU(?~(H>+o33*g=Lpt!%gQkj)tgH9@oPUtC<%5
z1D%VPBFxk4_AY1TUFTjTidPbk|K))H*ff3i+vlvrZp^Jc37w~$C_af0l9|)+N{z|}
zYmFG4L|6A_75$&ys9DzMOQvjnp5xQr`w9GVagh>L*6P*G={;jFpY5d_PCgU@J*ofK
zOX1>L10IV$u+s(Sy-kl1y|PtQy=H4n@@v>>uzcy)0Xz)#8`$YP(}fZksx2IX*@QfX
zWp%P!Oqy?fW9HBU_vqeae^XFyF%^7Y9imJ0$o-{(a!rbJ-j6U2?Qp*CI>OM_b6-r@
zSq(nsI44ssE)b7h!(fH3IQe$uQRp974gRoK|LdwuAO8N@gbG$yRm`Oy3m9Ng^#OhR
z3?|;H1mE5q<Tm2e8c&u_2%QU$*-NpSEQqUgSDCckNI<ws$g}3C<2_PbK^}({Zl!LC
z+ozhW-n`X!N98TyXj71jOGj;O_S^6g#l}B~hJ&TuUx4u8><%^-wg&(Ibrsnhzwc>8
zMz%U$npK@V`+M8NplI=0$(XWZixJd*4ud2fgL|`Ixz!-M+?YM{^mw^RnIABeEFy}$
zxdeCJcv!XvS1Q*uqsm9J*i20|LX@xPQw^%Ny|%>rPmN1EJLNz8C71K-`o5E1HDY-(
zfu*MZ_8O-BjmlI<CVv>~Z(+P*;=o<S<syn_+lbGA&QdWD=!qPuVqCt}hCHsW<HjhA
z*(2x7pMNFQ8EYlT_$d0p$zB$h%TP_0H77di=>s{o$-(Gol@Go|mh2wZJGpAhzJEMm
z*N|h#N$CcIk@@9Gip-VdO{~%Dt7&*kgbyxeG?y_gVMDg9;uXQ9Q!&P?lZ*3Y(E#l)
zW46%d<^ES(&7092KG*!*jU;{rOX3^Kjpw#Hmc*(1p=rd7(}PZ9Kh0z`FZV-JH#aC8
zb^SqtjY#OdvH;d*yHU76hXh0;+NVkDL~v+AX(7rI@)H@p1<Blw{$VwAvg5wDoo}Uw
zwb_s*<bs5TJ%224n1DG`LsU9U-c%r@-a}kAtjMQA>kkIC;G?6zE(Gwk>(K`fQL8zG
z<9}e~Q42c!LQci*);q1lu1gYIOWi$1agqx;{yM_lte4Sgbp+oo`TgIvF9VF*kN-!N
z36o;VUBM%t(FRd!EjpEAvG4Hivf~7e7YyaXJMW9?^ex$)e2;EWGH5h&xr>Oh%MC?G
zC(y({BNOLRcJY%Q>K~G+uJWj?C^fLtfk^YX=A9Td)@f3+DUaw+{z+q&-i(v&Ee}F`
z-D`vM`F*?dHS`jcD3*fRV(65bcT)C`8|$SxP8L@d=qSA6*<$Qy6>Mk8dsW${1!+?D
zv0jtCML*pmLDoLZkmz&x+7wP0P`*wfv^5iE7o3Ra1M#P1DN6kId%LN{3d^i(Cy_MC
zW93Ag7QMb1<XGg3KrgT{6Pj!^6*v}oY+R|*CR9heoxUwK!PvS-247++bWv!d;+8ed
zaT_*kP=`6JwO0}6=Q&3LDMb(`Js2Jw`RYvB84I~JopN^%YDO#0E4+}in?`6XQ8Y}`
zTGRaed5o+}x}LD`f*v?OmglB%w!?MrugwuzOnuCiJFwqY@u%-`LsVqFY`>;Fnw#ex
zlVnzV5~Nx@$M`33Ah=wd7L)m${fOc!&(P8ok8-QDTnU@kWEF&Kd$Yis9yNCN0Cy62
zgg-)sGb{CIfSvh5bExHPCa^FMKKy!QNTy3B2fP8_&~mbRjUSK6f+w5Vp7|AUW{(Iy
z?F1HXRiD@9!t<lPpmX34u(3Qx+J&Ba+82&9B}VRE8@$5nv>lI+M9-gxNma@Rsiqc`
z^uFvn#*?v)Zs(t?dz{=nVfvM?h+v2hOU1Tuh|#_IyOd*aB4bU;L7`&N@O@h)Bfj=_
zbeB^W^v7mLI_@ZTvWIN_d}#g0U+F>uR(hqEQ7mqYG4&tAs$G_5#yX+*-zJ=GoI=sf
z*JGaRifCti)cpNCb#3?pJ~X-&xqi=cl5=!{*pLbgtzEnKDOcNP6<R!Miexa>VNxMv
z0cz<|V4$D-<jvd|{xUJ->if7cKs;@LoKN%AuDs=WDB(`{owU4x*#)*81BZ{64ycoY
zm>Aa%eL;Qh%FjC`gfsO4mcF2)Ti)mmCYcUb3l#@PJUH$Lm7ck}G#^qiQ3LU(%DP^*
zDHb|Y!f9>=0^4IeVx2g>Rz#e_;$Mb0RJwo7D;UIIbWY=OyzCm+5qGwIy=2AID?}hs
z;ln5QL}zSBhQXy-J%6h@Jz5zCX26{NM=Ea<?oYh=OB07{Q|6@+EGqiFxJ9m^&X`IG
z9pXwwL@(TV>{fhTMQXsN2(z72WRB8s?ZW2%NUsqYmZwUC&{M3ZU-S>G$gvdePGc6Y
z)O6;rh>K~z*}uExwhqZeAC7~k$cAL=K2bbX>Kkr-RPs^Kz)6F>B&V0&hr_q~99a4F
z2ZaO4#Hqap2OZQB(v~@SAG#&#bL`fB%DS)<zHo0=me+8<tw3sdI_*TTZHrlNpsOt6
zAA0YiR5@E0`#|$#HmC%*4~dMJed>wYAKJs9PhmfAPu(ouEUb59Sckk3fa=!mw_TAK
zZ*8uLx#9njOp(CH{^;0d!1`cX`E<Z$a*)wV>e&|kF79Ti$bnuoL--{CFb_dLLi%@r
z0}dAIwTSBA=fuV(BnHU82xEc9a26L-nq{<86-HD|3_<Y;XX~GMAzl8skxcaT{1L-+
zBT?nYu8@aAWWq_@cj4(KLxyWwjGc7k4NhI7G|0hc&5w$Y8Yx}!Pf~t31Xp338dcwR
z{&8A6kdf+`pZUz1biyiVdM}r*O*Oj(ucx{~?Y?h)ZGN5~o}6ta6ijJmv9#jHZHJ`=
zgDb;}00aa~7~40VFTB?-(*w_pCr1cyW*mVR5c|&;5G>P?z~~(>ACZOq<r|TuuScDE
z5?&(`q#Sia2b-B#@Tcz3`U;ccpTLqm5N}3fDjaCpre9)L`2MPC@$qgVh1cVk){(d#
zlwQ5iE1(rXr7mmUP;jko*I^MoXl5w+RBJZvLAA-tdeBS*x4VizVjvg|)>VclQ0|pN
ze0uG>1_CXS+-n*?E~W)63$o({v9s}r1<#Kex5CH=polW14D)G^opAZR7;3h`q!YJM
zn~z2&a&q<i85UAQ!Rtw+7CL`VL$YH(hDYQrPa}Xtdh#fGHERw7LtvUAMO778%8gr4
zkB=8hw6!~1d*n(T#A=AsRpL=0{dy;7Q4g#9O$+89_rBHxwt1!W_9IV{5&MA^*Ay8Y
z27{&Pl)#fcp9ifO-(KO`UWwCBPZ6`$-;4hFU4c8FJjlq^9)Om_3fNx)qcj=iU~Rbq
z?@exn+~t-A>zwA-ezST<=3)N+DJ`<(!Gm#ef`YFUov_3G{Xc^p_|&&NM;1xmEx2t&
zZhz>Ve<<PM%~RH_ZrNL*ZgOkrMs0P<7f$mM>aU(W#J$7!wLE-Xs<<)z>Zu-M*SN?8
zK#8<408#}5#JU$~0Z^KDPD_eXxL<i0X#QGh+h=q<0#dm3y!7LcVcHS!uK|9Usk~o?
z>|KG<{`$$TiKc^&4(xCX!okj2wF?_2FnSkMGR?u4nMr`0KOmX=J2iWA;YzJw@ru}U
zk^|N9<}r79Ni&)90GG4Ea;I^P3634D0K*_eQ)_25v?FETlcsCw?OZx{Yz>o5zw&Bl
zcEJNPU$OK(FR$lsMtKY<DifSP<=uIUSK_1L-GpB44x~Pye!kXcugN01CMxYbZW3LW
zp~jT*)HkF>O#jk6t18Djl={|YsDjg>UsZo-poFM?tr<InCH&;qA(jFqZBg72hkn=9
z8n5ditJbV@rkG%y8`b%*vtd|dd_+P*2D|-M==tL(sL4v9X}t$y#v2nugVdFfnu84i
zJq;8caeC`8C6-wS^f6|`cL-B{at&1WFw-cyP>$|T7+0v@x7xX*hGXIZuTwGITpKOw
zRa4H;`WyK?dMD8T!`@qmMfLXYqJxSkC<sWGQW63x-AGAFcY{cG=b%WJv@}RJ2s1Q8
z2uMqJcS|>O9{j%V@9e$by|3$B=dXR8v)4bvthI(&^E{vRiTl1k&$0m5`HsaY<Yx!Z
z%Q=j&prhFFMDDv(S+P-p*c25|#t(A)T*Q0$kq!a0FZJYeS%Mwoe}jBcDF4B{k48qW
zdv9;=EC)XiBlrG^qqI(#UPVN2i1?Id9gn7A<TOW5Z?Bj2#Lu5Un;L3~O}IUH5PlQA
z9z2GYX}3oNj-B};m5rlt@8(wNQG}{w=57gD{<%1AHg~xr#O>KaZ@Evj`UunE+2=^b
zjGT$W0%gf1CjN#GF~D%li$SzSVSU*(M>#??<_N)KDlb0P=xmeumd~W3WrET7N6l6W
z5xEtq=~`bOgQG(rJEVN_B{1+MOZLyF*594;Q+3^6IVcsc%j`b6EfmB)eE9Z*yQf1!
zw1@vS`g(1NSt3Jzc%}r0aqElK)>s~t?G2Ju{&X`23Q;*0H{*1Dr|SDhff8Hm?44rc
zu5F5h6R%mXFO|;|RbJ73c@g~=8#5<4MN}jG(1G3QyG*=k9>Ts2r)U?sOc{g;eT;?=
z`yQ;eb&p04liiIjS;0KAspm_MK<!jN5YQdAimi517tONy(VEyCNzcuBuL+%`N;|M<
zgM>RE;=Se@t~WYsZBZ%Mzqq&Xmhn!B!|QR#1ue|Guutq}RU??ZsPq)-k=iB#>~xuF
z1oXB2kkgfP;4+^<Bt_pUoBPr{&FoBUjI5|Q6d~gMDzl<1xOpzay#2B^9|LLb4Qoqm
zG%$5rtATni_zx5gV+ls5E>7K%Un2LuKsF#b1u>s!Xqi}{<Rx-4K|M8aU)zT_I5tOx
zM7sTTLj_j+Y-%vXSqUW1ilBH7W1{t?sAyK9O;=QNB=^&JX&oANlXLf)EJw8|g5eS}
zd~9y_{IT1gNyQ29u=IQPY|{xNC5pUWjW(_w&1QgEqkfU|acb-1S)!X|s_4}FRKn1S
z4m@$nNQq1Rq6Q=9i=B+13z95OPBOU<P$vpm;vi8<UlT*Jid?iG9j~y{b^}NHA<__a
zGIPaiP|b*^e(vIWm6MK?UffHzAW9%!^((k<Bz9cVC0o|UL+Gc35#gj0f24?-yW$~c
zv%WzSi*F!Pre-{>#c;Pp0`|wy(!Tjs;OB~?UIki>T4u_}VX|@d;X1kBqCx2kQbZNi
zHV~`2Xi!_h^Heg*B|&NE^jNc6VNjYRu`03k0URc935xN?7)8{hBgK<n#IXc2b7i!S
zALJkw7yjK0l(i*=&V?oVdBMeh(4ATZF^#gP6o+xZnulctntk)8WDOrJyIBOz$6-Ab
z0*5--0H&zp;6(8;IEa`=+}XU+m~`Tnw_An2Iz!rssUA=~f{e9JD*RC_6R-=uGtzur
zs@ZYP0%5l-8hTSd!Pq0bECe+*A7=}y|5Dy!BpnCm8{`xfUB1gS%zQq`5==-|p^9^o
z4{Y%b+q5-<O$nM%FX7$~@Zh9D&Y0GBBAF(#1mEqj$Hp<za*EWR?7>*fGu{uW<Z1mJ
zWSmirkmC&Z7dd|!p?L8qopyTn{0)-w2qqAJ^w`*BE_}J})V{Q_7OLf3+ir4p9=YLB
z`Fo6~0mIok-cw7CJP6<`MKs8RXLiSFInXnUq3?6Z?!8}v>T$ycT9AT`&F(ls<7u4N
zF-9zYaw(iQl2LB1u(R<HDfn@1O-`|x^F5(rUYB3?HGj}6o@{Jl$cBDvB(Pu3zHjt?
z=K;5P6`H}N=6aOgL}t7&6<N-K4rMCI0}%S`fC1hsZ;rMKn*IaRs_-qZB~0S<aHr$s
zkehsu$%{sjC}y~0;$O9fw(*#3P8TZzC8>@-n#^Frd9bQ3a(v!)kDr~qqES5Zj*t*d
ze=DadTY!=TvyrgMc)dlHW~!#E<+AGdAY&-HisZ=dyBg$q-naWpI;?ne>D+e-x;{G2
zUJZ_a1&BpT3IX;&<x2K2fNV<5Rkj1WX39#p#Cr#kbodG${q@0-_J-TgwN1E0%y7#R
z8Y^cT)Adg-%8B*ICg}d`CoL2599|+rKVynAwl2WG$W$1SAIj;K^S)-3CKZ3(YnGUk
zG&?vXeI8Zdly>-b%y;`_#K4o(6TL`&2zBYz2UJ5Y%QX5!`^*=pFR%z=brkW6GskF=
zbldEliuOaal5-IAgfTwd!@8c0wu1e0Xz>1SJJXzc@vgqtg1b7#u%ORfUGz~ce5_nL
z`McFfmBMNyYI>Z;I}C~=MjoOl=OM`2_oP}wYsSifDkkpVf;%_u-_FG(7@ywx?kD;N
zwF{E>X3OsvXDTImRG&Sw=sVp6K4^(_|A#k}aG5YebzBiPr29(^a-x~k6_j*=t3%wO
zkR6Y!;KoCCZv8B$ETl%(IwXA1a{33)4!HFZl2;T}Imph+dBo*tKH}QwydR3hI-bGz
z+0=Z6q{sDcEnW9`zR`)?hslq?OT1G{8IO+27ccLm=V%}ZHcuGc>8Z(9JXxqI?fZsx
zCAJReH#B_zx!e|lVc@oci3>m}l$%@i&H*u7C#cq&de7L4Ee^TfS|nj<Tg_RO(ixOL
z5xc3i;^{5Od-JOHdPCTlYDdRvLT_uC@Q%<c5I;bflJKI5(LBg~3~`?dPdJ;8P;e3d
zRfZyUXnb#-Gw0r91Bh}<9qMWgNssL$ZV@Px3x9KE49)^Be<|5+HD1x<-jld;Cx!-L
ztMmaA*&$-$q|aHOKq_5GijJ3TVEN<i4m7s$^$<c1uVP1T1jPusRcrlIk?>NsD4zyo
zV{Y^-#3Rq!G1)qFO%n-|mQorGfb|?4*qbZp>5j(7k7yQ%jVu6LBUq#O$SM#@L85ix
zpJVp;v8IT~3BguEx*thJ%J7f++MEQ=+ya1|q9zmP_I-+7_r(q{xQxVmaP#L}hR91j
z8A`}UlHJY<D8aly&0AmsB7`!HERdZ>-jQ!D`KI?pd-w7i*`oyA51y&*yM0lJ3@`cZ
zO%1PTvW2h=c{n2)Iv$HUL}(<16G3;hJU|tPp+H|jjY$zk;b!~;8QbYS_Yfy;g5oBU
zs4?<6rTzHT_n#BbAqCs4hmQtx5Pw+eO+q#lkJn*~T9+6~JyxsTl6#+{Pi0x)yDuAQ
zAjb=*$b(-lv)5CCv$dm>pu!UGVG%}c428BHbc~%AOvHg4vDnb#PI`M=FAKUC%dl0A
zHYr0ke>bOqUB96~xbEiTfH0Ok3~z3gt|~PUdGQucMcLD394dyC_HXN8XXRv3P%ani
zknNxbAoJOAYw$7$;3c@9?mu=<NJ;$V^u3tzEV`Z&7znOkniB^nuh->hsGD?{l~s_K
z)W*s7#*)ZG6^%-l>{R}fqmw)dMbW5JL4w!puQ|Q6`AK_Br*?cC)j4`KD4QaguOIE_
zmkbZ(jQ0_G<nJsWH;+$mZ&WQ!<YW%4hmYd=?nHtLeq_RkJFtBa9zBL;l)YVBQ!_EO
ztF_7J*u-Pp)9S*k=CInJ%0@d<_~-$Zn4#TGM&NpPH|4?d01I@#vxDJSe`yc-_RqtU
z;9K}p(^v07BD%|Or-t#alwth*8^Lci^pxhAi7q>Em6F_eM;#keOZ$%|bRaCsG2OQu
zi<ACd-45kaV>`L}3s6^Enk%EKd>b2V3A<42V($`Fm`Y_y!+^`b8%hl^U^cTXrQ&T>
zwRfY>MmK%@_~PJ)MM`4MMpVevZvc-e!vk~7c2mZjv|h*Gq0}^rUbdGkcrR9X=6R7d
z7;UEVE&QUgKtAc##irR&`2sLmPfU}Oqv9EV!!ootALB-DfrDU4V_L{+Q7JU>AVf}}
zME`X<oi>(*=+-*Sr(>}*Q?I1(CHZ@W^hR4Gs6TJv00MnF=BKR}m5q#TZCdrFZX;?x
z6P>*%NFRX((-$Zq1wHQ155j7nT>7ooGtGWCtbHk`cO0@w$Hp$ITHFs&{hg_NbNVOk
z^eNr74QE$osXYt1H?86aHt~K|t&%y{m^y1d_l40NtkX&^sTpr3J#kTZy^=N+WK-mb
zckUwX6uF;vodg>N?#$KnX30V+gE~a^`FV@_TMBG$)ogf=26=`o<-x_T%QV4X>cY4*
zPiNs94MvwC8<aki`q#R7@|6C^ym7Ak`}nK(g%blE+d7UaucL6#E#hlE8!>r^_n7m=
zK*_Bl4ndPL<TI1{(BWLA@ZO;!cdjoiy<3_#x^r3HkaE&_&_Vy(RAT1`L(^y2J7fYC
z4V5>2vt9f*(vv6ESd>Rkw${UFhv&t{umc^2a_3TQE%Uxr!A^hb+fg-3QFwi$P~>{e
zKBZ{7t&?@#b9~YgOKdcGaDXQISk93tl2l5@^>ogw1C?83npb1gOzC1Hk=?^x{})#q
zb%KnPMVWU}I4pmv6V$0wovgmrxdq5L9vi$nTj7!E3l}9wKF`A}cdFzT(gxtvHfNr_
zCK87N7bSaLN2@dDb%!d(zdh%T5?;^5y{M3quhTiVSkxVi9*k>ttH~>1=-1tO+c<|T
z?e^FTm)bjPz<zo)NLE#qp$9WV!0b5K>I8|Nk*F@hONZfwUof%D{S1L|GCXRt(29$<
zzbh2_X7uT!Lri53HX9ZGc<mzpNOD77Kg)7r_zlY%Pmxyl)LBQhatO~+$@x>fo{U%T
zZ8}q!%xmo2>p0Ye?+4U8Z}W`0EG_gSv7jc&CPjIHLhP4-{$Q9`pe>I+`D-5Kd&FdO
zxow*@iur?6MCh*lKc;!-j$64cc7;IXLe!Mzk4M>k*UrX8LsmWxOU;eKX3sv=t82GB
z@L_l$VDMH{G(0d_Cn0><DI$$skGJI=gN9rC;SICTq|<_42*d0T6@lvl+t(B+9<6)2
zy<Wa?=LbV~IlX^SU99bGD?_!W8*R5<j3HzakSNE>l@&$zQHE0m&$?}{kCC-fc7%l!
z_&yBo4b_X6?A&^|RtZ1&gn#@w$Uh{C7fv|xWpRI>f@Q;drK|83IO^hGkg&{A<<!NE
zRq0@FQ^TLfQBji89X}=LJhv6gQ7ABV+Gs+7UsYCSdOsynd<#k}^QPKZU!3_kaK%f$
zdCg|)>D~u3-(!yk^`#Boj23oe&E=^*0-yUvo_cVQxc4Tpn%%<7=HzG}Wo#gTwi8sT
z8ya(zAAwDu#@%eQIG6<74^K-?ahT{7JD=u;|Nfdx$i=_Kaq(Mw*{W(u+qT0k^?6Ab
zua%UbPQ`AUEX4zw$w{2~rkN04af(C|Ckuy@%Dp)9v|7Dt4Sgoo7p7-QhktbD?B5rt
z4LsBo@!h9N%=tm6{h6Eeam~vSHsE`BSoN4dZOahwy=UUGz=`U6%T?xQG1BY7E;EHT
zX12Y_>#J>T;DOdm)<V-p2P{^P9^WSaQ}_)AnOn7+p;GxEu9{xuEn|+sLV3cLvH2Ua
zK$eH+ZLf%u_;ENotp5QgGfQ@yQpm&?K+Bk<e89?nCoT>zu{DC1G3mAP;{Q|>+vUqA
z+Y|z+hSeo$K25usJoYLsV1KMlXD0#Cggw9bQqS_L18|*AwUYvEhc0`0(XZx5XeDYv
zFDlsb05;mSa6@p4_;RTyBrhGe-{vJ&thG#f@VdM3E&x#6+^YGl<mR4$&g|<C=GBJ>
z->$>HhY6HFvSGHqi&BI+V~<V8y!Cz10Doy_X^rPye-6_=$vc*~DYO|YW;7r711hCN
z%XODv>2Z&Pycm3sR5vFC%>joK%lXD9l0EY3{oPL_X}Yjnn|gNx6Z&FEo&n<21T0Sy
z^w<2%6DzKae<1RlSefrStK~B}#J!_kx+3nRyq>@6Ug_;cE9)xKOC)77pTTJ)c(Z2q
zv_9*@AXPxu;~v5a&+aU|s!!lIt}ratn5&<zuC59o`0>&tsl!m@2WBxlTpK5b&i73R
z;L$I5Qo}_lYluJQ&W0Br-&$amUKQVxsrALql?2>$suP#2d^_=PG7u#CCbI1D5jrih
z>Bss>Zio)|53d0Cg1BD|mkv=-oZ;<_EBRN_Hew&Ftxryr`8eB|Zwp#fLMbj2_pi$P
znOI_?lshlnP3$iP69|bV0wvfcM<;7@v)hIdK2f+LlGt0Ty}hz>4JmU!mNsoCG1YFK
ztK4Ry&Rdy9J-mJJ#p&%BGuaNJ#e({woTDLjMEo%MI=|a@etd;3$#k6*o$}(ULXl#6
zaIrR6D;G=SPH`ECIHRZ;dC9mF6cqE2F%dna^y}EG#pnp%7(1)(`rYVv2sZE2p2^?G
zE5qNRA$@S=`Om4Ej-_v4rffRHO12@!*1z<%3>^pI)1-{)!XnhR?VlJ%L2-l~ZGaI4
z7}j$TU^i<^3|W7CBB*)t7_Y~mo82~y8ywLP3o~61hdr+?Avrcpq)l9P`anXQbo3AD
z+E<UKBRVL3o)xN{vI==%k{C4lJKb5iF;)9lAipowC49~$sk_iAHAc=$NcDSkJ0P>X
z1>BCL$K<ld&=BgI6}ZcanRPl8H-8fLCgYwaVf~|PV+gp7@Ho5Bn(b_u2Xbp5GwS1?
zScC33row4Tb9EyL^iZjj4@ppgQSMmgSB>1!)(Achud<`z&^~$xjEwfXG-LPD{Y{@o
zqN8bsx`fs0MG7ydEe2p5KX)i;y83FTHM7(Pmc7p7g<{8fY_zmp9rGQU_y5CGp2tr9
zK-To7+9!c)vBiGa*{{YZ`#ZO|hVt^Oo9)d?WUC;Ts&_7A`1>0K0Rsm$do(%sQ$gqa
zLR(3-zmT0p_NedX(kys(>idx+W&WaCfa%T#&9?y*#r9ET68{eCXJ(wB$EiV1%j87H
zB=MZQ#-aBtc)9rqS&K=6U|8*%@Jl}Pu80j3eO@>Le~w)0EO~FZkjyMnn_8`TyLeii
z<&?>aBGf9r!Q4h&^=LDDLsIi(hea5~?=p24<u?fuw79sNzr-`Gigi3R4Vw#Xc+H4h
z34$uIQonjk;73Wu?}{7xn!I0y&u~}{Ir>Xf?kp2^e`}PW=5SrTNFAzC!4=6peM<aU
z9R&RL4f7G_0~wE(h*X=19mYhCS2z`<0#ku%{Y*^9hx_{-ez}jOt<yiCs?-ZZMt%U~
zuhU;N?<xp5%bUVmO^4QCrI`cZT`fHQm%sC~DYXZbPs2r3jI#(`Rp@GW$(<HHiX4tw
z7Wq2hSuimSw@OUES_=Z@6AMeW^5fc)5USW#1)|Z$xto)!cF-ia6fqJ7+edc5%JEk8
zvC`RK4y(iO<>Y_6bE$6`kf>33us?H-k;QX)(9xpOupQ^V@nRGW*wJ7AIJ$n#lI||7
zS#A*;6cZPY9VNU$qLJWuX`CRSE=9tkOAZgU#yo|}OR+;#RrU$K<2|bQBL)0sh-ThU
zOw#~2b98D0;httOkp)Dxqn*;E>BG%2TtZq>cE)?B5qjYt$ak?&JjL%JRY(DW6zOFY
zx0Tqbe1Nh~y3f&!zC1d%X)H$;YcGt=UA5rPM!NmE**X~1>cc}BhA#qBdKzrbxq3h&
zOGmmZK-{k~?)2K~a+N2Vd@HK4+M!LjmN{G<k<OeAL<BxR*p%z;=BW+nj=i>T&#Q+=
zr}mcySEj9p1qZtSjC4>j<K5u7U`FUNy{aYIal&lv3Ena5zH#2WNw%@d83+LGa!i`c
zD^GN?Uz<W;0^eyy1Ei0Jrl$IPvke_`a{U*lQley!OfZ;=6{H`UxExNU)+1}oYTTw*
zAD))`l{XtK3B9oHt;Mq|Ed4T+6FM}~vLosGsd{Hd!XtjGgt0O^@`Uxp6sa_dej5Pu
z{?^2m{{H?sj~VO@bD(p8BA$;0!F8tI!Uf<Xt3S~dQ1BCJDFIkl-~zMFM4ypI1$?k3
zWt42kt6ZDvAvP5gJ5SZu*&hiNo-xRL0Hx4=W)co3n}t=orshMkR^o%<G*64QB$L7!
zjK4AQS<5xcE>uJ6Z>=Cy+@=1GJRq0%{&0W0?JouA>WVM;%vgfu@YinXUo;^_xVv(K
zy;=Rv;i)@3dndC7z4ov!*Qe2dI@EU)f#@GjEMssAG_1Xu8{REsxg~e)S9k$#D*u=Y
zK9*sV%H#Pwt?@bKC0IrMkDroWmB1MH;UTA5qYY#;oF78<beEqa_BQ-x?(Z5YHNaW&
zSvNx3s+{bot*$>`D5~wpf=iB}*RFn`j}r#pF`XG2*nKU;ESPQ)G3QexA#LGEneg<G
zDqQC71s|G&*ks)#ZRpxN6}7*>y|CW_7S^E=nU`x$6K?6}pr7~qf8G~c{g`VrmzW?~
z&1dl{;Txe;Ph|?<0_XI=l8`uj@0ZoLu?*)q=OMaU=F;cTxVtIU?7W1BvsA_@9LC8d
zs@Z(%eYrYgIs_l2jXWl)6@IQQ)y-X(KtDwY-l&Uy-!;;Zj*+2d)OtReid5XeeD+Ld
zOl*v%A<K`X3MRlv3Gx$g+bcaoZ4GmX1uWAr3lJ%Za|Cevil{9R)I50P<;D&BGY)go
zvJb${Gx(qjidtPVX(m37T+6d==gy~93t2dvJ!Vw@XqzxNH}l(k9veJ9o0+@T=}cME
zLR{}@0fkuIyVT$Y+R{U0_u(FCj1WJ*_evkBK%0e1imol!^FoVk7B2Sip;FS=hE>~_
z+QkhsimU8vxq{bbVs&@#wOp5y?n3R5-SVZU2p=b!r>O$tCc1x`KY3%j4G^bA*6`7e
z8s1nIhJ48h(IA8LyxFKoSt^wwFJ}_K!AjZbQ4b08Z&}DBz$!!qIg>k&$k~0LL7B#{
zo*x=MTTW2yY)|BSBfJ(y<w~FlZ}40l=cA|8Qq45l$U#gg3N~WVA*{J*2++%7rNpVV
z-K2@IzntTxnd5a5pwNqRLpr%_FhvYPM)OM=^UrETCRi+ml^S&Qh$$L|83Sm=51<al
z?X_9`<Z}x{;Y?m4!b&smq&I-JN+{u$BdT;6zj>BbGqI6>CQ8Z4SJLfLIXQjpF)@cD
z;&C_3SL!y77;22TyizN}_8MX5<b06^^}ZY}PVlf^y0P4qBkSMM#hzH;#VNTLsqW!X
z)ZisAy1M#N5aGpWO&0o1=F#I{om;xT<e5`x%)f;ea6X>5YU?*#eL5jNGWoYZD5&<5
zpLB1A$6yDl<MKd1LT^O>aFYPl+G5=5eAa4%ppDm)MUc@td~OA>y1Hnk|NS+Pepn>t
zM3D_c#j=(VIp}bU%J23xKtiI^omZ8}VPy?K_b&<Opk-aDTbh1E1bYAo4oF%8lV5SX
z6+K!phK03@6YW1K@Bl|)pBspD)GUZ1am~j-P02chD7_LwksbOuY%s!zrxw|J^NPUQ
zRDQ33YKspIe0HGHznzu}A;JQn+e>k?YJyRx&K>8F>F2gO1NB(3VOotD91G1ovMNHr
z7l{DfLiu+d^?o$h)HTh;t8XL}bXt;5cM2T|Ge3OoAC5(OCk0wwpCC>d(lyz16O+pD
zln}t(SzYaVb|Y>(WB_=H1MpdpxF57%mnx1G_Fb1UX~C6C-lQjzZjkt6VPdKav%R?p
zp`^H1IFis<UY>~v)xD#R75;v*1Q!sWiEEM{`kYgsG~VNytG^gC1}a83_7fa0EZ6d+
z?`C}c_JeENfE;+P$;Z+c?Mee`EZb~|xxtFu?03s8FZOiV{pqp=cg~C2G{Ivf^<}x>
z$Oe7Xxhecjx35T^jYI56BgPSW$Z^5edbU0)i=POY*%0TjU1W(q6W1VTOR6MA9>sLh
zUC1zfJa@K_2KA_b>&I%fUibI3{<rnp@yo08>+c=+0VZprGXj-Y{css>s70scT{wNo
z!3kisi=+Lej)=XPIu{Uc?t)E31UCo*^6IFdFEz@B>t{z&vYBC`hUF*CtR^LT2mwCN
zA54Og{I8eYWTd?+RSZzgF<sNUd91p)bc$82SyoV|2Xm{&E?QYW#i|S|JD0YcyB#p3
zwL}i5uYG<Idhg43o(UfJVz)7CypjP>Qdv2{e$vSchqeI=1ng#CE|4{@u0KBXSLa|l
zeFMZv*UJ|28ib=#z$^#YR4Q1t^?~});`9zV)yReaYB58|&vh9PxXc~E2}`*dX02c+
z<@en99-3Y)+L|kDrs;3R%~omE4^(A<P*P`%m{=PyJeC*$!e#CgRGy06NrDNPFYE(q
zjKPZv8CL@2N*gqICmDZzsI;eVGh<)7KCg6lhCl|2w16OQuD~(eH#@E}T$fR>_@B;_
z7@yl5SrdC2$$)u^DBF!?wdSfEpz;zqpV$9&#8BZ#BcENeL4p#^3mMLheOSTe>G(s?
zvgT2CKeu#Sv2p#EmePB5<rw-_=-2P1?HD~docoWuA{!&e4tolDWmphW@A1c;0H1<=
zGWJ%hgG1vc+(SL65t3*2POWU0xz}Ej{?B;Nw3Ib?!i|(Oh`Q;73+65mr{{y-tIO97
zA`h`60U}k}PnD^vNgn|TsjL)#F>KoOvv{Ie-A<=p(7p5Z@16tK${(Ur?k?1{IFwPz
zt{$~*H>7Z>5K)kwmbl1JP<)fGqY;?F)>_~wEFe!MMN2mXBG`R7nm>sFb3E==U{ZE;
zWQUU9;)(zHlB%l-beJqGA-~?02bx#SJGwiF^p*4hK#3!x$m|?m7HXTB75|d^X=2<A
z9aWsLC=D+i!+Fauclo_dxHc9xwuY`0Q=cH4v&Uh{3~DqcIysezoO1o=P2b2AFfU<I
z1>XXJ+Q}1e*!J(<0r?sm2NO&7Fpg2|Hheu*ZyoRDv&7ZY5-SIpNQwfiRjVV*W~M$7
z`+drLoOGKpjD$@YQmW>;`Qij?GkdRz(yTLUF8I|`1f}I}6n-O!c%gC3+7FsGZ{g3D
zhd~UJ)hi@&<kxRp7w~p^csjD1@BpX7KaErbZle^?Iz?Z)TXzy#vW3a+_B~8|+x~<F
zHIMJ>S5-QKOhM|_oEMMqi|mRVx_ICTpv73Tz2EtM&68uqVzT#>W#KHfO=0DxNY^v%
z+zrWG9m;S_ilEa7m#`c!_owIx9Kn|}OF1-Ib`aS7XdPPPOKtNGhtkF+^|R&3z;~4X
z3-63QnOqmVyN=95)GquF`W&=$tM}<#@Z66Q11nmAZ+ri?g?^RcsAG%)4nuDIK2Iwh
zoC}5UX_Ia|IB1aHE0+%UgB>+|4I~{ArcxAjSi_VZN2Jqw0YX9<h@c<Pad1t0u<&Zj
z;(#jT4D)U131;bU!djfv$rvw)GJLBoTqwDYe`z05HZgS9TqZ|_jX0K!*V%HeB6npY
z*9MU?WitES>V@2+cFG@X-iCR}&jTxRD~v>acdsQp)TTt8)i<vD#YX=?M3F-uSKg~Z
z3VYb!w#JU}c{)61fV~w{3k){;MWbq(=%pMn`ukap#wp;5sGB@J*gws|s%{pZ-T)&U
z_=}RC$MMVjeA;v?gn-F_4@!W0`an@CJM054V_dwBmh#BtB37^zw;fn%gZkfEfKrZW
z(`_^rZ;-_fxW&pCe6h}pIyc;)VVX8&&NNHJx7OB`+Az<&b09Zo`*OdlTR!7CNQ|X+
zr<Gm=seShC8he6mRJvI+PM{Km^G8O@NE=tThlsR&met#(-9S0GOELc}w|ewm^79(+
zRq;Dc0dlX+(O*y(h+T)gmS}3eB@CsJafPzl=*e7Sp9%nseu~&61X&RAZ^9^<e37`2
zWRViTpM2UirSs^cjbZ1-_QsVn0>~6#NY>*em|_|euR{K^(1#C6mnh#Q`-T6!6zhv{
z+1Lp63K!$@=BjI9c#nxvMtMW|@>J^jhs{#lVpmctudv?MOXYfpx_S?m)5RO6Nd7i<
zpcoyEkdTnV%za?Rq@AN_<yOb=1EJ`IqW&Cq+3<CSJ2ErXroDrQR8Ky-5$NU$A|ROe
z@XZ7Vc%vykyL#3+{EpjC&R*=Oh%GmVl=E9f$@`*)kbdIgHGcHWa_f!F-nN<1k-;rI
zGcs`CNNu=xYT3LH1<D_jA0pgFzAvV_h(3Rt^JA@Ao4i}B(}rd=Qz!W1^`1atYTasX
z(a=ZZ{u*?Q@rkJW>v!0j9#FlC%j=}3>51c9<K;5Rx|bYKy4nEhn}J)<1%2!{TW@k%
z`Wt%X9A8cA4??#|u<e`BO(9gyX~{9C!&AeFR1yQqc6-?&5N)n=IAm2KizUW%4rpcA
zx1{s0G`R(|)cz0xg3%%LbGZ$WMD%uB^?V`|@AKn{DQNNh5|EQ$4A{jhUok%Tgq2M-
zUSH~qL^#mjbKAFZ{VWK8pG5&#%_?iJ1Xj6U)IIK@QHC)qH3LR-lLG!UP1kLwoO9Pr
zxyIuW!JBIR*Hq{~;~p*^)9ViO^Zw09=*VuhGb(TY6;JMaJ{wGhu8`PAGRb=+Z;;$7
zmRnFLAm%^Cu!pYfPh&93pQ`hZN<N$Q;`RLLBJ2n259r<_@EI9B<h4KU-rh@qHZTn?
zYF#<|<kq9%P}gTB+k$XXR#Hzh7!6f<mAq&^kJhGA!(@|X195uBJQ(?}FJo%UviZ`Y
zG;>1&a^*6HzZ`G0#AVQ|>8_al;?4%Q$EO|M9|#)jUVIIgG8Ge>8a@X`AN7EN!=Jx5
zn!J*d!bx^uJFl^{dR(%0mJ?WgcV4fpx9a!E;e|X-wNygX?92!s)V<<%s2h(}3(Ai;
z6+ERkao?Q`-x1)27u{B~PRFQQmQ8j^LQDx8bGdgO^h6KR5HP<QLhNO#FIRBto)2I!
zfmC+$UGbt`F*io*6&1?rLRGk~1z7`t)J_p*mskw$BEUn4GkigJd_{=^C}E&Lv0y9y
zxJ0m>H86Tu;m{tTsc0((l}+N{1S-15?An>%<HNtlsS|V9*(^<G%^h`%L2Jbp$Sc}M
zgK~@yY3o_JeiCnH2tV<}ER0N@cM8>3;;^L0TNU6XAeJrk2Y8&+9AM%sOy!YQrjuXi
zdcLo_pS*inCnm=-0afk^pyz6x=m+-EoiFBrv5^fl4!}_rmn>3)PrB88PBXIzx^U%*
zFC+tSO@*C69#)a2Zp({q=M#54`PfTnFoy2BIr5qF&C8v`GdrXUkRS_suGT@%dR(ME
z2=cgYsy-LUBSWV_*t!{O1^KeVt8#3-=IkbarKfp30Dy4<;I!EMjaoj?XTgLpR-pY)
zleg2g+Oa9<3TU(_#mhCw$v+VTTd)w~CiuSC`m(k^ZHt(HirXFTF!G(RFX^(r7Rv-x
zo?c7F2s~R+JMTzP@O8s@tpn->2psaCi&q63@zG;-Um^Fi$qE<XGOKU=x*>T?KtM38
za=z=zx%vsLi~)4-EhD>-&5#v({aZ-(<ZX9r<$+v`)X!XA!&JG9ME(XfenG~`B#&_Q
zo98+L{p$=)R~Zsn4TV;emp#m<9r4^fy{Y7{$dPzEJFi^m1%G$*Fb`;cGCMOADi5DE
zHKe4c;YWRmz_?qI46RS-S5=PNcy*clJZC=u)m`;3oK}1B*)o5_%}I!VzpKsluh3l<
zyYp%(B4uU7(`k(Rg?R>pe(G*X;sY+(Ho6mqmp?$ihdn(OD2Ex)--;@8edhLR9qk7i
z`;kPvp-QvE#jDnp?3f(0zR-uThl%vuEu_H)XKlzsXi#9}%EF_dKN*R(<f^e;>8lv$
zq*PSJO@go57^7O#Pmlr7i<n5ij6ARJ;Tr%!cGPhJ($LbX1Hi1YIrPS<g4fE&y;@!j
zrn9h1&<$i#szPI}epjGgSf#)tN5#+1KX8c@@SY4Vnaf^34tp!o8uq)QkrzXsX?&{|
zoiZ@&Q^&%OInZ#_jc%Kll$=w#x7LNfz`h`h^^b{=p+^Yvl0vAa8d>}uR>15|RwH9Q
zA};l^fpqlNOo$zA^=tD06`2!CknlcIGOZ=Pmnulan`TDFKWip0{g1l21_PJDOk^WA
zI+hGR0mdDDR8qI>>aA9TNQEV_9BJxZCD&>D=Ld`%ITlAnX4gOzdb3iD%Y^>)WYDU2
zu3BA_3Bq8}U#J$$fN)TvJEuD}@IFfX#Vg=v!~e!>yXq|V<#H!gr0_YJ;3BUSkZu2d
z9jVkF(sL?0dq{tkvl79Ah+>==J$;|jk2HGWc-BI4Q>@DloqKvfa*nRw=17smp#3XL
z`a_FrBsaBvXYL=d?v<9(tRxX!s`F!!j-1--kTh0=E%Ni^?N6Ynt86KM27IxYDV_Gc
zR~3AyJ^QbULCSK1ts3>02)iIH0GjZLNFN+4tJnSxZ70jBvX97IJg!EjkM=VZJ1BzB
z_cbXygFY>hH*_D{U7QIu!c6xbmfD$|?$)Up5jCpMyQ$IUTx(q^HOYW7;Qb@gw38{}
z?p<xze!yyLax$SU6hFzWA8%WOJ6D!FcWaZ?CSOZWmJD^%ANyjSXHChEDvvZE_hwnE
z;G4pD^5lGh(~;)C=rIad))(&rh#&AV41>Zzy5ck=F=wlj9bBQVG6bYF77OBd<r%s0
zmV!~z>7d|3In9+2c-S%ry6^tcB#)(}ho?spEjWb^S3rH)a=+tWbz53^ss2s1t*)*G
z$A8y50jL;|z#|V=Tk}*cOn(y-iw=c?Z_Y6D-MI&j!!pV{&lIfOQWl~v9{}2-xOr!8
z`PQcbkO{y7AjJ}Gs&=*~d-GmrDni+?YTNU$h(|$3kXWqO^APbnK@i*cga}e^`}_{1
z0`J;{T3!IyWnwkbxb?bp(H2>#JdJE2{%zyllKz4D5^tySkqxHfSetMK+lYq#<W`>X
z@()&3-=oL+J6z^9JfYwWhCSgA4Vc*&vVmM{6O-o-DVZh?f@_{@7NJ^)oKiCT5P7SV
zDOr}?xdX<F$@m<C1*n##G}OSI)?l^!bCCG0K8C}Z9?t$Za(?hWv%8Q*e4neh;;zis
zv=Vy>{n?<G-ph%2g*-n#D!<c)GVCwVjW)wseI!qVARuOxzc8FU2n#>dq8<d@?Ls@C
zKV6J9Y|@_Svnlc27eN`HmU43ewOOv{0LZ3X*}IpMwLPsDGF11Xam{6o+9o!JI?%ej
zY85PP$Jy?`e$7^PS`C9zIuwL>p!pw!)$pIt1L~}%aFr*NSc>$KYTpyl?mxemfBtUL
zg?(C;QbvaFu|GcG<*qhg!YaP|7FWcBiPr|OYDfDG=lld&CjTIgh5f@5KfkUNB_$=X
zyFAUqHD}Hns{=41AOS!tW|1EWD?;OJlLBh#?ZR#nT%i9THtQMXcayuRDlIMWQ4@eV
zZo#b6l|y1dCu1MjQQ-v#A{L-?X|uiUUZj4&WpZFT3fdhweMA~~!8yQV`tTnKdNTO7
zGNRmeGp0O8dGU;>ld2gOuB%i-XdaOe%|48i`2nsjA_A(x%@}#X5~Y+d8UYFvK=6VZ
zvt+C0uorQt!Z%bPW(?yxEz?xApA45hXPvuZar`Lbg;pk9{RG8lV)Kd>hmHu8Zccyc
ztXR-Bkqt8LQ6KT{vZVonhAtf<7^clGV{P_bc*0($fH80G%VO@;)X70kVG>@6ImqU>
z<IVXL6T8{CKn2HdZDP{WQADTGm+L=tPB4>n6QJ>--m?`^yLMZaVXdh&D8pP9%<bc-
zrS@rIGpr5n<{r}JkB==m1tnGnFm4PeILQ;3-^k8QXjobR+{deYMWcRK#d`U=azn**
z>&6^*9G7jnx`Lf{e{Qe*zFYl@29A(b`HG>dxcbkIqPbyGCU{2mipr||bY^nkFZBS}
z>vtMj7ofF4`T}4QdHf44VvxG=`A)c~b)}Nw`cdjuj)CDV6|hHQ^y_ao!xsytfms7^
z`L}0zhy-P=B@WEalHuZJM^ps=ydNOfH8nvi8r3N}SO_g>hR@>qTDSHYk~Iv7I-6<%
zg@JsY-05z=<jeU#u-lZ7h~1j`RtFzqbe6eSYvSL=w%)e}p<@m($ZoqI;Ia4k>9pG?
z57j<1rM`{Y&6g>U?MG3Lo1Ad!q~_!iz_yYvZ1h}<OFjmdv|L56Irt_C28(6E(|D$M
z+<W;(oIx}+GMT(h$Cfdw@qy2}P?Q4F;<ATK(|Ni0DQ(IeU%0=v<V}P~jXwCn4S$3Q
zz$5|AD>XK68?uIzg){*bOGv=-NfXGlhS(186Q^{>NwNe5V#h7CIYlSA?0AEwVT(C)
zK&sM?&S10#d!MMRm^0U&kLN2D=H+1MzWxlQgru`+2SS=cAKV@00=}rA8rgPTY7bnI
z=rmIJ`f^@Bk3Y>h((MiVh2c9S8eCXy`x608w$72PzyH)BXh+Ny=j55faoayAJ0AFS
z4Zj0-#T=7@41=Bs;gs{r5E=I)Z5O1%C9WJuw<>i57iqp|kd6vx8rEWFgZ{P`5!LtH
z93#?Tx`AvK&~X3dqWq4v+O`3M$ld*lQRccB6nPf+)Cv3BYp8WMQcT_2F2}pDJn;L_
zb>!adhlBIl`Q}`yQx4J9(B*k<-EUG_JL<3y?Xs7k7l53e1vD9?uxj?fadH=u{_s#~
z58=H@<00R6+vNEY+Icmvemv>Ew3VTxw`>8pR}ZRyIfsq@k$5?c8-EOhf|Th4pMpS_
z03KSIxgRp2PRAz`cUcxsnnDx@nQL1!f-3W~BSq}ShC@LwMUftc(^RWK5vs6(NdPD%
z&EJZ^czrHDcn3#QFTy*N+wPZQox;U-7!yRNv{MVRRr!}{%mFc)aI%x_IQ9~UgwMh2
zEX(H9-iwdhv)$<^Baryfs-9$J@#%+~1?5vgmYir~K2#Y>AXApR*h6}v&0w`kg;aV_
ze8k2Mg5!-;{CJ=A=Ka_R0m%`sOJ+lZmF^x2EO(w-&Y@m-l@p`y4lMbzg3hK}jX4%j
zg3N#<+|8t*G7zL+X=y2kX(J%htH?1!b)a$78DlFp8(Ia_3jP{}`9ix-K+Zh({Kd8&
za82-E=Q8?f=1Ck@pamLajJkHzC7|>u1ez4H0ub_+?Oy|8ZS?o=-{Z>ZJlMy-4AZ<B
z0J>m)D<}2Ew^o*o`AYsdSpLPrZH*$QfR(~*&cUlRhf;KpSrkl`kllaT9|(oQGh1_}
z-tFlhQe_^y=#qCgkZ#rUS<23Rl^;(H#2#)g%ck$-OCSd1{E6Z2afq7uRIY!Ty4IBI
zZ_UnFh3@r_gPcCF`Ml%JuaXg|JH!<I4WxU-f>UOQCwSN2l(4b6ajJCvsiSvE`Ny1Q
zUHNDEy-OU9iojEGCEiV!whnsTXva&_&^QqZ6|YZQxM|L37IMz+)Hv~;9-LUD)q6ZU
z|8-RZWth}Z6k#H0{OESAaxin3vt9S{<=m)WKdl}Ru8dS>kB6IRN|}R&2#Oe4yoN5S
zdiv#3py$1XHSxE-UTzkY{$Jn?$|LX-WYfhv?ThB~RL(i57xk(8^5*F*-<%Qp#uMy^
zUe^tZypPj~FSh4^=AE?GmHG4lG=Nc&ov>(JNSUv%(@c#GZA6BL80m<+ieLRbgnQD?
zJW&u7q0#sc_%&Q)yFlqfa;Vg#O|FDzM!`2U`<fSsU25WGR>|Md4468;vrv0iH&;<c
zx#Mf1$#bi8A+Z1~e+}#-<KU%)<)5D8W<dK)RegL(rp#|SGqgC>Ry01Jun;~m1thRv
zoN!$#cRQ_?ER(2lC@F((b@uZSUxO=MV6u7FT1QGvqLct}SFbDoWH{S0{nj){tbjYK
z@A{F#yV{tNSf|MtyV{c4z$r6=Ft}67dY~9(Mea5E>HXob@02*-=r_fTZ3!<slAEz#
zJrfxz0}W@S90s6K8c=`xam6xX-&!$&*cy;AN@d8^WQ+=s#)b6GgPe^Wuz?9sgQ-Da
z>ohJLJar~7Pd8MWZoWlO0gD40L-;My&46)=kotMeGF*p+=?}tWW-#cQz8o&|*Z^H#
zK3gP4qf){{WPQoujh*2GC-mKrkDXa*GLNe~j+map*DS_q7fkBMoK)eFm6++OI(#_0
z#4x=;Aa<&a$MjnY(3vBwQpHK!c3E~Lb)#W(A9W_`x@CmjPnezxxDV6}D@5o8hzVa?
z$3pTL`TIRD12_e)yDFFe^1(|=r-KVNdgB`x?K|YsG+HHmYri(e4>xb*EYFGaJv!yE
zsQD*x+!!u4dA&huAaG5)pXSYVEmxy9H*vVDXH)pm1v2bkHG6i^Jz5<svl^1Bv^FCh
z(0Jo(RAQqIQw1Dh7fo*F8><F8F3-*S6+Q8Uo_>ntJ?<6VFp2dRlFNqKyYfkkGYcZP
zpyiN<R(x7Ipe4RmdydPbn3`>5>(|D*0DdB2xi=zNvRfix%_{k~a@BQrwg{Xh?}a!|
z{iyRWK5a25ql<D{<Nzo)9CRyAE*rTOCi_OX93Sj=P=x$)L0r!Ku;>E9(*e_kwIKuk
zGdg_i2<pJwHi)NeUqL7h`~^oY@b*P9jLn}<ET+b&hx?#Bq5K|PbDy~=5frS3y>trv
zRmWXx>Vdy*sSgX0AvQ0!3H6e{`T2L?q^wM53sg^~(rp<)5EPl843_Kjvzu2Ya+q(k
zHy?my^rdBw1cR2uiw__2s%|lrUEgSjev&$`o86S*79$5v4(oG&scZSy6swRJU}07y
zx)>&t#F?~I%F9UA8w%!%5k#T^%^$!Rr+*$JRnra}Wnku_YnIZpE%UeQefzMhel@R@
zCVOyB<_=_2)o%}~BDSWdSK>@ic0gJ~xT|f%9`g)^l#&u)qWol8Ob`9gNDU}ZE;n>3
zP)Lu>4=&O4xH(jIboEzKwn!g4KDQ4V`Gx+Z2{fVUoIJwZYi}ua{o=)y#8`|o>{~%a
zh8#lL!)&>W(4tvD|M9%<9)4jY-ZLK&QBmEx{oMwAFLbY+5v<dDjKMs|P-NPrAkW#g
z`RN_l`~vjmQfUdHig*{{xh8lsvUiTp!swIv3tpDP4?QnMo>E=0H^A%X92kD5y(uXt
zro?&WlPPS)ope~ZmXxdO&L}TA9D%gDYKy-wP~^-~cF@j4dq;;^+mQXooqM*`PY*4+
znOLaw1%1^56HjmcV2tbnFpdUYtqwHOx2)zwV98L6d3NCU%^-5UgRG(Mu@|UCN~o+{
z^@@b58cp)|sp48<rGzzrHd3p^kAxi?kI#VZXrO?PO~Gj#){hzyQqgqVPGu4cWJ2Q0
z0Zjg_t~CHeH^RF2h~Lh-?-q<`H2z>q$iy?m>B)23u7A71LNE2yl`e95EguqN0zmR@
zX>)HI*Ka{9xh-3(-h{^I_StuK?&hhydvS>&run5M`dxU;oMJsP79=Zvv|6Bb^c~9u
z9v2SF_(BV~k6BVE1WcrVmtpw*T`ggy>Zs45RJZzgcZ3Dw=@hotaZ1Wc6-aDHW8;~V
z8J?P9UtIWco{Y(8T|#Z8bfJe={TQC{%K4!2IJhx?R(OCg<BxKc8iR%HJC<Xqv9`$1
zdUADDV&1o@G1Wa=h>(utEy}dR%sRc6Gp8!@RLt2@FO2AqklU3Ag{3f}%1n$GH<>%U
zSr3#PD2FWOU(^Z$HM#y-k|kA>EpF;NXJ|_^Rv}c~pJ2qWoSUE-E1A)#PMI{QIUx}H
zDn02`1$+`t`XqK2)Y+n<V@`k@vm)9W9ZP_h6t%p#1G%yTg)P{Ifh5YnYCUQDNzI$d
zgud|(a?q#^yTMAfBV2*AkLK60<B!r}xl((r(+KK35z^tp*(2bFmULfF+8!+~1j8Kn
ze=$-5;RADg2B<H$Y$bS!OSJ<~hC|~p`eDER>geEv;qtECvpXsLu_3MN6Nlyzzx=d}
zfKqva`OR~)N%9>kci?bh%#m2{gF@?(2Y2HU#*H<{*;U~q+FCXZcIvm`RonLZaBG88
z*S!$O3sqhPjBD`E4$K-T%P_%Sfx0x%(@?Hq>I}1%%&j|~x~*FNK?Q~KKZ8vcc*{6O
zNn%F=`Q85I3pvN1lr;+epw6wzUBnIA{?lq|6VO4`;xoBKe3AYgZFsMpMOPJ?lgHiF
zJW*R-AN~gW^B1MOjXJ=>4c5o08I-=na`=REVW!%$dn{;+I>o?oK=o|Q%UED1UMYm~
zJ39xrT8umI&}gyBnJC0Z@Nc>V@tFh>t4I><;)S$!h0Tx5ari~B_11ioV|gJ^9|lCk
zUA!8g+P_)s(&JM!l*Z3?ArzXKoN&wd`6uLdD4)WF`%@cJ2NUm<ty*3eBKKzy=(<;@
zlQ$<qQw%WFbOADt4K$jP8YEh-YB7^^-Cl#jCr!w3Wl5tnO^Mr)qD|yVU5HKA<+$=b
zBglD`(ZFIm@>qX6TtproQ{JjqU7uR=y0z?_>kIYL6JC7Pbe#r?X5bfZn;|9W0h7t6
z&Az89pPrU@<#ptCpvmB!x4oc1EA7yOMNxTTCL@ARIp!M&0SY7Pg>UJ%wJ)P(V~)f{
z(eGUn1_CKC7iOTl==P6)Z9-ws!3uesB0#$&$t@7$(HEa|E6w$N?N9OD+SN}WFYwDP
ze&lPkgt%laB`#oBQfB7?ty4t_56l@iT@K8}6vgg*bQ>Q2RES?HIWVP}hBwrxh%q(@
zEPVLj#hoUeW<StSc#HD_lJwu-3SNB@g&`oYg@JTvi|nrru%1Rq8c`dv5zh<&b_gfr
z)yb3NUyMg@F*t*7X*}@`2KF~fj$yq@kE0mgoWFaE+)$5v8G&cWYun6e?w99ftZP_L
zZEDn;w<34X$LTW$_U*T&f3!pXqd@!Dcf)+aa&CEy|1{U??pTAaB;1SJMM^IA)R1RJ
z*lsxJ32Gm5*3t7NMn%59#G-3&yZ}R-6kky~iX(sxDG2}9KOX%b|M92Wb>ifW1%?Dn
zE>5pcst_tfr0A?-Tc{m$)xfJpLt~I8=OBw$*?y$@AngG}rQ@HG`PYU0^U?&u7Px9>
zfO~+rPKgsHUN}SW-<|RAH$@tc0dW4m+T>+{gZp1>`a1vjs)+w_$`<k8^`HJ@|LqR-
zvA?VK{7;WH)q=K||LGC1bCmr*7rFfRxn(8Tw@ZlqPd6W@u>V|q@84$@S^S@$`0ug!
zA7B6fGmA~g{#mk1QTv@wp9NSC(rj7z1}pJN)%yl~XrJBF>ve~{l>4E|b+`QC6IYy<
zR*RDDC_{thCU4!FX$Uu-tNsulD<dNDAo%!iEkLtr*)JwE|6Fpj>4|IH)Svf-T<)rj
zsEmy0jOfe}sJo`uy-z7Q+){GtJ16_^20eY0dj7xL0p9#S-`7Z22hI;oL!JoAx^mht
z$^mBfzuDH0tGEAKul@IsP{Q^$Akbp}=-pf*>FWM^+5h^Cyi!Cr^MAej_a8AEAwPgO
zY;rwgL0&^+biP<>7OQ3s(A7Qt_Zfd*n+RRCo&V#WOqap=50@fu@qhOH&&NP7M(5vS
z>eG2Nce96tZBs&=s!owbF*^6sM{f|wEhv561EJ#Ue&ilf!5P~}^HLDGh1C7~*Vq)Y
z7@QAjI*SDH`Cl8uvLNzYZfy;EwM0P^cQ23W(*wGAx_dFwPw#Zmprfqd_<^wM@&D<m
zizq26>0-VMQ90U!BI|BP2yE=f{m`3=+cz;A|F1{rznL&ddUwuep^bZ#>Du3)$TKVa
z*C<2GcLzpBR3?lekicVb0pLMEf4})N1omtdl1aS^p`x1q`}(gLgFuWn?I!-)yyx!$
zpEr{K5%YwK|99HQUt3$tPzS`Q|NSPP*kSyi=|~qf%727@814V&`%c7e=jGRe=V@Ev
z&=Y*;S)2Yi9en*emHiQzb6#9lFzXw{fV%6=b=SeRa(?z5^S7n<2@{<!uTEQCC<I_U
zUdc)C`zj{r8%Wb_Ip218;Xz)Z?_~Cd`Kmrh&41L}%Ubw^xX>o4yVg;jD5X82#I3#3
zSeBG7d6x)0Y3W8l`AOEOg9whW#GLtVgEifZpK~3*6ZEO}p?4mKlVTFC@xe?){cGYd
zao0pUVfpr4ndfrd_Jt21p%s5;;914t3|oH`WY9Ei&$@Bk)#<km4R}Bu>^+vQRsHxh
zzfHW_n{>=0bDxFk-34tIJNFAN&XnY^-WjSC(sYTP%$eS>@k@s1VA|53>}0ykpmh~v
zA?l{?Kp;{v2PMl_mes6AC1@Rbbs~4ni$-b_p2WCMR^-Y+AlDMWazhhN26FrggcSN*
zICWFti{4W_Ua)cxfvy7x?7Q0CNAE68S@Q#Mg-uh)EiE@!AgE>(R@S7{Giqd(mR-#t
zvb-ug*`wI~pvlj5<O;5VPN4vAs<jtk*LeB8f1XRIxbnSTHc{$n%tJ_eDN7d!WJ33E
z$M$Z2*Qd|Pd(QEVoT@Im%>D2Eo(R`@d7oW4>+c?o>`q_qOhidNkYR2h)|Z%kS6Vla
zU0!x*Z|rQoxvJKXk{qXtDIYkS=kQ|vs6BbKxqTNG41j)Uk2L~{&~bk5Q^&(dg^!Q_
zUQSNr0bpULDy(6kP{FofBQ92D)Haz{*G-kWn3}pGNg0ED6)`%tV&#~HPwgP;a6b!g
z%^wd`oFOn%-=ILU_Wy?KtYLvU*n2!WdAW-)zm(J7q-~PtU-I)Oc2=iS{F0FTUM3YE
z)79PCoBkwIIJb1S^j)qxzqk~y-dL{RbF@FQ_qpSk`QGSwCnep%%Lq1EeLgqDs<jGk
z$VcfFZ@%w;Fy37~57%M>Q~0PqZEiZI5@-Ui;;YU%@lypU>=PiH7Mrt0YrlE8WZXY4
z)8j&f&~JmaQM#9RWnImHWDmN$&(uHzFnvl6!_>r?MlZsdH3c@I`{0UfVS>=wbNU}(
z4j;_Ce{bq_rgI-t|JG7CPgTN3;uUz;WtgCIl2)cIpy(0uo{7coPm}dJU#6YPxGDg9
zh*BM`rJ_2ZJm;LNsgE0EAncs%b%PTWRGt^4$X~ydXdymqw+RzNiC5*h;nGO+&YO|;
zmcI9=8&8CE`N4<26Xg2uwwf}w`B1KpKt#rfs3P_jzC8K3#zVw)VI^_z>-t`Q6<N|+
zKjokec<i)EBe)L8c?t!&qnh}ae3O2gUD&c3xWT3V7_a3}+v(2i++a&)jzOMgU;d;s
z6aIQ@%lV!Pw`oq4)2RKI-|91NHGI4y0`6)bKludxuHf~&I-4Di&V61fT>bt&KS@RE
zqh~b0u-0?plJ?>+krVN9#GEOlKA--YPs-Jc8B#L2MpExj5J;bkUOms_=U>32PgOks
zw?mmZktD|GN9bwn9+sV6h^1*(m7<4<FPXJ$O}<MaK`Wl}h4;JYA?mmdGY4ERu2xR$
z`?n@{4(ImpHZJ3*DwOzbhSa^Lw8OcF&j<EQDBQb8>1Qe?@&m_~i@)}5EuuT@g()iU
z_?kB`pC8tkTnjwOD-kw)UrRN!8jI+Javg8GEv1j3k6BQ~85xT2X6{%w)^<{-Jh>;1
z8@@}yuH+f+k#Kg`xN|ZX?J-%R;8O-K+}-J^Y543nPMNEU<he`VSqG<M1pOv7abhpY
zU6-RpK#p67w_W%2|7!0oAEIizzTr_$0TlrQ1tcT{5v98YNd-i@MH-}G=uk?eq`Q$0
z=^VO|ZU$!PW?&d#fFYhecwX0izkk8|e0bIeV20VVkF|~!zZHA$oRz~J4|ksWVmN|&
z#O1&kKTWyOGWCN)meX(K(lfuGO-elEcj3)CL5+P5B>7ODC2{E?>TBU-=x5iVG6!9S
z@jpO_vaF$>98sbCki`kv6{e5_FSdF2&F38a^--zIro!RJOv===bf6CgSS4OgFZ*rr
z0r$nnddW{Io%zK>-ht0>Z`>L#yz#Qd-sO`Dumkfc5F8MjxbWQ`n<}l`V)b0=$KXWv
z;FekGhEgH~<b|JF;e-W5K-lcnj!r35PGCdI^tmKHVCX`7u4vT!)O|^}D!W-g;vTUc
z0wa<Q%JmZsbs|);uigFM#?miwS;IEd@1Lqgl-4r2ZXZ6A?-V~Gndr!bpYL2kCdLoY
zckxOUB#@SLXD09$j1-c1kxfX=^RIm`(&J*79L+92<b~*W<g2YZ$F7Rb@iBULkZz<F
zY5<Zg(U0AmL*YTT=FB+)K2{znT!wP_zU9Z9=L<J#*eprd$=ie7O&1T(>ONfi9Cx%b
zV3sMSQg;ObW%H}^x1C*YDsd1bt>~ZjeNQi3#^8=u=$%4v#vXIfSh9C}11nPPhXJ{Y
zpY4}I_L-DvckhGp>i<TSnK%bQ8#r%GvGT1S<DPvK=%4(4wo&&4{A^PQ$Um{Uaf?Z2
z@7%sYkcssnGc#S52Dw*W-#{S$vqMDdm&>rp&BB4K?~G0B4M|TU?uNVR)!sktgZaWP
za#No(tT#MGN(xL*3%#GdU2QWWj$b_H8#<5bC^s(p`2uk7Cu85e3bP;XHNMQ)U?cot
z@tK{fBm|%4&*1AhiiL&WdANW-^^8dN1O`{f84o+UOp$m6nelpR0mB)m5)t+u_Z5ZR
zVYQc=hp$S0b*{ompM55<A5tx%K+Dsth@$7-lnOW}hU7O{2}@jtHhNxGiZpuNM<4t1
z^v>+2_(S#K4a%0(5X4Gvt!(p%yFU7KZzJolw;72+Kz5$?u5QO<Dd1kl%em?~qX!Nj
zB-qYVsqKUb(Yg1H2Oi^tRNzmr$mrROiVnGl&-th|=5jSk3cuKvv&R@h%0>(!PbVKF
zBBC|ry7_2fkw-h!e_~HP*8eH!(>RcWZ5&GGh~sB?dL5okSgHSVvUO?la8@5c&8OaO
z24Y7XbA|HMkRNmx)jp`_Q4C0paW9K^rICMQYp{RYuo}9)cq|;|X76y=<c>Mt*q?7T
zXRcIG%z~l3Mi9|EXQa*g&2R8&lA8pz5<=g5#zMq14_O{>@{b5YNnp_BDpU8VuC}fI
z7*pkxwi@Vqv-1yPeMeHfPOdV(2xhvHDJdQ4ux!Zcm?Z!<=A;TXcAh@>&qAe=7+<eV
zAAD2xIttI57n0B}Rh;0cD(SE?8*dB)lbWYZ-y5KL-ABp-SwP^3N!4EuC&ig}AYtzI
zRK`CY5_TRg986(Yw%qC++YWMC99svw{3eq>JAdjbJ}_>Ny0FIw5E=ue#~<DI_>0F{
z_poP#@=jS`fF}v=O2W?e-Q7<SmJWegGcglgn@2OrO|vNt7rI7gp~DXPPad*Fp=>G%
z2?ZFBpKa-0wfcVIlu21>;C8nizAsw3J^IdQ=_nwvc3Lnq_x*)gIH-m1znL#-C~6)d
z{)6Np%qP<+k`nAmK2ri{I{MWPS52H?zi5t0T#}9Y)0*b<!_&dR`1zz?!y41B@HvF0
zD;GC+!@LK4rj;wXNp9N3#vpfwe{FnxS8-nDyMe*F)n0?*0E-PemRa2NqoY|7%yoBw
zJ+>xkephvw=RJf;d+SaRi|n3x<6f#6zH)`I=}3d(=b`FJ!GYb?d{~>}Jd^l#``)pX
zN#nh_j?a$j^mF<GBQq&n1~#XM(la|oaE<98%^Ed1zAexAqhkRfXCguDf_9ot4t0l0
zx|fJl-bO%*95W<Pfj3t#X%S$RObpY=J5jvv|J@nVg*k`6E@)6dUU=&rx6@_lc?LsN
zQqF`ce04W2#q!AO6mPTT{cC49`q<fSlUKaiHX+%2`S@N`CJge);p&&`mQhQ1e)PH%
z%|DR044Mt|BBHbNJM&D2i<lm1K}lg_3(7BzGOMsP$`t!+$^ZPD=%y38fv+rznkuFL
zVS5xRWoTfKF;biIU<^Q*$j_W;4;he8Bcj+8pGL=~-2=5ee<T564%d^iiQ{jW*Ciwt
zyv@%rVSwI!0<Nas!GGE|ZtrZMd>;JJdZiwfwXDu-T<~)x@@<p)UXrnfA$QfBhxco6
znilC@U*0m(d6L^iam7bPp*_uaUh9~`Yv+Bv4AwnWlJ8_>+m7qnR6l(By#DUogQ-|y
z4<J_3eNbU7>|rZBm9X&jxX;MK6yDn$A^MrX5pX&lXn9G=eth=7PN&9?fMOG6Lr^*d
zH;fJsLS!$vIFBpHG~bVGH~w+{^dd`-{h}~aB<Ex5)URK0GF8u48mFfLP;WcCm0fRE
z)D&vXk>%2-LV)$uH*f!ah7f}u#-pa9%J1-L%TEA;*zhDlq0j*n)CHzS@NUY$InN3J
zucVkwjN9TAGmnK4nlvO3?oarOgZ4hPWEYBQZz1)j2FgZWVZ?5Xe4-}<;2(8eBkL2)
z%KmS+@_)q}QwK>3dsfxCJHjTJjKuH3IzH1Ehiu%VJVoO`l037>AY#jBX3zbyxG=(j
z#5t+L&8|-4OCOMUaw;mU%=+;Y=LTZFp7$AFG3~{9(>ZjI?xYreI7KAg35uQhre}=c
zB6CKP@m*eecqZMy$jdJLDdMe9=k<O(>DzMqb|}CV4U^7sX_AUk>3869>D0>nL)%ot
z?~Ee3AFIK;Jb@<~Fe!tt;lO|M_J9IYCP_2;9XXs2+1DVOkr$bZ^`yzaBkzLJKZG*s
z=UgoxuPbiq9bda<;*z-gLUW(E&Q|AZ6tx^V1Z#*D6b-#8f6ZG5HeX#{?CiAXrrsmW
z0#s%r>sRqF!JOD;eLfc_Pdy-TspsO+_zj!A67m*7f8qZeu12beRiR#HIyVVvz&v+X
z|0sHhP+L<agO|F;I$~60J1Nxowxm+W)LB3J_e`!gOT^!KbN<!hMZv#-Ao@+ycZrS7
zyyVmro1tSz&iejZ_a`bu#Owj!p%~pJOwOB3W8@o~N+5_eD~#1Uzq5yqMxnR~s4Paw
zE(|jjpkJJzN`T)ZZrkazSUyZYdZv@p(|q4G%>`4PlE>?ERgf2+pS<o=0|YN40C;YW
ztvfIendB~&uHH=J;9jcmm=iP^GWtaYh?VUz7Mx45u9^1%4i~Lv2!QKHt6s<VoON38
zkKU*9dpcKgFy9~seK7Z|HcoLnI5ZIlZYB;#oX(ooi8)t^C0{4X6+~$wM<uWqZ<+@z
z_Y&t}nkR=_hYDY{i^tTq4;P*Rpj+cY#MOlFec~&_)%C!X<}ZLx;FEEK3uzK*J5Qyj
zHqClOT3mGWv?($A<V;m8$en+EN%lJ~CUlLclXNGTuft?vqajnsq2mRhUNSR?q>oK6
zIkX7L4-4#bd#AU6{|?mt<hklv$S{e=-D>$`z4t1)X$xl0t){-T^?7Rs)%9lE3pecY
zly_7rQHcK?;^nE~%Ob}7xE_snDJa^(-)S7iS;99#h6{hIIHYx~X<Q}VW}cUVn|mJ#
zM{z(Zs(=|XV!F1pXLx-gl2Wv5PxFWFk%B;Wv+L3gNvu3PdX&kj)m_%|<2zu~?x1+q
z{nP%*&DO%u&?iX^>vSw6A-;odg^mhC_B`1@5FYT<24J0%;HF~tD~ce|588>MD(yRc
zTTX)KhjnA^X4H`FhRo!U16&CE_N<a-in)uhGkreF=6xjsq=}z_Jl8zuH~{jWQfBr{
zLX`UF^{9+@tP$i-CzNx>E2Tl{2Y*Y=AiAjm_tf|=kc-o?0mV~r4v>UTq)@6p0mH1r
zO_-ngG)Og+J_EI^@Wba_NZ5tzN6h?k71Kl#lH8A1<C>chuq)MJfFlPm#?5rqnOcFG
zWyc}~*e;F*hA=iej*^ht<~k37>u6n}Y=A%=!TuEcPDu|fksG=$zPREkT_PHf1$L2a
zA@~Hl4TU6H)Lu5Mtcw1`v(4VM=3z&AvHG<XZ)|QNxbnY4V_^=`k&IVTmmic>*Bc;K
zuaIhm+CfBHwmxS3At~Jqi-x1`vU8*kqM4YnPLc<3l0G9zTGmbaEDgkB5U0mJRU%iE
z;&Wp&1_Tm70po3h^Xwe`dp{ukBW>yk8G!I7Rmh<E>M=rRsi{Zgp<QjLJ@r&ier~5L
z-9bY=p+}w;Os(VNI|s)*2VndR;PN{Rx&r1pQS94}TD7v=kN3#$?rQ+b2VfW@cbN5^
z$uVbX<_faZ8{9zF!Mi8a0I3*eyd`~m@r5x+9x(gzd%xmh0@jF5Nsy8rqBzn3YlcqG
ztC7ZGq&P5#d(v2Q&oVwsSBD77cTM{eP-q_5Jepdk{`MwuWgc*jhnp^QL%?kfucZ1?
zx^dMHy1AlLv-N1XIltw;KQ#{riAf5xU|R~HEZ(D)%>iKNYL+ka1o=>hvYlOMAm^<N
z=MkLYCBG|JZUd44>Ml!N2-mhWry8#4Sz=IL9s;!lxTuGhCD5|o!#}l{<6KkZMNVz!
zO^8EQ5`aTy0xdJ8=;U8lh4YLZ5=-D4SXh9;t_?FDS76h{V|e5i-d1kz5n8i~2b*`U
zbx#Ju4Fe&Whp=A*)^MKg$=2<xKpSN#KDj4i6~H|eaI!d8^wcIt%`Sh+)9829P{wTQ
z$?;xwp-j<FF`)Z~zuik3X;H2#4L0O*mG24NV(P|>I2>(c6@~^IieJlLhAtkS;wit4
z+L3K8`(07trY&5$L>=e&5~)CWoZr!L6YNaoPHJEAa@CMrSMeiY52U@yIkW)m5R1Xq
zvq0k`J36Kh77hN4&7j(SlUfs0mVrE;#IC8q4jO7+)MRD74viyCPh(A+Q)rS+JRF5o
zy!`Zy8q{LznEE7e%=v!spv~c*^`YZGwd=cv82Nd<F^DgC^80g&3eQ6X7s9!7)0><~
zx%5|04?A~*eFO2WjV<%33pxcQOS<|2cG_Fg{EK<(U7R!`Xru3QjTg*7YGcp?DeB=J
zJI`P{ZT{|lXtE&F*UbK*gLEaeur%HYKJQfIyi#XFzB1>43%N6I&OugRL!Tz%oaok9
zwBzrW`1#)4*G)(obS5=r7gqK>dtpQ6vm)j3E_vsZ-?;>mE4<9BhA(RYS0@u11DH*n
z7}o6*V%<K-BHz?+Z31!X+ugAXdwY0b15}N&JS$zH!B|nSXzvgb3;-E!P>t#$A|>Yh
zDM-xUDeXxOAy201#qo;3pYuPpL(3b2kQ|LP5otwdYjkYUP-?nwojW;^H#}<m46}+S
zsMO7RJ4nL|w<^x=wcFn8_vsE4)BD%W?`2x;ETgfqsBi-AhOs{L6tL7(0J3)cbl1Gk
z<^4k8l66UNZ?w0?&LuwD+c5pRQjhSIZg23wGMp7iDFEP#N9eW+2gD&zLqDIB3RcOB
z|DasdD!{gzH8nrap=Q#aB3SJAh(h>w@aP)+JH)|VBj831ik$=+HU2KUNzSb=+7Tb{
zA*6S&TU$rtQO`~fC%R0AT0USroB{3kGBnk`1fJ>q{_y=T&FKhF9MWn`jjFWpLj>UQ
zA<k+*dF<N*piWT%6|7M{Yu_xy`uSbN!{y^s#9A<7tAJuCgjyu5$rF7%zdl0b$VgX1
zh-hT<10trGGfJ{8I2gRnrT2yENJLe1D?c+eXl_2PlCVTDM3}Rnbm$YE#VOgiPFKu*
z=SeC09Zc$4=x1fyXp%5DULgnswR3OS*;;t-l_80DI(fic{||voz{&PrNX~78C&1f>
zl`$uQzHdX$(K>eZDQUw3F7v2;jNTinRLNsZ4qHpz`dzn}daP}+6hL!QX>Lf+WH9()
zjrXygDX$n`ebiYdCl?s-WOEiM1>7)S-^x>7@tR%dJGQy=Po|(u8-{%@!{Y9dZb9R!
zu8IjufGcXzILW;3UPc>Q)YYL*!H;sB&MfFGb~MfaDTX*Y58PQioCv<;Tn3!y_6`Uq
zUdyWv<(lQ7?X<sj_-t#4-2Ubb&;)P81^8%ew-Q#1*G0pn24AZSl{?}SsMldG3>|Cg
z+&Vj~qcjPeI7$SLKFd&+#Vh5Zb@N7B<e{G<DZ3U7gbw=}IhNe4I^Jxbsa{ui>z=A8
z7Ftm@invYIB{4Ho%->Fz{kr1i^VmE6WdOAQNBXG*wn`jai#w84c2+&3%aMu-J_(z_
zmRNLYQ-+&+so4+U4xO>93hV2w39LIie(DjjFYV`e{eND~p>9FsuvUrSIec}>i%Y<*
zFJkUzJAc#KSLeRO0)LVbP!)E6*OiOKl(iz=4bED^ZhPC}iO^Qy88Q}CONB+Xm#Qze
zp3)Tl6{m7OAWou;{{==~4D1{Om1L<oo2PdLZxQ0%01;?nHc|%ie3X*pnu_mal>I)d
z-W<FGBE?hgAE+tp6;4X2IYs?Aa;1RO*ArhP<*%B22A+itpbiC7iDAIsJ3MBK0p2_?
z7b?ogUtNPvbUfT=qeAqRabz}BdEn#hDa=;zI&3dEIoqrqw2kZ>V^&T)JhxAui3N*(
zAdx^AcT#*n1%6w(plw1?Im|;-L@ttU&`QDFOZ+l{3!~CvVEl+AYyH^uIfq*i3qU&e
z-JZHtpFgD`cwm7|A+E|5)1X9`uNs8p%S{_X0^fC+)-=c$pcM+y@eszSHSqpv|FoI1
z<JC*Q&rSC^m@BdF>t4)7G}9s}HW_gGn+*6okRPz-g_kB%Q#N62XJwzk0i1s2Wjq7B
z2WaCd#-Z?v`o<%~S#^QuB?f{i=u+lY1GUs9Xu+%%hQ1vERh;W*1hD9fvR#hXqdQJ!
zcE){#ldhAB7`*u$|2_SZKML|GdtC@g%A@*WGp<w-PZ4M6qCoS!Ians?0ZceFL~$Y)
zC={pZI+MC@Rn;c?_l`CgF(37Pw>9nFyaoq2MXo>Qlj<66A5o?Ox2pO9NU(j^Sdt2m
z;1VD~#1>fP<Dt!gbjTc9r)(8w-K5Je>Tywm0t4U#&01rcD~=yT#$-Pz)#r|=`=(cg
z;7R}iSTF%)<Hz<P?P@n_J{3&f2ILtpmI_4-qlDm|P7S(Pl#$u`r=*BSp>z+eg%`Nq
zPZSq}Et~;L_^}V#dd*~AR=YD^?mQoN2ePiU`TiY2vq9?=09`15&3%u=qR1sGU0vN6
zU0{;(FXa`Z0Q2AG0xTFh4NQloYar@;w%(EC3wVzMB~*&OOc&sTZPz8Te-B|$#+W%Q
zyXCg}O+4HkN&`T((5d~~RiFbE{||8RZ96Zp+eLd@vd|mgfK|8u)z$tqn$jd_A$^(Y
zd3felr-84SHeC-4Fgznd>$H#pCA@|~pmC=g;f({V!{~qJ6IwpJ$JPf=%bb3f)w#nE
zn)fHI)2Mf0h{-QOK%+oR4LB=uN(2s(e>CoL`9AX*_P%bO*ByetOa3G-|Ju9Rc7Mz5
z;n*-xj<6#~1>O~8jeA`N{J*C|_z7DHw(?*9w4A)~CJFG3w}k#;u<OIw%1mnDd#|`2
zTxA1IDX{zG9bgi}=g*wOjpfaLMrVn*!>7RK066$R?P4EU9-xlYJSuW^u|FTESgf{v
zoc1)dYiu1}DwLE`-U!I*885&zpW8v@{Spbatg`)!cdkH@3iNX6KYi-!bN}~;&*=YN
z(tqCm!0>NP;XiKxzw@~NFMqG<x-Zy5U*9NHcd|RBe+68zE!s9SGc&J0I56P9Kfnw3
zxR`f_557`Rj8lFqWAAGJ!1)$jr-ivh|3w1$&R<@m2I8Nor2T#MaP2=rizkg^9MJcI
z&5CYc#R;#iO?sV8g-dDf)qfcr`qHVgGM5w_p7ijeF5`B1>~WU2YM1$1?YH$w*!kh|
z+UKxuYCCgvYH>Mnk`r4-*}IQkXitcLkbmp=aWbs*_;Ejz{4Ytp43hjJzoxYhTH*cI
zM)cxmnLEJ)VU;QYb}(Q+AW+obxq)~BU?oSV(wy%!D}X-N;NWqbi3!y}DKBfF5{Jy<
z+08+RTWPyCuLp$OxSG9|&iofg6%yF6$B;XFS$LgIfXgpGB_ijNsDCMacSNqYmuA<M
zn##S9J$>x$!Rfr0b{sdUD93a?BFWYia$EdpqArDe6+0`Meq93~onGxd^C%gA>YTi5
zGa!`6`?j}7m*;M+&k(SZF4q8%N;0_5s7^>TEiFxRcJ?Z3Co(#7Emz<5{w8nRTwe_j
zZLs6%RC|AIQx!MlbZqv*JL(%?n1I2w0Fu`ah2|?MU|()Y+$03~wCeDzo5te)fs8m+
zbEC}Wy41-`;|2iseHp?x`JnoV5pQvS<fL%OmN@$+YR-x>7wmDV0c;x5yu6UozyMcd
zkv%`FJKx=3tU4c1KPoQn+@msD`1$vYGTqoSTK)HoQrI)TNiG`tLLyvU+stJc@i@!2
zTtxr;$i14IbMIzAn_2Gru)#2yqvRM>)EM=XxPdh49C*AtYWeeRW3szXyhR$*r<w{K
zoL*m^m>^n35m!P~QKmpT0rEK@edxDUXfDXzcN=&q_8L3e@e^w7Ao4kQhjyAVAwy6z
zO`m*p|77NJG=68~21#xT2aQD&-Nyk%?_?bw@9**4{51!=j<@~E?lLw@tuUXu<@iOa
zbRX6*6LH!mTU{&&Ue95CH-1lCP#U&7{(=!JMQ-fWg_s^=^P{G+H*t?B2*!*Z%s6pM
zQmh7QzZtO2GCf>giPPHqt^%;bIO2(^O5<6Df44iGq9(dRCY4&~>5nCe77L$GRTS%c
zV0|$Nq;#zm75`C+L2QQ5vtE9RYNt%wOVoBpt(`0_Efq|4c6K6@s_Fut(!p!49G;_j
zKivdGn=MPl$y;;0aI>kFPsFQY^6KDTWKYYdN@bsjADhrp&l<ji%#h%II3cx{QvPq^
z$IZb#EsJZ|UM}eSXMi<DLBKVGe9o{_P|xBvUYFYp6hru%YR_tG>gxD}J)IwJY;JCL
z4nV@GOIgA&1ZA^;Z22_x<st-ub!{zoL-3yLA093(t--nrFQ;sN584XJ+juqpK-pES
zxzC3G1mBq>j+Lg*v0+Vl`z=_MT5V=-it~ANwnZkVZEfAPrXn~QWFGjwok&l?3Pg<Z
zdM~?I^}Rc;>~~L=VQZh8TM?U`knnCL%ZmDA?<$$#Z=GNEVB9rekBFt4sd{r5myz|G
z)q}(bBMLX9m{|~g%@NPGyv(duwI6y(;0KE~1tm&aF1^BXJ>3644(SLxS0hFk*Gwt3
z8$l_KZ-5vG&sX1ijd&hja*2%N)_SyAek!qYjG(marbeGJgx7IHYMq}45j>I?yhU0T
zrBAzT9joVg@fJqpi30L88b9KkX5FB^7K4A!W$-@|u8#yHoc?KMGNtoNxX1F&Q1I7&
z97@>)Lvefu8<ykp63*SYw_Q7+9+|&iO;^K`q<yN!6XXJz!Fjb7MxiV6DLWb8zgokK
z)f)3$EJq)-u`zgsxN6~5?q+F|?8z>*x9*xdw7)-U?d<Fvpsa?q_4W0ssic9Z{^UK6
zD=|jEfA!GKQxr|Nud=RN`M*_>UhHgW6f%dEMnZ8&T$77FS35Vequ2D4Tw@a$7$cZ0
zz+&TgE}nKI=*7@O+Mw4hkmcNF+f2*6^L_=sKC`0P&DB*|&gzLGUPURVUCFBXrP9j?
z-QJ9wi!VY7)b2V_O6<k*I2p@Jfn<yILThZ85=F!5uew%cQI}I5BaDJS5TnI;Mb91d
zhaP!Evp-Jyy1c$-<aa${{^^+^H-k%f8RoK-2P{e^VPiO=&|Ov_E$VCew%wFHkkv^8
zoX+kDNW#ah;U?%=MR4}WaMkh`o7mdAH<P>#wS0Q-$nznl*1IQ5w9xz3M2_bS{a-n#
zS@M6Z!;D@yoP8S{D5hL54EchWr0qT?tk9O5w#hXn1CyfL{x!Ltvc5Js38P^2f-wyV
z8~-9an7N0f4TAjK^J+Aowc5!DF3xC6Otp<xoR~GQ3=Uqds;hMJVqClTVK(n&50#?>
z=}b)WSx8BElrSQ+QthF;TrZW*I&yE~Vypw4s4~6y&(D!$ph=)o+LiQC=U#69{8G~z
zUYJV0+XLsY`xwn%^(C1K!fXDrQ?><rRdb=wHU+EpS@#cDsai9xJz+uNY<zkd2KIoo
zffwvuY_WS-vK4MmsIqa>#5UCHWM{B7Q23SLKB?R1#nH4q87Ms<s1})P8<R-8-`y&G
z!-3@G(1Hcf2rf3JK$3rI;m%HZ|DgkazSA!3+Zt8-fQbI$mPPh$z6a~L1TMdqtJvaP
z1;l<hH1{~luk!~P=+j>?B8}>8sVMlDjd^>tmZD4d#tUtu->uGW;>?uP!RiN9&K_Bw
zRCpere#UH+tQ0DFm;$1ZPV~vTQY90HTe{rH{&AacYr=)|`)h>)&2Ja$TPFPTa}h74
zpE$OV`9<WCC>vMFx7>I4sJMsaIsN!0FN4|6WzIich`r>DGK}9+tN>V)Dk~WcbM4@)
zl_M3|rt)|sU*f}l*wL1qlE8R{1p#bzR7CqOw?OIDIn8Gbn^Tx~oTzQUB$P-P{cLBF
zvEfy;A2nNXQJf%+8gT4{2x)qARm5!brL?~u$QGUjl~~fdzrn`kwW$g(7a(pX{CYJZ
znyIycb?~>FFV+e--@rWV_5imS?(Uux96W{tQ3~&mIyy~wKXo{x)+Io>EBQeC8^O9u
zbu(9mmp#Utaa)tMp@=&=+%{Uj6S`^4xy?NlvzfK49|;`wCGBFqD+=)}t3I*;kC~Gq
zG*2yXp`0I%F3`bPBU}3!Sbga8G?xcf4r<yrmfq@c&%lM2jboQH`zVC&)Z4ONdbs=n
z&<L(4^mfIG?%}I}igtJx9oAAnAc>Di_SV3V`EH7tX-~9R6z`$h1Xv(eRkw|x<Uznt
zNw{vP!HkOGM)z%5lhH5}$xt#_y!L5z9xDxSu+K+qRqv<toP9wUAg*5t(&183fpjRX
zF8oa?wBn>`>@l9;@a{-Jy%D;c|3jz|h4z4ZWkqqqX<C%jPVl3KCN94Ze`v=(UsR`B
z$_TW_aL)#3w>;Qk9pGh^;M4SuE)DP^MZCSUXhs}kotuz5A;||#bX93NuZi+NE><J6
zg}U-eRzCRKK9W){Y`2{`&+wiX{vs4bD)xDz2x;=@sq<YKhKN0mevU=(PnHcI&-m~B
zxeO!1^f<sU9B^BkU!#WK?gPAYfZ_b3EB7F(l__C{s^LQfC2yV@83!1Tg=H|xd%fpP
zbULuySvA`r1`Z;1zjm{uyg1!eNkSCiSdf^+7}riJs2EixlG2ogVchl)S=NK<lTUs^
z_!!pkehlYHD<ZP3tf&wdsekl<hca&=Y}Hg-yN<Q$B276YDEx#r$o@nE@c@7uY@%rm
z?1T>F;kI1qTh(JrBD1!p%y|@-7HiFC?K6|C=3&BZ%ykbI&A6%r9(kW9f8H8yzik^u
zCYFh1_;H%I;V_wj_a#RUuTwz*Rngu0qjr$y0`0a{OEHXL?9Y@##1+e5=Et?1y8p={
zZP;_rF+@KUI8^He3k0YNt8v!iJ{2ka^ZQEx`{5f-+qbJb%`|__=JFbnFdE~AX`5oS
zwcl~_*6~dS_jj@nPcPyGEwA|#t^^RbU6!_!_RGg~i>7d84fs#>{iG*w3y{@jpYe`B
zd}oc6@m<v;+n<ooI%J}_S5UrNIUQbIL88($WqnUA`VsXJe(8#ZUslBU*3axa0cfD#
zVg?FQspv}5J%<_l2Fz7lv@%Q|*$m_8_xK(@&s<AZvw7F`GX8QaX+ysU&_#f_)n?dR
zY1qJU^h6heI<IS3n*_Q^s&-U6u+RrjF64TZWK7`sn}7moSd6W>0x5tP7N90SMu&0!
zTG-%&!kc?L*+oS^y$owI40N!x31ej`5ny2ngBMBKba*J&*4m-^*M@Ef)_W%x76xxi
zJg2lIHLSb9wb>&901#hU$OW!#l;!iK@1CMQA62X$rU(L=4Z_pJvyknyj~po(+1DQR
z$1!2f)AdT7$oMg^)kdpTW0w^wR6L1Z#g)GYlBxWro8)uh(%8LDT-<0i2+o`MINyKg
zLoVneAPeaon6ZE0+|EVo3hsRK-NdY`=-J*<A|kzAn31bVQ&|2{g($0(HgLmq5a^ST
z3dX#!3{%9)3l^1?u+jZhImTYJ)i<y}KD)2^E6|wEscRUoujRhr<jcR?OjZH_&)24R
z!tfBG$X|{~20J$ET|NY5RN%=v>}?~<rUnvkOTDK@;>cU~y#Imd&*`}5lhWd37_UpD
z^!v+=Giv?{4{lB!ro<tgxz34PTnBFpe|>r;k*(UO_f9c>wYbx~LMu#mmk)~#`T8<`
z>{jYagj!{*t3SEWOLUYnZ_-;PoZqWL#$B0}u0(j7#;6qrl`1$G=iUo6m*Y7Aam<ZT
zr$=#^hw{dq)XTf)eRXqu((S32Qqi3>?k&eHTMKt0bDESbm9+*31#Vs8*0FXgc@}rY
zW)$r?VI`98s7mpW3!LT=*@ajRGu!bp>avZl(BiBO;PX%${QbqiaO3o;s>kL%U#vPk
zT(rAfX9a;?A?|UdPuPJ0hs}vH^MbfWig;f1vfdWFRYa-3^SaA8Dh<G&q@o1i>Z5*N
zVH_53!4Sd8S-bPE;`jj8CyD-*-JCs7um;odm<wSt{p|TxwL3I#w^^CW4rR!$uE5@O
zT9EjZsEE=#q@gy4_aVz-wtT0wXT%|icVzzT;4D)0us@ggD#kJB0;ygmXpAr2uC5@s
zTM;^#6g(|O-gda8x1<vFhFqboN_1Eq^`dlT%RkFtVk>VBuCU`als*U(p4xgSwl(SE
z5Y$X@XRhW|2Dc(B1KMQCIYtigjo~QmfrP4D#F5&>pwTIyLv?w9<+)fN4`MqmIIEqQ
zhLpJ*qII8Fjv))TECB6U+ShanG?tLz4iLH6mHXl&M=@A`z8zkhRpOyNP|5)u=<^=y
zFi2%G5*`g6*Te$RGAUnbDyoYRLr@AhiluO5C+zCz5(GP3>n;q&(lS82mH;^b1kBj+
zin2&SB=;hkf~nKPQ!FB++gwm{N#32I^DvFMV9GHM4a%tp(l!{XvLOl}7a($1o0>)H
zEmg(X8gvdRksV2nD%#`733WEHDw5N)*6!Yx?#%lnD4(1%(L8Xdt<ZesP`}p@c-9fv
zQ7CFa2$d;aDe%wAnAkD{_S9oeI~kz@1#%i}N)sIHZeU=K(<AI;*?6jRPNpvFFQ2rd
zJ=IG>$(j4Q?Cxyo6_-nDRGu3%n{jV;0^pD0;VP^E9tVg{sg=F8DP;a>15xX=XMR^n
z0W-I+WqiA73}x)S{O(}~pyhP(y)aP@TPt~%^tawJodi2~K~9_zy~7|SEHtr2fpg?K
zviU5;oslrnd&<UzTZbs!D#gF9j=1S`EouK~3hsHN8lRAG+rUl0O)-3PbNlKQOO6YM
zic^2c|HJ!LdfC~%#oQ*k#3_kEOhiom)pjqe-9h2(pV5$vfX#ul$?Dk1>Q8x9m+E$5
zcP+=C$ckPN&aOYgG<uvB+DMHYe6!`M!zFZB6qp%aeE%mo3IGu-yvmRPT~F4+?RPJ?
z8HGog>ns+{%&G()ltta5laQgjDKk%QGjgY&AtHyH@feSeS>SYeSFqQ9n8j>7L>a2_
z{ChZs_qh|<nY`n5btVsFh;9Wm4RePyx1eN_poE4nO%<%c(@*L)ioZLP_Jrnbll-|J
zUcNZ?2|!{;e|xV{pWN*BmM1!qvtm$LG`4!7sW=xyXI|qUO{Y1QL|s}U;&pz4T!sO(
zwgx28ONWkse!3n#-%b5#O=f(aL!_i#PVafsT%G24p8i+YN1aZj;E(QW^?~~`q<x0R
z4SDFQ$jC8J$QN|XXquaJ>h#y?%YD1CZrNeA%=uro@8svapLhqbtbnv?N|#R<)k~DU
zX#MWMQg@e}4D!V==9zmSP_G1BkXFci;HMqfw6Ny539h<vpm9ZLhN8{oBliPLvy`HK
zh^W&kFYq}ZSsH9+s+98tEuyEZOVP~4K-7IX>cI&ncx_NrE=Q*scU`us<+|wnp;(;e
zru1niIxz26+U@Rl&3BY=!}xp=!qgBcXY6kJzE&^(V~uJX$$t3>JXw`#8k`+O#6Kw*
zek@m_#Ye7I9M<;Y+hY8!pY`0Z#>jem{}<V!C$@wJ;&o#+0q45IdF|>OP_Lrsgu3fX
z;GS_CqA$oAhD`nhz`8UwRdqASbR;N63U_{gSxb+$Js2V5TT+CGXJ9w>6B`yYJ+gGO
za{x9<o;IwN!wvvD#Btx~%y(>CW{7-&>LR~+N$C)9f9yju6BBbwZ?5Iq+gXYEmh@f&
z>+PN4w?Z?+ERv#DR*PJDe8|c(cCtmbg0s`gp%zMhS(z$DDGURHeJ8yPgO7VfuB=`l
z4La2AC)!4aEu0<&s2;N4eI{%%<x==k!kF<8k9yqU;}b54d+WR3q^wL-k83<9I;Wg&
zH7o#21-0y}etG{})DUghG3}|;Rg!X?C#I@yjyuH`I8-3`DoNPZ^iRuu@2!8a!}9f&
zery>zPk7OfdC}svjY6lpHTpzE=%Mm(^Hi$~`|>*Vqm2A*&U4g$xKOS}w4dugrrMn0
z)wu3xAWVzba`I^omu!yddp;DeS|^F7otHTTa9V<e4R+IT22~1X1lFQo-VbAV`!TxT
z?dZ9Gz362<L2y+=eBtA;F3qn|w-}%Mm^E&840SMXWcGcS(cG$=V=+@HRyENU#c~7+
zXqmp)*B8lPFy62tP$?<c=I|}7Z!hN$5PZ_CeHTOY5E+QYZH(^0&g(2)sr5jo8IZ$2
zuGhT)hIy<D1bkq)IzUglZD2(M#=Y@?@8tZv?Ds7aO|){5?Funy%SuT2Ra|^w-)$FV
zk&Ul;T>8R&`KL_KQ|^_S*9HWto3sh^ZAx}sIVq3HS2@x&0%W~3Lk{Ii^!Ug>0;<=N
zHkkd-PZfr8q?+pOi`GSI4tq}KiXYdAj)N=-Ky(GLPcuerVEtrvv)qZ|t(V;U<tDZ<
zz>X9`ST9c)5Wo|Ak5&+($>-!L1rk8x7*M-RY<j{bKCeUZq??J;E}8ODtt{b29OE_$
z1>p8qR;s3lV2@CXgv~>BIQ@Bb$(Ul7*7JvukU2m9EV-(&&y1b#|NL-!Y3ifFXK*81
zVmXCvBEHR)&VrK}wSg5Zs62l>{+T2jn7XY4$R6ra6f{)BOScI7*Meoo;3e6BqPVYh
zVRKcVAX=;?Ghx+4#H(?zJ>onrMhomblMJMmD(y6La-MhU2wdI6qK>42WOPk7g%Kz8
zN|M8bft|uU_mKqOmgNH+d;;BZBY~*DLMi89;B&<7m-_M`5cYCbfw8QBbF3wh-U>Gi
zsPbj8iv4vUz1OErxIQga>GMl@V=x~-P(z+6IC_TQdk0HhhMC+-NI9>k210aotO=E{
z8<%La-rkH*NQ+-U*2}7~mGTz8q+yW#yJceyTE;O?)P&=0$?I()HB}{l==W8A;~HW6
zsU1xxHX6)Clku@!Mq`^opuAhy-bQ&tBS1p@009;b&=8ieb}kGDZ7=D4z7rObj|q$}
zT`uQuu1qthaJII#+8sb5fl6nRpC=7zOy^!?NqMUYUh;W89OZ3liZMYEixY>HCryLN
zd2tNit1#SuOk}KgF4}s-=tWl}GB!UCQnS{SN_IrE{8slK!|3&Ljd;rpzp2CnP`F@H
zfw8tA(0sbxw3EqRgU9*ig$l6ht24R-Rs!tZmqAf$f6dxNVUqgwZudDNyqTr_bLyL-
z&b`Stuek;(Y{o$>RsZZ@C624%#h>EhVxS3910Q^^h*oKTVPZ!=(KCyj&|CEN0HSXv
z<vvj7BOcYmrPG;C{IK$`vGEsfBIbNbDk?52eqFTlKB49<{LO#X!y1`L9Ck-F1FWSI
zz0KG@dyqlC8S{!F#xxVBClv6_^|F_#GOw*obiU>7-B-V6O^k|!{Nby>*78NbKJqN8
z#_S5VUIOV3Eoy(}AHseGo6E{-65-be$C1tFyKD!^t8lFHG7K(buTo|+0pK2-bYW+|
zCb>OUb}w;+tsJe-<E<b^#8GpMt?mK@1WzDzrr6EhAsv;X@~2?y-s+yq;q{iAOIuzy
zX>!|gb}I7FF5adY{kwjSr%~X<s!n?U`njaikYSmc&ii2j=507KcK`q-i><q?@!sM-
z{dVSy-sW$bOCU@TzNMo=gSWT(>*WDHY9e&9mX{X0$Y(X|WdC6;O*3Y8b4BIo1G$TN
zyXoyNum%{}P}}cyH5Entxlnf23;6t}gX@xn{OE7pOJJHy8X5TjoK4*!sOf#UF#^;;
z&(8pIL#<%x;NYOOf4O`rkXSt4-2rK5<G^n4+lD@P?SX~S3+CLY5?=y5irsH*CkCU$
z;(X9EnxR7<5$G2cbHbrK*2sF|upg=kU;zOAD!pddR4yFk3W?C`ZBycB5cTmRwwMOO
zO}o{bwyLAwsZmSPTM}KnMX}Ecem;=NzK=m#yv1_&!88WoF6i&Z74=n)*%fX@@a{(j
z3^6!pR4d|!#ak4NKEYldD)#0&`+i~PbM$39E{b9=f$zr??#Nia?{VP${j)q%fG}9n
zeupJafB+W<Cws1-MREevZ(~Ts-xWk;d`bV%)8S81NR3y1vYd4CG0?X9sNJo48z|2-
zN4wmKA^%KyD*5qYn^Nn`F@>lDg83b*$a0XZgJi)&Iw`m3ccoYJ-hgh?jF%$UNZ>Kd
zc_bzw6eCnQwKDH-#vT0RcS~1d)x=n{Z$noiDwnT2CoQ!Ff8sy!6U5#NEW><L{go1?
zts36=HIB=JC!@VQh>TAmgNLOH1iDWw`}U1WX7}#Pqgz<#@qzf-dC0+*K6x`O?_1yl
zU7;u`BrQ6Y<<mZ!!OlLSMPsy}=FwCAaI+B|@las2d60n$wGGg~+AaeXQ4O+IL~pdi
zKiXA8D#e@54}iFlB`{kcfot}Pw6DaJ^NU8+a!TV{imF;<djr8%7hn2T;d?>(vufxS
z|B<zbtoDhD1$oztQlrJZM^*UGX#CE&VabuR9_+Z`6cMf`ks@#v&*7+}mHZT?%#thG
zYN%WpMtv#A+VGi(<N4L@ys+A`%oE&&&x19O8gK}iiJ95x-)2Vz4b!JM$xH@O&dy}N
zrPp9?VP0??p42<Pbx8C{JJR@K#x3>JO}BLVx`L1)_d3;2@M{xS&@km_1>^T`(KM85
z@B$CL5#H*0;=+v79IWkr+K;dn;`DZzPK2h2gR$E3`wb3I(tsE{FW{pJ``gI0@M-G$
z2{oLPU#E~!Jq}_HzDn3PAjf3Yg)n}km%niT-fb_EHng|rO})fUk!t1nrb58>QJ<0n
zdFeR)x=$E!EUM7J`m0#t`#pPR9C5DG*`C&oPHth@vo^3wRY}p<&dV6@yYgp$Qc4Vs
zNka#Xpd%yviTCLJE?>hU1J(trXK%{W{9r0id?&Q=C!E$50vh{x<O^N%Uex?Kvqg`N
zugKDJgYF1Lm;AtG0G1DGQQDScm?kvP6gEPu3SK(R_GMXne61l)TMct>EZaeRt6lwg
zg$?oPqFf6g(6qh+UfAzO>Qr3?)DWLc)aI(D!uOi2D=EG2ynhIZIF}ndzm=UT!}zY&
zp7+G`3TZscWtL0e<E+4uwhl|2a{0yMi|nU$zicAGPfpQ=$>;WVqv%GtWOJP3Oxp`F
zuCNK|s4D>{KLz}ZQ`ylc)ma5j#wjLEP!gsmy|7II@Ss{d=M4BTr(6LK&%zuhfA<(u
z?b1Ip|Ae-N>8-(OA-|=`uGZiTwIoNF*1M<%gZ#5rD|z}m$$7{9gKCU4H9ZIMu!b6_
zXMqk9?;2bu7nB%%+a->jkO1pfg_Vw#<Rpi(?yoIZ++3!zdX=O}T$8b$!~*bdTaV@c
z1r6HImpB%!-d@q)-!QakfxlVmb9C^<#i<90JM$#sK2-o0XA=~^izB>vut#z%yX6L<
zYM%#fKW4FndBBURVfoKa3%h$uKh6@=H`JtV^#+aK?_@Y!|6X8wMRP!V;Z5TkG>(;i
zj9jAETMn@Xwp<<Tn&uR%m^f!#9lr(R*S<HnGS{l`MX>K|8-rU_^|jI-<Eh=ay5nJ9
zuoDtlm6$ujX^FNUJbiHh9-rkp#gNg<d#gIfVlEt@=FTIhzc*@3OCW5`v+QKso_l(*
zX7BA+uX8`Ry6;62i%Z2x0?dNa)$1c}S7#|2Xhs>KCoJ7lsEqq&PHx%=h%oaY%DT!i
zjtj%r$b(On97s5#*-5YBpx@~He(Ma1<`$Zow(H-p{2Id9!10OgGL{6%vaa3^$ZN&(
zDx6Cm!NSTp*8bYi#I&Fv6B@!H4?7E9$?2J(WTO)A@uo$$S6JCutf6{-a5{ALlmA}*
z{B7ims?p83$%zN4!{_vyccDu61j-apuwB4dhAeY5GUTM4gSt%YBU;_cDF;efa+}|U
z!4S5IX!H(n;TcZ;9#%=-&FJ#0?P|BBO`NYYH}i!3`O*DnHTg<uWmlr7;b8FdM2hw=
zQGH%ko3xMh$!C8~C-#>I&rA)43$~!e)xr(LZF={B3v2;=IW`&e{x@Y?xmypya`-@C
zcOh#~k^iaw!a$d4XSeX6Mzsqb;l*`M8TAW`MEP%N+fvZ#{jq<EW3Q;G0f=G`<4q8#
zK9eF^9hUI4&~xwwN-AG*FbEgZozI_}wUXwbr0C^d+j{b=<jlvc+-=k1wj%z3Vr0{A
z%8B%c+m$U1^WA}ZMp?m6Qs$qNC-yeuZuf(uU0*0`!1V9I;fsb=Bu=lbo>|R%!5`a@
z3F;@UhO2f;yk7U#+oXLeeQ9sKkkvz7NS!q0hBOWj`l+9Zwlft*@a_>|*&L_6?vd{~
z(LVelZPQjB&<k+6B_KQ+U4gs>GxDfjukC))zeXv0l=KHXW5K6;z5m>#%`4pcd}3Uh
zt5tCPj4kEI;ENdB5e+Mwi7QvbU!`&pPiUHNNmsj!%AG%L)AnqkEpa(4QQ!9)ZoA%4
z%KG05g<dt=X9eaa49%MZ;Aqbm$_`qyVp<IsO|M}Oq<96XK0p}@S8EmpivrjZ-~Uhy
zT`@=P7oV7v=6^W<SyX&nKc^t;-DG_g2qBNhZWJiD^lri@X19m2-11=kz{gXY5jn{N
ztovO{6`^*DY_aoX*-g3Eg;?7*tZDO!RV5thk5Gl~Fuy}`EH}(K+%E43PN7*nXJWcA
zbk5&WWjdp!VAJE$94P7cwB*Lk()7P|Q>k&o>()?G4x_C8beP46zmq^>@ymiI`~l+4
z0VW5m7t>+KXUiBZ<}IxmvDwCp)Yq_!-Io;FA|lkYU+Kd7t>^sTZONrpJ(gEvU1s&a
zLR$&bXQCFp45>aMNLi2T=Bza@roCKdQvX#xkqvQ%=^tKx#sg9FQ<)Gaw<0Ff<-3gu
zCtrdeh8W-duT>gjcZI@;%}Y}q>PPp3u`up}d?wrkPhY~?L!Bo%e5uodhkTD<)gox%
z-CF)$qb?Y-?DU+2GQ@;+djfb|lR$!yUeP3Lg%TbjHyvQkW;$$Z<WhiA1vyHkn4wR{
zd0M0QMcSRR@9+#YJcyT|KH7#my0we8LLw|p@xs~#Gv@w~pIq3&D<+S<UyrSKhf+N-
zw%F{eTHyP(`}*Ah7yQiW+DJe?yR$?8XpaQ{F1lBaajUFU2<KV-;`MckI&GH}z4CeH
zoRH#)e!YD$c1I`1d~iSupL~(p)o%MlKbCqwA8{~8W#8*!E$qx5=WYzu`ZPT91p?QV
zLViACW`$AqO*~3!c9LQ`6}!|q)?41?#H%Ms2&lR9HO?uB|50aDr-i4^kzg|rU)|cl
zDshujn>2f1=~}bI_};xPfzw?xuELFuwDzj9uazDcVeLlD%u%A+=<Qz-a3c7iRzLiW
zm*?;_x=QC_dZSsXJhnuoYFTLra=5-dS-Rc)Hy1J&e)vU#x~*KT2eCccCm=i<U2%F#
z&%q;>ec!yZJgao(o}`f;z*~zfmbvr?z9T!vx(wRD#rDLK$j?<;&40S^{xSC*snzg0
zv7U1&)7s^<X!Sc7QoO|Y2Cg?WnP0LMZ)Rg!UEqd{JS^^|i$8=@`j(-0?ny1r^t(ay
ziXR8dQgka-p6+D6__4&7`s#7vgq!C=wv~Q+R5tGyX>JSGluNG~<|Z?k-1jht5ia(q
zD?;!Abi_>UcZtQRq3Tr?SJT%@FU*BjBW7C3rq;uwWQB5Z7u{apkAcC+)e1Z5!$g~y
z-zq)oj)I7od6<<(2o3OF1|C#7Hm6A~PlOfAq*3!G#r;$D9lM8OH079^VXt3V2JSRT
z4U~>8%kwAcFC2B6*5RZnyb`Q!8mPUUfW@=6vxAe{PS7dfFSk)HS1+$K<jEPixCV>p
z)voFzNg5p;on-i&4w1Ak`OQi$+0?XDDt&#?&`Zk|Xj17^U}zG<*q&l|Y3ZJ~$0Nzb
zrDJG@X#L|oy@&xudpkRK2M6H-mfrVKo2aqnWlZivN#ySCz>%GLV{)wtqipzen8I|}
z6ExS<`(4h{3k^mH<uCJt^NYoHccWnE$rIes2<YC#(m>2yimnwn0e^%ZtyAOZ#@`5^
zHKg%#DNFUt<A<JY8#s=P?ga{33;4ZO*9){_&gkq`3Z@*Yjaa`PPqUp<5PW4Bv2*E2
z;m|iTaZB%=2Y97yN6-8PRv*kq0DS;xBY!mXL0HYotL47CTvQqOf6WF23JW;-GKLJu
zz<vgJm^Jh|2LKI+jZDdM!HX!9M4toJ+ut?1xw<d%RtbM`X^GYkg-T-87bhhQr1AUO
zzAQlNItzZQ1%S*}ChSlWAn89Hrd7apIXlAfX__Kju`0}{03cj|(o1otYDO*&%mi%x
zh}-Pz&Qb9FvN4H><s3Cpuv)*ahmx<xSIP@s>r&yrH;VEh|336G)&nloy1@H4AdaCG
z|Fg{4))N4-^rHgJXHTjqc8*M8lDT7nR`WFC_%v_V+kcei##N)?-<w0-HB+$zXTTg0
z19IjvvY{|V@$FwgN`&>=0J(iRbb4Ur<#c*72O*^JWqp!x)zo;EYIN?9ICs&2sVvwy
z)(ZaaYsESe!k{1XH@<;X6g<?*xGF9JF0myzF;=s`EzQH3Pv@^;=N62DwjtKlI-BmO
z!ixR|dU;VDx5S*@L&Sl-3pZR-bh`g$+t#n4aBV8PbE}=1>yQZ=7xPg2nCRj5!Y14(
ze!@#)%oX_Nbn5?TG4NtI6VQ=4p}CiOMW%ulJ=NY4h#E?kI>G(|H|FBWU_@wfsREZK
z<lnUb_t8XZGCik)28m+Fa(v*I=9s@{n&l?-hT8K?o->(WezXrU2~JjS?6cKUeIA+n
zTJ;B#cODBr0J5MHVK#Q5+C|@Eyt>YUy|6w8pbbcIj4M|5ZB<5vO+Y+7&z0_29Nla;
zxe9ezq+r8-2IQj^G(?`hDFpmEOpT2v0Qrqf^>(o<_G9*BpH4+uacQfnza<4?SHbGb
zpB~NnUYCi!E!Q6ekv|iGLKpfakyreQMO-P^F@Q6PznbJjjJtfrqiy>qO<?*f9f4HE
zk;MZWon5a?mvSw1y2tPPy<_L7IEv8gP6cR~7&+4vuzrhLO_V$m``zbQSOt(6J{0&4
z>Kt?CCsC$ssgNS|V7`}1(xh0o>6O3~Qz#?gc{1lMfx)IP06S7aLj}jjAA}wPFVao_
z*S3hcnkN<k{==)WsVU&M4&Y$iul>thD%_)vDs;T(+LME9M>|YiM^_M-_O-E~%ojRy
zFqn>hIIgG1^_bhg0=hoc-yg#^SCyWB5B=ZoGTZ}7b^l-5;QoK_{r@I|sw<rPye+zd
Uw{(H#GYBLrsqnT${L{Dp1KB9EMF0Q*

literal 266446
zcmeFZ^<PwN)HgZ?f*_%QG>CM!bVv&*AT4E3(%q?qw19NC<j~E4bc1vc-7pL_Fu-s&
zxA*Pyyzd`yKIexseu&vzv#)inUf;F0!76X$urWz7K_C$JD|zX+Akb4^5a>ZG#$Diw
zNqn~?@Z+Jayp971#Hw)nb;rfs@ihqa4D?F+rMhd%-h$8f*mXY!<mRDsibR_{X7mKw
z1LT94hNKz25+--ra`S@OzVNO-dY+!=jLf9o>NQW9cnlC<_}l~DDaJgv44=xIm~_0~
z++#q2Wen!i7+{MX8802se~>T1jqL}qXQ=o%%Nu;QJTJa@Ft&*Ri{&f3qz~sa?xY}^
zV8<db5APGV;Y$%HLzF`9Ey@4KtE%wxp!NT8<@OD43<&4{zW6Tb|G)PC3J>1tR+52Q
z?^oMEAtwBHZpC1x$`AE)s7wZdMwF;L4__kBQUo26rk`XOUp%{;)Sr(xq9+#MyK%&H
zpEq8a8kEub+QRwX|1H0a#Mz_;->0=c`YU;*Um{sAwmocY^mvB`@ARla-RLio$l1BM
zUF~Nj8DKDn|5B_Jw&#Y*y>!9x0aF~l!$c`7&n1<6R099JC0;chf_Po-%#!ETt!T|M
zR~z+4<4K;+PQ@8sH^V$PyJXmh{-OMqh^jI`#Uw1|ZWMlt`JuzSJa1fIp#7olUH>m_
zt*K_jqT_@{a0fr6-1?^1g#Xsk{(1Rs;I;miW^l1&<73d8VtFF)=kog|a*=7LF^yoG
zJ-oKR+l+j<Z_7K}J9S3RTL0z#zv@eDiVkbuA=}l<|H}_EzPkcRd6p<y41L|#;C~5O
z-FrXv8fu&QSL-vXF_z5YFpu~p1?o2WC@%lonW71S^O*e4JK%eXG)OLH&*y6K=z{D&
z@-hT*7BA0HCbfS_^R=V#%@#as8P%8xL7x7(ETJ(8Nn?fqXZL`erkCo!%}OaP5%ypI
zOJ^149rD+zb^x83yMzz_-BQIc^eNjr$JMS>oBYMg&%f~^zIqtMwEo>`<T=0uDZc+j
zjNMbKD;+q---ChP1%M*pF81A%UoOZLCrjDpE5R1KI6M~rlJ$udIdt^=AvrqK%f>VL
zufLF}#8@JBVKAI?C>uGV>%Upy2szmk23YeSWh2C)kfZ8<750`tzV!I5T#F%_HlC)c
zzw4`MZT~Q>wNBUmRAbQk@0~%bQlM@Dme`iZlJNTx9#l~1i1WX+6bZSlz_7#*tHG?K
z(6~RJ2fmN6S^)|bI&*X0On1b;0(=Pf>AT$B*lS^%rt-}2KOFRZs=h~RpnWMiwFxqd
z5Omn!ZG0$Ebo}xk6UnIe#Jn@Ou7+X8o-PTJ2n8rDy%mq`g*+<oukI)_xOlGSJ)b)C
zMiP|{-TCC)9T=aqa6cpWdaqYkEQHeBnZ<X~$?Yf;d-o>Nr1q|vzOpka`Q`!&ncvd{
zmF(Uii(ulOh~q(Pw2=nXQQ{nGU&z!*V(e;S$Y*Si9wZT-$Jcm=E!Ptzo6@T4x?8Gs
z#-F3{=a?hrZ_-5E>~A8Rpz?Z7Hkr0pcyU)Jres8fp7;M_gl~a7z_~(N9~yrdvf5&{
zCa<YU3I?(f);|jp6dJ4*emG5NXz4qRHR}ho^Gj-9BIj24|NA`XT?i<`&H+#b1whX+
z_`SjMce{Uyf?qBDS>Y@DuHx(z;k-Bv9Ou6mC*;0uSbPk4pGm|=)8K!&Wfn+h;e$ka
z*ua{&5J-yumV?yNfIZ&6M#&cZBV5B_lYf<!1_yPWrnM-{^n~ia#TjT1mx!q*)@^>s
zFJb6|<(K~a!8cIAwQM)9s8qr8(}FkC|FS+AP<P!lpavT2Hh&lA3Euh!$$xzXaEIGd
z!pm-AOWSY>2BD0Jt>sQZrQUyTJ)(l%Xx(6*YUcUpJ)pLl)V`m{$D><Rnf+Fm(x(2?
zTZs=AhRU&Uz*=uj7TkgwyEWRvoYG|Je>EGpLlwpy@^CHP|HYLsnOlRc+Ug?Xx1PUn
z-f5olc_$Jebszn|{UkyGA7^&!<Dy$0<1Bh;Wc^4Eu_6XsNS*h;M5>^NTF;-K?RI9t
zrM6}pqy)tN`*7<xgKze)H;|gm_cVR`?lO1(z3q5+xNxYPEx9M68E*O4oq}Fd`(~Gu
z#<o!V{p)3Rwz^=1ua|6%|GI|^G{EouypDv<Wn6{hHPjpk%>1nQBe9H5J*U&?|83zL
zk6VAsUm9+kzu!hjG|~TeSJHi$avl5M6LL;e18T0O^Xd53R>R4N_^NO~zD(xp*P+P^
zx5`YPxDZKU?Tqgm=nh->+DVC=)zXXk>jx8dZmmR3JNp>$#jdc`OESHGop#liF2ni1
z>)t=dZ~e`9%g9L?oBiE_!Y0HRU8ni>|8ONtH8|9Ry7+oAe07@OKNf<LBtddtQ#`GP
ze?7<5Sa@mk+a*qM>Caz%y9DZe7AO!;8|#SLqKm+!-Cch_HbM?Kp;iKo;j3@~ztzRB
z0a$zP#y3M`uj|o3oDUkQgy;9@l^j`MhI+R3QUEWM+&bTU^DS6f>y4pqLd>q}A{+_p
zpAhLb8n0&AS^N(<gWM%E*a=BH$t{oDvch#IpL!Hq>_)^)#AE!~T1|cGbugyhe`cLK
z#&6!#^q1GIIY6X&d?xN~qYtp_)}+n;Pm@OKULxOIko`Rd+e?t#y*;1YqG_pW<?nR=
ziLfj79Q(gDzLS93phy9W5CRU`jhND7|MREsroGBg{FZ<18Su+O_ie}ZzaH@`(A)Z#
zYW=kwFIJsPz%kh9jqjQKcT*6jH?Vl7mqM59$N%Ufsbz0;HMk|--D5L+^ndv&MIa^O
z?EFVaA0xy0d}!V*Z81l~epfqn8_t_O{g<Af1JQJfvew@F7KV+;9hldqFh9F}^WSr1
z$UI#736>#)zOk~@gYzFK<vPD8b^Sw70;`_TUrox0xYfA7l>Yn6h|=Bv@hSgaGyDAi
zU;F<)4*)|2<W<!`CGKK~u^ew2ksI&nhT^04r&_p($*cbS#qJy6_w&9f%Z|PfIkIA)
z`5)1%#R5f-FH`A3FIpDf4&%-Ht`$0Q?*oWon&8>`#S|~nOnGgyYL6NdKRcCEjQ*&R
z5cEF8j{AR-n-TN-KzLAc@9g^2y(6IAvz0CTEz)2^?WwT)%E)j0py*zsE1!)cm6mK#
zrnIq-L0bTMop<O4%?QG3U%{=>KnWj@TD&?Lo?#1-MAba_OWBCtt>)2}kFJ(WVSFV*
zH&II0>0Z05is5v!x_Q9jH4@#smOGyN{gEocI;%LaSJ*aBe;gGvaU)+;`q*d?qv+9o
z_wa(L1Zm0p=PkN-63ia0266R?khpCw@H<datoF)Z;e?M7@*F>1Rcje=5dCcNcW=<@
zD-b%UbxqS!<m{g#tPKngGYil*2L6&%6Ttk?P6Lgw;dNK}xWU=;EA(Ek5qC(|7rE2h
zlVz7Zp?=KKOa6t=)$SRpW%@UJ>OeGQqsE29yKBhg%lRgS$&?%pk)QJfsQV8x4w$c)
z1yI;uU**^_&~#?;zb#N_&Uq_UApP;2ob!4?d6x7qO|5m0fh?F)Czd#pB9b5^%j|GP
zV%twB@oRv>K(y$qY`l<tRjzQdAKI40Mi|uI5t3(zg?^D7Id2PUYHGfVGO|1kvU}v|
zxx5uD)fwSjZWMPiT~e4frl4st!n$i5l;{O03}Fo(eZec_!$%VMvDeJgVY0v0#xi1}
zPUuw$IvHH-#6Agm+X34O`EgM!ox9u|o>AQF@&kgI*;<nSOQ^3Vz3)BH`?rNgB4>b1
z&^d5^^eoBaK4lhGbS|g%?8LOZ)l=xw($eMTrF|TcrNM?bpS8-zAubX36%~<LX77m6
z_9Dq!nZGUg?G`DO`6vQUl0XLFvQMM}ZDkJG_r*f?;m==9big~+ye|!0M+9QF@QHot
z3lEDjTIuQQ9^_tXl)hx3Nn|+Mc4!swqVdO*Q6(0<Bx@3|J?Y3i_vdPM<4#mU7kKep
z;K^UQL3e<>D@MCK%4wTgNU+@rGemAHHjPf0fjZ05kFYc`8}hYy2sPHr7;=Qz0=S5F
z%*t(149n;kZGSe#LzzF-OYW)Bc6(wm$oJzb&FFzAWasCSGtA7UaZYCBA9z9PKaap)
zxCIi*NPisgKkk5yd^FukG&Kbd7ORjexs&a8lV*^B(f&9be$jaRZ;lI)=PQeMD|sdj
zzK#aR2XZDSQV`b*=ym&?@mG<qEP$#NyVu8%DzmOI5;Ux^tgX?J+|ht>3zd)S0Ab3~
z%&~<{S6@CnH%9+>D`Ou7u8o#~;zJXi{Ou&MmqH*V##m26g6R(F7r(}?BhmzZKq%x4
zG}AP7Y-0qhX;9++-!TJ*(O1{galG6C?vad{cG(zh>VQD_BM14`6n&QqVj*IUJLsaP
zMISFzekZ_1ew%8=7rP_IYV(eC$tAyK)lfB+RUkL+teU*}Xi>Nn+s0^)r7QMx#r+f7
zlR1;1yHrnA3Frz7EcE6c@xFpB^2rV1ns9`&Cts)X_3vB{t{mQ5kbC15U3pj$FlBIT
z*1cjLI(t1@J~iYMeSE?D>8$!@C$1BM_ltnn)g4-Jb9C0d8DU6la5$l(#d>b)a|9-0
zt*aQ~^EyGH=cABR%4rbWgY%v&@KD3YU*ft|=b?rz?6taNoHLH7y&(rQZ?t!ZOBA{c
z*7_VaZfpGcSU*kz+;F{HP=gP;phKpbgh;;ztLuf~srtMFjZ`!Xg}ryRc_AO2ox#k=
zd#4MHli|4mAkZfozzQ;Qu?@%>r@NU9xf97(-{+5hUEAgBDU5Cn1=wgZM4JLW`1+*3
zMw~K$?<}Oy+&Mtm>*1$#V|~ZFpSS`T<k5Xe)lW549-!lo<R~ISm45=pq&}K+{+j;~
z9&mC)?XS+vmr|!{^iB0S?UP_>&&#(vZ+pZdlU3yIfqV~7N<?M8a7Fh*PaWAKXU~co
z@^Macx{zb2Z!6e1)c!{HKONkZzs*jDF8c;Elsjn#_j~rPskpjoew)p5mwrcEGx+A|
zb(uVU47G>|i+SVsS-sK}J4>Ds)}SdzN)px+iKCcfu=9cEeuafgNx=)!?6QRgVZ*`j
z#cy8rJ3Uc^auEY{&iXEL#xl)S#}mf^RHPz4YZF5>;6aG<6x-?N7o2xIt6&Q)+hQK<
zrrU5VKxos5E;tM1bi$|F``5Zt!S*#S>P#Hsu;;pXv`REY(PDcLzG~y;w8($6FkuLG
zSwrU(Gk_dt;)Y&hIwCU<=CV1iu71Iks~T!VM!gz(bac2cJ~Ss|hqWANC=*=!xo-uR
z5~|LH?iK}iF&@+mm9Fpw5Gp&$SXr?23^6V)EoU`oiacg!!31U8`I8~>8qT<WV+`qf
zi;7Hd<e+>N%QL`XP_Z0kLnUgr)qf0TnhtYy6Dj_AL>juUszFWUMDjzs`1_X_S|W=@
z;z*xt4eow-RNE*xq%iG3VYXF0cIcq@f+&pT0VqMe@HN81?2|%@;Xr~I&#~dUF^&N_
zag)Gzh1|CVS@(k9x(SjOd~f*CKwuM#5iBk5kM}?f!rA4i1On~A3f6PmSS2y2X@lYr
zB7A2_nvu=l@Y)VANy&kWg5z$x2My{E$9q#c+H<2>W$A*VbvJ3dG<AAU!UqaX^qKiI
zj_PVi%;K7hLUL?!=)?spvK(LuEJ9ZjrsXH$s$b=tOWn6^Wn>6nLFE<G#4EBIEmd1{
zY#WXbk1w{#NQRp2t}etq5GT_X?l&hs$Pbw(;Y!`Zkkj^dpUa_XmU~@u!`2*e-=gBK
z4$p$TUqtP=Uz7%g<k=Vzp7!*|^5Z1Pr|GCN4?OqBle9PM#2*f;&$^Y6BDANt&_c%4
zrGU>_b7UzhOLM=t$qfWL9h=;HAhB?2eREKAZ(Ym6#s(M5M!WtB8By;*2K`BTf<O#Y
zsEr{#5U7=<WSg{|-CacvV3lFrVIUp83s~`$c;$Xk`h0nD$k|lWGBIC~fiIq90+*Lb
zz$^S-ql#$TAiy4RJUuhpxUnMcD&R&B>v#jc&XH`ziQ6JLh)09(H++rMUL1(&*GAab
z*a3Keqa=u|55ZNe&@-$`suW~qYAWU$sxa8oh)V2Jb=I3%bMGq1e#QUNeQ*UYt53_?
z>=Ot?S{UuvVkeCDt;H(tr!^2s31oQkhKmm^Og%B<<Y`xq>jro#uC~IJ5+ms)pT{b6
zl92y;__*rnREsR8?X^WQSwz&Xp-RtK?V0+ygNOY^X+EO3!Iag>OK8&YObSDp4gAyQ
zxNJ>jv!rpVrObZs^zNs4*SZFTjs0$lSwDL28zIF4Mz2pfr8>b@L~oO^M7SYV50X|7
zlZK9Cn<s|BTVhu$Nu7LCClZ@cYOy7WB0feFilHtGZ-cVh7>q;i(tnLSqgX_A1z>85
zG{T@h2Q@ckOGmRs$R!&e*Cy!!xj?M$8cI|ferSug@qX-`)xC9Ls=^bkj5#W|i%nc_
zpXcLFQ&Lno8rr5u?Q>KQ#eK@!s{I~jc}(EPdg~Q$?Rnc>CnZ_D)6g-Oz13|^$3UVf
zBG0`<!ij6`@oMwUS^ZsR=v`0-<$c?c3M3LKX$Y;kCz|QCy>ZF!q?B{;;ktQ_YR>uD
z_`C(%A_Uw@<Mq!A9p}XVEVl>-%~2z7rPb`s$QGEV4(rffD{{_wGt(6K1uc$}ma1Es
zHl{b50cTyDltE1?@s5n(LmD79Q}V}FzI#r)q7xy-{U@9HGQ8B<O$6j>+==stJ+-TJ
zhG{;pk*~l_uU-e&{^S(OY#W7$+$8m;l}S_=ratq+Y?Ac<4Fajn5szK(?8>>IXi;q!
z{XE?Y45~wb>5GYSTvs0hxoUh&IdM8}U=)5UCRrRtaDx%kUA3fL!&09+t2U>l^FwW_
zx^q}8t3Sf?Z5H@`oT9M{I_@5D26Bd~%YFc2?oCEM@;PQlfV5Tm>-jOB`NDiHn8G#t
zrLL*{R>4GOT!U|ef@Ny#4Jv&0$%kWGJ(-Vldz?QmQP^{mj6XtEAq_ze7NJ=+Ww~bh
zy_qwr0_hGB!b5<5d~!iV{o13$0ym%iju0?j{Wl!Lry1lXPuLF*eiq~{nkpM}e7lX<
z7I7tEio2p$tx?5F;cFMBE<uf$KF98#P5rVaS;-9QB~*o}Fm@?R3MN{m4^i7K+e3Sq
ze7i|(hK*+pBzApWa<CM!qfz*$ZaGR3?<<=#FZi%cnyRiwzA&`6;TwFy4ccV9+~%kc
z^Re9l-@r?LGk(Xix$Yjr>i&Va6PC>u>6zdVIOg0*l%`L`VA$k2MMGtBE&t={aA@dI
zZ~52es`eDkLj5+GYSq#&;hL^mVMEBSp0kKt#bs&P#tf-0Ht#R1R5b65#cK99Gg}4|
ziI<myg{hfcoS~`11m`7;c8FuU&%%vb*v7TA7Xwzy2b;$sOn5|3s;lWHAI0N0*z78;
zUu&z^F@k&_JZZg?32Uvv&Zi$Dc%3~aJ5l3xvlve1j_D*eFzJ$E+V02CFl2pCCHdoO
z`6^r-NO%<@2{5TUDo^7|Zi9;;kPp*+^|a5z>%)mwuC+BYV~mm?xVkF7A{@<p5@YOo
zaT}Dvs$L9DF%ok;k@-C`G6IxDcx_c3RnpHN%RYG(7#O(nDB^SQ=L=;O&k^=q_Rg>1
z(f+YJRL1B#K`Z5~7q@!qvglRW7R#dQ<|xT(Am|f0+Rq?aBe@Yn8A>Mdc=koW>z@xt
z$rwaYm3v9}Yx^LHO7+tq@LsyAnH}7v>?1>4;^r)T+kl*fB%Nr@lP<KSzjM=C5Ka;-
z{XAm%rwVh<%Ai_-dNn*$0?Ymy>wSvIX=5=sOfZue(NxClzo5)Eyycog?#}E_HLg8(
zFi`lUn)c(KH}fTnwRcFjs3=egC_`ssW^{!v8uh7*3p2$syf^$)5H@oG;~T{)-Qyd_
z)c$6>#RD-jU+Reo6H?W5PvSTGCt?S%hKM&`+#4)MR1bKDb6z&?#v2<MuzEWVu8cIg
zOhg@T*Iq0|HD`X(!R8aSr7g-MnJVqaJlfK}0y50Q(ILXku~0F${oJynxVu&cPGY;L
zHj63IBtvyrB@4H?$njeYG{7SUjQMz7&2Kfk@r&D?eqoV=8G0{#@o`%qniGc3BDxNv
z9TxK6Ks`^tyemXzz{0~kpbZB#rw-iheC#duLsL`3I(6p@o3&{a298%Z!l$S&VACq}
zgfeeup-Zg{G~DnGZGXHrM7kHJ|4{mZH~T8>_WZD9U3P)f_1@WDWkn6kq)2k&SUKu)
zIB-5Ikn-X3ucUW?4)e@p5~<GN@_c2U+<QmW24t;9*AmxC7%o#PV}<iclO||x#k!R&
z{v$Kh&npXu@a!x6?A1i<f%HN9_zH<9i6~^K@XlaYApl88JzRQ^X~8kkSlZHZvzPY!
za|P0!a?aygeJgaT(W?sa!Q*n2Oq0FEdvU&g`#QwM1K|g8O@*Xk&uz9e6x0Z~^=(g|
z>=Y#@EnvLIWD%wmru4c^Ul~IbXo=LtBsp*bf3jg_o@_x}&sCmB#nMIT*btK}@SgpW
zfg7mT*<*iwa;!$YvaA7L>w=5rA)Z^m5?O#Bj&}!l9JWaVL6S~#1xR~QKMR*fC0B4w
zy{Yq~;D==umkAdlnnx+6O+I$jL0QFp+4!UQ3R5aQ0zE#XpWxA7?i~_0KID6e6zI7c
zNM8b~GhY1})*7b6_{T|%Vsi@gG4OhWjb1J8L{__}c<JF?n=NobeJkvrZc^_t?XLSz
z!5e@!+wb@7v87O%JeI`xu#S6XWNb@ovp`|MC_?BVSTL}2P>|_6QLy)B6pakX%K&{B
z68=gf7&woYf%9a}ezD~?27?o4?Wy^Rvg39t4mNUZj3!PGi+>N-j;Ansem8JP0ZO7?
z#Z8{Yig@FEo3JhI8Wy8KofN|{vW6lti#ZSa)i=Bz`pZwObr1^&so)6pSdkWf|M{}W
zg@r^b9IPp@N7efP3KuVgc4C%qFEtOrr%C?GF@D(r?J`s{ROEV6w|lkG^rjyYfMuPg
z>ZAd?c|o&ex;QwqP4+q7(B*UxDkL0@uv=d9p}2C#cDl2=A60DVSAowZTimj`sq^7!
zukej07tp!-GYcTx?!XsD3|7<Et>}Abg<IlQ^WJF}rhO=nFk6$W*qFzAKksDXrMDKK
z17v>ME{ISa67pURVJS^#S(%EjJXzHNirOtTC6z4Mg}@o+?7dl;r<Z`0mAgSRB;;-~
zcR;+8RVMtPB%_1*dWvLck$Si7o-dFz(dI!*75x*~g4g?Vlgk|jm%Zt#Vw!Bk2jZFi
zforic51X6w7f*Z}E`FzqGd)&11cL7a4xC$8dRxx!8lkXgr0R<LIBruyp^8jxzl%Xe
zeUpv4>@WCOrlmTsoMp>Ru8s!Ef$RX=)ANIw*%tpjs;6@z2FEbe@;(5wkc--x0jLlP
z?ttjb&S(B1Mhb5c5T<zp?a{M3TxlH{Q9)V*o{%`={MqfUfJcVYOG!!HqcQ|Ce!Lq1
zC};ybxAgPhlap~Rl|EOo;;9!GI@@-JbUEEST9+vs1aE)8>@}#L<w}PuisqNP_C_9D
z^-9Z=_qdIY-F&%IIzFMjSOTj0a;~dJXAxrcMd4~|DE+pu5NG0U=ETtsyOB}-5z#Ur
zh$ZjyK@t2SPBXV6@8Ov!Ue79WigT*P7k+e7qO*`33W%oc2t5&cFQy>hqR7U{#40$e
zDE9EPv(FxpLW##P){LQn7Lg7vF_lg`ewLYHuT!m2i%%_~4tQlpp(OG8LNn`oU@eJR
zV7mpkY?53hjG2^s(r&6P%;Re}j+VZL^Juq=HB-Y%p9On=9e<fH0^aU(+`E-(<{Whr
z9&+nZBhRi^w}#K4NGgcIGcTGKtGl%Wicmqs8Cbk9vjXAcVO>?q!Nv$#`O-<^luA*Y
z$^1U0+0%6L;hRk9uxX?)cmAt4DCGLa#OcvYu{}`W0J=7f>?~HRh|afZ?fBiyylZ8^
z3KSnsE~YoWWBN`Sg0HALvjftUjl9z4<F3ulB4w-^6*THiw1D4$li`9gsAv1Ych#on
zBi4)Na33irI_)U=Y$c34_@3UZ3@*>K#96Zorq0YqyjCt1BOE52oT|)sNqp4`;FnKG
z$Zy5@8akv$O_`1#Sr?j3$DBeZOqn0QnDgGRIyfJPz5HL;NX5;sMZ{sRZ)X2PPU`kr
z1Y?L81(Mfv9Z<VYjkRJD4koUzb`iHIZ!irBZiPbkHy~qGm#33_?*APl*!7Xd^C<Z#
z;p5{<Nen=<<a1@e3a*qzOUrjeK0N`z+zwS5Ic4%bPwbpXr5<AY#AEIr)|@?*N!-kq
z&LqPg==Z3yyZgxkkW|>T9o&tP(FZh2$qhbr`Vp=qXo&VfGI-dE#jrjzyqEVAav<^n
z$^&^o<+HB5>fFU(eW8$o)^SOk{sbreUPQ_xO}fRLlUKYWJ9;;OpCsm=X%5heeRmqb
z$V!!p0PlePsQvNMhQ1Ad6H{|^wJ3PP4U-P`=qpb9CVAM2PoH)4Smd?)O`O@ZzZDPX
zix#9;SL#Kl6qh7M<PZ+a=x2)bD`x-B2@R+I1|koHnC;$GqOv6SK=FE!N-;vaO91iq
zB_*~0h}~4^8@3OljkeBJ!!vxLn?K&2HgM?U58)UV3!vt(sPz^{;==1bE{53^pcBVH
zjnYc1M77N)OtWn-x4;rDpWo@HCaz0kDa7GVLgB~E?=Y0`X2ey9a62)m_d?o2%C97v
zz4t?f({=dfQ$-10BN=H}&S?~inJOEoLV@vsC#^4wl0B}m_eH(D0BT$|EHPpl?yb!5
zv|r=SZ4R?En_xJ^m*~w_?n_{(Wp2_*V3A$A?hQP4&s~>(p`?A`K@e%p{CJs*mzO{`
zi+u$nEHOPjRxpDlLV!uMqp0J9lAKM4%PLv?+jvO;Ab37Bo)DK1XW#(%8vunI%dP=j
z9H1Wpb~Mc#xnw+m$Xo}%yzYb6bU2W<G=vl^pYp`2*TG&sKtEm8xGzDc;C~WsYI$e;
zJ^AQ$7h`IMYHt$S@Lg`@H>>#Vq_mRHL$a9M($L03|5PM27C0wcJh@s1gaK*4afWIG
z4<#o~jh*rj3fzzJ+C({uz1>es$MfO2Ar-5dT9ew?BcK8{qAaV*_CwAf+&=7>^Jo59
zb|}JL(j^u99I45#I@n~e7x!L|2!DrD+s?EwHXpH7W)B_H?56uX^Oq|sDg88)9ozBY
zY9}U4*iDrpG19>K_5MrBRwFLgWk=*$F^8Ux`?}yh6tjErn@#$@$L0>&l!ne+Y<lB!
ze+T$JKWh5~DTozP?20^`IoY4N(F36?d8Q`1WO0j&iW3I}$twRmz+9fm8$aD0xC$4a
zf2Tp_`|5+~MymZx{%-bR$5wY|A)d2_?m|N&%Pj4#>0Zju2);(|X)oQ)o~j?aeUrp(
z_!Culpo|EhI6c}JG9;GVW@-g6iG!KzjRFW{`M@i)x!G$On<buu1!@6qae;B*KoaY%
z+;?|{H%zP+TU=HH(;#7JUZP|WGK8^C!^Ok{rQduu-X2Yx^`abOY@ZFUjYOifJ@4a$
z?MVY!S1f{K*xUz2N=Tl;MLxaKX<Tcs)+wVgmvEx`4*2kEihrtMQ9UOQkIOc)189|{
zHS7%5r7tpu&h3vksu+|$xr3|k;{)TsgndS-p;v`)Mbt70C^~!cM|fk<{K9im!jq4u
z!sSOgZjDJZp0*nX+)Mcgr%V+%g?iugzVAtkuUp#6b3I!Uz{q~jqcHGZ&vx(&TVxg^
zkNc;(ox0Hu+uX#LWboF?6V?%BvGa5(3^K|K=U?xRk{?H6D%zyEfmZ44Dr9H2!YkV6
z1}Y4e$GF78KU3g4e{oEZ?s~F>reyR9cl&EzLc+FpC&bw#&E%<`wq^T3PUpF_fEaa&
zW>c&SuS&?RLQaemPqk#cOB4;xmahL9vNwZt*81)fMZ*K5QSispgn93bWg4B;oimy%
zHkBQvt>n&+RTk!v`fVs$iBU-=X#Q@X%XbDBmu@ZVWM^|zrcqq4UtFn8cTDQ&@RV5h
ziVji?1GR*TenKbJ^-BS90sGB#cd4H@uRdT<i`aUfIP-ew3}rX+xTj0=5b@`rwfb|P
zFT<mf4X;hH`tRHN>l~W|aHeE=p>8N1O=*yX+pT{QznFx3(LAQS6l!stLkh-kmYn_;
zhD8+POD0SiSPh_$+@A>0x%xmLreyPd=g|i*{GHk8>RCk5=R2IQjBX^ZJ>i5>81>>R
zu;OPoS@cB}-YgBmFgV|^7AmT)@`@P*3V-hVHEzmhW4vcD;P}|x@)Tk1V>#ysNy+^O
z6q&<0sp44BYYNi6S=^62U9(E6KHa(GpQw(RHj8fqQmU!=T@QYjnFuWhFgxM*DK2h*
z2;|<}3v%N9G;P~(h!whfh|5<NPe*2<GPF>ZLy_<rdv9!At%-!ieX<p?02;DDGG+@G
z*jYGQ(n9TiYqBCgjB$Fq>Lks;P^TIckZal-VY)K-DEQWEDbceiSutu`ZbYYud^@qt
z{wJsz7}&C^5y+s$$K#<JbQt4cV|Qx!;{iZ-xmzGdF1rjhrVPCSd(<yg_Z02<Bgc$a
z@5q}UZziva{1{qdY>v!j3GXf=v^1boHWLzCO?82?>T79K7Aq%i`4@iVQz*9dv>c>I
z-QVWC)A=cx=Z3b^C-An(I0toq|M7_3ggc2~+Z%0+I#fnBZ!YDynKw0&Sa#_Cs%=+a
zyFeaYUO+>h(OtFgyN1SfsC6XKc<0WJW($Mq-LXPr$NUJ%oyoKzBfr?_-psP5d>(V%
zo3?~a2Sa`RHVjTX+8X0#pZ1L;h0mdFBWCyIj+S*)$Dh`CekU#c=-d+w3h5fhFBh|V
z&^2?wocH!2H7n|^fPEns+pD>SW}fVjeu#tO0p<-G4bK`b%4|&&wu~M@A$a;~GKt5;
z@DgmTBeDTjmxyV;nLA$BB6YeDzws3E&3=J}I@JQZuAbMQ50Sw{eB2lLZFK4m0o3JD
zx7>el3cqYo-N(T;v`+;<f~VUDj^4fu9$7ePK=KTiLj>%Lj@KK#ua{M<({_nAM9;77
zUs1LCJLozUfgffTabc4<>6&V7m@4aPz@;v7HR3t%f>g9(JNw(mCvHOd$9TjB)8mh)
z0SM&iGS_%ZxXh}(tN9v@a(e+{C-rMFNJkf8IlZW>VbPy(8#UYcaV_V{>^y9IGHWze
zExn;_Rku~82iB?BfZlq!WKjOzZzk{eSWFipr>B<F-enslET0Do7RY$qsRHpN&03a`
znC{AHv#C>;oRy=JPG{B}_4#Co6edG9i4Wop#|+yqg@Jx6PQJW<2Y{FWIXTeglyX>=
z|3=z3Jj0w?!3Bg&>E~Q*Tx>&~@&W2^Fl3sJmOA}h*9*_g7UQ4FsPY1-R+~Js>kubU
zU8IUhPQ;tZpnhqqpq%05ZR#D8b>cTDQgdMuQC_OF{nAkIMiTD&#dTWOz53(m*D4>!
zP30r4h<GDN!;!8=&5YXIEu14Zej&(bH>>FB-%Hf8#;jAhEyep`Mc!v4XN(^cMPeQh
zX$YdPI4MvVY8eT0h4i&J0~lguv}K*E;VmY8BS=X~ib)LsmD10nBEncL-hG_CIT82H
zhUAC0j;|uMUX}XPqtH%|K2Oqxdlb&T(D{s<{W;JSX@olck#ZCyWV4{}L^zX%)Yn98
z>^S3@m1Y(rY^;oPUZh9PR(QXBLmw24JKS(`SgeRUtgw-5BTX|y&Q~ax><5n{F?l?w
zr#Dg4a--ojXYYzo94_6Nzc81LG12qTGVttL_UWrT&W*rpTSjkiDHVuLhCG%e{--CD
zYGELx2MxTtx_1ptQ-m5E*7q09oXpSsdS~8b|0`V_KEX2T5}EF0v8y|LGef3)_2+B&
z<}h{dB-{pdG00<HfJSV3t-~Mz1)I;Ufs@}(^(UH|1)oy;$HUxnF3%hQ&<|*O9kq|)
zIzcab`X~4Z%xcsp=iFXv%uOENb9%K<>m*jW5?orG?lOBoscY25Gb8>~qxLJ%%HzwM
zcS|@yBE;ejMB?W@7rR`VbvK;lCc_NHY!2p^36Yl`2$7mg2}6;Y$-w<%!@;!T0$>eC
z`9EtGx%l-RZ=>wl9_6;T%X`G)It83uhX&ssXlPi^r<a_c=*SG}JKyZ^%Ox$iyq^oa
zN9})hmdih1wuap0Ayv(Uq*-~@UyeXqT!ueo4?>Bl0*adlQgRBxrpmC26*&gecGRLM
ziS7o~v?O<W`@HksLS^Ho1k~>ydWf}XpsAqOj}@9S_v7n+p^dJ&fDoM1!$3tXupXri
z36P$M)mY_-x&@W`*vKk(7jGK5%1Idpi=ri2u3t}_awj|+eMuM}$R2<vMtTl|-@uwK
zO|U&qh8k@)Sc9<!?3kRkfqwH<(5PZziBSGefxztP;<m9TV*u7j_aykXy8cYlX`){`
z!ZA#)NFP(6Zoag-%#sh-9i;}LjB5R2d3Gh^l*j;UYPRa(?~~S=l+Q6^F>m5y>#f#F
z&=Zf}#ZO}Xen_{0V295?rtiJKt)y+&@OlGX^@=&^h$G>j=03W5ZtiMn$n|HCFZ}V%
zX~7}K>G{>m8=%(j6~Vm*au<~?&lh4kLr@Wixi2gon+xqXM1op_1nF<OYCBEt$kd}W
zqIV71ysvk^;)?yWa9bZt<S|XLFuHrGM9`;C%(y<rT>BcwX|ciM8ms!vtha@-iPvva
zB@|AK?8n_=z&x_f^uhofa;z<-V={9z-9Ysgo5Xx$9$NG%+3(@Lq2ALY2llQFpwB{s
zhTb!L+ZFXYPBpxSR1A(i7^c$sO<|zN!z7FX@zotUX!+ZgUb*LD=^Sp}Y!WHMgij3{
zhM%(Rgn)-GV|-iykg;1EpHp=(yltE73R69@!8@k;w$FkrQ%bS8ZL8e2S~Q6+?kGJE
zX7m^G;dKJLKNp{1@W;!lsRo+U1zQ@*MmO5LHKS(qQyZX!93WjyLS9}DqdpEF@UJc!
z1C?1^>A=n&fbQjhwfww&9X6dUXMdM_8XXjc8Ti9%(B&HZ2p64a@**xi)Wv^*CRS3b
z&vBLc*Y`V%kW_50?d%GVOu1y_Ggzzs0}a+gva?r&;;n~vT=`uNDqbUk)Y|z@P6P1b
ztiA2kOxzrVU+hndRD4|?Pegr(Z*BQu?*q9)&(0XXi*<J%1(=;=+m8gtat@NCN7S(>
zy6rV+(_QhJVrrO!OKL_tZJiZZtv$)rpx^s+(>kI2tAlVw8e~sO%Tu`@X7=IoY`@5;
z8Uon~p}4vG`Ed+9_W1ZZG63&(mcxYKzffO6Kg5Qhf9a>R2o7^<M%0TGP_#nlHW|?R
z5p`tt|L!JSnNr0u+>$qY|2w8uDEhnZd(*d`mVXZz0>MsF*Kw5m$wwJL4o}WnQq=c+
zh_)U%9^R?ABA36GT<UnFdi6Q|B4~IM7xY^F3|FNH^->w=)epZ;7lbK>K0rrbjp+-%
z`r0dPO{S!O>zZR165^UrXoL84g$pa&cn1Y0*ZG(ypCV>~8H=>EsJmJ)Ql$QwLT_fi
zW)RiMnvM3*j()06bMl-y(|xnRj+&X?J%r7)NX^~aygn0jI<?AdTgXOFSOWmbM2Mef
z?@TJ=kmF>k7CglQ>m$rADiiuq1$xM?GUbS3_s{TB-L84ojNZ}&{_F<qH#cY5oSR2U
zJIJk6LqxMaFEB|1Oipxr-JE{a;h)kL8&UdARJ5?Nj%Th^{Zma}wea>vYeBW*PrOm@
zt7GpDYF*=wFHjeMmR6J$9p$&)-Rnb@ZyjuQPtt;X3vY4d-QAhP1<Xs|DzG=VX<T8H
zn7<-FQEGsNV?)VaebVS$4H(igT|=Q57wB4(mx6M0;=D~r!LS7}Za0k@ARqMt(pdk!
z#QYH5hI$F~*7_;v$Rf|%^U=X%OaktPuQU3rJi?}nhIKX{y7L&cA{}i-Y|C8K<O*0m
zWUmui%bn$zb-pNF))^XyxVsU0;qVX$HT+YPt7W*D*Ndh$T6S;nKq*q6KQd$fO%|ih
z9IHSp>*JjQ;Dp|uvJWqQ6T>%vtf4(?#Ws9gg?VV{8;o8~D26W7MX!H2laUN7S{%)B
zCa(Kl$tpGvncCU2Fyp@UUdmJz4N8R}%2>*W>Al3~PYRQsrFyb?Y$@d2jHfj6{#y6k
z_<^CNU}_t-zMA@Usv@0D$<`oIO&JUI(k66P>=K=4!|StZCD~ox)Z5wzi05qPE<@K_
zl|>GcTnF1agp#hv$+Q6`^q7`8!nQ)@-PwtY#=XTsj^g5q`x!bW43#Dn;aO}{<0+w!
z0iV?#z`!*at3*$gMtHp*khD{1lk69FTm&V0=f2Z3L@3wSW$PK82q{0DwV>K3#<@SU
z(1@s}+2j#ViQaS(!<!Tp+Ss_u_TkWR6AJTKm8_p&C*SqnoSr+jM&W945Jzg>VPh-u
zFrOndbalSAmG6!lN>3GwkJ+5{($YTOJFmu07=MfABdCDmRO0&a5f^!?J`j1U!?}il
z0h*_G1v0$0fx2YfQJic)u_k96E#Y^BYvKZi@QY^TPM&@BP%@NKa&^NZX|DXUTz4Mc
ze^UTY_d;AVOy~Ug<E<er?^~Kb$ej5cTm{gSa5(#8!rserG{kl`Tp6GKVi~)~m0fB2
zTg;avmvE*o&Qx%$?V|eh8QWW%Us77VX%T10;(<YasGfU(*X3CN9`F<q3rA%!1Oh=m
zm1p^uub8R_P9fp>2Owq)8NKJd=!O{FcoD3SBbgEwm0S2Ig7ZnZ9WYKX@F-NBnNSj#
zr~y>Q3Gf2v)<S9j45w{;+j4b#gX{u8-B$s>uqPr?BjohD$tXqb7>yncv)kbTP|gqH
zPW{;*!Eb&t<I>3mz!~{!ij9>9wEFM+Sk><2z}p20Iv3sHM@%X1F(`iydIrzFP236d
znc=7O0^osyeB+x8h9}F2^hZMpTq5FzDLVpOKAa@0k%HY37f<dwEhFQB5a@o5l~)!k
zK3#dat*a(ww}D8T5NE7d1kfCu+B3%pgB*FWYN4x)w!J*Oww;udaJ+pTCuKJ52VJ@!
zb>BqGd_x^wX-BgjjqzvI)RegiP|%cHITQDhw#Rz9I#=^eHV;bdh5!R9Yo3;=LKO}Z
z;K}`l-1(YH@PbqS$^adUa9qhAMx){E#m+_8XajB#U#J8BgC3`b;}S2Mu*K$;qGQw=
zlugfe{V?*9s`=(uK?|z8b`(zDJ$!J9NN?O1XRY&`pL<x9vxOz@b9kAERa4qR<6pKO
zbudO^m=j3pCTIkHpz?K1Wh!K-nxhnKT{LGu$}f4cg5*23d<v2X0779QAYLN&rVqVG
z04lTLLb)sMV17R7X9X?!DW!GBtJMh%n*>#w`q&*VCFf&0J+$<#r|6G!V%4o-@*jah
zJ<f>GT})kM_kij~YWm82xu8zcEL|L|gtk7_!|VRO-;#e89lZAFbM|`d-aYL)9;1b{
z^__^9R&yFP`R>o!I#g_*jW}sRzq|1pQ1iUXTuN!C=J3ocH1S_qH91E2vvFOU`#@`&
z>6j+DA#{tW6q#%2DgIJ@Yo|8$)jKFK@GUhTP_R&{5JaL6kr1XBXzwvz+uU3RB8Cy2
zo!&wxpEX{yWll4K&mW5Z8PElK#Q(sD-avC1C>s=kf0^nS)Q_Lo9moo~(*RQ_ESu8$
zw_S^mgc?VE!MLt;e}F~Ojc;%0V}P!87k8P}+sJQ4fMS{|7m-d{Ke}4u){(uaS;~PE
zLO!&slUL|k(W<S!Yx^<v1@L8<fG_J6xO&OqCE0vg-kysuF<PAuAdOSQf)|62(`gM9
z^cDQ2O(fTVs%#*-y=Ql(h={TE6pZ&}*N|(fB_x5z^vyNqS9r})8WFQDJfOzx!i-^k
zEx%DpTjzHmFrbvqmUMDQi4iTDb(TGh$?MZ@amotd?u?!$%zBFX;~7sRD_N(;RK6=<
z-dQlo;A5MM<w+`Xffdqh%84fq!Su!&OB<ynmW_du;WLYk4pD;(1likjtsJqB6UuU_
ziB?tWvN)&!)jl2+Z*h~|SdBLn+}z#W^t?7qM=6Voc(Vh|+;iyO<;Nj&5cc<YNL0<~
zFTUGW7u53Mhuu29M+qGV-teFqEN8LG=f|6AMpK2UzzkhcRo&G|cn#H}Xw6gZJ9RJ6
zTWRThzcLkjoyG`<40CJr&G2X-a9F36fle%f$O8xBwljYC!{;?0kS5I=Q&9B32mmy&
zakJbCJuunv{oEOyOvAHDp=u2`V}+rFkHtSu@HHAZ43|fDPFeX68N&;0mB~cP$yNLc
zAKZ<V&&D`SsC2Op=VB3f!W5D<FcO~fSOLBO%s1yoUI6nh3L;4~luJJZtC-Yg82E>p
z&IenA3ce$i=1jS%fOaX6(Tn<;10xD=*5zrR*!9RmZachXw{z%lHJ!uE(ft89&~^jH
z!7olji2G~QB&1>-6PoDH8EpICkW#UDT&?0x!#_V#qGLnNwE3^c3Tp6e;ks{cg;$vm
z-?hhNjq^$r+q?$018$)^9w$Z<q1?D~+y^w@q*d2;F{@rowN%uhG%Q#9fOz=YLHS<4
zLE<-eSyNwS%@=9S?~$^~sk#q%C6{SK2R|{{zv*G+;P9ehtG@88a+=&)Z(CCdAmq&$
zY&F6kWF`0D(xm8jb|N<zb2G9v`t@f_AgY?<@*wH$pmx2<nMLZhK%_D%kH45TYW<Fq
zAeNYmn-ke1lsHG~Ybfy02(c3Y@IjB}7;139o~4)L+sR-9l}{{wuQF`+I?pEw*%+fC
zkLQr@9WSM_H=Rv|rL+w{I=#o#9{c`wYT^3hmcgmCtikRG54Ac$mMW*$>V-Q$fGocS
zf+6)ki1LEN`!CLD$-=dt#v-)hVGDJc`5gXuID3nYP1!Y&q?oVXZmj}EM$c|1@Fg?h
zFD}f9AH8tvPRO0R0_L`zpVOkwS$s$(&({+_(FHQl#q@quLx*~=Uf*orpo(Z<5Dn<l
zI03WU2LS;21HFYD%#ODm^^^tb*3c@TsC!CYp^;U3!rDje!KkJ8W;?%n$+N#ey|BdU
zWc(QeB1gJImOPr9%ka%=L4+Hlk~97dzAHI>czRhOX}-A%i+XOXxNbsE6L)@&?9EFK
z;v)EMJqgxn*OBma*WJtaJ(sowG$~O@R`rJxJupvg%jCq=EC~!@k^O0!CIHP*gR0hZ
zhieN-yK5}c4_cc?nB4i<6T5~zy4tULI`dbF%)Ew8vxdx%D$^m=42?FGC)f$RPZBSP
zsQuk(b!cmHD~|Yq-eAXDU3O{QVNwc>az3TGx20DPsz`pu#5<P?7<!ik>;eNHwTFln
z%uGFjPM^!18WRl*J4AK2CiyEZXOth_(QRKpUca(JNwt3?C^cX9K)oVTqTTDT*5~Za
zO!^IY=BUKSP5+?H!4zGJG5TB78OmmOV!jID<8|bU9BLA0ii13EJOG}Ba&m@}!BZ=}
zV0z8Zz`t7W$6h&II3agWSX35vLQu9AiF^lUSPR=qoeDc>JK#zS89H&58{7TTrhwQh
ziB!^js-IJP<{l<8viG|S0kTrqSSo4!pLyK_nTg#YS5XL$`F3gzE!&sy8=akNZwtWM
z<j}FdZgOjJwX8!JHlF}Mgx=@e@U%wD_0crf7zmH4R6-jCo}%)Z#-+L)$c>q54u4oz
z7?^*2vw}gh{X{U^(H!s=te*zBKN0m6Qfp=uE(B*W%2;mO{5K1r{z8N)QRyieD<WIi
z$A0eSC_nLD95WXHnQ<3k1ekCR5m8+@#;h#A6>N-g@L!#<k1(Qh%0-l(U<#r)y+4t~
z0NU(~gl|94p3*dnIOMHYFq2eh>wfI%;Exj_$V+&|+gJNCri%y8hqMu88#2YhkI|wh
zWKPYAtDn71qp(fyY}@ue0h|T!L<ImlFcPNctt$efwJLireHH3ae0+5%H2hZ6ES1+i
zGI5NmQ+Yq&ju;36FI$9KzXS4xBGk9{<C~S`CWMWCG;75%e;=4CfK%w_N9X4$9Nvqw
zWk9bA&cTkAZ$3DF{#CrtLe1WBzWT7o%(g2rUb2o3av|q~H-sC(8fn7LuKRXFj2;~)
zNW~lfy=47ujjy7c*F|{Tm+Q3F^9Xi68UOT$uGg;>x8?9k`2MFL@$>kp!Gcnu_ae7t
zev%Iw@}t@gvYpX0L3IF8RE$`^NpHTyjVpw}x^EUetgr53D-;$WW&G!=xPswh2`>z+
zR@B8lAAionNiUcA;wy0<4YD03^=X_S=6f)0R~M%@Ll2r?E8z&^#096b(irmBo9Gp~
z^oRTL=H^!~h-3;+SO}XRoF_EyAtLX245JnsU0@3fjgj4)0}BfX2o&t#jMz(CIQC96
zKe)@Q$$Sw$z!`o(G~OR4z54>kOJ^Lv?PxTdpLR^>Bk+)$Fa0y$dm%k(oP~M!j**il
zkz#v0Np7c2f?E2#PUhYWUdFh=MV%_Y#pgX-oi)(M`3^g@(Qzqo7IG~_Jh$HSHHU5a
zk<Mg`bQQAJTt&r6RDLtvAdr>pIJON*v#1U1eczt?DXvMI{xFN3dC06!^X=eCj%EK_
zwHcnou^^XR$-*Qi1&V?Qqr(n3{yiD0A8!}=_ebxE`@9IrI?AMmQ{A(nQp_I<h_vxy
z5Mj`Ab>1@iTKixTqaNs79aA?RWitx#X^V6ijdt|fZLX*128@kXR2$n^oo0kAPl^}2
zrERp<#w}a>bqh2bS1FS!JM?)>#`=KMdF?8K(v~Z%G6mgC-ZsxhkDH&jYI?f}H{R2A
zgz<S8%M}9E^i_w%C!Pgg%{2sA75JYmmf@E~s}eRC4uR~k*!2Pn8LELXkjGA9LM>Nc
zK!cf_z0X4+Z|N_xG$#hSe>y|DwI?2hCA!75A7tm>ls-@S2=xi`ex>t;$9FoQ#o79W
z?`p0ggk(rK-}0FM=6b*3k*<boH%8k_-*RUc41OH_LgF{@4*4^OE_>FDTr1oU2P3&`
z)br{!7jdfW3~~qYGPxEkO3-gYR5}LUFF~L#X}N4HUlU%Z*|0I6{flnM@?t%VYM9(x
zz+vJV0Zv)acRgADNzB}E6Xby#+4-$K{wI8d#dwxt@v@x<eRMEE(_~-cLIPWqmT$gZ
zP+iM1w%h9N>t{oJ@&V&H!7(v675w2H7@G;P=~zG5U|wVFB_E=PtwLc%+;Y!I-m7w9
z(}wh=Iu%q_@&M1I0MTZ)cyaHW9SJPi`Eyg#I+0r){@JZ)fd_O#!1l$=bXu<DKcl?z
z*0<%Il)Mo?ENA8QdVzh=(a|gMTNT)3&R|Ct&yqhyBfu*Z-L)b8i$g5=gxD@vS0yFs
zk@@Z$3tSL-NcAG$=;Vrg9=A<#?=W?B^VB^FMY7zr#=1H4Geps=aHRnCrsSuag2)bx
zN1@Ob*<fgk`8tB5>#O>TFv_u0*&#7y_1oq|OG=KR1`oag@<COprc#AjXOdqCyO_lm
z%*CK|9JzjzT6ajc&`xP(3N1}`e=)E>!+BHw0g4D?@usn(7A>>Uc8Q=@g!?|_-4=~c
zxi}6mq%0aoPP0h89=aAU!uSoE?evI&-fh+>IMwY<O&>oB40F;ycwDLcsPfxU0hNRU
zaW7z=fugftu#-2za{?|4UG<{v5F5K&`!t!H*}9v-gq=!iF+8O2(4d(1rpRW$_}tP^
zA8QIqRZA-Lz;Xi?e-*rSSen|26HJzP08J-3+j~guk6QF_(RcFcjPX4co%KL+ws7Z8
zPN{^iYg(?Hw`0cAAuqcju?wAZPsLHT(ve5gMJBI;C{;3ojivAwNnNuKCoLbyiHJM1
zVl`U9z2LNmh}Vt0?A!#35c&KuRRaAp6pKY#R5Ar4LGtX6CNW)E8M{=%&1N}Of&fN;
zP5}X8g<*>je6E%G@Y$yhfPqeEiOr5BuM&DMZ4P%u0`n{RmG9pc?7enVZULa1UZc@%
zbDS|Yrps|=cZ;)S>smE_21Bg%#Czh9*cr-*YmZ-bL@nJKP*?R(l22~17URX;Pc<SS
z-V#-JDXI@~#(W_MKj^TV^x<Vtyb9sUL=~?-+E%<4pYSIwJ`;T~L*C4bekuyhOE>Bh
z6OJ4kjOAtSa2MJZi^#re){Vf$w9HGneyUVM$VJ{-z|Sqy?k^|dKOHXGv&pgUj?2A_
zmKi5mI7OP#>8PhcqQw2O6Us|SelDE$Q<7+29`yc3Wt$o-%E2Bjr*Vf<2`5eCrRbI#
zbGe*Z3b<~-PFf*&%tjxs#9@!FQy&f0ZO(Hdwi2;mO(MnKsL0TwB0^$~P5tBZ{8AFx
z%U5c7L5Gr82$pN{>yeYlxX)MaH@Fo(hT|Hg#%SPXdl7$y3kopmsCrfo)HQvCr)kQO
zBTm-l($9-}H&iRJ{w`ue8yemgK10TW?&jYfoR)>qWXJ#6vR?bhCel&|W4D56^jK<I
zHP23YXd_Fn|KkRDE}`xw4v)FRXw6$kz1Z$0Qg+;}KVrH23kVbz7a>aodi{B4&y7d<
zW(L~$(aP{QzB$E#ZsW#Dl$E}gp@ys8>@TU#;dZAwoBfc?d!2phm=@X%o2+|6V-~Ow
zT^B-*OC`caYf-i9X$l`=DmkKEn0|T9R()}FQT|6`sq1|mhuzZWSSfl#B2I2Ux|Qr-
zg-#?f5}4&2Id~jBq3ctM4VJ>&(a|Ba*7Y=WmbHF?NqcFtwEqbR4WYoaq?qOl2WpsM
zkA4uL165&feFCqYJ@Z=e&d}{Eg}lM&tLsSAaAT?VoVHlU%*8}+H?Q!ikT5!{4e8ls
zo<6CJDtB7;0}T@;e;yKYKHTw;3XX=XBF5yF_c;F#dv6uj*4Flm($&>MOOY1$MJZC;
z-3rC66!+q84XzcWXi9M}?!g^G@gl*cXo`kl!68`AWWDeEt$n_;Z_aPu?2B(+g~?3D
zm}5LXp6CBu`y|3zU$%5cRJwN0_L3JjX;NM_{;rN|gbnytg5g`;scyv&XCGrF*3D4)
zU2_^kVSo}#9p<*7vF2k;cToCB`un_ceD%E25PcKn{(E!%9D%IqXNP|U`5afRO)V$#
zO?)OXL@ClXQ@59K>&o8AA0DL|#SJGAYhz_qXgu`Z+T5;v1a3Gj+w93mUIeZ!p4i}O
zM@C}P!N{!`8$aBW#hsp6cx}~6LBW-^g+MPi8%|z(qxwCo(9B))423vrH-#WIFqT^?
z=PE(YgKi2f2eX_%Z*M+hd`jLPC!gU3e)j5=OPmn>Aa9f)zU6V?xRDklXiJCO4U0Kl
z&M;~1mDs>u5=?3OYG?}O8GDhm-ZnA>9Y%n6akm$_q8_sS`8pg%A`eUyi(d+`WsO(g
zL91+Gkw&xo+hQ3KTRR6UP!mU26CMvNcw5e~j9RAP+6jQt!{%a=-ZYDBg&AsYxLe#v
zEgJ3PbUiNwV10_WZ<htwY<fMJTO*>HJat$9>d;~9Kk&}kPb?duRwO$sPfe^1!>8Lf
z^4}@XnbuOXj<~(_U+aIlU;>BKH$qsv-<GG*ZOAJUx7w<WXx&Y~Do2`#Me52LkU{+#
zPZE^iEaGH0`XiqM;dLtU5AsKG=<hU;5)m`J;R5=!?_Q4t@rVexGQzdO#_w(Us`$#C
z-~*~a0b%utcrZj2$Q2FNB(@$Z4$3)_XAdA$si*$z6na;Rc2t@MP+!|w4XA2u=QvmP
zF-re)B-lnw;|#(J4s+kUi1-*uAzR&I>-J>FS70P7WV)j{dTlqoOW<o@aJ0|`LjX~}
zbNz8csLf`o2xt8k)g_=|qjFg?)yZHE(Nfe&=-ltcTB%x|rTBJSK^RKLAWUf@yn`J3
z=sgc<l0)Nh=i>d?WPI=M%NZ(2Dz7B_tKS)moqo7f>(q<XYWggJ_YLJt0a4~#Zb|%&
zoww{gvrfS2BJWjsNo&!%i^tIpMS9xDFr`h4d~i_XdUyg9MXPF@=7A*Xrm<Tt82BCS
z<~9W7jX~{-r2RU{I1L|Is@hkHk6#3%;5wdCh-EFc9O>hMf!>Q=sSd-^rNiJL_P8&H
zhv!**yC0XYjFPRm+K#L=&{beW_Vuf?t_1j{$vx_^lfuniX6M#2?bfV}7g1a7vjR@#
z8@33K<M{?&h9ED`k=O<AvXW<Ze>9t|IGMgFT?%V^9i1K~Hg=W~C(~l9lG=AXyAhj|
zT^^SV*keKJmk+VV0z}z_KA!P=p52t;TN8xsz$aU6G&(h30_SHfq>xb&Rfzvc7et~6
z?6(82wUI06?!u=PrEVcoXfWZgU<<h=l-WX=@RSLK5ZXZIO*VKwY1zU%BwygSuxWW8
zZud){51!%@K#H5fN+IR`8>Ok(v%3*`2GtvThD~nruCC*n4nYZ2^hLiMvJ9me76H0)
zb-oEG!HQkyPiKi-bySw4>WOg_j&}}tf3;1;%+&{qUrk*eP+zq<&1}}LZd@RRH*RrB
z(Q0V08mk;$mZU(RM&;VMxmDHH)@ncOZxw|nsu!(cZ`7Qd87zEbnhm)2Zur@jbf|AB
zPEkJGqO5D=Q<*!Pyj^rKXf6@Ha4-e4MCV#QdE$sm?TWqqYo+RXAqO!CmNYd34|s0_
z<kue52E9CL5SjC4S)kr7aBk$_7a1KsERutd)wb$Qqpa>lKL&IuDWFS#qRi@`%jofp
z<y|fUWPCdTr?C&{Hp;~|=(ik)J=w8X-g1uV#AgYiUG}uJYGl~*A|Lk1D8bcFXF@d=
zc5=J7Xic)re%>;9`{G$ye#R$wN$T)GQ*CwuhWgFL#cZ>@uSOH(Gf((`wxy5Uq~|0=
zdIJv4;~`r_up)9nAQ&SSF@$g_F*nP<noznrryDg5u*F>Wcg5~R{pq{AShMo(=*S>>
za|YC#1Ues;NNe};a2bUdgbI+I5M_$_U&!58Tj@e}oR^S|<eN}mjU%$MYxZR)xrQA*
zt_!m0Zni>qd!vF-^OVncm|zzd33TmlyC(&7*C&-+mp$1edx+`e+spl&DzMuCn(t`m
zUfSrAS)>cKa*1L`o!^2<8QD_iY_aV^Oy$PzpUx)BKdK=Ml+MjO)sEh4+s>1?SC3X8
zmCi~JeM9UtIpprgR0aC-dz-8}u6d@D1-J~HwaHT*-RE*pYSmAQoMG%>r_5NW-H6+e
z`y@_!;xv3y!;+|{VJfa*Yu<^41(U|YPMB$be|~B2h}v7yI--7S!bi3Y<;AqUxue(Z
zu?k0JIGYE|SQvFqP!U1ExFdC<j1$3GXzJ;y=Ndecc?{NMcii5+n$=i1-5TMwCeY!r
z^_sq!8K0UvuOt90#o1w<#>#SN&3@u}a=`Z%8%RsAm%h>Eha5_}jlM0@<{+J>#wk$Q
za9q+4eO9TbEL6;zOh<odIM;^B^Y)kfGQCX9&PpJ2(sb~3H41=7R^A-`Sm^IcU5xK~
z61yv1aP+1$+4$JK(YLoh4;=9Qa!``K_SFle!+Eyugx;qPS!6VxnlRF)EyoJgE59jE
zLv(p@K2t`6A=jaN<(6_W(`?Cf%xrjK-@iP$Sl81)SvsYpjjFIO9Yl)z%zdrlhXO?p
zHKxb-epI1v91|{X%5;M;xf@I3^xGn-!uuCiuA_LGi0&dwF2SWLOKy~HMDg)v+C#%j
zw<$j6aUH)Z+mTp1neEMGpN6dwH6g2`{>Q-MzDv6S+=wk3B-NPA2(Cyad|a*O=ORgH
z<)#y8!YuQdT;_`E69osDuelpH#(Q?l&PB051t96cT`3XA?)@B}n7meh!RwpmWLFWI
zm}=Ar#}xrI9u%&WJey<{N0#w)Lg9RL&)21IBfA~pk<ky$1WTL?DZmGo1TzJ83O{{~
z-QmifivM|pqu_n+bde6?!o=@X3g2{=y3WFPrP5|cvyIKPrlqw~ZsA<bzp=U6ye=o$
zG=lc`R9r7%LzNP!3_=+k*V~hV8*ux<;+sqPrr(7Wla7>I8IHln6C6^o+usNCNpuxg
zhjQ6XL4DC+by10#vVQOlIN0|HTLjzE#P8+=<Uol}ZW&dmU}UNsfO|5r%@LJ%?sPNZ
zd6zzZu{*o`b3rMRvsh#xSKTrYDKWk+^zF)vt77@R<w+LOtc5dnrK!wzL8WpdIFtWI
zrCt}0|HroxyR~<`<d#c)XiY@y<T$y}<~ecDtM>fDk1q31oioyRStS7fXtNkJkSYx3
z!1ap^l+7fYs5S?UB+K2oONERd?hmGk`fuFSB2M=FJP-r~X9P4Pzow@J*0y;SP@a`-
z!8aMub|4D1lC>xXYq~riT`Y68`Be+#V4PiB!Uy+X=U&}We|E#!a%M#9CQY7x_4fW8
zTdQrQ_+q0YO6v9?99uT{$fs|z-fUW;J}!KU#H`xHgwK)Pd3!hU+)SVdHSX_i(;6j~
zz9b<K0yT^*xnL3MnX(wIfIw1@U9u-7JRQtZ88|-j=z2VVV|paR`#Fz0mOfLxU#lrx
z49`g%#=|Bh@ltG~WVH)|nK)m20u&q6lwHC;Vul63gyS$6G5b17a&kda*d-z6H_1@5
z1(rMqz`D7!h_n5>+g*>PGT%MX?s%Z?65D4QmBkM=UER2xnHddV^#1Cj6FplH#xu6L
zK$@0hyXhmK?7Bq~KPCVC%86aH=VP{%kx}CwA>G5j<^7%SbSLjymeH9)C!4CiibdFn
zlQUvh8c#`MAForOcq3s!>13<xQ${R$C~r9{V=Z|S8J}aTKA;EOMpA#J7g|>zQ9M5z
ze(8z&w9KTq^FdyDDf9d?^Zj<&^s><Qum)6dNrF*??)l|lwd!(U*Ld;qqQ`Oyvh!g2
zrFB@u{^Nn5Z6m(1w#_9EG#s&R5VCg18*^rDrv}NcNKU_QX!?6mk@L8X2O>1HT<PPr
zEAK{hP}GYZ<}MLAGC1*^7^i9iGd(8*;QTWy<4erR_)V+-8H(0aN_wct?UHWU!|E!@
zX}7l;(Qtjz%)v-$_I`lQRIFr+!&S1GYkWX%R`NCnc)eqYf%m}psCBobjL%9U-O|k<
zSUK6NQ!Tq`GmnG}%MRZ=7%342I%1a*z3Gc7Q*lzyJ$EuOJqcreFBU;-X!1fVPu<d_
z(et!{N)Is-uF-J_FpLl>isdyKdrNDNnlxx0+6KC9x#lWHKo<1dJg0n_xEMy07S(eV
zB=x>dR$nLCE%mkGDf_vtEqXFKMu{FD3v40+4eDIh7QdZyZmAk#V{s?FO;zozw8g5+
zht}5uConMMdVlLf_{;FM0p{!{^v3>s^$h#=iowCbdU|>^V)-;Z*bf_5`HI}0zscxw
zo4cwUHq}ZQA&klsl|1SMhZ{2E$E1tcSQj!|Cn~clxqw!K9xh9g0=OY8AqQohX4Xc~
zvSSnakfG(^ji0khAPCw)*K+sAXf_P@!57On-+URB5URJQZ@hAJ1NOkOo?2b2r;{1u
zFSyyKov@V!SooE>V!DJ%UT6|aJ3ATKsb{5Xmkb9@YGh-~&rsl?t?f8*bLBu?@q=UA
zN#se%aEbY;*T}k=>qz_6)1XwMU=9iZgX9GkN7+b6us}~P30R=zPkH!fkl$pY5lPIn
za0P(A_Z&7Dqr=XXHz-<$`l?&yd~@n5Uq)r}fFbE}J7M0eqyA4fVtd`7uUl*i$D$>!
zu)1>4M428t#x$?&-O1kIP%TT2etDn04v;BMDAU9z^*JwJ38)fydIl60=D5yY-%L*#
zm(ZmxUS^G6H%bNPj_Zt3n{J~!<E4s!mD>H+H!qOK!GY&&l_c#iXCGu=Ze?>wRYr!@
zK{{?mJk11*d@^4CtYv}tlmm#HgIKYmF(kmR-<c6Q_1>f=Zdwt45g5+s336Q=Tc7=$
z|FPM;Q#x-<9PzJL1cdqVF`uw9!HloUt6vI<o=IC8SPE>HBVnWbEq`JaYue`FVtuda
zZgnvkU!aqjv&1E!6D7xt3+(0cl;W}!x&5K)m8PaE4lm>mV~ETJTp_UAV>R*4)-C>8
zQqfl~?dtlOdkDaL@#xgFh-XrVtclvs3z(@TD7CR3%^(^Zf+R)Y>>6mv+E@3d?Q(3H
z5@zcbK9*~W+KZ2E{%2pW_ibzDaIjgA4XVQ+;^Ft~?1L~C#?9#Q(8OGx+z|x?7tNux
z3Y{rA>He08!7a<iTgyG_!wcy=ej7uZO^D@q8fcQ_b8zm+7GxIpxwtB0=oznn|54kY
zF#+QS$*3k|WE|LI^vGI_ZlUVR4y|hoGrXPi-AMNsT^?^ojbz<m4>hE@qig*0U4IDk
z1fq*bjZd!Ot}lUv|IQ*#boa*OC<R)P?q)2z;wr}W7L_e|E|V&+#o4~oJsaRL@l)Y`
z->=+oT?C^Y@t5V}v1Ev2ZgmouAB*_m+_dYHB<E@0PBnNU^j?L-07LeW$`}iwr)14_
zVH~(_4b;fKHK&gp%6(=|K_xFKiS<Jt0E`-Yd$zkMP~?9$)kTGmY7vrr%@aZ~8M%nh
zAJat`(iTNxd!=oe0plqf(mreRJlgyb%p_JaD{4O?ke#8ZYdWsuzOZ_=xnz3HlESKI
zA}~isBKA>N6Xx}OFxl!THZNym8UI7ly9F80(Qv-9r)ob->6C3Ru(xzlJYOk>tz?);
zfU}ov?M-boMD#69)-SARI$%cv^e+#(2)!hIy5GIR_aRnOTv{|T|6_~$WW<cNtB}ge
z>M=uq9*f`Xb-KUG5}(%b+=Gwuj=y$<YPanaa7qLaVO=mI9L*_;eLo2z86cEY1^X!I
z(VMiP7XmdA+LKBn>wcQ)xu0z;>Th~u!M<tI;bi)&@wggsySXLiyO3rc>)9(YREZOK
zwIASu>GVEw5q0d_NM4>a3GX5No|*yL3%}i+r{I$CJqW(J*qb#K_PU%`DKzH3xrSe%
zP?<sIH>=#7;;1y&fbBFrGYkTA7=<}6u&l&mbXRTut%bU+V6(Z7-NLi8o^>{%kr{31
ztt3G^*j)OPEt!mxNBPBgE_v&Pj~(~eVmgE#7s3=W$ue%1HawM{+@stoq{+rL@7_fB
zLci~;a?C-?T^O9rf>dyfNKdRH0}Ql-*v&}qVR5`j93Z|z<9;aQ^>|!GD062-k0P`D
zHd(s6!>`UHTWY;Wk?F0ZK&rs6r;*)f>+u?4gpl^LdC?P3`F}+7T`~({AB>MV%iV-0
z;)jS6?lhsjW>9*(YQisH6!cPiZgDVUSj0cD<kW!mLtdH=!k!?sRHWIjc(wi!mh=+I
zeh|}rY6Vx!p^=dkMnme#{fs$XID$dT;}0#IlCRslyNhS1bl|r<4E$YOpw&u7+4hdq
zKfhB501>M&{kqj^`DCoRdgq<WU_Qg8E!Gcbp6Y(5ksEMF4-)AsL-=tj<_w|1-}3l-
z;LV;l<ZAodUJT?9J2SJnw%?=4a}4(>H^rzVC%2wqeV5OS;l%pDs+uX`YJ*CXI1Ig&
zaMay>IqtSux?ps^C2NP>$N_bIL65UHELsx9%M9Yo{)78j*4Y~tbPLO98Gu+8O%im;
zny9FM?yd(Y5;&y~0}(H9ZuD$`Du-d(ofTACTYC&R!y*rL?)}(0x+41za+sPt7Phn_
z;HdcPG<m{Oj{`_11whk#Y^^jwI2_;U>2U~0pR<Xlg09P{_9$+%&b}%L0+5Zqt2c$f
zG1E^g-qf#3q1s0?vDyku@xU4C<8f<L#{{+<Rv&gp@P9xW$Fu>Iy}D??MM7nPnp{!9
zk*ZHshm`*d3q8Hr=rRPXxzxYZk|2kXU+~X+lt<LZb>VNVCb6)-lAx$LE{DKHlSox2
zBAG9I!T>+w&`0Kah|<UaVa%E(7OXy^|79HUyKxdqv;b2SIN{sk{qU4eCrohNp1}9+
z=iY1q3qT<w4CqQ(9QRduU1GVz3M}sBiyPd3uf~T>;II$)OesGGVEHLo%>QSVfLn3^
zsNUTrP4u+xKQD%b#eF79Y4Xog|Ng}P?f(nLKL5V~X#bykSXlpet^tk69sqjtFI<3s
zw)g+4xf<TtEC5hTy7v`H3}=pxf#*LH1wKU+&;T;!zh9Ikwd4QB=jE@@Jl}Wpc=XRB
z(U7~djQ@VcOaTXMSOBP4Mr%qgkZLcBV`5?w!x=I&35TOnyWQqC-%wFeJG}cW3v&XR
zEd1xPqx;{Ee}_%^-!5mh|GD?SU4x4M+ACl?H5O*f;9Ji=czLV)dFH8L&Cdh&G+?}b
zueeh#z&=VFT@7wNh~d0v=LnyB2Y!Nu752}<{wThfm;g2e_@~(dKL#|^F?*-5>LCce
zGM6)V9J>?&f)UI$2Z7L?x>{;H4C`g1{acUCx7=Q0PQ%Xy{?=p5@L=dTBY=(mKVt9y
zhv@n@tCF7Su}O_@J^QW`Xub!gH2DXI7T|~<_Wwux@-JHXA$%BGFXiCz2z~3Y!5nD{
z4=I>nQkG!;M_~aYQI_TXM+W@EiP8*@e*g2=zdn7h_@8_Kvj%|Z|39JwPr(*|ob0&3
zt}P592lGQxL#Ck4Ub-jr*T@N1!ScexpML)UCj(D@_q)SReiwb0CAI3fYShn5YtbgJ
z&A$LVX7_0Aw444PI+aem!*7<ZS7sE`=7PG4C-R^!XH9PnR!&K9i=B1Xo5027D|g>x
zTf6;#eNW);?^!c|mwf^zX{>YUR(H(&(YC-G^S0LeT1F$OE1!}u7>{CM?b){l3IA79
z2P~}OtUKGTc(9Y~Y{P`le07}_MO7rDJz1kF_?pI6E_U`y<opcCyh*9;nW4+!y$_qS
zA#~#F$^Q!n_WV8e1A+kGnE|5vp|c1R1NXNUz0uLQoVV}BufG7Ji{GAAzjdcIn8QzZ
zOSAKxu0(XnMKQWM&-cD4*~luuisqYcBN9MhgIm^nrZc_0{#lXl^nd@l2=FFr>k%Kd
zkRA%eYAM1Zs2MZ-YuhsvZ~U{ej2Zp9!?pB%;fpRPq0DtKJLCnIeXgpNd=?N#os?DX
zb?bRb;vNnYRI7_#fgqfoDXdU_A?vV6!p57;h)=y|N7wDg>@ZJ2D}ea#qksI@=#}l)
zwOYuNd~Xf=pIh#?a@3f(HYm4bUH~wLaL3)m)?Dd!zU00gr)Ac7G6f98sBTEhWl|#d
zr80PLBU|EN!@!qMJ>AV<-gr{o=HYiW?PGP~V41Ru&OGh3g)Ist^+lG$5qu?%YS(|8
zGr$sXyWso==qFGoW9a0fP%O41uzYa2#+}6T1faUBjUQC>R_X;BW<<>Hswu)~U=s1n
z4_DiwyC_#5pc@fE#)b12;FI6)SQ7^y23X}utUjRI=O5P<u}1^jR5XD+ayA7>#J%=2
z*RBc#VPll!HB)!D8bmr!qPq#}c>UURHCmpdFVQ)x_-9A@i0-hN>+;ND)`MYMJ*!_L
zBM<QyD!J1?J(Xeg4-|JXmZk2F%Pb@>lCvse*HyhA(W+da?wtio5%#tCeX-CFpgo@z
z@;*jLH;gF_m9g7&^$hrO?9{N2SWCzpI_9S(*=+yu#%|rUXS#!^H+#{>_`Mw3hRpc7
zv!$rHo)ESBra1I&K0jM*cXB&FWfSDub&-*|I*e5B&o!`U{ZauuDp4}mR?zDQf5cR~
zPhMM4g^yUPG|u|w!;XcBX<pW^B{9ZqVcX&RXEj-)#@Cu><%mXZPRxvj8h!h>{jH91
zHE4X9UA6d;*F(2E-}Wz)@%R_3Wi@WBFtwF*zK7G3de>*62(#;tLzV3T^Jq`nt%~4n
z41)RgG=|Jf;#>QEBbPX4^*HF*Tk32DtXhJ4C%yCQ<~N=z@f>zq_3usp9ig5Kp;9TR
zKw5QiknH`K_ZyRqy_HwR$3~;Kor-=fYnB?bK_LOn<Cv8Zwx~b~8ad7FG}rwKUq!4B
z<Mek>#%ih8W-#NzHl+_6c|sGPKf8aH=}uEmr4dz!f6?OpM*kPEzW}eCr3rt|Q#B7=
z&8n4HU*Z`#RV27&rMT8Comzh{aGk#aFyi`S_U-!QV6USd={DBKbTdymRfaN$T9gFI
ze66i4?FQjxFlN5{Z;GL=JtWLO%-jyKgAaiQM0DbhhwO}|%SQ#x<mw{AP)7H#-b6V^
z?bEU!9>2O=UrHx?ow?A_RGiw!x}>x?{(JL*cX%-<f*jwDCLJO5=KUJ?#gDLNO$5Wx
zHj#bxyY8?FhR|jLr!S((drj86m-o_S;T&dCZf(EtknsxH*O!GwOvVj+TrKI&<qCqX
zen*Oj>9KSo2kWy0;-0tG3rSODPTyCS{kBnAzFX3_L$|$p;I@nH^Xz5NEwim@?fOxz
zg6nraOB&zHAIokl*e;lpX$#@+u-z8(g`I-4kN7S-sYIFQTc74U$q^Of60Ou@0a?uz
z&1UD`rf81=uDnU4htFjLST0klc{7WL0{|s1|J%9vq3;g}+(WCZwi3rffy0WJQfDVC
z1aD&M@AD|cdC(OuAGjlsJzm6<HFHFsyT^cpMw1vb)UL%OA$;*8+4hdl<n%0in?KDx
z{I`gUWl1~Pzs@#T`vYJUzXQDvY99vtwVz4J>?PWiiUH@jOQfddHbItN4ck@fhNfym
z^V%I*L;NhYXGexJ4eWE*$!)>Y2|`I}Rv^u~=-v2^sx}$5dfj_=O)YIt33gxwm*4An
zba@VPH~YE;%1@D*nTHj=cML@yR)IVCHyE5E0cVrHV{Npj21#bdzWwoY1qoTlU;K^Q
zm-4!PZu|Gs<$*9Snjk?Lrifo}z8?m&oFEu`k)^Nw{3!!l;L(pyl~UuXmZue$egMXC
z4kVUZ|6mWgl7P)$A)8KjYdZcWnlDzo9pQg`yFaGbFLK6ZhW0{#1ZvURE-TMn=5D5{
z<6<Ay#bP58l;J;>J0xuvJ~PlE&*HrA)g_2Y&gAzq_6`=fK2ahF+O_YO!opI^2RsYu
z^c^EHM_AS%tfi%;>)&CR?J3?&4-VsC%<gl%G=;>Wbj!`sCAQY%{pD*FO2Z1n?|W7V
zYEo0PP4?C`+0<s$!_$HCn7iN%Yy3cla?#&=H}E=>L{cVOoQE6Gj%|F!7(mY_Q3{D{
zm3Xu;M6RBMU!VwO6o`BI1Z&5sBK)(V3AEJFL+RY^?>)=tC!Z!7ef8-q+hJ9()G#!)
zkw757<VNn4T6~^QD%cpYUV4!p)y0bst>U_62pW%3qX0TbB>6MYQjx$XADI65<+uwa
z`SAsPQ^wm7+}Tv)Jiequ3(qh;h;hE-u5n9elQ%jSeW|ezg}7R52O7l6I<UFV1V@9v
z&&PBx;-%9WG~{o_ZR$m%$tfM&_QMW-){O4~qA##9;Kt6AYda4Yy3}|Bu$6haj+xZK
z+eFJO@hUsedBc2TrZk_=PDywN=A=xAU;N;vGUGh8(lp4c;y%rMn>8wGT2S(6ENN+_
z&>(w6>J<_9Ex=1g_5RMF29%#i=WblIXK{(`>U~eI{{AeNEV96a%g@(krOI$&j+<&*
z054DAS<bx=g+PGHPzDHrZlTO-i-ds81<$KN7wp=5Z}>p?mIX7?)FDqbHo;olt2=zD
zTJ#~2sl-5S(?YPR_$ywj=e|LwPABrh@sHzZgnSE6tS<XfHG%UmrdzRipkTbnA!?YV
zmYqg3A%>x(Sd*CptfhV=Dh?P4kBFGbV9_`_s@i2iCAi_7u1$y<m))T%K<I4@{e&9!
z2l^0l-x5PZdl#w?Zz#&E4R>OEoV2seyZ}SbDV*dF*mun<U96bj=UnfA@OjMjT6o^b
zh_iR!0NDa*pe1)^n}|fyR@NI(ikwLw>$8=op6ZGm^Byf54F&L@8N*1HeC$R}yeOe*
zeF5%mQYV%g_h0TR;#vX6*LY~Ea5PnStdRqy{aWEQha$0OVUi2jf1%0u_MHnOQVq9j
zVlQu&V2qj;efAMhU8)`1H3EEV>rp_JkmSL{K#wys!D3C3|JtxTxZYRG^irOGWl;mu
z>j_%J^}p2JgGC?|>E`NK*-f-8r={0u<}kyu0F&ZMxm^_nBN((+6Ji=~?J-hP)d;iN
z5(=(;-^~&Tr7X`a5(UM00646{M;evQUW0Rp?ESq825hsS_o(I9#6RQ0vo4dt%LgmR
zjud!y6pgOI60Xd}#}bIW43Z+J-L%G=(+rdECp$vB_ynV+zI8&+GUfP4mjYa268B*H
zbo&>-+-j1_j$zpWHGO~NY6O8?O+VYZu$|G237z+e&|usUiK*A93dnW%fzzu<pdZLD
zq96A-3WG<}NU$^CJpMqL3L;n5(C`C{txRzqLXAz0?Jv26GJf1dX5xwJ{j(v9AEF6m
zSf_cAOIkH^cmqD4F46$QCkk~6h-t6anrHifU<M%E&Lh*Uy9HJiV*$fS@y`IO-Z<(x
zm43J+sjY^#F~QtVaNEw><r)Bvn8~j@z7^wU^He>HoTY_zwclEwISOJEh_9lO@wfV%
zPNx5S-;Taux$AGMf1|&OiRa9<XMiAxB<9B7z+Bz<?U~=G64pCy^bM|^xsk@^?vWhP
zzVbsI_W843!u`3HzuXDm`?)#dL{SOONEm^<HIGFB_@VieUc#;hId5iyCH3rUc^tFy
z5{GwF&V!luPd#tV2FfED#gam++$R?^V!8_nlV-Rlu7N_0jX!-CYpfJ-%)7&y2XX{M
zs+{ToCEHD;!N|we)jWH<kxCu~<iMF8lP#H#5=EWS!k~p<RN(XBHb{=8MD>>9$Ocl6
z!z6wfnKIXIW8?j6vh^KGY;MPNj=JkxTIoYd)L6X#Gp?(lTc15THI+Y2Gj8IBDsL5M
zqhUrha})}+mjgK(nDIJuFlg5-OZaP(&x(%YalF8YWN=@2@^BN-!DQe6+{Osnaz5~x
z&Xm{kSKhsfwa(AYSJKTn<h9<MIOgY{LC=~gkV3RiRyHXW7BZ=~0T_dK*OZATG3Emt
ze_hVhKHz{1UkSDx8;Pt?2D^>c(g(kqNEMT9DjG@r1M5jP7QWpxlMQ@U@qmj1Ht;f#
zAQ&7_-EqEbqQ_|HUAfrqzd=5wuVJ0+onhD8|G<ung9jZP4S-K18d}TXqz)sW!~!Oj
z%SG>-6uFu*;fK#d?i7M>KRr}HLYQ2JDRCRe_K4;iSfv58sQOivDL%O~PD^@mRa?}g
zw<d%3c@{2Zc%T)EL?K9fm)`Nk<ZE2pzxvyCM7nE-RmZ!86Rk$)+=s}d%1_?}cIJ)y
zB=8Sx=GsvVFgx@zSG>dj4*gdM1Oo&_N>@yMdmc@zph<uVwwtD{Q&j88+94eY{P#E4
zPbB_#j)-#&hf0k8!m+b{O+5(*Z!0l_ru)lMv+NG)i_u(~8TZiWf+LM7QWKW#;?>Ou
z-;qGF)U`eE*7CDL->?bE!tO9(n?Q$W8OG5kx&B=ZVpM!H6RP_m;M`0A13A~(d&9;p
zzi}pw>W}CeJQ1q!#giC57lTVh^@j?6w|!sZeKix+nvK3%_xkHMl!Lm%dftSZ)17Vu
zswW-HWIAQL!DNxabKK%gPhsPJk|bfiHr(Z9$v?vYS3e^v#B59m$c3jjAI#sRXNeP`
z<0Ow><J512Va0eI{GOca;A&$P3S5hRIhnNRF+F5-1|>HWufmjo)vitwR$@T;7&0z=
zlgl4_K=rYKZSXo+6127h_urbHOwJBE9Q#JmJ$i*@*n5`I-f%Z<o26?zVA_{qODih|
z0(j@T@0;cXJn!+hid6>KRBgxKi%DY3Rxfn*xLr@{VrqteQUFSk4T29~*#Lx1f0*Eh
z)H0c0M%k42upF2HZPqzeqE!SS9p9iTO{#4)aPI0d6s~M4cU{~Ju<Ejs(yyiZj*cj0
zZ~(=W@(aL4w*pYX@2DkM)93tR@L!JPy6U_z?Q{I26u^oYS84zrrM=eib2~`B@7lAL
zMIiwy{6_EoRlyWytzMU<>}i$ECLx^Ch;z8x?-W`f`iQ~*8?n6Wi>qi?4x;X-*+uPB
z*XtLTO|Qn)zuEe6RZiXK-xmO=%VKmOKQWhr+PUfFeR&;p$<bj&DA8tF^zxUR_rIku
zd(CuGP-HCYS}}{Z(TV)gj9L2ij!vwOn(OrWF;qn=D`3&lABv|UpfR$YoEt`elw|Q@
zr<Rk=ie0Dn$jBw9IPcP>94jpW-tocx-9m20BjUS08BHK}kMqr*r#%ZY`cq%4gA&Cw
zf{mJ=k~v2Saw*oCQOXRw`WQARE`{(kVGm^R8kztgXd+^|HwlxTxE8nNgRL2<Tr}iB
z-EY<&(64LjR0e%09P=KE%fTH-<3_!(=&|9<fO=1=*mBaq;bb=W-shF3mZvwP@v>yO
z$0~p0&!+gJ4K(nreV>RwyBlotbVkb`BjCRhub23^7sq@syt35^)sDT2tB|!9yaN!8
zM0uQPV4$7Y=>i|1BRA%=Zjfb?2V-ltkNt083_1GjINY?mE%df@#rl`>PBXB~1-}|>
zP$7I?$0(q0Z(tmN;05SHXLND=VmF2m;<wq5aS`LV|8j<v8DF%e(8GK7^$N0QR+pgy
zAL~2*pKj7SDf1Gx=CfJL7yR%3uD)BleqxBl{Up>L-!@w)5R<Q@Gh)mk>MDg_1vu4z
z0ehcbA2(nn1E_mV_h`A2Cs9hMtY_zD<S*iWj(k}YT&=DNUdPLf4F6*JYW1ow?$1MH
zr%J>^5Bh=cgB}WMK)>no>%6G6Zu=I`%u&fg^)>EBjLp=(!zaj9f%wixnR_@8UNg8G
z-PHJ&C?y_1e-`MeeknJrT4HU!YaMS<YH$BcZ0!ADTj=<U)zN1RPl{M}Yh&B%lg9P4
z$8sHh0o_WSgD`&>8txx<o3YP&7!i|3Z{e)AL?;q8JuP4j{g9taNEys7^hS5TFUKQA
z4#!-9cr(^Yw$5pefl3HAn!iTJ>p0p~tS2R{b#V^VXSr$4y~`Frgg=-|MKWXB%`NwQ
zS1!gvYb8x2qwO}k_V~gzpN}OhU!UvM1_%H4woP!njDz1w3%iLDul3HtpP1|O&BFao
zPj}%6vpF3kRBQQofxb;R<KBmv6Z!X*6`)tmg<^T!wsw_!F==57o_2O^cJt+Fmsn}D
zQpEYLuK_oZ$ryd|`&SUyUtz>&T<CnJ_Lv33p9mm5j%hYO8soo9xt&EyT@kcTjJ?JB
z{zx3)q<*2L!#%Fy>oYQ&+f{Tsu)HWB6f-w>fdTrr&@t$F{>HscaAoAdf0)7j{tM6A
z2n%>!tz5{rHQ$|yWq(61vOwJNl_{79KhuONdOY^4frTQqbr3sK-$0**?d-cMcpt5$
zPZEhP>D%WpZoKD2ldm)e%9&+($Mf+D@6P~WOhBVGkLEM*4m5D*91Phm-M$R4dNsWV
z8WlNLSh{pap_eC_gDJqqF|{JkPM>H*?hGJ~<9$gB$i7dfLp4iobBmBGIwiwVs@t!a
z?o4dZYJWTi(2XqVnf`Ov&B?IJgcD5%b>iM3eF@9woVs3HvqqI2l!!25JfaE_b#~6L
z!&v~#3VCIQh^7Zmaui+vWcsg!Q>UF5k6+wWW~qk1>F=6wp=;S+197Ei2dxz8%&Xi>
zJA=Wp?Q``rK_n$vxpE}1ii78#5KK~$T)r55@pi@o<m^zKQ(!l6=`?`Pec$HY#^mI-
z_+#H%EE_yT$A@JdH?58>iD}$m#v7s_i0+A8`aI?GdHwF9-H}t*=X%Nt1FWqk<_Qf4
zJdI!aZf;HO5u5w-2UC49ql>kNss4i6Y`R|`zvZ;Z(>cYW5qd9pDiesB*bbcy=Jq7_
z^LH$ozua&wq8)7yQp;Py@*7Dt>&Oy^+H?v`;{s*fprealhuQoy6C`XdS9zI@oL8$8
zp>%k-kmgd|!!k%4r26vXvd4>?7cC8bE1nPF^D}~zrmvR!V{II;{1g&yX4r-rXE|pb
zge1)fI}QhrlppXtP^c$koxPChbd~0aIR-+}ZyP3QuqdnT4!;p9yWA6O0c|wI5z3G^
ztyL@NsV$STr;65{{@m;&>{9@;z%C|!>Wll6lasgD?``V{>^~<R^$2i!58>wKuF8)^
zi&_Vq9E{UJ6ZYLEHmpBK4W9U3AGD{RfkUTS*i1AW8_PLU8kYs;AbY%$l74mhQR&!Z
ziAIzk4A}536jup2H{a8T<X%KPz?C{0DVbeeIWG{<PRshKrW!&BgfnpH<5P)bej>QS
zzLq>|e;RHO^6AeQ!lL`y_QiG5fW<gqmFpxZ#KSYUT7bK`%DHh&7GI<U0%#%lb}=;o
z;2Dr@X&rfM$`C9f*F$gUj01Z9el9;EmxLT%7m$`pd|3L$Ke0S1pFk-Jul{X;TgK3v
zIxzs+%otO0VJ|^JhZia~)t2meJ&TLViV+%vR8~e`{Uwtdc_BwkhaIU~dG+x=$JCY`
znhSOPDi1_r@7p9RWoNGb3TD3NJox&Hch=@vqdb*^MN$~$wszlZ5StYf?8jhHk$}hb
zxy`-dXCuMWzps;z%{(ekejc#Y7Re)3iMytrEKo%^+8G#i`$OM)=g~BIQQ5s&>CDl6
zEw|FS5<$4{pi<<*MINFlU|}i0eQ&<FnmdXBuyF!|js~Q_A>30R8?w-++BsW;o$`KI
zCxv`P!93`LJi1mF^Oy;}po=uZ<!HL>pkvjG&uJo*4<PzY>qKlx0T#W8ZKOUc#h)p?
z%SyqHMsL`jTD~GyXXJC773}*@F2FiCq^hQKbHbNy{1ahU5`#vTa3&l19BuCIrw7=2
z5Qi#<?`5wwz0eGqk4k+GKYlWEa)G?}Jw8j$yK+@jM(lSS_PVxT&+(Ehzj>3lQ*iGC
z1aRp61}`Gk;t7)3U4CEIjAx#^P*Jd@@Cyp^HaM>r<WzuNID<*;13VD1tEn>NRKT?$
zSQGrt&M^RHG)pek2I{~4b{poAjQUm6e5vC*@E`!$tkB?0nc5;=?G#b%%y5<7yjgj;
z_tvFM)zjU;5_UhZWb-#Bx665Nn7@F`YI#EoA<EU}FZihI3)xD=tl7<e7WTL6PiZ1g
z4;62P5zkdO{fvh<N{aw2`u9Xc9U}gssnve_^7eVi4Q19x^r{W%qsjK-@v$@!@nSNe
zx%SAAs0qAGD-G?zqGIAlai)zrrFa{Eyjk|UnYQX?%v7~?3;Cut$C}7mZflUXqY6ae
zUidMn+6?9w^&6QpSqitTy8-xM15nMANFw9ALsE%ubmd?M=KXEd{>*RvgB8NFv)4vi
zF@i~#_dI^I^hH(@n0xz+@>(zLFCi@zyez|a?#;8(*l31#!XNdU?sypWbE--N!4;V3
zVxnDc4y$K5i4!s!-H@27peo%7?6E!dEXgIsLpDgW=%e3umGPDxPA@3_+|n_kWP`F2
z(Pr!Vm$ZRdGN8`zwPxPX_e4b+70c&7UY$$WRCpIyqMVVrsztd#L*JE7D(1&Q)xOPs
zhZ}QS=n{(eMCb4WnFN^ld;ZuJpTkb7W1>)sSDIeFJEFK)+;{*YN^&$^;LB#B*;U7#
z%$<u)e}%q5;}kL(97~3UeG}u&u4dlZxy0cnq|f=HP3drA0Arv0nd{>yuy)fjZ-@{+
zytQJ27rY0r10BXqGnbYpR^N!>b2(ZC@>iS4d%HxUXmz1c=UH}JSwP~_ya#wERS$vp
z9)m0jUSNwd_6t6OXWS&8^wbQ!ZJ`}pMu-c!SMSHJ%i@4?o$Se9iR$Hk6Sn?+-OPs$
zkc=N#KSnS1yv9MR+Nk5CcB#gwfb*Oh>NbHtpHaSj8L^fnCH1YxAwf+@kxM4;=|gvm
zuET5$3-Lpx;fUO(#sZF{Lgp8dh5Z!Nc&E0atJWpWk=uF_7)H}y!oF7f*2nQotCAis
zh>C1lPs2SDN-;QE80YqQJ$G@HO6nfFsbufl?~`}AR^P2NFr_`{O0NuzsOPq}%1+td
zc3EZC3Tf&qjINvcft&2{ro*51-Zd3U@a8vPJI3|%ehmL4qo$q_;&-jZ4p%;dESQK|
zS#*DloxVRD0+>%8)&(fWjhh`VS-bVgps0Du8^qKJoZ>M(KrLSGC-S@31D2x_0lTiG
zoPidBmauy!8!#>A`2JarrrbCKjoWQi9KC7I*#x4|m=`ogK5NplH23`ms2-TP=nC+=
z?q`Ojs~`f`-*0pYy=79D3R;M3%}(mP*S=m0p2BIfYb2XEq@p*3w$D6S0LrU7skA7l
z>66oO0MZGV>(V^i#J7nW;~atA8_NSsT@=9a49%C$zYDEB&5(kEf}}ElZn6Y))1gEw
zps`Sjkb0_xU^@?;6&>F^u_8z<75RLN^=W++sfmM2K)}gZ?B~N!(4mem3D;J-$lVi?
z)PFzmJJ)J28VshD@NhetdB=cFP)W<~ANTv$t+QqCF&IoIS!d}mqcV$2TK5&opqDCO
zpq95_4JbP;Jvvek(1}}Hh$TZJM4f%78Ie`ik<Br6^cYEWM(y{SDa4H@S;dw#0haiM
zi1b^LyRw{k0Y5K;1OVOBgBe8})2~k+l9DY;q^G9<DTPWSa>ax%pG8>;d?Syw@>@KX
zmlc|rX1YfqhonM6lGI+5P-s2tH|L+NmS=QiTIM7JHTyjMW*lt~$@O_P@s#S&q4{8c
zF83rRU!PiX!lQK(Y3q4sBJs_zeSy^wJgjwcSFE!d(k<<A;o}pn&q}vz95E>Dc->xu
zI#rgUn&=EoOWXkoBIog@>S^w`#2%X^3S*|H&|F(Cs0f*gj|3)DGY7H{^xk$LI+30@
zTllFgA*6I)&X%E$xlW0iUTa$3umA~ZJtI+}gqW6;&W51V$_E+>D8Gwe=+e^+DPO0h
z#_o?-q8N^+DHOBD?c5>p^+Rf-02y6r<*8Sqe+E100SM~;keIGR`v6D(yiQ*bukZ(X
zaUaW0aNpeJgmgnk07(C=>pszRaCK}=<y0(S)uf03(&Q=z4DN708S6}b+ZaV@xbQo(
zF|0u{ttH8M{CBAW9qzzVd2BX0!Dp@0nEi~l5$EqGq-Ig1crHO6Zl~*bU94tfdfmTn
z)!5Ew_dQLLg<@~rj#^reXiV#U-Q4vlxTbzZ2+0)xg6BRY<1noERS^!dmxzf8B?QK3
z5UGcB<#xirE}7+RK58>FKfEX;5{!urry}G&mkAPZ8X2!;Ib#if06ceK$I;h-X);m$
zK+gQJ`;?4}#j234M*l}j_cVE&@x`IdY$<Oh)jl0S>o8$Flq9o9tu|E~1g1|yy+`}(
z$&e>8esx?zId&zMuI;xDbJF28JdBAwXr_YG9<2$TUSCob2uAxRG5p(`Z@H%;u#*@M
ziGg`VL9<0Pt5MZw3To>{o3B7cHWn8D0J@U$Z^dFtWKEfAX@=8nvM>|oy35h_?NY^`
z(wQj5J%*0@EUXGjBk`z01qOUgVm*+~zC2)mHjkb^)nPohVfkvbenN<|+0DrZ^&9F%
z!~2n<_MjVY)3`0@GP7zuGU<G69Sq(TP7*h|=q1V4|8*qz0YWSg@LGhrt5kkaN<_18
zx-7ZesNS-sivwGo<Uzo?x>PgcVsoB{{09pz;*cWwQ-pQfg*4R0)Ppm4Eox<BRB)CL
z>%$BX(lyD2;2v4r_ZFY|Z366d(nyuv^nA|E2Irs(CxU9;Fo4I*VUxqVt<D~!-|gM*
z5T4(C-`lBVYNb<yAoSu)UkCD(y)msT*IEc~H@ocwWVf@)d#tWwdUMp*LRnsCQV6MG
zheybj(JuK}Y)o-<Z6AaJh<<(_5n2A6A%G1$CS6b;9TQXEAamJSu(09jGWwD-l0XOG
z(Sd}h=Eb(5O67036OW#{8!Ds#hf^V42Va2jxvDX5#I_<WGl(%1P$57APOmS=l*-+d
z7KbNhx837m&xBZ&C|QAKtiAS3Y9YGX5gElb9;Z=aK)V&CF0In?_E_qu2?P3q;578D
z^wt*k=b7yb1oeC=qfvtsBUB%jqF#Yuny(O1DZ>~34_lXnz8^*gnKXSY^c_YcHG=0S
z_OUWIuJtA}Fp&tyi)*s6XYgPCIov^(@Xs_mxMQbB*5MlKOkS09({BRg%eL<o!(Atl
zR2{8O$c7vhxJ?#?@0q^>T%9iYJ@dH}9#sR(3iV-b`$8pML7iz-QnW5W?sN5DSR^zr
z{b)HTIF)Dg>n%-iRu1L^i{5I|G_PUH=+B-l-;+S~b(zJOL0lo$T#HhgYykkKrVOW$
zI|h#{U?*`S=KKlPcRWDuyHWsBerQu3)g*9&tRcfOZT0yp<eR^|nn|#J5JaUAGURvx
z(W?+_LjR0|)jv)&`}GG2sn7T-V?0RLkkM0nS^BfJql$@jbIXKRY75~rySRz1nDbGL
z^pP;F*Yeoc+oZBA55g$iggsm2ABAM4z>#r5FIU=5SF7zBf243~?%O8mXI^it1DOQ+
z;~u+*Hb<0zuz<6ptduY5S>9+J`jBJ92%zY%payHg=p_jPY780-xHq6&79pOeGLjIn
z5N@=vZ(7#dF$?PA!}QvswFtnJyu26<PFT0-p?`Wf2{;*Q_0Lwl;``mj721rIiWL%I
z>bEcPQBCA}DaspL`99g@2JEJ}qVep5@{x-zD{;9E6KoB37P8xq2<@Mb=auJ&nV6YT
zY~$CatUQTT#tuoMz8{P_$5aiPseqq2Z_Af?jQ*P1#QzPj{+P%}aN>I)Um+<Q({j08
zW~5IwEKd~)AZmL2X6w$}eloW6o~U(pcllzt&!q(Wd%4}pMBolCG`^go#Txd;#|D*!
zDD3+gJ`0tg+~2W#%P@AOP8)20>>Vt48h}J#cH{LtAYwntYmY``74Rzq5mT=@LkMXo
zp|vv2I@sae^MPOuli+m!_bM!r+gL0|tM?zEsQ}+qK+pbCJ^{uF4?209NwA!f{gWhw
zZW-WuxF5Q9a{DEKeItV9mw<)8_2B;f`zlF`C&=lchhzfo1@Iq$5y>SUw0cpXz_6e@
z<H=+6*}FY$cZvqA)7GIs1nBl~p=IIAZaSz(fK~a-gHQE(TC!|44f3Vw^Xw|5dZv!g
zguml`TY(o$?hNDGWsAn+%WPDANXx5CH@ZGC24w9vAb^?7i3Zl5-zn5nb`@i$%nf(I
z*L)HAJTXdhxz`wCW7kzdMLTzTn&>VS3_3IWh6IQtw!8S{bT*-NtvPRE9!u1VpfpZu
z6@|zU?NY@IeIDk>O;bTB{KTe|A;5!OvjF^8I_BsElXkw!O%;ZHH<(BC_JeOvVgV3i
z;9u}JR(&UCx&nzYUyDyBwrFV{IkHIb{){d*6Z*Rp;0NPKus}bW>wApyq|j#2o(j!2
z?iib(I7nk9g!!yo30(ej4|)?3qp>7ld5Xnv=i+`TZvnf+TUqeWYUM^>ca3I0`s`>^
zZuz4ebkwgP2+n^W2r+0l21SX@(!5JDY~Fqs_<(@(i*B|$6ttmE{m9(9RrJB|!=WQ!
z1G#1eChrRo-!5jfqvP_BkFu9;-vdoneAbS&r2g1oqSYsT9CZ)N;}fub&-mWIGNoA-
z9zVz|7Y)L#>bbbs8kdy{Kwcbj>mK`cKbwZ^?lf7|YIitP4;aPWv+i9xV8hdNu~Q5-
znx=kjV%s1Ww@KRJ0rB`C^Q>F5WRvwq6~H+_NGTVNM&#ZZ>Z!tiF~k<L7|9kOt&=%j
zcJ5d$J>`omzYI=u#A+3AzK4@~(&L+K=DD&|+Z+)=c%&)Vx_+q<IzasGh0B@+BtLe*
z?+E=B!V#aUHMJ+)#b6MRELA3}^7HEjLIF?W(`!Pn+%m<C+Wfc*pGU>Ni+is0Nq&46
zS)|mXglv-wEIWR%08Q3mX^gz{{i2$U6O-k$9W8Xun@xQ&pG;5uJ22sxlgW;X4!7&y
zLt?-Y0T8%?^xj1iKzwJlchX5z-dWTEf<ZbhrtJyV{e{ZLd}N|@uuF~=XweXCA?9n?
z6NN-%7G04=k~-KCRJ&rCjG`+^-EN1_Ts9ReE{<hRmwj1h9)Oj$v+#DugevQizbibE
zR;Zde1dpZ+drNmbal?|!rlI0IXa(x9qz{5)cbp$A#AD~eZ()l1d{e8?drP}55A^Cv
zOL>u0je!>CWe(Uj6*|rUti>lgcu5JDF{-Zm<vUWJD(>JsNSQ`*`x|Zz_shP7Judk5
z54_Q$pg;WRLjXbk6zHf9q{r2EUTm$%VxNPX&hYKxd`{j=Fp3F>HRrrS;*q$x9PVV7
zjeq}`o2K%Oa<wN?6(@)PGeb~DLS?W-+v#PQd#t?6DtvWNWkN`~TX^+Nv@sA@B)fG7
z+^5lOF`{$6><)67pH`b8gx|ht3Q`p*4oB3`hX_|N><5TT;G=NpGXTrV;D<1ZAvTre
zaSjk?(%!Zjw*YYHzECX3tnHC||20SPtv7-!pFc3q4tE)lv8a9D-vH%k*Xi{n<|8Wj
zc=H0e%1+n|#4uc#6+6+`KhtK2P-bJ3+fQ&<lx6KW3p+n-7uLV%&lx#|A3cfnXoc;$
z;@R7o2sCO4))8|UWVav9$5lytdhXMo;e9X|MRTS2!hf|wV&M*Xl@)M3ULL(QGBlyg
zMI-4A;bK_B8J-lsrf~QP-qzQp*4Zi0UoEz4T8u<HBEsvoThUXOMK%62`MG>91UYrx
zeEI}$8r%J6GRsZ2y*q&Xih-Ws8*7~Uu9vYUb`T?HN5}nBM~8%_jrl4>z5M)x`TYwO
z$BwI7!%)eF4QTr9cc0Y(v&=^;rIaA3Z)xn>JIiJCfZ5F~P5)b>d{=1=v4e|kQK-Ik
zI7(8Vw>s-<l#tE7XqQe@UO%|i>1pv$6cV10yL;rf`Eq2p_$&0O|9axqEma{v0e}6e
z*Z2n(4bzAFER{wsO1j|j%=Cts1a(?+avYT;c7toA23PCZd>+tEz&r5ySW4QXr=Ivo
zlqQTL2-d8;%9pmQkuk-6UW2|1zr>~jZW3C1U+$SLVYo#P7})ySogKEdQdT^%4$p3e
zRw%T_fjM56H<1!%#q0j?bz3AS5Z@VDN-2Gw4X&9xVV85KN>#PteVwN-E|Ddnl;*Q@
zs<##m?gH9(7d(lz3WNf*hPRPZr@MG;rO=Wx#ONq#T31t>-?#+Crw!94HI=Z<s@&OS
z3WE9!SUqhfJsd2@WiPL9aOaDbMAYg+WQ}jiS+KrKiC`YoP5I}Jc{IIGK&CY`1lyWu
zUW5UANdq`dsiK4en-IKl$dXn!lUI0Eb>1n~UbTmdPsN;S%z0owgLB|T;L6@{yhbh?
zHkE_W`OY{Y0H^Db=m;?-mjR_x;z}8_Rpd4BF`osYu0}gV1Xl7?q=H8gqyYZqt-CXE
z8h-N^h@HKjgD_xer#caAF#<hug?M%H_@9LW%GQy@j5ZVj@MjB8Nb=4;Eta$xkU;v_
z_|CGbChWR%cD5pL2^GAV^yu)PJXxyqI<0Xe<6zxKZ$)H5FRMmiNAj**j2`ilbS86)
zt3v#GvJ}*E0{x#^<0*mQtt8zy;%E<Qd7KpsPO@+Vz&3iZ{Ty=8>Q9S9Vj(Q)vD6R7
zTr1G8<1z?FcCqsoz!(F8ZFNtPX=a3ld>ysor4Ip42*6r=kv19vfjst(4!5dI`QOA=
z3%(rF$l(0TU&f&)V)j^Px~-@Sm&2v1j0d>&`&$;J6k`*lv!+CUKJA?{Jr`HqwCk~d
z3Cb<z*J5yKVF}FBdtp@6A{?Br{7|7440Ny<9Il$XxwR%MP`O@4Va#z>vV$%gb#4uT
zKIKOvbRmS$W;EX^|5Gv>(Pb&j{X6ZiBRC1r!ex3b=IvQ!3<R@HC1hfBO7q?6Z-y^f
ze!Y8hEH!4l@&B;*mQihe-TG*%w8e@;u_7%LT4-^H7I&8dfda+d-Aao~f#SuBd$AHA
zDems>?h*olBzN<B&-tJCz2pDkj&Z-9{eck{VJCa7xu!kOn)5^&A&^8bQmk#5IZ+R8
zT1eZok{);d96E)MDnhqmgWwM65>|dp$LmU_udeDt7vJ9M*Iq*HY)XKR4ZSp%a`4_j
zHmB!bxq&|cVVKoPv+bc$V5`ITR9|avVKKqd1`eSqF}{v%QG|@%Uz^_!EngS|*5G4D
zUtdgwRKw0TizaJ*H;*vuFe;R!RjYIedV5HOKX+{>FnW%lb+^czTbvCq3AQ$?0K|40
z=bwzwofn<fg-lp(0{qn=P#FMGGgkso!KGaN+tap@fgs_hQzFQ=o<97LIuS?%mcV=@
zlBW+y;BrpA2!OQ5K5->2_pj7feJK5hxQ$>->nmW1lNE|&MtS2h7(Iy+jPX`Wzw$j}
z$Q)9ZviXrQ!5qGo4Agnu1p8?w_fw<*u+Y<1`@sx6y3C&bFz<vRLmWDwBU1eP)dEKh
zQwc;vB;_qrc$~l(yi1l8b`z2dWagq5Ei`~#J4$dQd!-U*GCC~0B`-Z#Fvn6>P#>(G
zSUidu1w-U|+&<%S)g+kfRM@-g=A^l+BDHeGSJ$X()xrb0Qb^^n;x(5PV2Eq~=~h>R
zA(_iyslEbHt%>=ie2m-2<}gb$pLW`&5-!zO)%h|_rX}n)1NseY#0BABo#v`~r$~_@
zS-1gOj{jOqo9|G1K5b_oi56T;YY=!i5vX7MAm7>HPsz&Uq;M6JXl&oNr#VqWHZ~p?
ziBtDn7Nr*{mCzTXYuK=2O{6v^)$3GHoZ8a**uAxC>Sg9Q!Mf^mh*P4UaL~*_@dH|1
zI;ED&jv~Vl6=BTk6wHg$vuQcCB!2{O<vLr_?7>i|)D6~KN-)8<v7Y+6v0AxIC4cL_
zQXA!@m(505f+|U5*#4?Wn*!t=&|4s<2KpA~t@tfAC`22-yx66_Xe6C=jnW=>{$x6r
zM80-uT+&@Uv)seW`sTPLR?dG9Hz<?l2?GP2M&n%|)#Aqs3WAPGjJ_|>_dwWC5vs~M
zGa!q9dAB!^1CDR6*_a|zOgwpX&WdM`!+-}o1F*k%-N_1nnW1L3O2iv4*<LE2T}7ny
z|77&dbtoD;GXr+Ea|1xhaBkrH$e=lqxBdN>*iWcx&C9Wbe>{LEvX-`a(SYW-O>&&p
zAP>j41O9qFluAsev5y#yz2M||-Jr-`;<c;b)&PWL8pYX)JCu@&Ka`HfD~Bwj;iic@
ztzhD3ro0*N0EW(QR-u|JaP7BLw9dqf(YYK*s@|_buMZWb4jRfNMd!5-Ko-GghTxtn
zcF3^yB&|i^O%R>X51}61kje_V75iW8YSG0BKU3=DMtO7gk-kx#UqEl_9oIN<6JV@g
z!9DTw2XsQ3dB=xeg2Zd%Hs+zM!kERUliKWtd^{$_4nR~rYm5r+6v4wCh>3;@Uvc<)
zTR)mBCSJUD+kJp)v3fB#Gg<e|XrZTvsWwQcwF+B3W(g6CggmL4h)cs;bXUK(Q2{xH
z^njur_Aix1>M!w&iADTwuGUXRUrfkpX~#iNIxwai;XCb~paqkKg^*_~Fb))hlAY}%
ztP)}omE*E9-5#gOLvNqm3;U6UkV#E;KR#boYsTk~3JaI(Q?@tEiVLYJ^=^osg2W02
zc3iLm*s6gHOHUJ`3YF#-9i7C=B?57t%VR}8w9}nzD8a|35>>6MFaqG3{36z~PNk-I
zCc;*EEy~JviLtgDb#hkmAXMJw^O`}tp6g*(+6tQHShktr(BJrA#L1siK1=%O5&-q<
z>G#CkJUC;6vGkg8I_Iylut>VL2hGT_#?ZDW3^m>qBZdr*(O1~F%2@S67!V)6_Y7Zs
z?2L09Ci{OfLx2RR(}1a*9t<$uWj|O_4N{>5qHj{kO_zwKs4%R0&M)%3<vE^L?kE5C
z@baR*_45R}@U>@F>+L+&Cvjuu&$guF{7-(!!N<q+#lHz&{toB$5+W2tvDa`)6^xV#
z#e12$aM?}_G%};|cX+8A9i<t29&lcj&whAwrK9QG$wQ(PLB{y;N-_?VDnOQP#_MIg
z_VSq`iKV+g=u@IIx<v3+<savrvtKpXYp@FFJH$Gr#6RSmy;T*$Jj7H~ZrZF~*XMW=
zDPIy_+Eu9fDCbS;YT^%5+cogO*Q1&zfbL)^UL1%vYU@3QBSm4)>+)wy{27Y=bIA1U
z@Qnbrx=WZ|st|g{U(ldzTiNfVdi_MY0{Xk>VZ<nJ`PDQjP@8J%P+^JNGskLnX7<aG
zDE2eko!Kw;ddic`zRMcLwH7D6QhN80oly_E@{#OnEH{t0zv*@u9s{B<3<k7D$3yL<
zcW<veKF{TQyVKB6E3<h}($Weyoh>T%)Y^$#3wj_bfcD=t9zA)|CfB?Rjhqt`^1k2(
zOE>-Gq}kUDb59aoe0;c?WdpR5Q?f0_#-VhcfV-EF#xvAx<EQ#kGLAxZD729Ubp^<Y
zp%bcLQPltakAco?%3Ha3g;_)l$dhFN1|z4KWr09=;VkFwt+D6T;89`dHA@08+}{o4
zRX%<!e|dYRLQN#(7&uqEPTEJ$qu!>l#Fb71$^kgP|EEqan<gx7`#VP}<<sfecZsf6
znx8vyB3;T1%$u&{+{gbk0CZ0mx9KgMBPLW9(<1Dd@n;Tj;mALW-hPypm(67Bt8Lqm
za;3w~Ibt$4ByL$qnXX7@GyYnh<W7Ly=8d*CfZTmt7Je}G)Zbh@&uI>(?g2k`ke*|a
zTMFIuQ$#*IuWW6>@n%_Exdfb~2zV5aOIB9i_c^fb{`EoQ6TQX~I2;5v-kC4zYKXti
z1Pb|o##k4bpNa5KSJefVD;p{}p|`6wb?1@Cb46{#SoITA%2{*z&hXatlQU@y`g5L>
zIT}4<(i8uU7f*jg!<y-W^q2kXInOyof1u1fQcsnwGf{3MS||lt%n32p%*8!m_DF0z
zkV*KhT+W7bJ*`^BIDGWPkKc}w+{Ak&Mhmf|Q+9kC=ihg!K?A3+W*PJv9}U}+mU+4C
zg(bw3)kH<@FY)|_*|Y=-(#*~Gr&K0B)A#NA+np!q6BT7Mr?Dr{w8qrg_9<z;rj+hM
zT3*JB2l;PkhP0ahM!M#g>hTK2;p~+DcrkMRy$Oi%NA+bNN>i~o0%3a3K0M}6N!-cF
z<X*+CElG*PyAGvI%D2}r{O&R!iz^x7{4OaC$j8C>ui1BYB+HlpPn5-n;WI~)Uw{H3
zFTkC1`5-Nl8H3S%p_ps)R@zGr;yqAgF7|0hpp-+WruN>Wv5GIdyFPDf;HH`}fCHL&
zMb|Q5yOCHU$ayUb+d(5Lq+G4YzR{0)j#W{*ODO-c86St%_a*vA7munuq8m(l^`r^A
z+LV*qoU|YuB$nj#k4qqVG<TL@u^`@)(9(d7?`z&R1+6J49KKln8?iacraJ*3DgeCm
zA^{}4@wd}rKtoQ_&Uh-r6=A~+EjT6eo+PHB%CMsESuIznVO0QhzfQIH5hrEeeb3Ul
zX2AuWP>u&EOHE-h_kRL--UPLixh)z$9v^rjx8yge3~CYN-1r&3$h+~Dx`&xM9ySZ(
zm)G@E!Z$!$_`Sy~DfUOQSe_Et_Lli$Vq`j(J<5whr>LzpCVQF%Kza{v=2yQ-ej6ly
zX#VRyI`zM~0IyN&;rVKuey7^;xZn7E*N8fB@1V@PAQo_Fk-cpfLb$>7au9ev2B57v
z__*v~Cn2i7=5Y5mXOO5Q3ccuF!0H=!zO3_oU|9Ei?-|}T2q_I+4mg>4xG9txW+#KW
ze|~eX&(80c{_oBpfX{MHG<*>Zc?259yUQ2g^pKa<)8=H5`dYe^CE*qU?Ayd!LIK&(
ziCjJ9`NlFQ?Y6(e)YhWD#(F*ni~qzj!G63(4reBMZP9WO0uEi&@)t+$AL(n_BXK3I
z!xHg(tnE%p`V;1l1im`lrt~;-1_|6<FPLPD{)rCf!<W!|SA3f}Tsb*(vR*P?#`EnV
z<=2;@sB`<VF;+O|;1tjt>5NSL5gjojs7n!O*=KWM%$78y{KDnOfxqQT89^@L=MKSL
zI&};jxwY&i!JQgI4#yJaA-<h+Y~&Gh0?lx=CO$*_4v%8f&U92tw3~lvYR`66hR4z!
zkeTQf3=(%{c;3W55OrZ71D5bv$!=~_bz9O720n4065T0LzC63~)D$+}1H_pHpWXr~
zgJfKu&$$akJ7N*wM*EoG%HZ&tY47X9(V({A{AaotjT&!7Nb!7oD%d+`Ke&f4rg_h8
zD}oYRwT@5!mc0aCO(bzf2)WaFA}^frkFL_b$!F@&GEqj`p)PJo1pROAKaR<gzOk}K
zje6qK3W~yS4z_3(&zir<YnReSn(+M>M!+V7)9n5&=;?W7hatJ{<xThJ^TDbkdl7gQ
zJKHrrYob0-DQhJ?^XBp>gZgR)I5nhm*+EIlJ2sdC^S{tJT{qGDWe-Z4Z(#Jg@z~Xy
zJMoQKA_b$)^QI%(tnq)QG9xGJmIr|cg1ekw?rLZf70?>`KG${XNEcisk7fW;)avtE
zq2(ge8YW`?9`8_$y&5c9ljgKx88A`T{vJgke+;hbKO_abQ-GvkdKI>ss}^ct$GCZW
zat^R`B(HUAuSPUW{>6tnsj6C!s}Bqee8F1=ZZ0md%i;cWHPWSZiyR;4#kOX!HSCPV
zgQ=-+o#e0o3Yw^4jvwPm6ZFOd1G8)OwD8$<xggayWwQZh{p%_<I~HUA=sBc*Rm{z`
zsp5HY2KsdaT79RW%4JmAd31*k@KpJZ)&R%E$N9{B^jk?x+Q8PC?0IL0K3nEcK}kBG
z@PK%j62tp=ig^Gn!>(&x%Y!ruHl%(N*>5q{w+paZ$`<3%R4xlAx!e0=thWg(T^j#~
z6Sdb_wn+~`$cQt;h3+eMuZ4_?&mpudf28=5(sU5Ri>G<H`%tVtw2`6Fi;=(z-y&*T
zTrz`!pWYTnkW<d2BD?Y$ylpKk9y9wTO&<rHYHRnaQB^mNt(@+~R=FJW!&f~!(|_L0
z);2n5NU-r%c@LC_`$qhx#QULS`(Ns66b2MFd$grYqlY-II?dYxC<cZ5v{FYbh3i3B
zP~De;xstZHA}14uxT5+`t6~K`_ZvcDgtehLZ|!HOt#7@O$Jn9`HxZ)O75-P7>u0?F
z7fDN~8AC1Ro004d)^e8XBzGvnr8q7xQ}nU{DAbqP<iwS*dQwfjg#Or!YBZZsvZQ`n
za-ic5x4l*(@v|1J^;%lnkeAjLZT`Jai$|tKNd1hhXGh1~-EC=&Xy;YK{K;5M#iDkR
z^+>in|8p@6Ky;#f;&&GbaPL^j2FY#$oNHS04WRtqYiE)$*{zqYC|)5$%V=ir>Jq(w
zF`fjh9nNAI(&p4<2a3tcyv}9HGUZ8e%V&kATbDHQ8=5rD42j)qoi;YM%1mC2A3A@N
zo2>;~u~|KmCv_i#XAh5gnKJ991+R10UmoaLa$74+w%sPA>GU1pM}uj+6=5$~CNi~n
z6zFIW=sthuv%S^-0+oUOS?s=dVaX74bwc7ejWoZfPpU;ni|BXk`T?i<PnZ|uf(&WF
zAVs6PAU&t<<!?DGdJQJhK&>61UH!JCZLj4QrJV#Y#B3a_x>D!%5L6g+W>}WGa8<#z
z1=}`myM|t*yW*29<yZXN{;M4<I6LXSn?khxelHVZK9mY}rKZZ5`Qs<>M%Z<zb1tHq
zlZ7{NL5FTX*Axy~$X3XZd_XJYu*8*nPj6!K&2^%%pifPPS8J0gCChzJknnmqNxd&x
zdkO8=QI!#vmx-xXHW<+$VnXYE<9i<PR7+U+W_{P!)z>=BjJ~sFWwkPf#rDmssAG_f
zUZNkN>9dMY*L}RA4sB;}G#mB8jZ7{DEeK>wA8;=VE0PwVfTODb66CDcXXkH3L2FFW
zUHA`44F$zfUp|&=2j|qvI!7!)6DCy##;7v1dGvD&#S&nh@5y(%Bd32dKcNGl`b`N{
zs+?ON1@J;MB`n(yI*o8;d5eL5CcJ~0f}BUEr-)C5cBB%bF3y-S0EBmMc+1KI0N6uI
zuuo<luio{Y_p%8O;O{O-6+Z$6oEi4B7z!NM`nKUKYBWaj^(qPan#|hT+Dl%GCCwi9
z3j=w4&9&vX0J=Y|j<C_M_jaF1{P0NiA_Eu8>=#Xz;Qa5Sl3aIywd1ovI5MAiou39B
z8LbPpo8jexz*kHvG~d7*`*>`C((Kjkt8fHrBm}ej4JPFF`$OP2Ui>$POP`TakIem0
zUe&_)kZtK9!T3|5ouTC!X23ArS50O7kT+N~q6VlrE$i94W{o~A*}ocSlW9FSHnHq|
zpp(jCbiHy*AP^)Bgcp#(;dFew+oEY|cg*eQA?*{ELmBWs0^X}q!((Z8GFLzPtoZ7@
z$8eDHQX=7QUhcx^tH%?Xy>g+}eCRV;62%xIUSmMLc>{9`qy;V=NA)@K%Ny)&gETi^
z-wl9DVKqqdK?evCR<Sg5XmK9dG<YN=f=kkM1M5y-^9U}&_H+90x|vvn?f9!4=~lni
zybsEf9rS-Eh;wwU`)CfvDHO{J{mvJ@X$Gy?YGJcrH9U%X7`Rx2F7p21VLWuOfrmHK
zAAHZfwlwb^C2w%0`@*4*r8;iyM6p-E;oHl0ZLGHLYIkbiM&@S>sn;zenN6mOtR}Ak
zw{E$v0xvjrq)b%gV)k2(Xfwh+2@JC>d=W+`hWG<eF15=4th6Hxu1@+2l+ddK^6H{q
z1Fd@^7d`AauMMXA!7(q>`B4#CE(FOFfPhoA#xmIpf;aW%;#d#otA3ll0o0cJqmU^Y
zKJX#7agE19x}fePP%9H4Qt^L1UXEh_arKxPAY?Xd<ASpGfb8z}oruH*u1bHK*I%D@
zUf4qr?UQdxu*eh*O!)wQF2I~YF(k2xQvEThJ@XTTMTBfc#h`o#;_}+gM#Zn?uv#_4
zN|?*CNBJOs;{Jtq^CdH|rIc4U*!|+<+XNdNL^^m88xEZKJ-lBh(NqO(+zM2&L%F*p
z7nUg`RY!1OPe&Oc^D*z6=lWMQ2tWY>Kp0$lzMOQy&A)XFt*C{SRg{VjZ@(5BYOroo
zR~&2rmf#ty|0k39H?;+TRO2k1O2ogX3{v)~ZBa%I)N5)CWo6-RvunRGIBDo$O-(ri
zw8X+k$IchL)J;eEo8Ax3V-nm&7S?~3WYE}^-n6(X*4;l(<Xg<K%k*7Ofb>Cq&g!2X
zDoM^68Z1J3flx*PgeAtVQ)6{7jJ9PsiLZ*aO_O3FrkVwsU7}h*$Uy+z<v;dEJC3!T
zt)@*RQDbWlmF<+^<3;+eFR_>KhXR6mt4o2|X@5t>y@#KWdn=G3c6zFH9N<B=-q)Xa
zgC|orvEX)`*J_eLIhwVb|5J1dd|Qpqo_yZtv-KbMxpiM?FU2i2?JTPg|4u#^uMj)U
z;sro0_}V%*q8(FWp}9A<-zn_x$N04pfIq|^X>#zb8pF?90{(s+(#r3fkxY58-#N@u
zhRj;DHTwu?6*2LjnE$T2lT~sn4irIYw8Fo=e9pJCgV?EQDdo#0sL086oW^EvQT2w4
zr_T@YF=xI2)UJp>l(u=<tc8nbQ%)%h?d0Y4&kvi4Ab&cIZ(|Di3I6f%uN_X=XybcI
z(}Bs`r9}eZ2S7Bqy~GetfpV7Wq;H<-15_%_g2p?;f&zlZwxp3soX;z1r+Z!a8=0#E
z7RU<+9YLUN_rd?EoktD$Umy_oe}2qAH~rU-1skOT!GHVxUoX=c|0!hm@8bbb0DSuQ
z0pRujj{WZ(+=|5iKRHM|Z~=<CeM-FC+#GU&V8eH_%l~_m{UdLGagxV&fTB77d5JNZ
z3NGA|@Q{7~huZr;?_hpg036cze_a7ey>1XdhKJjrO=no?hj+^Wmi#UMm3-Sp`aj;A
ztm-WA&&~g1ber+E6w7~J=iL80_J5s&TATmTD{lo+N8WZvE@T0I2!BzuG%)qUQ+2{S
zYfpz|F<nIUt`KZ7!|Op4jCw&M=)4w#uE5kQxdV&q74v^8VFuJ)R>QFb79M0(pn4Wv
z-Isg|$C^RNskN92pa9)=mgr35A$u?;Qwec*+t9!1$ln0g!^@Fnav@TOyvD3yCUy)!
zsJ%1H5oUIW4I~!7>2fOy|9K6|-qw!#&uiQIf5*11XO!H5Y4PeYwAUHgVH<YCjPKwL
z=R%Bzu9Ur@Q^17&+a;!)`hVT8|CefT75}%ff9>mSO^oFbVs@eZ_9Kh(0}WJn#Qv=c
z76>$M|G$>&|EK2tUyJ_#-E#exmC8pg7~WVuTA4DzBr7yHEiZk0if*IDScZ7r2b~jU
zzWBF&@aY05LJktEurl7a-g#Z+A$OxKZhj{~&L=U;(n;^%-U#`5tL40ug-r}AE`crq
zo@*Q@4qZ7rFq-SG(VcWcc1bwwS^48U9Q%i}W!fV0!@}!z|Mu_a_is(m`LCCqQ1tDW
zOE9Vec+fGZwIx(XrdtoFdsH=bN0eeokkcxkKh%4hCYbl$!K8EMi_s<{*Ti|!|G8Db
zAwX9E1<<H(O;4SL&`^4Ok{;2Ak1ye_nZ=vlY{4ks=q>jK`cL7ael7s_YM+yt7uSB*
zl!9la2?_l1wR}PkxT&0baoE|`heriynD&Q9{?3Trr>$827}1A~AHC0A?68L&Maci>
zBm4icsm%qM(#zwb7x2~!16JxT9p_8m8KG|}u{M*ytxg;CnaNNp?K%T)B^jG4)d7D(
zQa*cdQfqKlFpVQ<P?&0WOqJ?SPv{X$lA8h#t9AY1Z&j+A8d9BajFhjWyNQw&+6S|o
zsk`yHNGFq9Hot_0xqRG@<q5kxvnCkG+#_Q$P4$Z0#&91eI#S5^!*eL>OOtmW<Ew_>
z_dX>Auc}tlM|0f4ySI7C;Yj9+@tB|Z=USK$Mfp8`u`r=5JT$hjv3pp&b)}&KR}lYp
zW2Mo%^g>97R_6k039)+?Y)PJ4UY)Wbd$m~yVith@+tdFu55zS8xDXp(Dfq%O8!9l)
z=g!x0k@>B~kH4?-B=9;|>E@F_vXk@q8Hm-i3zR3?d|W!`YUKQ6(>5K1os67rko^s3
z`h8CkDxVbWrA8i*bcvV7)N2@ijoVl&QO#d^ao3W|y(iOR7hB5Wf~t#<z=*Rio@2oF
zj*5%VyZ>2<e_k=a{m0t%_PeA7t|eOu^810ozZQl#6}e+n-_&>M74jDrE7b~V_R0p7
z7MSJPTR_<8>vKnH_H@%>UTB-jW+(22I778OejUANOY(fn@Z|WlDFA8mMg+y=7l;1|
z`C#nD8BT9&5OX<4a$i$9PYy2o>!K&MJj86svB@g|Il&C{Hs^wuwb3*isrF6yInCOc
zU8&?Zya9zrv;Rk-N8tNPzP~N~D>n1U?V-}CgqEiGPXD4T_rxSu<aYaj#(T!axLjUh
z>Fwa~HU~xio<e?6l&a#;Pr)HwK7PLiE=BHizKQVc8RybhT`g{}5_S@A=);sJp&L3o
zl=Q}_^#j|sBtdahZSUvBTi!NPbi}b#Fr^@{TvXs1b?eIXwTu(ebn)?TCgo#qNMux(
zSW6q5EI=yy|I8^f8PI@861S~NWy-vH)T45Ic6Ooq-LAeMx388SO~&1TFFU)X?sFzz
zQ&aJ<7SwzTtmSo=65mivRA-MVl*_n)r&5s|#gD}T(|syDFLxzYIeaG<xY>o(wXJN^
zZzeid?0tFNx@PM#A7Y3>_YZrgd+;$Hn<ag@u|*Fe+OUghP$8*I#fPsX_A+4?XPmk8
z^r!80!JVdlS`?!0=$mWviYO^Iy*|<qGL41$?Gaj$<Kr74S7I?EYGED!ferW3oX#?P
z;$LuxbyX+a78M1>E+}B1)&E`qdTwpKG*^z!hY#y@P2K-^rp}~0A3Yzs?PW2rh~Qo#
zNw)x}>BpAu%`Sct1d0ehee^GZ5)a^XP<x<P_+od2Fy(K&(1g|tYMFh3gOB4yn@KM#
zDAWH~cCcDN4pW)mu{WKcGBJGQO6|YKo#0X&gU^i=O5<uhl!h|tE4Ep)H$u6?!kF?y
zLoIeY4=gmsxMCvXnl4DiM)Gj!PAR&GG=vkn;PKtz@hX`+EI<3;+@+QHMUjm5Z)prO
z$M^Vcc{xGM-2Yl7%jUFMwfnl{on>{0QC{2W+Rx+)-;7~wo`6itN}voIGz_yNss$m)
zv@efqriXV%;>KNFwzJJdXDEvH2;Sj{3+xJ<t2i%@(1D!X^BWH4zs}w3EGy}ngJ78p
z`)ucJycnO%yeY5;F=MYWU1LxSD}B*<xv{`1;VYxOT8Ko^errU|XLEt>`J=oyiSB!q
z3h=Abq7_a;gIUOlce?{<!zK%QEmyoGOV)^(+Bv(Xv<BXdbC+Af&Y|g1A!Aql)`mS&
zLOx2mrrJm%#<qe$_}m3Tj#uq^Q1v4u4rj*TETcRmbOqQed8SfTw}E~D&SgsAS^*8~
ztG_48axY*1O|{xXMt-Q?D!hD*X-+3ID5mEbhfF%vIyl`xEXje*JvSmd)|6gMDD4cJ
zyQo1Tt}`bvXmfaG(?tHNypM?B^_VQ1$up_f8+5qb71PQ+uRo{W#1+^PC#9}YrwU7t
zxjM6We^h_>YPT~|rJHM@=VVXvkU8K>-?sQzt^=@{7>|_cyUe|uWHSw;Zn_JLL36m_
zy;H}Q5L>gcDCA@@w8ac5Oe%KgxC8k5S>tdUg@jO}cR7@EUUCIxW$#@ci1qB1%BP#Y
z>j<56b=jO((e{;|O>7T!8+Uc{OgfQUk|`|8UGzT!5x07$sWBSsMCxtv)HQ1h*upU{
zLZKjL5)<JAX>i+XW425rd}IJ+a<$jefX+RV@=1c8URh(-7mNr9!t^aLSIa%D4R9Q>
z5vo%81gO!w@^M$%enWhoUk742WM>lupl{}N<Lh*OYej2w&7-`lk=s>?1ws#-fR-rJ
zEbnsWrZT#J`2!E(Ny_4erD7yCo4LsO<H~9!xE?W~^<XKzsXD{nQYfgZs;am+2hpm|
zqt{)W-lbbYU6Ry^ogw83Cl#Z3Wc4t5k9++Mupb=RiI3h*Xl>XP&gB~^L>t0RyEs$J
z^<u(pdrMH31~K}fl1vJ!a6VF&`dn4k+;&}XoLRZ<Uhzh8WNPE~)<hbaffxK{W_^QY
zqlF9jzBxp^Ws^)@6^6q87YI>j3eS9G=CvMwdPQ^TC@<qhYDv^Uq!e`rmm*Otu!uDN
zh9J54S^GJ6<M#(b1X=6gk~Q%eG%nr809rLm#90C|Fm(U(Idu`pkS-44A(y{oatD;`
z$A{6>_=#b8voWFf9N&w#_KK?4#>S@>al7}j$`)nWES_jU%aM36??}iXpl9BEtk*bV
z;xZ{NvH%yH^X7SsYa^vf^P>lG2NXk4q*&jz%~R*fk=JL<kP|rPQBzIR{aJytMm_%X
zyJD6UqBH1~SE4Iqz@ZhlBS=zaGNlih$hG)&C1%t_SX7jAxBhs;-Cu+mqgL(Z$Ah}7
zB6JVF0QHZB<DBzOK3w{iff~FMWr?G|>Or7F*Pobb;&~Y)0$soarWN5t7a*V&FN9ZF
zecOsnuV;zusSVd)zzO33ApjI9#4%wwus@_Ad<eM%YVtG`Uo79IpjzdB;0V}j4ZJUw
z#dQvMfc*ZydIT)r>v+-{sx&_iJ<Fu3eNg}|#KC5TF(ejeS*DVv|0!A(qDq(>_$jDt
z(}1%Z{N%o)f}u;ls|L!wiA?KxsAVd>0bsJ61O+TzH6`q%bF~;lyabBR8iw-LXZuK}
zN=YNa6$xSin^>|nq<i*vqV<cSd3qiH-8`OVij6N}+op86_+&9S(euK;n+LYEi#%8I
zZ1^i#i{?qC`+xfK^BV*E?%V;5EmkyxEOsueq<KO;n-|CK#oPQ$kuTatg7+X%yfJ(+
z(UP@r$jZ8Z4Jdy}+!L&nBrLT=n(%u>b7^PZWZYfW(2al!oIZGF3ZWTI#Z%^<(4<w|
z4;yKCuOGxIus6^08`+dS=HdpamUrsLz4P=3dTL0Kk?rH3)0HW$lJy8qc9`E`obc}I
zWW#2BpP&|5b6ExY8O-<A3{INPFn~G`14aJ+QyEZ#TlLuzvQ_dU>cqKV5Qf>q)|&0u
zuE3zyg91Z7M#A@pmireQ(D=$PUh5pU*BwpGb0RW|45IwcW|D$1Zk?pBoURt!Mf`uk
z1HHd&vj((!RBeT?Ar!aEuR^AZenbW03TqSky`<3Ijgeb|d}?%fyF;d>JGI{xL2A2*
zl=_akqw<zX*NUU3=eW<NNVZrL|62{yTme0gQcbxl6@-AnD@Sv(PH@s|SHcN%u3~X?
zZEXx^q><;yd83?S_nrpO5f7w6SSoi`wNqK#ltG{kx`P@>Wd851AnCngd%&kEXno10
zXlz6z=dzlg>La)Zh1CxEHq$g@g@06c*0c)N{L6(yFcw~;a_5N^X^z~vN{PSsT54_Y
z`laibhAc%sMe`ea`qa?m5n0yYw&Io(35B*wdd_`W++Ep@Vny+>niS_PNhUiv<2V_>
zY1{*|MKt`#Bc(L*@7s57DAS%s)66h@QWzti^R_C}4CuL8e<Ty8XfM_9iC%!icxc&S
z!RA-{C@%+6Z7Y;U{J9)U{l*@Hy)yc$Ps5w1wu#vI5_m*oKFoux^YeK*a@PGv)fRl`
z;oXNBQFSM+QZ>@Qs=93t>a=rX0ai7=^i430Rh_wk&5ku8;5W%no4AB?nU1LWs>wfo
zPor0N9wuWh4W7-X;-~Q{6s6NQ>05b8uv2_g={`IEd^K_7PS*ED${KTn?gi_)jH+)R
zU;7{AZ8)7aE4FEui7tk%AU?x6CGGasJ$%eL_KswiK)o~5%N6WhM`brFd#;*wS>X-%
zq9@HEwEEL$&a7vvr;W7HW#!l$Q{1mq+Qt%wmT2o1aM}GAQcB+;GEDKg*cvfF$MOTM
zp6-;EAsP$)L~HFk#Jw#MVX%0@w%1npyx~;BI_7<iZ_BBPXP$k+=W#I?xkeWr)cKL^
zP>Ol>sw0ED)tP=`**&E0d1gHu!{q)qw48h6kFa2yO_+eS*WTr`Bd_@&Az%)l02#7%
z*<^XumNA#<55;;h)Yfm@m_UE&a(lKC1k{DOyp@Fb+8n4nU8FXpnCuLM{k0@brdN6P
z`+qg2hMJ}3Q$*d4&6~NGY~-9ilW2ddz?rcWt5HCfuxgaL+22k~XuaqWHk9{zWBSqc
zy!Ti`QD|uL=E1tHRC58Ri6wh#d5O@Ey9%3`gZy`2a0B|JwO7W;cYJG*GuP!_jA&cd
z?t3F0Z`Y}iCrg<%<N4*S9s1ae?Cb+fYWZ@E|479Wa#z$v??#{H#dR}5vdE<|=ttd-
zgWJ7^WIVk`Td=of%~qmH`!emJGZU9eAJ~GcEbFOeHOK1|yftxC+yUz^|ESvGmwmg9
z<*st~^4r};Mmj-;vqTAdw~Fi{{&7qB!MFN^d_0WvV?<-TtLcZ<hoyU=+MnIRKF^zj
z0%!dv>S+Hkydz*RgL8(gYg2~6q84DYK57D5WVL0YA0J3y4o~|V?`D1WaeGDMM_`Zv
zGz;d06`vR;#-MKc3n79xUGbu&y27H#7r+q`@`xUtN<JS?uDJ4>tBBkaZ+X4+9m>Q8
z{He0*9&PERC{2&ud7ZW2Pej}AYeZ*!HT)d5L-}s6?t<yF136qe&mR2t05)AMn%ZH-
zyT4G5o6ib*(q#korr+f$=TGh(XlD=`Jjd-zZ4L^fpz3@aTYJ(){hkIG@)qu@pPI(I
zBW&7xJvcH0h{iXb-09mKcmO+vVsVex!{}7STXPSXZa{zdW2%wLi&pA%k6x^Lt$#^n
zDBgjDCg)qGngI$0u53JI4}(wmGa`HpJa(Fn)fA_=6~}5WTlf{H;>m)|hTh?dcj*0b
zd+B&J9-f*%@aUU{X1>vUwA2f@V2)*KC`F%RH=l1Cmwmx*M(=VZyQ=dml(xa!$tLwK
z(Z}=;q-_7fFdv)YLqji7-4{GZqV%=3p`rPfx;FGBfH5;W)%yG*yU^3+%7d(w>duti
zR_)Gc_F}Du`f3?5VLnko+ybrq)oIE!J=!h9vgCA0>p5J3h)(bHM+X-UT(J{b0Oh~W
z$A1-jU~3h!R;d0)HRj=j)?u9Jh?=Pb5SH7Rc^Wr2Mf<Bw_ncYI8NU=&5g2D%1!9hJ
z%vWI^AfQ}<+sP#>uOx<M?_1Rr&^|M>cbxlesG?~;%%Ktp*(JIlx7(rYP0Vi>meTH2
z<c|~iS@g^?r1mTMG_LG}{&tZ9yU~#?F6#T=&COF`(Muyos3Lj#;NQpsY+J-jbXUr1
z-2^7-^5EzOyy<&Kg{4y&t#|Y1k_f}Cnf$H$1}j6?a-qRz^Ty-tf#Frt*;LZRK!+$A
zt&YeuTpKbptg~ZBp9JB1C!{wbVY5nOzOHL&rN~EF%MS_tV;l19nfK~uE7EP|$sgsT
za<SNTsfV|&VM1pSbpW-w<@|wIppa`*-loptD6hoNF53^@&AsI_X&!H~zCLo7mOR{{
zREK}Hfs=1ySqcKpeZIS+{f&XVYu4rXeW#<lAE#d1Ide=naU=e?y!<f;cqAJ3L~Bgr
z`ucD})X>O~u+^}Ou;;2BAVeuHGHQ7y3LCx>Ow{<ry@Y*f#^k%fy;I3lutSfvQu*Pu
z%>$1lE=36i^&3+Mi{j+n#YmM_vg>vEd2^S?8>QO)E}i)#lq;%wlT)9xDQ%s$eC<Xj
zJ9ZV@hU3afD(Z+7)3WWXZ}TgtsOsOJ^x0i#_cb>-2MCzcMb~|fyO%#nIkNDhtA22c
z5{)!tpryXwGV5;kot;Vx=sjWBBVlQ~_ZE;(-F@yVMH?mw?SuNZ!Ik#lo~5VyqLuk8
zW}AkmBI?a9ixtUUT#c?lPJ1PoW5eW+C^5$(a6L)QikxuGa;(+G^zBZI=hL|)smKX&
znXzcpH&mV5O$VC54$UWc8iIZZEPW)U_l1ma{C2mpIy~Af<;8Us;WizWUi?mfpLkB&
zzjmR)K23XM>M$Wn>q}=@1<(K@WohBJ-O@=+3Fi@ZfSTi4NvNy$PHtaQxX8`18byY{
z5;qc#8{5ENFXDHa&Y8Qr*-Hz(X`gpS?Py;KK$UeXZ(3Y@|Clct$Bs}z9%Yf2&-llF
zd)6FIqjV|h_9k=^Yj^En4;58Q2|--zpTr$_jI_GGzZH$27LI22+u1GLzb|ySDkY-7
z9%F;OITRfg$sBFxY-IT?@@l;&N6UBS%Dix@HQTEJ6C{SdO>=@&Xr8>~o=Oe0T&OcF
zYW$nDsRUOzsFs>w98IyKzzR3<%N<Jjnr~Sw6tUf|m~i-xBr2AJZQ7-0xe~R+We+$f
z?Ak+_wOc{S!A7b1`2&^a4((QF69Y9b)H;ej)6w>@KM0MIan<z1q@6b%j{=Tts@7nH
zEu~5}<u2^xZjPcYXO(E)`ja!RTy6wa;1iW6V`+UFhyJfknQ}$3Fcehf71@{97I?)O
zatTeZiWgGhMr7BYCa4;h)|K0KdeE=3KFF4OyxJo%;XF(9){jNx`uN?c^#1HUXP;(u
zL)-@hJ3xmB(|T02$XMesE2O5wT*HQtU!$gu&}@^ozmO8t`ymOX1o;~N5-hOl9qAPd
zRY;o~UAIp7>Kh}kAq|uz%T;I8bdWLsAgOm#gZtcGDS8ggc;wOEcwE;AuPdoK)ixyc
zEVHnksTSItG#$qUz1ZlDdbj81vAE<xW>9zhxWPegfd=YQUcS3B|4eks|9qk7GA6Ig
zRTHh1+6Zx7V94~n>S>5D4?cpp`$1OEj!J<zgjrDf!2SHX8Jl-^lo~@<L${^VhP$9$
zy89um-t4#42fJ>{2G@IH?{>25qbmr2Bmbo?1B)$0W!3{lk0udg8`)6yI6sG=&|lc6
zGKEZOI@W+a2_X0zsKzylKF^>dh;?Y2yS4&Ht4HJRdq8>V>3>^oRZUve3v^07j6V0}
zA`a`I4rwc)sW3FuJMbNMw)wogHb(2~@{Ta9i!r2aMKWfp!oL*KJ47&6<Eza4*;UlC
zwkr05_qz13WRc004lSS%b}oW(sqmQkdH%E6Pyzm7)-cwAuW9?FOIs-Mu;X=j<kI`S
z9leSeld;Dk+7qx!5u=55!o#`5YwIH)1ZvXE+TUh<?H5~}mUe?MB`Dz~qNt(BY`NsB
zB>Hj5B$zdrVOqYcAi-_YomAbbc@JkGBRBt7aj!a)9i#JShvafzjSwO;>@+J$^#u)C
z+7!7A8>%cHCfNEbo)2(-%?%M=OZf1SzZNZ!!IH|IbmBId@fw(Dw3K2hyUY0@TV+?D
z{4uil^v>oZjfZA*W{lrONo8!zH;_z;s$XP*><3|tGCgaMc%FUXLIu1q;kDUP+=Tep
zo3+0g$}aGmkF3JIz02Rn-wRdtsq3pAj?jQuVlvNhHTR{{gRTy70zi&vfO|T#dPBAa
zQz6`s?{R$tj+0!~#JEN5v#9A^yv7F;#g;XCzl<-BEexp<Z+ynFLmv<q`=?8W*O%2X
zW+b6PCIa~t`zPe9HI?1yo{j2r*K{ySu?GzJ&oYI~Yk!`CT99M)Cv(+k5pZ0G?^0|D
zQI(*!a(F4*?d@YeBEn=aJqSNuDt2~ydiU4*SBQ!-gA7@+JH+$6Q)Ld}9Wjgk%BiF7
zi)5rJ0^JS=d#tApvp3lInW$`EkKIio;E2C$V%NNPqHv|x%)Js+J7$_l3iKm_^1|@5
zd!GF7Kk`W)pMyO$K7;WT5&JRt7UT7%i5o^kxFbh|v1BncO)dkTP0TJ+;E<__&0$H9
zXWm{mxA$1S$!AN&YsV$i%u`JkuvIA^zMimY{tz12#@(!U`BVZsrDm7rIf-TU$-{6Y
zFL2<&+6VJV#_g+O*J;gzwd1L=p}Msz-zIBJAjR6HaoKbPVLs<4I%VD`v18)9oy8=t
zAgkos7JJRl%(ckN+1JOh2`P5pxVVo#UZqy>+j&7qF?)8S;^UEH9oC(Zy=c>U6IFBM
z0BtR<i&3*Jw6m9Miik0t6bksAa6PWybtC3Y^EHlTI_HyfB`QS3ql@aZYC-;<F?qgG
z@}{kIwq+{ta)(oySMeSE)Cy7_HeZ*=qrm-x|9XGMr(5`{LlbSx;A1DWY9EhoVZRpf
z1d5tnRqxt4O@`^79z?+`Qq4@Ae=OYJTrYx+8A%CzL76<3<F|{ta0$})qPfkOUdJ<E
z6qU)2x|2)Ogv_Q0UaXHXt^Gys_quSh8rPpuYyT*)zz3Q3nj?cXlX37qA(wBo#rQCJ
zlbbhW--{y4=~+gfkAewpXUsPDjP0!U+~jF^BMWV8_9n&C@J)${?`GY%=`Ot_jg7u>
zVYCx35atobZ&<evl~I>D3hiXz%W<Vn_mbJ0%Q5DI@VKN#S7vFD++j{e!#->HTrD`+
z(fd~2)SLF5CZiTAQs*EGa(>>Hm-C$~&e+;^es6NKo`N2n4{TVc5NCQH`D%frqdUDW
zx(1cYZ;aM{i|x9c3;?Gj86!c=kJ-(%nzZ6kcWQkq2i>|0VDH*$?WbeNz$HhEM_x_L
zUI%hu1bP@ehqpyYeD}Enq}=){VzK+jOe#6v9F0n|6wlfYVD9b$4Lw<+`Yzu#b`NJz
zW5;)eHALQX{#c9f{FuD@#l-Q1AT$X|>_{1tXm+#Gb?_=v?6bwt85y>#WkKx476QjJ
zF%;*((?AXUrR?t)4=fnfbQRNv{L;10b*sr9yQd8}c99p0rf<EuVY=RgISyIowxpsC
zdlkF5<*u*0(dY`=VC^buprIp<?kVhfe3#{tp~S$v&AYg4E%kzisv;rPXYbE?BmDcG
z69K`y>WEI6!q&H3VC5&)!=n$~euU1)8$Nk$TXADQYGtjVr*xXD9>!vt290eJRm~rY
zbh&@o6_@L@^ekvs@0DoR>McmK10f&eV}S=+FExo*ZvolHDVOxfd#~CCO8;DkG<A|}
z2+p=~0BMM3AZ)mNEmVdEFiB>zpO`04wqS&l8-S1m_I%B4!=)%o+tFD8&HXm_B)60P
zLo^xXW<RU=5KG0Z|0`_&4MiE{xxz!lc%;UZk;wcwkPu32p$jHRfZ!o~o(hl8C(Xp)
zd)lOI*I~c61kM<N_@m-wUbc|lsWTlqO}<1HBOpn^jcKONSQJ!ZvPi&pE}PTD-ZcBY
zyHNopNXZ0OuQAPu3gy`aC&3JoU~SqaL9n`v<u4ofW&n&zJdpHQG{A~ecj`~9YxO<&
z$yZI^kxiUjeIjS^Q5yDegE>Qg7d=|w)<rti2#>19@5>TgbpC5S8XG=T!5m}zF@Sb?
z5(&c7NURFvU0Z8w#T7IxOoxai96)4y*4m|J*G#5bOWdH9P)M(riw$6x7<{X+<#FM-
zr4>q??fuUJ`CPny)e3l4rXeodWGcMwUid{u;j{zA|NS%%peD8S9HZ!(XNidel7;h8
zxaS=YA>XvQn<~*&EPLD(l$2MO1Jg3Xq7^de2Pb$^UI<3msmVR*#mEW)ZiBbk53WfX
z=$oabtWYsskqQ%&Dlk7QL262hN`sm8+0KV<?b5UmyXUcVM|8QM-zJ&+h=ZueFX!01
zKx7oxW9p>oDcDtgebHIF&SBLN$<|5vXevWyS4lT=d@prmuH3$mqB&g>f1-Fto?PlS
zD@ACNn*TD#Qup;t+c?)7L^8ghbYb}V+a!*z^`yIf%&-3XuQby7oCA6ah({&$^F42!
zKx1;wtwGkV$0bE^cKIuj=TE+Ou+4f)Td`|hqF}lA;wFd`^a^M_eGmP64JXFJ)`;|d
zdzoMzX6J6jh~{ldPpN5}{nIM{V+&^66rcU@pjnZ4O5J0b*1*UQFAwhk=!e6ty#&N^
zFo9M3ZW(*?)min=wIB{&^(2=wvL_>xpM!m$2i+_5NM$2DUMwnEq{%gk=iy#+P_K4z
zos@1yBzYC`n}%Zrs9kiwZg64st@5dd%Le-<-~JP^zw^=E*uI@UPvqt}3?}HRVA7(@
zBjTVu*eti9@BeyJT%t9CW<}+$O7`K@hVdY*+hD(eFlOnG^&Vx))Uj!PS48(RXnD}4
zsOm_?<-m6U?a*q(y!N$Qx1x7a;6jl$wDfUS4D*AZY4<=sis-Q%BBSyA99MZ^R|+Di
z^RZ<sUyCUR=^2IOLLER+WFEq`x&FL&=5T6%=J$u?q*p_8qU?InfY0^nWmh+wwM3Sl
zc5!jBwO$EX@AxCwk0ki-6Pm!}O{AfEx_@J>z@nU{tF1NMq-xPzp#Kf439UbI7~C{7
zx2(8$ch=Wle!t^pd#!5r^`fYuUN?*X6dCp*wTm5(W^EaJcgHj~ZR{_ERB^E$;viC!
zjf_1?*7YU3)M{Vaaj$^2{4}JQ=jl>3Eh52}Le@3qx6K0sy8N&l8-7DtPd$kn!<WyF
z#erD;BW0=`Zp6g+K|d!S7eSSi@DCt(kj@s{R7=zymZB@#LIYq_L^qJs<m7oeEVr6W
z(7Rl(Ss6=u)+&#J!J_j;Jxc}{jK$x;Xc~@t^lVL0he`5o04(_psCsM!vnKB>y+&UQ
z)<)fUgiz0khq2D^7Je4gA_ppj>i@(L4*Gkp-b`r(NeW>e5PvPtV~;P?{*_H`3;)w!
zFf3keV97_$+6152E^GfFRbyVlMWbLqAlB386EpH+HRXG3P|Xfy@jbZNYz=+e{(fA(
z51dHM4SMFDlE900lKSlVZr1Cl<kwwfGyar~o&m4xkpc$o+yzr&e*5mG1U(^ICa(+J
zPRGG8SeD!*sY`OxY<@6dNYG!HWTon|YsoOZp5JHb=QA7Hy2B0LoqJ;FZFe(BXMl`t
z=Kz&3dCc+fmN`kd<J;6Gg-lps6;I;EyJO=a5^0rC$AjgFIK^qKfL_zOa^b|MFaq7N
z3>7i!sq=%n`_Tv11NG4b0$#2;c`BZhfJXBpVGS^GscdjW1dbJ)sa=VC7Lk}f7s0cg
zP)c&N)~tn>cuk1Cy9Pw6%%Kgt=;7!`lPA7WZ_l0uZkyb!6sMQ5z0sIBob$A_`l%Og
z9YLbDziN{AywZKxx7u_6#KE4CkUh4^uE|OC9suTA%qb9A$<ITYd7pyrsO?ZzA71OZ
zBK&<F%L^CYdvYqOj}hWgr3Pp3AT<a>@;ic)Ea@peJ6NQWrV2ChJ2G&2M>A*|8F*OA
zcHpiPQmVcdy)@1ICF6X2rGM(K=@~;6rJ-1Ij+>c4#u~)XwI8`YK~gIrtgGe^kjRN6
z;B#Ax@FB8Bc96;>ALg46;Nz?S#h_}P07mtOb^RJtq+}qrbj0D9gkWc3oU|6#p2x14
z<y3Cb8!JmKx%`YF(n1;>VvZxNZ4Mrv8h_Gui%4GZ@|rG4g2twLPTHvlZd1^*3X?hp
z^|qerKRp4^IPbF69t$DAQ1`**sn}!6y{oGfRaefYrk$|<GeIBZK0bSgu?XXkEG?*a
zsRd$vav3EV^^tl&WcN=6`|Q5VWT9`as>DMc7sDz^S8_6MiD%W8PsL1iy<ex2!>SOd
zUk*_S?&g;nf9W~_G%~~O)k^@GxS?hQIJV5cjPE-{9_lsxdLb%~|E_sP$|8?^UVFi7
zU#Fzvn-YQ@C*l&?q_#^Cy3<7)OgUNbOqZN?OY1K|sJOSAj5rF}Rb29=K~9h6J{&LK
zL%EogExasc7u;uaCNq^~ZCj_BD8lp9f3h6B2VTBBS)(V!3qeh#XfYB#uKS>Xwd+*$
zZOI96Rc%j)4AJ%xw3X~Ph7S(!_htyoi4zed)Y7rIGdv@S9p2}^1w0a!d!GA>OK^d;
zcknN8@Vt@1Po8zN`2vf0PxN<a6f;CadC5FOa#@x+#>T!cZ>MJEC52?(edfY_oIT~|
zo}RM<_@IDr-m5M430&~hL>Im2+{)jUpRL#G!5i|&xV<Nz2CvSx?j0!IaYSQ#wq8}-
zuKq>_=I*EZl9Tj>=CuAoqtH;al3uYaA#TX*XLY@LTSk{5Ek_AgkQmSH!kat8<Xxnk
zkij%hxT!Pb_O@q-`D@R9TNUr^@7l|eY~BIADn;)II&E%X91lzMjJp?(YJ}r+=1-?)
z6Jhg{wx$l`hL~bQKi#Apu7I6oeN0F}z8%|TPbb!4k233$&qYYm#aw4c>CH;tl{^xp
zg65gVTikZ%SL5{2tFOgu{|%<p&Iu0PMy1u~Gi!T$RNHJX(he=Z#e*`>@I1-NNiK^%
z#~oP`fOeLxr#2(K=MH9uMoi|c0RT$zS_7vmY_c|5TnyOoFLIafP(_OTWNXE>KazKS
zmoG%o=<;szvh_Ns$k9dtdULuqRv-%DBkr;A_y|V^AFJ3iTK>w4ISGjoO+=dZ?SJ^(
zPzer}qZe=X4)&6iXkF2a5rwW+%|6#<>n4pG5lr^0er(m&n*+I=Y_`bG#Jw+8uBK>e
z2-xXE!6f~lxSWiN%#Yi}I<JHbuls#Sgp6L&_}y-OSH&@asUWi(dS_3u2|&B)b9Rv9
zyLGdGP`E3O^)ph7I}=ZGy?JLLu5jiC_#jW#-VV%F)pD3T3r>J$Ba3#tTvpATtfZvA
z^oksRH?jHZ?uiK`xU<PVof&IAAGB^86AK=r=&&MwajP6~j?&X~{e+!X;_GMSjOs~$
z(7NIw+4U<?7$*B#Fo2pYXf>K*kw|he5dZG|13dy#iupzs5I=zHBtR2aiiWBk?EU!5
z5E4hiYEjzyYawo8lMK*O0*+&@d%p{#R<L0EebS=$xXNC(-4ZHRkIn@A?G4dLm2PwM
zr4{1mpK(F?41@lPfET@blf(-YoNT$OxH@Zyex08L#tpe45h5Pt?Zi@~jBk{5+bB{7
zu_&nfJFk8zZ+F1W&4rdUu)h<?g)%rMN_DG#x`o7Ianq^mgw<=H^2pF-(D$N=K103c
zdF7TVt;H(G%em=hdw+MeU9GJOJ&@d-xxWQ%2IQ)}E_LFH|Htt_QA?7aFTbB?(+wOF
z7Zv_M<*t{_M{;Bx)RkYOw)~<J_9f(Kbg#U_KLISj5+~mXVhNN-N?E*x=LN=ymQ7K~
z<4nKJxfM74-d)Ia@R4YWo~WuEA1@Y&S^WXtGP#Y}Io*@L)}e_>RaLPuF+b7Us~ttM
zkNFF|Zf!(6|J`?}OL3TQj6PD0v7+l^iJp5o<c~S?ifgDuFnw(Hq5;9<3YLz(FB|80
zW{oq%?MX@b88L{$j@Rsx51%L05Up^et{Wmp2V2;Rm`Yw^6&D5FCLF&ojW-<vPKBbx
z^`Jio28k&us<X7NHWMv=ya;%r>Tn%dDi%TUF|+nK4hadfuV#zUO8JEo1QSNOJCp#w
z?HvHLM&dXN?<-1kE2?MH6=j&PIlpr57dd9A2qWp3{WYPruIm|KM4LD1vg|Z8cua8=
zSC_H#e98ay1)|kEq_mY=4gMaacmMRixc~=4nv37XCmmTXl!;#EBsJFP{od=mcd1%x
zQ*mR2c?k0Sd&N$Ce+zc^?^Q)y>lM2K6Rzp-B%7jr{AFZ#<d7bSS^7%0dU|x69Snaw
zLXl3dcybq<e&I<byF=+1l9swBD4;fZFUx0p2~HBIx?ZP+LG7<N6IFQwq12;0RAPCB
zx^H=dMzux!ZT<J>mwRb>*!rYn=0|yTY5y>>E8b!LSWEFk*sGEaEhu+nj|1gBoLtCr
zHw?otH>wu8-TA?c0P68pXx20bWhbV#C(LhW7jU2Hlv4vukRPeMzsDzqI1o|kz5p%#
zlOcrTJ>dI)h<od(sJb?8cnriML^>s<L0Vc`=`INYX&Aa;FlYf0=@g{9n^EcR?(Xgy
z-i_Y3x6gAw?^^Fa-}(+~F=yuNbIy)yU-{eDwrR7tnFN@J38tN!y8T*hnQAe7WXWYC
zJEQHb7x2mqMK6TjQQF>m<hjC-9MwpPB^g5Gdi!lzb~5^|$N}$(!r`jW5k|}kQC#^7
zE3=UFcIjqj)Fi;XJW%X2q=<kM1^OHF&$7n82Jz~iY}W#oDESovM}cGGY3Pg15oEkO
zyc|isF?J6g@2M&T33`s;^)bAlQKT%n94bDx?0)chdibepv*nPIHtdnaB1}>S(Fr+@
z2rf@*4LF3L`vRtbC~euWVWPRaRKmLLx2Hp`4st?&KVgDg+kPnXB+5b1x}{`*S#gWw
z`YMO#y1hgH$+2XAOw8FQ?9&nbNTgHNsa~Ab>>^RC{s1+i$&cN4_>@n*dxx@z?8}F4
zB>JAeS9G(<S|8Ow1Q&G@p{0*Oc2{;)r+mLw`5q>~YhzR2SM7vu#p1<P))bd8lC-4G
zEu99JqK#pLjAWmK;x9ZrLcBPjJh8>yvRn-j>IgD|bU3ySYq7JLLy19(!<X7ivZert
zuWzhGYmY=s?U(Ekp2+Il<Rk!ZVahSHkYpiNMgu^;Vxe1B_YrTiU|-k*iv{3D)Xvsi
zgyEtA*bFZVW(&@?E{{8wTv_oa!g;ShIok2@$h(0RnYlXXK9IBeL&Ke<;)BSC`x_$#
z+Z*mBqO}ex+LO;9k=pC5s-bCz>TQyy{)MaCK#%vMv`jBY8OreUVBzXt-vrP<5Hrw_
z9)gM|ugU`$TD5;Pqh+*S+fcrdg>mW0d~^<mg%~x0Z~Q~1m!WwWUaHBc!c&Ey&uwO>
z-U*-H7+3*6E!mMYcLXEW4pYikgoN#9eTPQ70EJeKYloM3HPu%fd-L?;Xp3v)qBU(_
z>`^#9S?a5HfD)DUxQ2%;5x!NP<R;<@p?2hF<&v19#*4ia%9uS@r3Fg?KL$os8W72Z
z0EdQ9X_!Wygs`0bwPq4&92x|1T}L=oO<j0w4^1zdJ<k(Z^7CHf-i8`t2-20|_}PN1
z!4Z*bH}<08xD794;d?*QBY>3#*5f)0JNEdJ!1!!ygwd>z24nGMXkMRP+Tl^W|AdF0
zrtlm~@M)9F<jMBu4dc$)Fah(-EHd{S)QU%IpX|B3Jkqw3#Pdc7{2aFO_KLy_;JIU_
z=9;kN%Uow%^bC}*WRTa30Gt+g4WGY{^E10EvS>VakDeHck+=922t`Dso?gk49{BhL
zknT_8o})xu3cYz*%|<m)>r)?CtO>jqK$HWI{W>C(c4?G6f`zzhk?_qj^!$s_R1>3C
zdoj+N^vuj$i=12wtPsG*nFRPPNIu}M@UbXp_Gfu6tP#;u4wg)j0pHm33Tv49WEY0u
z9XZ7at;fd;_AI8U&yiJ}v7%l{52!fii8FF7LIEyV9?jVipjpD)ULl#*Cp}S!e(D%+
zs+t2L#bG0_0wyDqR^uv(Vs5zsK|Mw=q{Qo^Y$^r<3604IZeO8lHtqb)O7Za$4l#DS
zw=%Bqr>)Hu&CAmC^B?RAS?}hloxuW9*9=#mA_cjZJ7OTm)Y~l<P?}Y4Uc43@c6$5J
zJXgifO=E#3mwE~~-#xF-oa5XDDb23UovZc$MR8#vvAOZMrg#RkW}#Qb6eQYMn929G
zSX7Q^#_pQca^94ATW^JaRC{vgS_`GYO+nAqTDm59eB}mt>}>l0iwPapMHRRyWg#BP
zQT-a>PczR2o8<B|>$)ASFkr$POGaWA_pmRWCfa131Rg*rHv$1@z@(1mR<-xLaO7T2
zPZOG_?!ss__HNRR2i0)HyrDW03hYiZtE*u|^SU>!dro?Np1VEMbcT#wq)q?B;PAJg
zj5fiXQD-TX=%PlX*37uG>qO!aT-SFH!_&{U)?m&-OaVHO^_`~cRf{Ty`h&nrkY$>c
zWv1CQt2$AbYpp_f;)*OO;kI_@QWC%}ge90F$9noUjyXzY@Hz4zZFKIKk)HTtOMo{S
z?c-cWt2lhd_Z^_m>c0gzk%_HRV#Jl>Me`ZaPS-I@?Umvx_7Koo*7?s!grz=+E2l}y
zm`8meKSQ(%pIjlB-0QNHe=>qxVctvJXV#~kT)Y0&T`c9CksMug<g!ZPxfn50?=5Ow
z?!C!R4O5dJKD;AbV5-j);m_H9@XKS3H4P#qWV>Uj_`%}xWWcEV6)~}ny(|QXHp)8w
z%P|QaEwUXA_o^4=3^C6a%Wf=CNAP(&s`<pAXvHb2(hkI&#jt1-2|Axu@yxG5(=4Z6
z@Z?=&!ACd8YbxK56`sti9Ijn&u5w*uz?NUZO4gZ@!1)GZw#Nlusy5>ihT>1=q^5qt
z%s3gz!Bn+%T(#*q1#wud6*Cay7_B_FqcPU8slFY#_!fXwr#@cfhb1(tfYIZnQ-d@u
zd&$<^%U0VN7e-^yJY4S`e_X_%{>c`L`T>pDT!VNlWx9633vX``h2S3iv*YC9+qGUt
zvJ8dZ=Fa!;6A~=jT0B89yE)q+sjnZA3Vg#bYL<RZ;C~#naQc#i6M${Vk6;eP>3f}+
zs;K-+mEU3F3hQ#rfB$u=QdBJH^S85WO8yuXoeuJt9;P~AhMK-j((nMGsitJwa6JGL
z_Khr~+P_@44_9E%L0kd-0ifv6!A>DI0}mT`Zy!JztvSjKJBWwaI@qK&QFa~$<9+#P
zQXh;VgXB;j5lQ<=cp#kCdk$#BdrqeL29?jpIP(omPY#;B06-<n(>ebT4?Y=r#6vWH
z#vn8Ls*=}KrUA|Q^?fhHu8e$H?VD}-l2mhv%-7KhXyatFu3XXNqc;R!KQc}x1S32G
ztT=ydD+C?bhy%M5_Qh@;!=z{vH%0{28Pj8cpG^ITHma+Xqr|?6q>mH6=kJ(RbI5bw
z>Xv_MMukA~uBHfgHjh}1@6glw&i8y`sYV1qe==WlOhMV+4m}LvXRTL9^PE-ZQzfaz
zKMSZnEo?R;_Bzrjx!9Q8P-aQ6s2Lx0r*Y7PJMMk4n9PZ<Fs<5)gYS*0Ly%Z5J2TZF
zKKFIFRaFo2#b>%0;|eYFUM?#aGJopj^w?!$&ypm?zUe7ESK96olJjt-bQ70A1l1LD
z^ZBvpB6v23-Bxg;Ye5dYr|Q`ALOjO9$m8^8q5)+J2;c+Roe-?wWN4;`5OXCS1vXu?
z)c5OLwa4prnLaA(7l}sV9kW6QH(yU9yk>?6f?kTvf$-GxC*kcJVrH^u{_}5acXrgX
zRzrP`wXWDd<J>8Tmn8%TukBU~MJSrz)SFKZ7Qlwhp^1{!Kv8*1nKEGdH=vC!4=mNT
zF1c0JYa&b6q*I8|qIjc=ns>KR!#~#NoNcHqu@XsM_*{ErMrjkbaW`DF_f6MIM09S9
z_p8ZHU3!&l4e5F%Qhdu997a6CpzF~b<j!Ae%vgm;*}KM{Jea^9_>z&9Xqh2}j_KWn
zB1OygF8PPySJ@@%6a{@46M!!(dYQd9?200UTSlkiGP*+hnENZM)tiy<ke$Qe=vkjm
zi(AF%oC6Obka?0N_{VFVFlz)I!IgQ1^FlnCafkQboN`FNGj<HdqN2eQ!l#j?hO4Fe
z7GlA^OCS7IH^?&%qN>i1um!EzS1`I7n7Kux6|i%<EVjO`knMygnf%}<G$AP+#8}%|
zFEHYR+}(^iqlh=!8SuDzCKJq}igm44ujJs$acs6)x^v)4*5y=EaXBt>-br*>&i!ul
zI<;@6xN2+0EBmt_BVDye+l8M+yW#>8Y&G4WRC^6e0^7TA{^_*o3dH1LMrZTGF?@VQ
z$Hu5%fLB*Ekc>phH;2@3NnM05xYk_a*uqxzR$%tnDl2xqY<c5F<;xG3WIWV`16hsn
zbcb`I_l8c^fD`be>#3L!ii@$8#j08$8pa-He0w6+75jM}Pokzvcj)n?eOZ`GiQy$?
z72b+_k-M&e`Re4b#92@ppP0IG9`v^C1mWo3e$cT$3kQp?W*^60=1G;WM<+xZ<obiO
zEMO<H?%KJPC0pD<<%8k5SIUlgQ#(Ed@e~fV?&5B@<HnX-hNDT`c^B0_PXyq>DD~qC
z26)1UNrMrXLLO7!F+SznKj!TN%ETcN)-L^<PdflImJrI^W~c2!)6BN=t(Ib=>fz6Y
z1{U7&v9_-UGN^ntlr!|AQ!~!*mF=T6RLFHus4`?yM(4^l7)eEaQFzK`(MaUz!Rfr{
zY%Z!8cIaVB?O9BqwjkYGN+rsJcpegOE*7{L`B~(%0%?l>q+Iqz{wN(cM?-OzChZmz
z<}+sl*O=3wrD2JFj05A4N|ie{C#Fk^8SWZLG8XTz9F7OU9T<>S!X^0h+pA#1Zt&HA
zq36*VW1F~qG)9ltWvb&d2IaC#QK>N5IW`WB^NsZNM3EK4Z_3#e12(hdt+Y@{_#9i4
z#bVh?NK((v!5K4mOJYPrk55c1O*NVL076|YZrA`N(7B{t5a|wuR~l4tp26vvR5b0U
zN~Wqt-FlP1A8}CLbUt^F%d@;roer;Xd}Dvs4|hjMXW9V00ybP(q<iuXtUk$MYjnZr
z<qU^UzBs)lC%-9s>~6c=ptKM~{9$x+#c^;kPjYfnqTHyT<bc{WpYaLXeBfuwYf8jm
z3twV<H>#q2ln2gJ9!<;%Hbjw3WQJ>_@12Yp`n&<p*(FsPUsPS+82LV3*NVB~j#!zz
z-<mgS0@qc&5;fiZ$ED43Tk&7C#b@FQn=V?FOPJW)FbVK*RwnkUB}styX)F&Ve0VT^
zsV{K$g?$u_V}Y4xgG;=_`718MAQ3XlE|=Av`boN-EI0K$*K?$f&4yz5CmNpQLD@~&
zR|`i(^rXrqXAau_fp`adQrG5FzvdQceLtrl8h&icP2~V^jMg1Pa_wV{?G;nXc!XD`
zK)jo%_1UM4(ff65($alWeFt0#i}>viVmOI|%ZhfnQgLT#*<3q;p;98|kA1IWCpwKG
zcBm6mO-%hh*2D3{Lw!?5yWJ^?2S+xvlucEhP05E7<(v+SCx$gRrbkH*ku4_prsD2E
zNM=?so-M`hsX%B}EiJg4_p{<)lxHSL5J6yATm>JZSJ9W?yFh3O(r_R9*A_5%cy()9
z@M(_W$34RWi_55jB>9W8oGfZEddFrdKNugSPQ*-{_Elvt?7Z;KGkA!9Jy8vnbH$#h
z6n^k)u>_G<CHES}5w5#Bg}QmqKI)jdEV;(e$iURRLXUT6Q2m5|csGd=AJKtUNAAUQ
zMKsU1S*|Oa-vKm9pdA`qR`b}Ri|ZY&Ol_ITvF6xC6#(Gp<yfF~S@Li`sN#Gweo<)d
z(UBrx6g_9*=S%NK6Vr_jIdSNEc)4Al2O_SUBAwO!uq6PeKH|Eo_56Od+6h`#K(vE(
zuE5|cV2`75@!U^rujxr2vTuTgXsz=f&o1V4Ci*8MujNuA$TU?WCo4>1^MIq_=y()?
zB3?%1u+k;y?)U?(P)@%`NEm_{JEAZrF{L&ZFU@CW_tcQb*gi@5)7lfwA!12h(-Nwz
zQb3V!Imq2Zbm`Sg$((2W56o~qZmV8VNbcP#{(0`}Os_i16mv2MCzr~#!^AZU!c^GT
z$16-%W#)S$v|opwvf86CKiF}+$ZOJ(7`fcejC)=qYbjFU)KZDA63%1@BD7@$ygE2{
z@<vxq3^64D^ZKMPv07u#nf>Ab*VA12+`cq``rFrcN_d2$_uWB|VwZ4h{@#QpI}lnu
zc3ORG(nF%f!|;8nx=0PxAtDAqpasM)Dk@D(qk$F7pjgj_B7Cve=feNh{axg;_N#uK
zdNhv_ql==;K+IZJ{-c7)5xh)dGx$bgcxbIW?w~8sr!yQr(Goqsh#`dYgGG-(znsK}
zZMX5$Mb;5OspaUs?iA2w$kP1;<9mGov5p!N;Jg4ECei?gWnvN=W|cTqJ}_9bqD<G8
z=Gd1t;pc~&<bxkDs!~TPuI5ziI+w+#>DaWnwK=hGdeq2aB<fSR!RAS3a+h-RNfHs!
zL7Y}WmDWKVk$tDYerh)UXzvo&l!?e1H}_gj>Q^&iBO@(waRsv=I}{^6vEt<1<f2--
z$t4u6mIL2S?89q``SBT1ui>r-PMVH;2-bosBW;fw*EuVpPYvr~U6>^~3q(g$<8Zd|
znw5j2Q`yGN1m3;fwFv?XQX(?4B4Q7QF_|09MVZkgOukbDi0za^f!K3%UI*Cuc!ogr
zp{?e$7yWSyT0pd_%fjA~PY#cyCP4qw_6>^q!J8R@4ZVfuJ3%zuT_o_tU3kB$6gv}|
zYf1!fP8vEKIv{l0F<wXXfVCiJYvxz)J9XPde`7MA@$k+YT=i6_uW#v)b$jMwO?b&;
zJ;#p0w4kznLR(^S?oG3Ket?^t^twCHJ)45&7wbUpW=u1O7NT(F{2B@-bS{EYJ6y}L
za#FsymouPDHv9B#c4nphRKWrH4e?X{!oe^rr-8YHHM5Dl^RJ_)T(6Ms_|yep^VO1-
zb~e-E;bT0&<L2wt*nsuNZQ$#AG?K_1Ty9{2TU~r+yLj!FMEiBWdu@~sy#RDgrgB2B
z)1Eo%-b8$bACQ<5$06UzOgr6EvsZ3;DNdKLrE8L5<C;0Moq38gsH&<QSZri7!3Qz1
zj@#%pl3tS0Y`pxu4q`Mvz2Cl(WPnstX~+!Mtyu^BRbB>|Sqa(*3}EhqHASijCoz%K
z*{ApGq(+3^@D|o=0LS`y`-j&_JiAS`L{yj3>TMV1Z;0^-MQE@hU6`_~4nKSW>rnib
zjLbY^fd3GJPTdJ3Rs87=hyBvpy1%Om#bC_Q-Y2gt$Qi3CCgv_~J9sl$|6<>9^ra2=
zTa&f04j0YYg%>Du;dW4S+dMH5*ix9cQZVpi`ostw5Oo5HD(>TI!q>k<Y|s;?<~KPN
z`?;^DKikocqEQbQ&?KX%B3Y78^vhHec6J|bq7!@oNBQiI-6P!LPD5}_j?12j@4TwM
z+}|&+%fg&zN!w+xjwnlugsU%Bt_4JRD(3BO_6?SH+#9Q88o|ab%Oxh+<mX>kQL|r6
zS;!#1A!m2QyK|8c%XgY<pGaXdGjV}4DJ}1`6AXvD)UJ^OyoyCw_W3JEE$txWxU=nc
z<MT<cL)fl++}rKlq33Im)+U0$y|ss+3+4cF$>XT^^t{$S7wtZUO41RuQ&UEKu2k*_
zk?!vI>Sc8)oU!56;|DeH0uG(HYjQV1BoG7jS`B;o;9>g6ESG`e-lPMZeT`IHqiHAY
zAX-02UuDc}QnJG1@D&N@{%3vdLLJ)#6J<8f1N3js+LA;Pv3r>iQ%8p!rXAh)^UWT0
zre9bn%obvdpuTawTT*c1cz$gCK)uSrV)PS2lq|$*_^j2&GZY`v5w7U$aarWD!+EA9
zopl9SQ^<M!S~FeZ9wi8muysUTI_Emcw5VWltEf6c#UG1`h%78u!i;L?I;)g4d{Wmg
z%P%93=B=`@*1TyQfH8VTZ}*_bZeq2)j5D`_<0859+0eqXsQ8ByMPu|Kk6TW}9q?Y0
zD!YB|g~rSZWbf73Hp??^iF(F}tW3M27An3K$UeT$;J$A4u>@m$DRF<Hyi4APl5&|m
z^rQ~B*i`{D{C?97kwuNjV}R2-;Gn0+2U>`oD$wY$i_{xI#04!VIhT~XvP&D^)xJo|
zJ}=8nupD#vI$R@&uphKP9@Q_eH<j9X@Qsv=?4a|0P=BCTwBmbr1C`+Aj*ASJz}_<0
zrAsZUa`X{LiPpi+(He0UDH;C_uIr|B6;II|vn|d{;_^OGlU(Mkr9s&dQw^I)*Vza6
zWJ%&*XtvU+kb6RcY%gjy3aAqvm$I`lwv{I2-kwgU7;{AgGj#0*nB#M)4Q*`UsJE&G
zLauXh4C}M+TU-$P1jKeYcjH{GSWDG<wzBNvdU3h84r6j)`e+2yIQ<=cZkOepIWyii
zS7a-~Debu<Pw7uzA*d0Hyeb|pAoX9Vbh6UL7W_mYU;?BnE5xK#lMEQj*6Fd~w!S%=
zh9j2Z9pL$;Fff)_fzwU<@}3ndNB-Q2wmoS@zy3YoEmY$Cac_%B4h~7|Ui6Bnl&Mk4
zz@{A*0&7bQ+dtYpe5OUGG=l>|zkP-S8Q3}1AH++`)GM%ZQPla0dpYTUT8#iIZlcoB
zka8Nkx`J6Nfb`MyhNLoAUfi1(3b!eElGlasXw`m>^H1*as1x=o8rXVoK?vn@R-@x+
zuc*fI7e{Rq>ZnqD)RLA#)heiSur+FS7sNj$13nUx7$Smpz?srnx1Ez$;VhiyD0#&R
zVo;IgWTP?BAPw8?UrO0Ruurpn&}xdS4MsI4drGS>4>sq@s_!BV89`1|29ERAHrFnX
z9DSJ5sPm`&NMP_CzY3%cr9wK+M*r~}5U=7Am~ExRJ$Db7zQuQLkl1y~(7C|qvVpBB
z;6-K^<nC<7H?x;Tj?5Ofe9UaF(jCZSF?tGPZw(*3YzswR50D{w`lM1HnH*jId*iLs
z{^L5UI+sQgvfIM$H`cNyaxpkLwhgzG9vvlly%a?iTxp|RnFb4f-2EzZOcg7zQeXKt
ze^4)4@`^CzqtxuPe{$bXR9~%4dau-86A3ol$F>d(am%iIOf@cub)fcw^*?=_{egI2
z-ft~|)CsjU)(lmMC1Y+`J4bWfT%@~$cuI$j-t_s;At`?7bdYbwjbQqSgXz-^TyKSa
zR_W3Pj$NtgGlXo}YemJ4@2Xzp=Df=@$6lG(&on7lO73z?+0jLJN0KwwJdG9w3hTrn
zE@{q+mDLk$5W5n+f`+z8zwRHLNi}>h(mk@G5WMV??zfDRsKl^ktjYLz%8a;n)wi+R
zXhLgGCq6?{%f@1$rno|;v-<q>@@yanO5_?lQnZO^iipp+mtF<AtGBsY89g9C$a6A7
z(&c9FY`5C|h8~qE-DRJ|lsLQ6$V1*W&CS-mcvuS}r&i(~BN{}U4PV?1Fz<=eoKjQ$
ztnKWaY#r15%{$)AQh@BhD^PQa`b8N8Ga`zSmYu-t5Y-<nw`s6<OwJiOs}1)N1K3d;
z#cvSQkZ9Os)vRE^1(|M|rpJ@=<mA+kage)4=3n_Z$mq|H^*`JeM$II4<sB#R^Ys;R
zNtm)WS-Q8Dm8Cb<KRn#rbS2DFJ4u;Gu>_yC04tanI-{E>uI#!r+GxeKK5<QXW?;`q
zR7BK74K2e8h0dj}wV6422ZG8z#P)Pin89&l31kfcE)ycOs4FlH_(EawVs$}QtU5tR
zTe{3|lb9e<m~jMTc2c4h%7#Gtu)$Ki4&YFG13ciZ6S<AT>bWCU={dR<l%;1cASDoL
zd(G(6Lam~_r$xIyF@@$WGggX%0K%I;&}C<US6on?Y~lQ6^o8+JNd${Ijv!5pPDy1h
zZI*l8_&^TI(@iDfT(jjvnN9V>noZb+wq+oMT6nG0@GKy^6l$9rNHST?5a58uxj4$`
zsn<6eED-wq1cWaX1)R>K%4{}d?{4Ky#H2-o-rfC@c#W{oNxW;4ToS|k5wVlaLLrlP
z#a&$57KyVyS8t5)_XOHjReBGnb(AeGGAD<<v3^3e{oiIe6qa10MpJ)GG^{EWKV!lG
zGjP?Uz<5BCeIJ#TMw#=c;dhmdBGrU87$A*0fvg#ckwn##a|AduX$+&N%*o6}LbMD!
zwA@=dch8}^URwkJO%ZT45$wR8f>QUuL?8xz2xRH+gEBS$8X;ML<9*onNb6&?;l@Yt
zmg<EH{L3Tqb&o`pe9v&aMa>_2;Cc88DwPqB(eNDJcb1|;%NsorGGo^y@GK0G?EMaw
z#H@wf%MIl%3MeQ)`z{Vru3Y7A(X+-4$NSQ~VQdHdwE7i}n#8P<p(MLcUtu9VJic`c
zBG^D)e;2~s0RkC)5a194NP|C9?+HGqv4i;d_S@ZuJ?eboE6Zr2UNBp6P!+FLWY99(
za9>o!S+ks)KE&5of~z_+{@SbCQb2>%UuzQM=kGo0X^^PL_rpc?!Qs<}BMMe&IoKf!
z)B#^_utWEN;zz*|)u+uLpdM!$Hg%b;Ekf0l7vudU+^jNp#l`qA7R2NabcV2Om-s+s
zv9BuoZU31s7DSgDB>bc$29<sMUzJ^<HV5AF<#b<P`je%2v~YI*5t~uO^hzJHjDT|@
zrF9?{or=L^yKt2^k7?m1#mB6x=wy=yE7-~EKrJ;(HyVUFt|C_q4xErjk18mLZpOdc
zey4MIveFKX**%c%)BizxH#iXm0$~xos>msjB&kS1Ax!}5n12r^@SF9@HA8~FFH*`9
zH(SN^rOBo-&F{-CW`QD!#xG1c3=6nJO@Gh=PVP*Tk>>z#XCRvY&;1A;v#!6f09U`J
zqkG+-zB-wvbpEjljVUi?xcc_rhcoA|)t>+T`>*f*AdT?<z5;<rnf=TW`adu0YLJ7B
zQ2&>s5C7k}5C1M2Ie2^e@ALou{8o%jtE(xAIC+N_&mbWzdAH~8?;@#SUsW)P|2X7G
z@8W$caI)^0Z9p;yf#|=E=EwgoXoy+$f32+Glb_4f{=WS`K0QqW|G(%Sjr3t6D_kN%
z-klf#-dagZ@Q>wpcQIY@$D$LN|Kc;q<m%879o-NRFe3cBI^P8T)vQ}DC;<G@ZW|}C
z3Qdh1m8f5?9uQUAn{ez0>O13;T9?(y-fw(=KY?2MKj#IYy4ZGV2V?`geCL!>gO*Sl
z0Q2T@Ph4oBj4my7wqgtB`=`@b0GImxq@O3PGV(XL`iJWtqaBktM=tNbc^qL(cl6};
zgWts{-|Ss2j{D>IoHmPeg)5Az8?GzwW4JGLrG5Q81a|<d`TNVyPtQW-uUE4rOj5&!
z#0}TB^7S2k^xqU9$NTTbV3C9Fm}-5*coDX2*Y;`gyXV^P=R*K!?{_-jU+;OHyM2xD
zUB@%+v$oA|p)YR%^zUDX7tl@qKL7REs{Q2#*R7ZY-jDE6*)lekoBy;Dss6tU(+Z|1
z<^*nz#d#dh_+xeP6(dHBB3eIDN~U7F?=;|$;dt!DkS7pxZMYxg_b~*X4c^~UBNR0H
z-8N6Q|KuBhA&62NLz<?UG<dzya-A%<a$%G5a~uv<JF^XB?Ck73Uxf|qtfrV8$_OQt
z`&P&wi#al1|EGp0{~9~mx)0Y&*^gsBM_g(Y9KB=m<{1|!;bzDk>3MSAfb6RJ)#==>
zjp<g=+N0mya1|5zUBjQhg87I0EyU)0#+JnG$vSlX)%a{QQJs$4cI??ZWRJSrBERpQ
zC-{?(UHzPn3F6PVHQO0$yG#7b%~sp5+GnGKFon?tYM7&y1ApJ?3Ebm<3LZeSdYJT{
z`up`EeT)GKXNk#FJ{pSmpYgiS-z5KsIzGz&>oyy3Tit|h*CNZgoXFQj^`ijKf12a<
zuO4cZ{}~l)U%0MI7tYX9F-zH|Lczp)F@p0{^&dj{`Y%GE6$kgd<SVwb?zE@ZK9z&;
z{qq)^f4@ZmaDuDGwhn;Rq^Ym^emDG&6Qy`OQHzVdd&y6Sp192P&3pZ<klybSkN!wL
z{p-g>dMwaKErnPu_hd<)hOFJL?|CRLv!wP*7`@&Efp39TLVAZQBKKSYpoQPh`gx)h
z^cKkqRQ_XIZyy7^*}|n8P~LbW2|s9+%)mdxG5IetUAT>vodSBwmjVr&ciUjE$WUC`
z!5)=E!1<fSiEWpmIw^R;^oDp`MMvT@lK>I2E97s$#DmWX5)?neuegoWzZ;|1PW{V#
zoH_mW=5fj5_nr@H1VrY8<31#<{;wrja{m!v%^)qrf3_tZ7Z92R-*x%<Z=sSZhAg+O
z0k>QDgqe&cs*<f!#5FhhdwixZjkXZ_^R#GHJpy%7mgagx-6DZ2dLw6s7Q5arf?~0K
z7DS!3ER)Ta>@8{W950e{Z|em<)W%!8mBfk8NJgqkT2q*Fub?K}Z*_PYL$V`Kle|YV
z+gH!DSM3oJ#$}?JL20Caf4R2#QOAATORlOSBSY}#uyo|XU)bFqwG_Mbl(opmc9FO1
z=nMDeKa))`%~%&AzC3cS3v0m%M$J}>lo&6xK07p$jac<_I^8cCD}z>3hNzY6O&hl`
z90-)FU;ZUFR=Y0#;bY6_8Qz1X=d~6!H;%$Ld2oYJ$6f8FCtZ^t*VI(t)ddkV6e+|{
zS{yMX2;KYDCy6$H)<WFa{8ue5Yv^v#(!F%ft9z3A*W^rmGycys^^~zvPNsWVyJiKO
z;FyFtMXdZaCdL#;|BO!J1v{@N`TK<`8H=@?ItSOXxZVr&pJV6gZF~8TC>Ygxc%4fK
zv|cag?9t&NiHEhVtTZP`KyY=0S^rs5QxTNxN8Yi51#*)fTJjtuwfy~NvOnJ($W8~|
z++E~06p)4rjnQDLpF~)s6)qxvk#^$ApN(-CsD}ih3oWbHIu8YmDEgeaW^H;8kBz~O
zl@1%UxJiE9)6Mc9_w>(0f+V_jwU}xnJL0*kEq<JQb8&S&D<k5!r;PzJK&$P5@0Vyw
zM5_2hN8F$sSie=jPd`=de6rY41#)G|CZMLt#NoTYp84TLy1%~+z5TP{f&~B3auGTt
ztNML#+3K&$O;&$&jK+v^hvm=P{`|)l`ft@HwM3~17$(1t?kGeT-&)MXP2I||3S0Z4
z7@!c!_;@a!Zo%xue|~dyC?~rSy#hPi;p?yj9Z^j3kKSi*eTsIOXUKUia?k~_0vj*d
z$%~n|t9$C=Hk5F24!eN={C1iQ@-CO3z%St)mPpL*HT#$6Td;x^@t@x=27<;e&*_kV
zm7F<JWv{veVSWxie$Uz8b(LbjGwtM{pOhZhrJ|I*cz5*|uk*|-|D${RB&GkIvpG@g
z&h+1p`TJ8><Kb26g8w+a75|T{xF*t9i6;N^=!^az9s57BflK|51k#a%3MIqMyIN^7
z6RHNzva+HLrn!mQzn-7&f2Q;MP0Tg@6o821TTjoG)&3Gl(QSWB_ehysQ-^=_x4&*o
zjQY1q&z$qWIzSTiGP}d+{tCK_i%awb;ISILerMsFRd<TtJOt!q(aP!5k})4+TXz(n
z;|dZ88o&CJTijq{@toT~44iVUv6@NS94m3Aoj^*MaB3@q;r-}I7H<kaAnDv>$;bFv
zF;DJ4r#!a^`3)|a8h&PGx$^u=$*Lbp+>aQ~mAJUA*kB{yf)gf6UZVY_T&(`t<Yn|L
z(oh`O^BSm`!$$1RuTZi5Wq+l%|IszZ)qu*GU-JbZ9=(4{^DT&f&oPY?8sA#42Z7jC
z6Vx7=7sbfc`<?mAz8;ItK}z&^aX|uB2%3EV^BLt?(G$N%<L?vW)Kau*u9}On8*>rF
z?iWgPF~8UYL~#-EkM_x?ohdf<^kV&R>*Dxe_qE4Ko8CQmjrCmcW33G)<Xb^9F0X$*
zkh<ZIH%5w?Vtwl=vicNvEtr^yD-(_MS0RSR|0tB3IpFQWaXC;rsbOK>>LO5*KO+6B
zU}Dantfo)A;}>7mg>95_u*34x(|#B1_>Y33C^xu-)I4+oq^=uYXa4mp({F#il-Lyi
zTMuohCy+fOH+6rNkc#`0S{eE|elmMh&Z>KQTcv=GD*j^Oc6a|^&ARyM_7L&gZChi5
zntOpN*=3j80l#I8!=EfRNc9X7)LnG)Jyq)Z;1xIS52$KK{2FaR=|Afq$EX|}$Ozxs
zF5R9Rx`%VQ8*xKC7@Ki@W8?CjK%|gX!fAg5HiJgFm2<GkU*aPD;U9HMEE=HnBnLeC
zr2$p05hrRk?$EUzG~)5sy^OzFsuA-?OS^mi-=|#)zJ?|g3Kd@aHBCKTZ2n-{+(KXy
zgm1Z6JKxCq>w4vXT(AFsX6paXyvG#d+s=BUh&(|;PP$qLR{vj8Sou#`lNu5IZx$k0
zh4KFvzx;P)m~&E03qpWm5m%Ylyc-|y??3SJS<SDplEeCAWOZFKo&##*3p+T;%Kw#o
zxtq${W83QXOzTK!1tseu!(CFav)gvD=?$XhKM7vO6!Wk4O0mxAGm-UtS{`r_*DV(M
zEj+<)D<Gf(%oR$J<R7hY({}LuZquaL0K7fyKpl^T^_HDBo)VN_wYvIW%yT6>`}&V*
z-X@biCnJtFKL60M6Y&8L48BHdHySacZ3YGd+~#l|oaau38PXO1BcVJy9|3R7vH=T~
zKmKK@ND^gFR^K^;X|-{?TGVyW3>L%tki>TtPG2^poUlR_v&^EBaEhYZ>RO_T@CW}$
zp3EHyL{}nfOC&Xu3rVBw12<4l7DeQ;!kvU;WRe<><<+dQSw-U0?{!-Vyzw0dt)HDQ
z=XZ2L;{+eu><DSbY@a^4`!H6IAq64D65*3fm$9o_$Zqn77$TAd0<*h7gIPZbi>f4O
zk(v>HRtG;lX}R2KFh-pjOmL{`7tCIV*E&;1(X(pr1d0!OtBid&FP*20k}@5P6)Q~<
z%Gu7j1u=GHaN1#sD><OCBqz{O{@T#L$K;QQ5_#?g+SP+(UNiPj-wGOK$wQh}`qIpb
zzC<stte7}Fgm|XXCOmoYcZXUILo!pPOtp%86>fwk>0Ldd&z$OEn0bInSpaH%rQM3c
zaffCFRH*n~%3#hBYB&kf5u#;o#EzQ+!%vAe24RsA_|=Dr;#12Y9;;!d1H?1K?0Z1O
zN_LK{^;qT3Lf9qibBa*V-Jn-e2IO%55S8ET6ha|EMx2Z<pU4HE3`ktgtmiItSl)zm
z)AuyTzIegmgsVyZxP*~?m)6#qE2c$+DKPH|*x+{dY0`e+xDz~_s9EX6sX>lDfWn1S
zh=b?+w9QXx74+Mno5S)Q@$XtsU0Gsed_JO(KKWi+IAYEnckj6gQmQF3B+@^;#wn_E
z^c1r!hd}4W0*s1<EB(fQPY`5?U|W<`bH@kg09wL2R(t8F{>tU%tRq=swB#9lB(El>
zb|k*y_2RKGw0eJsYF*@_D5++{rufvpm6>}7kwlf})a5-!ej+UDhU9(*g^AxgUW5Kg
zWcVN}oaH<{+F#P&&}KI{m{x8var{E+*yvO2)2x&qcdtQ~odR?&7GYnLKW@3T%nu|W
z2D3I4hh|l4txfFuTDshph7`x}yYD(|sB<DBL!?qSYi)G`&0K6II+NHb525F~pE;^+
z6PZTMS?s4bGc4{Je?91eF`SaR9-hP9YK>?$M_u;zghgdMXcco0`+SDbF4heyb;nvp
zGAdxF6%0pGrDBOGFP-<g#{@2YB$s4f5~^F(tyk;`oQlJ^9TE2Ixhp@$mb%V6;IS@y
zvlS1?B^M{Rcf2kaqLJwG;#dbPbes>?@~Y7-4)-|6_rlE5D)|zvQ8%=0Yn$zR-`YF!
zsZ`ObB;=hd(NyRPhbeDORk+Oiuj_YZYbp<93hoi%F4LWwbwnf5%v2jTyl|9XQz))w
z<#B6^+VlkvDpm1A1Q~8mLm=GJ%H)0c%pY=gLNy{toWmrjlHWY9V39u+BiB^Z3~PtR
zb{P0KX;O&kG2Ja>FlD!}p-i6kh8^so7m`v?O}^wH5aNhIY!C!z2SUH>B=}~GMwO20
zVRClc1z|Dhog+WU5T73XS~B|AD0P^+Dfcjk*8ZT?5q-8P7G*X@ZO7vH1dA{hB=Id)
zg;(TDUCv4^_A=ELFjM3qq!{$LR<(w*rmERi3UUPRf#G>biHaX3b?Wz~c632Kd+MoN
zWoq|Py&%f~95y#!T9eila%Z%+O$WM>kdKxJL!?n^L{>x+kDi?m9i}0UQ@V1D&9gkC
zE3J&Tw~nf9b9Gnz0>I$#?vF=RB<O5E)`{X-F#dopbQ_u=P`FWV_V_MN0?$ro#lx-Y
ze51oaM+D*%O}*7iuKmV)a~!C}weQRhYc*H7L5BFaEVD49oh|zsgVCB%EWF~Z3HP%X
zml{RTrj3^V;K6a_cP@no{8H<iS!B~-2UE){(ZeQGM^%#RHm}V4j{N;cK)P-x*s%^(
znTG~_EBo7C3VA{GR$s4XLWa66nEO6Du#@@=o#vX)py+6idKe5_{>Edcr=fTZtk?rR
z<1@*rrbZej0rKO5ebNhRq@~HkNk_Kl&)dR^9-Cp6i}{$WDUKVmd;EZ>-m2D@FBglt
zWx$AL#2WQhPi7thiG1IX=ayVp@86!H<7T_l&%(8;qdH=TmbkIL;;NSfUy-WZ^5a9U
zby@1Cr6t-6Q5_JR>TH>J&yKNT@(eFNSPaNHA96h|U4pq-CkaLE%R*fc^=)MqCL#fz
z(XM-Ec|<O`<2t7mOP%N26>W5|aXy$odyV_%bSkUY6B1{}+R~+Zv&;PHE;vs{b@1Ff
zfpOLu(tE~Ft5mn*w*A*923#Elie6<VPi-a4p1#hB18>D<w>qo!gW*)VPkjp>Jd!O)
zWmkL&Bcs#@8a_VQGmdStAV<e-zq!7lvZ|`?t>#c}t+R=$>IDJ&9bsW%tpSdlAMRa6
z>u(aw)iv^6j=YJ+R7)1eSW;^=?vdi|s&WE$<?<s$bw~Yeo74kd?ST6n0kygEiH&EE
zgME4+AxS*(#I@*!B=4CS=g4AE$iHKghDZfbyfi3U3D^>43dWRxp<5Uns|v_=L@RDA
zXWtPB5MjYn4(TvR$`76`OwN&|m>n~3-IiA@Vdb+wq_#}ECPx>_rQ&;>;X{2lr~ms?
zjVodQg@kw9dCQ_l*yXf0rQcs#+Dwth_~%g8t_AtfeISZ`0+o+L3??as+{rUd&dLd<
z*HFl>W(9+);eDnv05q^sg%m&Bkc`(>{qS5_gph#Y@*`Q5Sj~`?v6FcUn<qB3&!=#6
zXeY{83CpW81JStHp7H!-a147xJae+HcX4v^9=YrHcS{TKobN8F^p{EtPH2Hrr;X{3
zsP68I6t+*u9;QJ*a9c0>=%5V@*&~|Cx1>^E2`Uo3R5Y;^H9SDP1H6entg>@(quOj^
zGH%7if*>7Cl833#56$C<{Y{dbs23RBo90@1i1^7K7MutW<_8;xXbetI>*5lEL7eQa
zS$}g@MsJaVYA)PV%73^VlBh#ucV8I6{rsdrc4s`0S$<zE>dyN#W7@jvy+NWo%@RK7
zN&O$FKj@#QXG)NTwWh)dT;p$#m^T{=lRe2bXUrmhz!YA=dk&;=N~um_WAbj2;d?ps
z-ky;<@h_rwKS!i5v^mTrvb7+oh};7|?}vzIcc=JPQY&FE5LKt!L0_`+HY4LwMC)0(
z{E6#@`1&}>D((j*MS1^HaukXdG5Nwv&-@4lU4aH}tTXkJD|Zdb5_Y;hD-O7-w)=AE
zh#gfPV0v_DUa|tc{`NJyC`3_T-epgz<NAk4r5r*(C~(BTEMvX&40)b6hII&yRg$hc
z#3bhZ)S@)o>E1Kwz*oT-H}$DK{$Qrl9n0dhg*J)L6WhjRtphs9>#pA>S*ASci41kw
zYh4bVPa3OrmwomIBr4iYc4iqU$u(=^lb2~)sax~$aA5H`e}1`N+D8{p5;1$NIKQmx
z;v(MZv~it)<h1@1C2P`rz`dp}qRTbbyf|F<o%$%afL%vs7%P=>UzlpyLK<|O`K(bO
zVBp~JWNu7GDkYZ_4x24`UUJ_!_3Q|)X~prm@=#wk2Q&yZcjD4=GDT)_g5BpwAKhcE
zv6JD(Pxm;+4~D;%eO{U#t=W1p<;hX9xIzZMN3<NEz~OMjr*&K<H}%FU5!zeQu*dl+
z<KRM8)<shE{`h7YDJ0SU_OvGlv(tQ!w2?+y(bRCgN{8hgImtM-c_w^UMdz7Hj#T#7
zjA#*PLy=a8)6cV7n75402ZkBvXIZZ_8zKz9kPX=WX$XT+k!Trfsb#U=ZBh!MX77(g
zZ4<so+JGk;PY?Es8;LS}m|QYn_vtWC-DmO{2~i0<U0>xhq(_XT_%OZiZ9-XQ?t6vl
z_&DwsGieB&J<E!|sFC<N+t7QTyi7*I)vt><Oev@<TL~x1hdNXD?;&hU7~a1SnVL`1
zS<Fa~Q2MFF(0RmR(1f)<i6>_eR0LkA;)sU4`Ag3PsW@7RDz)?GJmK4U9=5Nlt8A$*
zTF^ZvZ0xb_979Hj<Ii7VxAd#dMd9d7!W?f`9qu>WeckISRBnsGo<o<NA;8qEKNoT2
zgYJO3-;gzm_haIrrP}GO;*4rwD-XNaW?yR=nIa-NDMIL-Z-08pc3E{cJA=<hcX*IZ
z?famD%Z_V5J6+uNO{LQwboL7_iB3hbE5csCYl%Lp%DjdBa^)<K(8bcArGgI%uPAgy
zP?&Nf4(~nIlZC2nFHObqJTBGkv>#$T+cb(GblLDpe)DMQuzF{eWU~CYQZ8ifU4__n
z3xlJm^R=(%avzjsbMDP_Zw?fUe`r6MART9K8AWZ#sw`Dm0SFjuLtn^-e&Y5}bhX@v
zD%9*dGvB^#fDEiC@4E#+x%#XZ{cP^!`x~~@*C84pPHmjDX%lzX$J=|-WwAg!_`aAd
z@)5to_M*zlTeTank7|VJCw4;<%uO7~6EWhr=b=BY)k%|>4dL!2`o!AMpe27|H5xFf
zKTz^^UIHTEmN;gDJaZ&%j7*vj(^tceJIV4y$$1oaIdL>Itez#Q&4`e_|43Z>QG)96
zg}D6wLp+NHWKwJ8k&>d6!=&6gaY>YcphKbi$bBN3A7<m<%kZHjg57Qp6gLX=xJAy{
zM3s@}Rk&P+=4VsRN%?kM_uesK96WGOZq#8HnQGK<pK85~5)mH)YVk+a$k?ZJw$8`o
z0pXvgoMPgeuN5;N!J{YqPi-y(OxC_!Hg7e&jPwmDI@V0vwSB^p_kMes@^T0Q`L##D
zN_ghl7a$SD4VS5T_PoQF023c|T6$A`Kd8M@i<F5cnap(;>?_hxX5|PoH8NipH+-8P
ze86+`m9E?!r8v574R+WU+7VCbezB*Yfk%4K#5BQH+dNaIFZ-!zuaDElrL=k<sspFN
z83^G&byBmIMIpG`-**BdzLyi}Zsq|52;LricSUo`?KK<!TwDT37KUiWQ&zEC7=U{i
zt&f!N2Z9DdUpgO>h$pi##-FCq1<+-zZP+4{;Ef>02@Df>tP5HfuXPj(Z28j*03p>{
z72nb}-JJatGhX{?#wvnOdm{6>1zFd{nPY;r=q-zlmGddD-Cj<ygM&+Irg!9hj*keJ
zW0dSioa{q8FlY@s`gjOGKs*N(JUn&AmN%vxoZ8**#Y_;Vzsb$bP0=jfa?%}He*3+q
zi_pyQooA=g#z#0eXBtQ4rc?QkH_5<YHvKkSlqps)Xw_`S=FMU2Ty8ul>~bu`!PkpL
z#xF!et3&0kPs`9Ac!$0+#Mn6ss<F9-zoBZgm`&}H%q(lbMn!&>2W^b_%>L>P`fez{
z|6^k%Dh<x3wkrEf>}H+Dn_{9J%`b^PfbpEPCpiRdSF=oaxVj``X+HKI>h)&4xT=((
zL8~H&7bUTp1_6&`NwvvNrzq8Du}C87U!83rbzhCPpHX?3(|w%FL%dlzHzUitue7OC
z{FO>GH@|~Ky_jOs^XjK&_%uUajm&%`|M+Fn4d56n&NBDlT!e<vRUK<Of1jE85}s+c
zL#I!a6PjUAktC?;WTQ3tELrVT%wR5VN8m*2uq88#R^qh0%HX@ZUv+ZZsloy<xsK{R
z_fd=mg5l@W9bGtIp%er#m7U|RTI@4XXrp&m4z<c^eM&)7i!A2zWnBBd@nOdELz1*n
z?nHS;!qbQ$8(IF#vw(oFP!Yg%i-MQpW6yk!7?rdJ7GOLfuDoZ1t^sLKk;TiEw-o~e
z_6q|RMxCN)fB`**IoJ>6+=!l^j;>z&s<%QA66V;jB-43*_Pu4m5msF~6*bNwMl@Pu
zxFxef(7otzv_RzLp8^#*J>7%hjPg~S4%VAxr4PBjM!kPLfDbW#3gQN8wRD<cAhHtx
zO)o2hjm^}7hzqx}U5O%y4vh^?X&N*n-#GF(oBy#u?&2?|B&@lrQRyOlv>eK6<P$xD
zo%qgv@*~{XO{VAlTlAQTnh;h2S*}lMD#GrwBr4Ckg*SDc)c}XWh!ho{8>wn7F+=uD
z<hPfz!i?|88OtcnNFsXcuqNp{R8;7@-dC|A&>rIuDQr_c{?sC}Doh%3dxV9d#k*d)
zq9!0o;e~sPS$ezN#(1p=10Fp0nt3-i_ZFL{!iO&CvkihB*8O)rJ|5fKz?SQ{2nc7Z
z(z3|##$iHTQ5?#@SLR~8))+DLcIa4&U=#Li;z}gNsm{rMBGKl3wu2^{;$x+?=5i1d
z(a6mLTN-+3L1=vA<O^Sl@DriiXM5rrU~ncbkQG|SCq_O!Ckb=0(Ry_##&R&=p--Q<
z5a(gF)j;PAtH}`XI4SL7cgsE=yePQ&aef^-rPvk-TP`XY$Cs1cimQoqb(=HFV^Fng
zVdq>VPAEUuE@3BvLFrfKyn`93qxK&N-JYp5!=C(@wZ>^Q8LvBrgIcrmL!z>&N!6q5
zWifKMX=jscfWlfUmOJg1|N8mYOdcnP?ArP2u+Y)Q<<!AgrfM?H{FzOp+~&lN`Af#)
zRD%liA8%+7jF>{VtMm^jAG(K*iWPJmHq3H*6ffp#Sl~c(UjWaieB3V|OnE-=I%k+4
zHsCl~XxCP!r8%Mfh>wq7h~eCFc@o2C@8l*>pM1nv{rYUBvKTiyb8S&m-1QM^yn9xL
z^fhkzGt`qT5w=QteE)}+IQTr<*(#wLstXsPO#!#~M1U$;Z%b$M1Y7>(WxvqALk(=#
zX+Lk4Ilv~Bg^C%0(!QbEe{4<_%79-V*I~#S)t~*;uiq!r#E?FeQ=T`W%*noCKAI>6
z`hn@AW~2m_q%w|>Z!hQFw@Ut&`Q9&32I7Gg@m)1b>6VYYRd*}**-z3iqD|@!485ES
zxbm}Y=XnrQmWw{7=*dXN4SF6k+Ylgp+Rwi5BS!^njS^-SV&+C&cfoWOYZ8o6KhIIj
zD3D*u3uk)y><!hxL5EjCP<tA`mVKWmb}=2oASp_GtCPS)-oaSy4#Czq%<=L~-XY2H
zYTSbVLLefYK9f9(dwYN<jd1lEn}*Gm-{r)X4qs>Ibje3gJUTbRl2&=!Ta86xF)%_B
zy0c(gfZ|b{YE3Liq~*NTF+Mi}*W||}e|eU`pWu4lDL|4SRI(?)D=*gg!EAAfSm(TK
z_ZndnGOXN)LTtX#r*bKR;a+j1@Xq6wwKg2>2uIOU2bP0{1$SPKD2C4;&xctFHZvD)
zy=mk^hq3f*37PE%35CAsYCdaNT4^!=;#=!ppp!Ec5xH|6lDMOQlDoBksM&bHzdqrx
z+R&jaJvvb2wwJr0bFA=6wh<7(csX^I*~w?kI}9milW);)3E%F))RnAi#KGfs*p5om
z=3?dI;wnt6s&f6FW2Tx%tKu}C$wcH%T5D;!5fkCCki&0hwU=7+t^nzsE2q0P^fu}H
z*J$SkZM~)rncA7daUob@j+w4?K^s|>KbB*h4eV}AOtcM7!nvUnJCy$Lp{7!gogQ-1
z+KuWIa;Qd%X25`e-Fk|(^mt)$*6<{N(go7^dtWvnM~c?j;PcL=13T09*{!)Iq=oRI
zugY4EyyuPNX21aA)k{tnp~Y&jn4ipYQWEj?;t_fEu;*5?b=a5FqkI+p>&b*4lT04z
zf59CwzfJ1)@eXN>M5@qj;LGf-!zo;uTFl#iRv~8aIAW8B3Q@6y87Sh%?~DEeI4t^b
z0*qqP46!CF0y)BnYqx=k_R8*P%E`c@13oW`TqOEXPF&~He7ic5{G*Z!l^b#oH`w{%
zqMftNE$(E1WA233U03alH}`tA-Q)LS!E0?a@<s7aZ(;MjvS_olM%5n)cX9y6AwBv`
zN7?EU!~!+yX2m_UoJYd5*FxdG(z3@<P~`!##MceT{1gu#BB@b)BK?q)Pxn61-UKWF
zA5m=+vwx2+#1d6$3v<`F-Y*ZVsZCqW=!Z(yr{9P9Jb|fI`kAqE_GCn$YGt<h6^-rA
z+L|oPaq2k1Oa@wWqPH$(;t+*$J!fGT4y-AsI&mB+^luMxqd0lqholEg-PgJ!X(*LD
zSg~g)qn0{b*-7F&9~a`20Dm>|siow$aX2hkCq#=!-m{^kW!uuyYyAtlii)}x?-seU
z#qAk{a8*}4Q=}AQROR7lno4e_Q*ujxi6ft~g`_8T$YoWxh#&3=Fs*KX8AT6rniOYw
zAf#ri`zVHy_b#QP-wk(1<;spU$NJm`4kBtr3`pX4e1dP*E@wL9uEo0Qmov1kBtfg{
zwQK#QI*jd8)8G69b$XFPP+9E`+n4K*_8Jdp2)IiRt`u!-dLlw_Tp942uhB&K`ur`A
zrw5x3gEZNp8z@ImlNd|$lHoXOcY0x5Lf+lwRC4`^hXLWZS2wimpYW90+s^plSZsK`
zMl)v`O(_~);v1}w$;u*5^$Ev+|1j_dJvU~4X)?mFT@#a%1vs>VgI${`sXmZIM4G+&
zUfv2kE*6hf6W!hg1o_d9MrM8x;T5UC^}D0zB7A!$)1cYFf72ud!*7`g*Y6Pc1)9Cl
z3s=m2kv@NZK!U_s>vmq`8{QsSW_*00k5$Te<PFTgWt0}LukX<(oxU=Yxnq#`QMpYe
zEe9^&3ni{)_IN~S=Lxhglnu*sKM@KImWoIDw<1xS{9hibJY!U|V?js5Bmg0hoNYX5
zr@_(v#Hi~z+o=!mon}<fHU^HLe>G7$Gh!h^?JNeNO<j!f%ghgFI-9Vbc;eIGYLBkq
zViB!8H}b)d6P_?+H%N|!yL|(kQ0SGNZb#14Q*;65lubr0wkRno5vbtV$@03=ai5_m
z^ga5InX$=gmP7i3iaS4iEKkSRmtfp1xA+ZYYEMdHd@2X+Eh-+LPAu-v90&{WL<OPY
z#uRL(MwjBGCzoB$aSmW6^x4}pVkU=l#?=rC?LdNl+VvqmBswQyW&}f2iBy|;xyP9z
z!%tIA^_)pstUNj<<D$sCSXTD7awn@;?KdVC^w}|53|1XM#auNNwRWn6hM%OI7Vad=
zl|rY2lElAAagWzR#aT*LzSUnNxlz`{agt1XgbYcn#m5JM#Up<^f!G(&@9`Z>#2$Ph
z0w}}Z%F3q7+-$V)zfhvo><d|H=Q$7IRCCmFzO4OD9&E|2Sga3rjD(5~PFB&hn}t-=
zv}QGIR@+<cC8tbwe6NWq?L2E)EqXn<^}QyTsJbGW(bp(`)XgAY$z7ayf){$e<Y2y*
zYJ8YKD?RS!px#h&BXFpB#3gxS>C=No9AD&D^Z21=Gk{S6E@=5;@8!N&_4kxU<#4<w
zF&zfLS)0a1b1H$}Xv`?~efOnYbF=m(6YKwB>&@e#eB1YNrI5W8vLvOHeJ2c3Aw;rf
z-}j7t>{4VWNyrwmGswQoO!j??u^VJJ82d8J%<oq3=XpP`-{<R(c)eztx$pZrm*YIn
z<2tVvLL<wKd=8$waM(xIf_wC^xOm2gnmdehwu8Er1@FG73V!BDiKOZepq0Pl6W>xR
z&hcu`rKgVn2^l-cFVnQ{>UZ~~bK82ui+#^TP}F@+tLLWeTWp&cH(w`Lp5@AES~mms
zAyJgRRMH5v3Ln|AL*p2nM8_1O!Oo+f=pHcvjrXKWM+mNYeB!A|Ws&ahyAtti#?u!)
zVA#HpI@hj|0`;+q=W<W1_DCfC>IyxCC&GKQ7P=0kC^$hca<_OueE6Qz9!oc(k|LX)
zy*`}(@1=}3k^xQzf2X#m+{Iq_YUe3E62;E%p@)EB!NByq(iSamW}S?y$zGP!@MrXZ
z(h2jAtdwIk?_bZpkA_1$e}9jq2sM`4H{5qVERm{TQHP4k(ZLcR7b**ZuO}FLc<eOB
zs^*TD9VP9bq&ZN};M}|Xq5)>0sQ!XnQ5V5z6<Ro&){2qe@r$|=3$V^HaN;Fl3D2X=
zNSr+kj#wwTKs2IF3`j!2gu~iPI6VIgj~_rHUZ)}Uo&xz`j?!lwuL(%D4K{VSf>8Bf
z*rQU}$?_5l*nE8=$eFN?Zm3rvr=()_T`ted+IpigTX*IJ>8OtJ7(*K${=^jlxmXM&
z0_0Acbfz>Fbf+e%E`Kz($)jWvAFs_GY+-<4Jb0JVo2u&}STOTYp!R5}srexE1GUWF
z_;}%6+~qy{uz^m|+M{HPA}$Q{k;IQGDj?rSSlclJK%m@^t<%MTS+dr<_{hvyIF>c)
z(0N;_02TgHp@2QGrId|ZAK&WVcM`Y}o;xy0J7i!KD=y~0<9Cwin&CP$S+PbFLEnEf
z0H;^g*Pkn^WFl>{KBfI_y1in;aIYO4xZ#bhVGr(h;L$V!n67K=DbaagAiT|+;H2YJ
zzp&l<!u+>SIX)15x$8b1(^OR}a7#~xl3(OKF)s$y8GZt~oIfT$Dm_KE73FW%F0HO!
zjlFKyl<6<v!73fmV5ESjv-c@{c|xF1@%WvnSy&J+amDVtGHU*vQU3Om@KWUT@WQK=
z<efbz{Ylb&Kd`~~v?`$;e2mQMcbHQY)~HtW=LqxcoP`<)wfPPa^E!j;>`LjT+M+G*
zR$voHEkXOr5n|nqfS!TOmaHA#KP->+FK!b3<yf8{3FPG+`#rSJYP^LYYgdshPKnjv
zQf2*9q2Md+idyI{5B2YnPD07+vxxF)j|;&qHAYdaZm%-9S@-B_exFR~NaJ_kpYj_&
zbfZ-5$1fy0eHG{@wsjcmOJ#1Z3-5VRx#fm|0kN-V*(#N=vSL1fI*X-XzrOsw(SHkb
z6cy~=3)`cMEKM=2cw6o2FO-_xd}G2u8=8^~4D?R&%msMN>dEYv6O=Kc>Ky@9Y6r5Q
z6VIXD=@jB`FGTqO?b9qsQv#<;g=M2h2iXGl*)%vfXuuEJ8=)wFnDwQk4=n+4MB9QZ
z@8F0U>c2Qmz}*j4FPVkRY{js$ZszOw&1f5Z>8H30qa+=G-;m$*y-5x1q;G@K-OxXj
zdo~=4bb^p`ZLI58E#aN5zh|t>nUa}J5KC3_M3>A~x%V8BDv00ePh}-nDe|rWPS619
zn2r6?o`M8V^H`i1NFAy|8|)t6t^Wd@{IS36P1j2zj+(`9&g#5!EE^HAir3zw5_-0#
zlK3OJ_tYjJ;TUvf{pEv1%QypCd2^1<I~>;vw6vCiq6R2|Uof`p07&x53xzTpS+cs|
z&yJl{oB)&2*=811{YZwvA9`4zt}H9#a~IyYOsDf?xJ_81op592OU>FCDs>y!0Ohv!
zD`)lbIg`rH5dg*$&PoI4#f!eC$bbt@O<qzRA>^G2Mg2#JZxgq$i~PHH1lN;2HSCNc
zO~N6TcpVV&<G0Jo_wVz?W4dVJhOu8R0Y~M*8qX>s0sbmdM%Tr-eXS%=Q_LUe&((Rk
z<JNARp6y?JW53myj4XL6ay>f=4SntPL_++UKE6K_W8yZT3+2^5Y|aeye)F!+$lGHo
zIZh&$nf|P+`(}dpC0mWI*#gh#_w2=mxo2DMar&~=rZa^d3gTV_JcQz+jL-^XmTk+&
z@hn2FqW}-d1V~Tq7bih?bhWy)?E8$eCvG;3NvH8C+4X%V9St_0XCfu}jnGddUjr0$
zxgRb&Uq=V(i7T4O-gsxyWaodwqdC$YJgy-SI%b>KU`6veHKfP^iDNb%uY{C2ixo`Q
z&oFQCS8e@5S+RdrFI-pih}Nl?d3`7sw-sX~K0676R5GvrAg#Xs8sfO1eK^Vkm_cDx
z&msL9&!f?vA6|Fsb>0NdZ7_7?lD&_}(Q#B}LtUkk2(sgJ(FTANV{}nx04*<7_*3B5
z4~=OBGWMwJJvm?wXv8nRW?QmnkNcl#0Bl(VNqMY#sZxLHjr$pWf`{>T%e&5E@mJUZ
zIRve|SbTVTjIJ4pBD8Pe*=DT%Ezk?#jR_gsCEqNbnsM*&Ki?m!v7>GacAeeM;unR3
zG#O~<AC~$rJR>4%k0emu`S|%Epbl(zBNW}ACPa+75hBI*x=a>5_@NGy&vJlXOSB45
zl_HD{Ye+5(>+DRZ3$F#iIa{5Hs;cqk8%golX?XXj&)HHb2Bz1g)u;OhFcCKSP3rAl
zMH5TTL|rCW0R*>##oKf#7iUzS?hT6mT4%rQP+z_nlUqm2M<(8896c6?B}e@wcStWK
z!KrUZtMiS&{8(JRUhcsj{NCXR&K$YknL5GNTwh;9z3N9wp3bR>ruZbf<>T{$%nULM
zKUN=0-XCT!ey9hTJM=nh|61rO2lO{{%r3Cx8G{W8EwlmNyY6~W(mQEcz7ro#Yp(AX
zGou5?4V)s1%TaDS0T7j%vqesW`-KNs?-};Wwy0NClWEt$R%4BqvmL(pRB)acCyW~o
zm9+W1hVmVR$9e=8_XeJV2_=3(2;3BcirpOo(S`S2p`xZ9^?X0luFKX#1@DVyyT#W)
z;vNRDeRq<3b7dJ-W*K<}b$Ec~zYf0o99XA;^amOCaew=3&k{q=uD>@|SdiU4V&{Nz
z@h~_TcUlj(Fp!gf4o5yHy!~<k&`k~Tt$b;1d7Aa<28rJwyj>CeM~=jLcsa~7Z%;_^
z_f3`si>IzVQxMV%MC~f~<1=yUi~18lhcKxb6Hy8i8|r;q0_Yx-M2-;MZUl#?Re<Ar
z-{MkgCF{bB@NQ=9*2C+Ziw$?z)n#AhkzDxv@j1=@=7y{@q+;rVd)P2d`<D7Sn0fZh
zAzEct35fh2;w7~<mMMP)9vjnKZaWS&Y51NM9R6aIxXlt~wsz&|7U_nYud(+Tn+$vx
zg<duhzt-Y1zqx0FUp6uJa(cHP4cNn=+g+L!)Jb^T7&5O;F7LJXm~bdH6n6v}PNtRL
zO{U6eFDmu1f1=j_Spfi)^${-fb|LXMAAg*zWLX1$O1~vgG_#dp_%*hm#ONopaMtq3
z;vi9W{=#gq$%Zak>v6=n)W%kiNKu1WHo$XicS&=qy+ZaDBVzEnsr%clt*y%xPMb#l
zd%vJN!yu|OknrwmXSPXke$Q(wH%%%ks*oC#&)Vt`LqE&SVh2pi7L}B8zoCw>_NLlV
zBRbD8iwo)Kv_WhlP`clcOL|9>j;eDY`hi*>Tnzup78UmSVa$Z>!y&SjotBh^tc#j{
zH9eF$eEHl_%sRO9_|Eer^R^8;U$<e{J5GQomw=n!JnxUbAczG&wLcWS%zxfGa}<kX
zl6u^y&H79UaDK=8!n?QAZavRtWRSi3V`5~(0jTiyWeI2#z#21Z2^8$X0K)}T`W6n)
z6tcx#2NW~zSFFv_&jz(>hg-Uy`F>LQ_JB_MDoh0sP~=q8!n7JuaijMYj&|OcFHJZq
zaRB9r6KTboqhrFxxc&)w7M*e%`StG}PiG@`3?kuIE>9&k4NY`E6;t&pb}&i9Ch$p6
zJ@6{n*);p3@+}cVcE@qwNLLFU$_@CPnYWJ0B`K8to!8X$p>E(__f>8WS4M?_Fqt2=
z&lLmY-ehXY>yWX>cb(@bq^`+K0RdRlbyixP<s{+o$Z5NPd>cuiPp=FZ?fNv~>(^NS
z4U4KN#>FW0R;4to`yLudUgoUHvU?@E=#e|7UmNy*>h3XYOb#cC53T*IH^QBYi&bmL
zDf;-WDsN0o``;)51R0I{(Knf$P7gslW4<~Y=+{?)GBBNB=k)WeO`OfIkrTlv2M|XN
ziJ<+Rv0L6Yj!oeZ>Nh@lwnNMq!|fChK%1*C2YU67bFaiMR&Sp$>y)HwWaM6CwOcAA
z2dB$;ym37^nSTCq`J8zeEz<ZY@(VH+)J=X%=^mKv)Z;-sh*o}fwIa`Af~D&PDfVL8
z0PGRbq?Q}8g#93Kxh>&|EnPe&_-J&Z(3iA5kErJ+ljPwqVMP-OH*3qcz9}icwmzdL
z#@y>@8vRP!4wI{+N9m%23x>a?`NUqfXQ*p}(tY)Dd_2ZVqpN3>e6>2ql|JH2l+2Q{
zGC+Plpvf{_09bSKPY=8dU){O-gFV=>GeF>)Kpc05(skCp1fZ$ZeatweU^e-TC*^LG
zf-c9kY;QJ;7k&4vu2ahKb{0OMg{f#<S9<)TtFs}l?+YdV0Vneh_K5pkvmpXMDkf;U
zX(G5qUkb}eH<;#oLU(ppz8u#t&5}<Voy2{q!$e*B;;At<Or#H}2S!J*fldJB#U(9E
z30HaXYWZUT09pVchY@R>qeSM~o@~{w&F)aq!4&$T@uC8{J(g=o`$IM<E5N1ZiXQ>`
zPaOXd!q)p4o!~7y9{{?XC=}n_g9I35kX+c@?D3dbZ<nG8<{SfEWwwq((KV-e6i_6o
zQ4{*~I-)<O!g4-&+#*c8+?cjA#I`5h7?}0^y}aTUQo0fNXxeR<gPFWC9B83N@UG$a
zr1){p18uuG%czqztH5Ja6KKJtZg+i5y-{rUc!3+?x_Nx<=YDx%%#^YB8&^z&1%nmA
zxTElGwTCWi%}k=w7Df4`m?1?{5~o<4JuX|+!|@`~Pc6|$Z|_nTk00l3K%YY%`*|*q
z<7muYeTZT8TP}bpO6FAPhi+1fsPu&{i9mQ>I(&70piyE@0mORkKCsE)TOr}p(4p~=
zLME0a9ZNv0bRjj|QYfYCyz9z-u3s8nb6I1wXp<+?=rW{L2(m@o--wa9rXSbW?#BHi
zS@ljbW%$zovQ8yN#isga9pC8vpZtD5oaG-|D)s5nHF+h&8$Z}SgFUZWQ1uMlxx#)+
z;*~;jsq$9NsteGwdmyF??X_MLh>oJmieox^{vsoDh*xQm9LXgy%iVdE8xU>vT#*~T
zfqSwQ7MDkt@_`|9X<(6=)>;s7>fWXKtf1JTPp^D0ZFb=JZ9h>rn2ne<5T~sn!F;_m
zt@##M1?}_F`y7G*#CRh8g@ZCcbSUgnu&8@_6l;)FFb6lSeiuE~oOqN6wDRvrjD`rn
zn~0DS{vu-2{_XV<zXZvJq?CacRdu7tDO%i!<&moupA%JPRus**>VCG9S9}>+-vRN)
z#*cd;6axTx(;RhW)bDh+?z#AzkyPNs81J#US=n8+@E3jXolJEe1EIT(plVwH-VVK|
z2t|3#w{|u&v#qDgF0})QC~#$+obT%^=B)tx>1JhlK9v37jn`ym$i~Hf;pigj{J1>G
zv~d?k9xcwwSW}h|z(CGEouITu7J}T2!RcN9gg#{+Wg>dp1*EBtX7z&)&Xt`M#N19N
zBLxiNM_P!k;VJM#B(Cg8lcH$$R%P~cwL9k=jAHEd34I7eugtWm5rs-AMc{Kf#5Ui=
zQ~R@4OAWsMDnp;|p(RRNu9HpA2}1Y}4ZU=^T&goZH?4oJ7`<?wXCi-v{oqyjC8*HT
zk+FPBvB=MIXq*?hGh1|Ae{M@LZ3drd@&HRHNdZb*GWNI!WT&0hLU-+0Z}e30%1pU^
zhQ}%8{H3iK1{h2p@zYuwT4Y<syVtKq4w2PlvH4ujeYt*0%Nt8zry;Db)amxU1f_}c
zsD<?lQ-vzj<c|8q7fHGwv%P%(jFZx;|Jp)NBnixe<Fy#CE!Hqt$LqccLawsFaFM?K
zYC9kWe(D`L2((<^)jJcogJY&xSDyk--=6dCR;!1~Y<@Mg&D*Fzm?Cf|TG_tq>^E~=
zeDtCXd=JgVd`0BRoty1c*7;gL7E3zx$GvQJ!-D{yOk<xm8d!fDb96kIP|jI^`k}$}
zh?6fGq{itLpYR_oK!AC3Z%oSa{R}oHkZYO&TQl5VpG>61LwMzbK%Hu|;e&RL-q*ii
z>m6f85_r@O0^jGrsqi~!86|L<9e5aiPsKWFdg;NnLzksq*(j?ku)|JSHAc~E@?-;P
zUT1Yf9T|xE^io5qHEZH3h`*4k<l>zZDbY9Gm#9CK6~-N0h#sm(P~4bgZ8n;h!<UYp
zdIGjon9N$U!q@jH1!(Oozab`ikO_cZkcr-mFJ{ObMB6gt%Wyw^!t=OPjK8#&;rE)2
zIKC|+D9Zo<2PmXT($i1{+q+`}XlgToU;(eg657Jew_c78&<113(T{UA;0YEAZH{J<
zn&O)|kaUDZ6mAm-7Qv1_#SHJ4%c(jAoC5vv7@3AEH5ov@z@O-dnpgK0uUV-kd1_Q%
z_}j%$Rzs#MEp|9_$X~aJQ(F8a1(dg9xTzgG*1O3J$*N>#%~#DYzlr2@d1^5Lh)B*A
zNPVC#wU3NK0TA=p4jX5HUCwVFkc^6{+XOoC;<sKi>znpgG(lgCe3r=U+5y?U9=5R+
zi5H$9lY^_jZyTUlxXF^!znc3T`#~a;UZjLizDP12I69=?XyV3G@FHM7jX-Cx=^p2~
z5ainzpzK*In|auc<DZ;5ei~)DE`>0jKJ>V-X>CI0Z!wTgIp1RyDf@b5d9jP_l91UH
zQTL-W@$xk#Ye-3c00j^_TCJRFB`J+kYZ(V6p{Lu5ZWma<_>K_TZpd7{fP29PT02?L
zrfb?tWtp$&8#LClFW(3z$L-A%jI<cIPgy4+p+OkH#=$S7t*z0T-=uGNQd{J>Gvnc;
zc5{iDglO;+6aCMtKJ>W9I}d-#ld-eFAZ^XQz<x(v+2b639z&F7<tZ?2Bd?_ZaoWHp
z>jqwfhO%G6(IivXFY|_BvqSDG#=d!7Us;$C3s4`WH5);j289MSmp#U}uIo;a%D5k@
zE_}U2XNxYEW3|49c_Y&93J5a!vGz`lRzsODtw<XGXxzJkH%QqzTo(%Dtu&}LB%T!v
zy>?k3_3KZnSp}GiY!85|$l*_C^L@hsy~Awssb>a=4<|tk6Th$A2IPbpKfPVKrkk?J
z4HEi3JmpK00K!F<xWJK5fyl>Kh03}G;_r<Lf-%o~zUn;5P7dyV|1<Y_>l5^YV_Wif
zEeb+CpSsq>^o>(rswAf~WNv!gXOm|HdIUq6cOjd?cyjJ~+Me8DO+@Fr4**28@5Ge~
z0+Fw+j_`~bn9W-m<JJAly1B{Fn=bl8wDwqaz*xGYWc|)$;M`WBE@@Knu%bmBV;A~N
z?qbx}lLql%=Q}_@AiM|j;`+}W)I|CF=V?y_p{XaKhDH^ENA*5Bmqms#94#l)HzrPg
zJbS{(%_L72o^W9IE5GLc1oI7BZp{;;{8Fv*$xcZ?XTJ$`xxiv};y>j>y2U%@vAsGJ
zl+8`D)d#|CZc?$c@XCPo?67P?tc%vc<cjH%wLWWwuiWHWr?73dKE*(hQLUKM?o+^v
z=AAGCug+;ID#J$g0Y_ARJ=ws+0;K1<TVWmP6Al*z^j}OxU%tcFo%uj=v0&A4#Y&-9
zfkHSjRB@N*TDa{;?aCc;1hK_U7Voud-rCn$Bbh}vfQ93{1y88uipq$7RLBjm5&<vG
zOP+QHsuwajJ_=a07c<Jr(z~l7YRv)SNMVhn4u|RPKyLe#qJS>r(u?Q;Uz2$*4<?9x
z8F#`q$nNp5o2>{kB17Xkj*j}5_*a`VJ2^ix|Icv27cxvP^1Ke`k-Un5x2or^uIHNj
z9qltzYv+<4w&t-p)0<YH(a3&;Pyo#G^KU%E8+*e$3cKqAerV)jDTVD-$Prt9;aNLv
zdMBu>n8GYDxmju<^s{B%r?z5W^V`+<cy_-!zmu~ddb;Z*574s#%)Ua9Sxd2<sX)s4
z^ef$|wIa#9#8vK_KiC32U*A}=F%%9Id3(2UZ=j=Lh5w=E@tLe_qVL%bK?}P08HPg@
z&}GMhk5?~VU>Ot}ox!PUFX?_L6F=gp5DU)hzK+U(Jkx<%3OXEZsm7N09|@zo8DtsR
z!ZOa%xVVYGTJ`0!X=}szgUiv`swQMh0XjaCm<L)jG%wlrH>U)v1J1V?eCCHs;W=K9
zW7vqkv5E))d*B=c$_S_%#0jM@J;qvqJ<&6Vx+w>;-}gL9^cx((BUR4MnB0;*Psi7i
zaGKwP<t4E(`weLL!SYUN#v`fg5ykW$V&2S%j3&I<-Y{VQygoOs4A=>ItIutj`YOoT
z3hQZE=DkXc-Z!jW$oRYNC3)_wCXKo!jo#Md<!370QAw;_{Yo72N&svO=N~Y<r3Ua5
zr4Qm%t%uikgxOb&$+KTy)6i2W5Uk&s@m`3ly89-Ko8ICoG6e|y{yVS-+?E`_%)g)M
z6}iUU%<LA3ySTHd#r22i>TIuhS%##&X67#9=ww&Y&a0ke@-h$M@Ym%dQ)B>_k63kh
zxt}t)Km6T}Thf{uH#h&M$?e9;Shx4nTC{kl^-9%!hRkSj05)>lsK*!{-Mf|peXmgJ
zbZtVl-Gp}YTvow)glI<tp?nUopM0ZV;>Ngr2p9KpUkD=1BL&|J6a+4NZ*^R3grb(r
zEj>RtL%OoiyX~*FK>O@61@@iaImu4Rd~<b}<$AG^<-%Vk47iLjABtKA+u}Ns#f_=w
zbAPm%$S_E``ma_XD`{=No@C&>ytdL-cN_>_0TZ7bROmr?ltW#IcyD9wdLgtCd!TE~
z7>-QxN5?QTB*q|Xx787k#IasmfpylErPEPuv$&1l%7v0L$H#Z?O5KnNYyp`(glWoE
z%p{NlA!!ZA{<*(RF9i~<b-;Fb(B*8s5#b0SgCuan2s3(by1coIN~mlZa9sjlBICRz
zgMZaA`d6J9+1?JuyXIu5b;Qbt!}BlWZnQ^g-8{TN^pZ_iw6c2X2$%=}rmANX^5-#v
zW%};##3_ouLHTTt4-T%r%gnSOt#LwqM#oh<ZjQ5<8ST+WNi$71*zJvu>)$Q%chL+1
zO^}#)+1dVn<)PCsW9L#%bJm&d?JFtb&Vx52E_u^D|D0F!8V{51N!&9PR{}suXbG%)
z_`P={F3Ji#hnm2K0C^>j_?_Yi!yAAcy;6Xo=xRiWa#YlwDu+Vu(_6f>RfXsn*C!R{
zYzFWd%AtmURX*j7`YQ56A|+b?*0sB`B0vE<jd<x;saN0_`6@f4;}@tff$?;+qU>h)
z)6ej$3E?P|VM?BaQ{B1#f4X_ZLWW#RYP{-$Hk~d~H-Beg(O+FZIde$s0X7hOToI#;
zk5z{~!w@)c`EVV9Cj;&zv|bU-y*FqW-_Y4-bgR9zY!_N_O<`%!3Ba6_E(Gr1fBJ*@
z@6O?Z*32()CVhrM)U_R{kYo^AC1FbYW7y7;J@;gyu4d<k-w)M!5B0^d>hj4+qIQ-5
z;Ed<>*%H6D8-YwI*#dy(;T}yKdvMvovR<#E(g8}-8zd)eB<|+4wc|$EU7yP7Hu63w
zo$UgJhnQ<QwTte##qytvJnvD!|Flwa3BeN-1ZM@L_Juc#*hcOjGe?PB5L<}kA<v0a
zQDMnyd819TXlVE68JMK(#zOh0dqLHSfV@GjD?fsm_yjTXnQG_Z@*40aII&eX^>n()
z#PB_pi`r8D*ojSMag=KhZ@6^LH1LaA+$r#If6ZFF?9PA_k1hVtx~jpL_aXk#%=wEo
z;K1s4T>2}l_Nmq5E7i6fRiD4sD)AmT2JCF3!|kmv6e$zt89Zbyv8(!3xS$FXmQ7DF
zFH20e;vqfnILxf$OHTQUiY|CnMo8%1=p0~`FGgHSoUKdHD;}W&jPhjv5%Tc(&T)Ec
zi~Q1p<h@z+pG(+yIB5PyTWU5h8q1rM+9g_RYQO*Bh&{^M=BY482eCv17i&DAd;>NK
zx!z<;K<L(eph5&V7H5rwW-aGfCp6|URY{-k{=33{+N7k{ZcC=LfmFS>8QvXbl0LPx
zQ+~J2O7=ljh!8-Cy?`6-VA_vox_C>6$7(+{LEAUQ<qG)S8s%D;gwrus+~4X*u)fP`
zz4F-7bu0?tSGqxzCs(54rag4F?iAhP4aEYpBkVhHrD3NggAwq$vfEb|hGX5I8p$<G
zK_EZCJezmo1I!1orKq~nZ-Y>f|5_t)5B@oaxuYsa24f0Z#PXNFRED%OB<?mKGTtNc
z2nM#wfPjM6#cso8x86SBZe5;3J4)QmVo{hgBXE0X3U0*ndSUy`08Lx0VMT+E){N@6
zN6?g0#SB@u)w<QLHU=>&<BD^U`sEO4;|WgJSR&3ByCbXrDEVd7p8)5?(Y7$14g}K{
zW(EeG>%gwOSVu`WzN7s;gvocx!RNOO5)EOHjfjn2)XWjMwn$+6H=s#h*@Jeb8^88{
zftL9)eY&OhVg&*PxgTjht9EnlLp4(T7^yF8k$Ee&yZ?1@^VA$cvB|7g0Y?D;8IY-;
z>uy$y;JKm&i*WbiTvmTlu^UtUTX6pqS&EVZ$qnKM>-!bV1oO6^L%;cNfqX<*KtArK
zbEv_ohRr6#WTPQl{IUdKznFxx)EMCPeE@a&VyIxaP=b9T%@uu=HdU$12~_hSnWPvp
zRbbSEE;q9^0-6Ap9UrV_td!mPr^+l&3*T*d8WmN`zVs{^ggLel)9^5R3jsya84?Vy
za$XX)Ns+iQejSpR#xwIHyBF#S9xp2%agcKeTXqF@5?c;Eemb;E|9mz~wZ5sYU`Q4q
zI)N>=uL&oyuJPjoNi`3cRBy!2hmn{(BfOw70mNUHL!aK64=n(at-AX!e#5vP5Y9s=
z9e-jXkYeLh0hwyS#2Ck`f@v;|qO1!ROQdItU~BCXUOvETu?oKnX|D|hX4mkcCV@{l
zmnh@U)zFIcybrxd3+-P$o;RdaZbI0L8$5b%Wbky!WI2~*&`%v0QDn+xvnwjn7CF7z
zeGh0Yp39uoN<y$eTj<<SX{QO@E`-R@^5cZYTiqTnT5T<j<Tz`E6Cem-z*J3V`Z*Od
zWa^w&<N~k?O1WF9TF*>SCe9Y3@ms#;1NBCAfMj8|E)NqP)I1px;!dQsu@5#^OrY_v
zF7)t)rxEs66JP^@o68w*{H4VoEv>Zxs^^20bi!I&qq)%l^Mwm<Eb$q9Nm7R!X-`s&
zrCe9fLWWitVlt$54|MT?hg-7kLuYWB(?lueC(xYk57Bz<&&sOV6`OpBNXb(p*_xHK
znjUu@u*c7dfYufb#v_!vT;*;}3cmEUAnTu$6TKkcEUWgNhuODon6>?=qj~p}f~M{j
zLz$R$G25UW9b+M*hH6>^TW)4Z`HeN54Yujzc?_HI*`qxlGrYv7fjwM6dZ@{+=klpP
zYQRNak&L}@p=I`9#s<9xn#F?hXp9M)sF>!!<8`)xt$7EwQ{1f7%<>!AZ5qU;qtONC
z#k-hwT=WKw@!1Healqy=@D{%lP1qW<8G5SUlNUQ&x{p^lT-KGTI@-+({O#5eD|<9a
zYJwjz)>Q(&%+bHYW?<xI*&|n6AcN0Zh>h{d{!WX!>|Qqo218JJv8BrDzlcPEGZn#r
zLQuDm&FkQJrzj&}Ie`I1cok?<N-}q*f8Zx6D&(K7%fe7m;0}kceq|giTp2D{L*><e
zeR1#C(&!>@kam6jJDK<p1vg*S-@k5lp3U$2{9aq>QOEetSG5e_580&sj<=839LAt9
zG^tGO(SE*x5$MG4Bm~BP)<16(c)ACZ0$#~lfWfy>+b08|w9P9E{L%!=F)DbvSN&PD
z6rqh%cB%@^hI1j~ZY&zDqqfPG7TkDQYUGO{4K&Aeai$+PBNzkdB=my%Z%7P-8wbrt
z`;vO%NB6W+^X4>KOa<O&<5q`Kz-Qj(4(N4*S^O&hEb!Pt5_LCSIX?mY>G8?+Mqt*z
z=DL>9S=G}!l~Xh7hW_X&^>)uA<1}UobxNjH_+xhy9Osdrn&X?Q*|YZGEIM8JhMBc}
z4*Y^a^{lXU>`q2U|1%fiP1K`te%v}I)_AkBLWW(4r!%REB__00zwOEeA_mmb*}Z*A
z!Y7m9wLlxc9;uqyW7$ARw{>2)a&VOeo?^N4OqXmK(HmH9jbAg_=V1=`jGYZ|oiwU}
z9&*~Au@OX&FmiA7%GW?QGye;T%_?j#><0g#AquD-QwGk=J&s2WbWQqc{O5Ijm?wM=
z7Kk6Qvh}^enOCi-a7p{@TOTBwu)LlcMbaQd>Rso4*w0{r86L=ZMFa5Z3SzHD5qw7Q
zhoPHl2**)IuzfgT*B^Gm6pK3h6&ZL8v{OT~550QaFunopYHyKPDCBG|lrS53k{qv&
zMI(-qX9=)@sPW?&47Mo`CA6O)OPI!Zi0-DT&%ne!4)46d;MOo_u#A)80!i|D-vF#H
z{HP%}pKaVd!|zZ^W{}x;XLX2Bar8Uy0~mI|S}HBNA2J5qhoqm=?YxDFSY!VkO|Ylz
z-s%vW-;$>9UBcUXkK>Ku1!YpcU|kvKgXBakfo~t5Xv_#d7)9!4+{8BL`7%#nt7xbe
z-Tme)4ESM#ryXcT@COEz?5?`30pc*I;u-Y?(E5p%XX~twVtCQS{(*J2X<gT^%I-BS
zm+B>X7r_L$i9;<XO>wFACFfw;Psz&JbyA7X!Ze`n4hYj7_LRl0l|bCsiqx~c)_8SU
z!eVhCt9Mq#*s5|n@}xC>ATdL_<#FoXc*jOSdyDt)wcS#63AcEMTSJ}Otej%v<=?|Y
znVK6=b<IoP2CZH|FY9wNGZ>VGy^}hnjxIm`vU%jrx4~Bbctx!<QbIQtzZ*SY^IDBV
zlHVh7NW9R7>e9^s7ySlc2i`+I0Tm9Y44%|4Z_eey9bz1HdgT^lNX4eii!nnjW(LmI
zZlivc><fCiJ$ovvj|lf<XO)|y5W?w;G%-R;7>Tmn-awavWP+sdo(X#r!5oLf(BWAd
z3j8lHw8z5Fb_S&PB8U9~9bwoZHjuhHf1vb{V<1V#YCm*YA=sZtb&~=A11+_(-e3d1
zeQJO?#bF2~{4b9Bs;Y4CsS0a9qN1b2Jork1EVd@&l<mILahf{r2;+6??c+00#fIND
zabI$e+K-_j^c}@Cq+N5)l*V<tWuSar3THd}lf!l<R<JEH?fVDNCNi2ib8Kf3TOpi~
zQveS+e{ntZ&x;-&0?ot{?{!WNqt=*Nfu);uqJZc{2!98=w;~+_#u|XV2<)16HQYsf
z7qP_;fYCVpcUlhviUOQRt-`+Cx|rCWXw2=m(H6=cyp77h9;_%%BQ{h5_es+TwTJzP
z>hJZY-w*2OZ9#9Q+4-~PG6m_=V!n6>Nb0_)F`hjDqRMJy%Q$wqCZ_%i7;7ed_Pg@w
zo$@MV9oZXn$}DynZE~|5^85Rz!32w64;{6rJG=I>xi;{X>1GhVlITb}bAQ8hmF%EM
z<<vo+`d~yb6p2{OBZREZ5_(Bz{T2o|ohEPwG>*4{!GXg<gf@~B6vZTibEe-hGM3O?
zfsi=77FJTUqB^_kb+S-L<6kJbW*%8m#75ZeWd3$h6*KEOv*eFXsj#tMIGOb}*?fXH
zkbQxr*rgkQVZ7Iv%L?G99r%74#NIpt+na&!3Zp{p9L*lQm-U42N1t!(4_x;RI1Udb
ztlKDfUPs~5WM4KO!e%|T(%(k!#<i>%r_e^1AFu)!0JKK8_qhi`GQ3VS0|oHnKTh^9
zJARxl^<N#P%Mz@HrV9hAAlha{W8BN*pWOVa#|?vi*o*RHnE?Wf8?;NxHHTo!Pn<o<
z`XyM{`>8g(C>OHfKv|)>rMU_if0YB?N4zMRrwD)Q{FBMR3>YQD9&9P3X;;ae20m#B
zRYZC<>$g6L@H{;91nYT`8LDLvHXMA_Wk8E0CMG_c@oD!WoQT@%^@<#VOry3aWBRp?
zS<_4-3CDA@u#<LRVt@6rL{vi!2Clq4WC9ml))@#d(hdFa);)cVrEQRcI<L-lGqr&~
zG}35nkSQRPnG?Nn&6tVG(DZlk=pYc_-S6V>&6b=d7H`cWBuN7K-&0xRDNOG#BlLU;
zcJ&6>H-;uZpY3G>9u(YdlkUH#LD20Q9YwZ=nsg78<R!{{rI^>1*)KRW)pf&AnDYJL
zhw(GZZ$2R`&q}W%UO56OOe!7MkR@)q2g4OOsLLK01s>NCwi^h0CTCq31paseg4kr8
z#cj?W(I5^Q5NKXE^eSj99WFJf;WZ_N$JgZH`)dL)zRROm%U-7kEY;D<9Ies#S#^Rf
zG1vjw-LAal(P7REoGn?Y|MpTT<>?X_oi-4#81IHAAHU7!eei_6QQzPCcs9hfR9d(u
z-T35Zhy3xraA_qQ)hWHQ-|eIB>H3T7z{!2;PjjrZ2N6(?<lgIC$;@NJ1*qF$>zlg)
z)}{{2_O@s3Wi<sAA8irS1U~N+Q~wglw+nhJ25?#~_F!855}@rLY+Z3vUMc5QY|W17
zMZ;I>egqmr-<^4Y1@k)o-9a9>uPIx8v`E)cM2XnX`2#3S&dwH=LdVjlm|Y?T^g^N=
z=)+F*_qgW)DabV7nx4D-ld-o#Xs;O<%pc!@6@mRvvk`t4LoLUPhOiT?ZtnEb_wW6p
z!6sOwdN4`_B(?OSv3VO<?&L_L<PR=Cs`*xOhJIeYt7hA0RA?eEjyBn4RX^+G#N!71
z%IGQuMFO`u{98D~C|b@Keljj2#&4UP8OVwcEW?j7L=#8HE)j}PYQBs!mt8-hLHM<U
z0{53<2dZ%@Y#<ADi^lz-=`-fb^!z{o?ddP`-aRA;9b`N)_L-HHee(FOr3tp7V&sbB
zw;I1A%`y0?DnlItSAaR$x#=nGs~8VknT(x<RVU>JJ@Z|^YXjRjThmQXd2>1-<J+Fz
ze!N&gaV$!smuur%n(i_X+lbA)Swr}F)A#EtlE&Xr<3&rm9Y|u`|ITEC;I2AXS997g
zGjQz_Kg7*<_Z~)$)I-ME5P~8cOL{*B^Jmm8eO<MG|9%X6*~OD9H^(x1?``(j&Z{RX
z)E6XyUGMKP{yR3Zv)<d=d$OnH1F_@VFxdmTmMt2RqVn*ui{)f!zk6qcP^mTMOow7V
zCq^3MNIWo3jdI2nh?A0DObw;arECPzggEzd?qLBnvF4n55vJOYtyfA)dQMbU+gZm<
z2x|iYYr3*KY;0!@@D=qe@ptO949O?_Z;hA#Gwd=~NJitXw*@~}`;0fOB8@(~uuq6`
zkv0P2e7gN$Yw}@4RvHX8!yF=bda%X%J6$IG!Q{lm)J(?oPc8ieSCep!&nY=LHDY8u
zF8e_Z#K9Peq>1o9Vb)md@7#}$QO)^Mbg1C0>Xo4;EwjW)DvT>nT@o4H13hQwY`TNf
zrsTI6-9wE{9kawoGu9hkLf`y(K@P#0K{&`D56J+*2gU@(P^X%*J7@7<*nW5#3cDqj
zoFr_g^=zwkOFge;EVj(7iblNMdwQgjcvz^$&>%#$VA*$`!h<NDrw&f;2F6Ek^G;!I
z9u#G%(-IkS{rA2}?$to#SFd>GTMv&pKol-t;-HB3G^@`wSUT%77ro`j9<v}3Q`Bbu
znR})nN=eQZxhG!HawzlWn`CuRfpT#o*!pyjSDwtp5abjnf807Kv{1`rQ`D9};;8Bk
zRVbP*IRw#d&LsJ@CeIjZfyjsY`%wooo6H3xB1bzn4`V#odh&sTJ(+x&m=D{J5(N(%
zSL`8NE9AA~Z!>$^*xwSHT1W%<f9wdpA2R#sYy(VaY6<1!zFKP3a0q2Po<kkUo&;Bg
z{$$c$;hrB`Z*CN1aQ57XxhI!Ccl%=wqtp8S#KfsV66_K*8MlePrTk|Q^xFX7it-Pw
zrCZw^3mKo)!3<8tg)|q%D2EDHG$-(jVHaGgk_u6!RouyV-P}po@J|g!Ix4~Hc2`MJ
z4$-Vqoa2GB<zZ_u{cm$Ff7u$h55MZ`y_`pmdbLiN*)6%E4^=^|C{>1qI8r(W&67=b
z@QAdUMdB&xT4oc2k=6Gndup|h4jXkV3IrUrURcq6;A4m&q6TtfE1kLw?krxtw4U<q
z&AMv=jaxkfs9HkE4ztOw!!N_a4v{E2qAx83%{ghmcSLVB|GSbtmWL`D9s>QhV~Sj=
z^(ondA9QbSZlaQ8qAa>8-^JLA`NPCRziF$fBYeXm#e2l92L$t1INKCbJq4tF&A-zB
zUayr1e<*@E!XPSOA3HyIxgWP%06+*bCz47KIds%aJA-$$Y1zAikyA^!-kZ(5VrHwZ
zO@^;MfEj0o&gXG0{Wwm&aVc%e|21=(jbte_zw$r`uXT2ifp=W#xJl8mvc8xq!AohC
zO7!nfS@@?AsT|vB5*I^W`icUrJ8?>RdrLkJe!d5Ofh3PVb0fcQ##onpMx%N8ZjCe?
z$*xAahq-e4)}-{ka{2~8?c2HWLf1k!9Ct&f;=~6d{6v4%!D3EeAnRZ*9}$bzY*kO4
zSi36X@cw<4PYs3#xRsdvED>mwnE_Sm#_={P*hRASKBW;K%lT+Gp7-4I|4wWpA~GfK
zlFxo5^L{~(MK^K+xVHY-h<V!t@R1ecqJ6Y7d%W}(w`M-&r2s8(qWagO+07YQu9<?=
z4c1Rx@;Rd!7S8*o)x{w}Oxdq0G}gGvg`8#TxmhNtbFBWO1<+YCWe!oc_%?Vj>UOD1
z)vOGUq638vOQ;Tb-PCXYSVGb<vN1uQBH&*477DFC)hR+XLXNg@&h8^@M{d+oHB1S;
zXJ7vrJ@_#Bzw>?Prau8ic7vaDgI;vrgf!Gz6A5_a^-nPV$}QGIG1`Pf(l41D6g94G
z8yeR)H#Ro6obp~3SnhsLQ_Y!~E)~asYJCaY`AnGUb<7=0Pq)lTRu9@)@P|PJ(w10s
zndEf!^EmRkpI522#DwGRz`I;CH*N1)a#b0MPGyXgLrDgje}-=f0{7ASapzP0t&qVn
z?UFzjO;_JMaCSCn$;adXEyvcur91lCM2bb}2=4|a^@C{HinNf*e-0D4`|rJ&#(YYa
zzo$74Mq*whw_H7UCQtA+`Mm&>uqITr+0Mme<pxmhC(HY<J1ynewQ?W&DZFpG+R(D4
zHJp(!$eJ+dGYBb4y&C$6y&33Aj%m`>tUV%Kk&SLCWOThlS|R!YDany*A;t_m<B!nd
zZv#4y(v++is<3|t9TR=&`FAEg1<=PLPf%!>aS~6ddi9Xu*+MncdhcMXgShLu)Kqp0
zz4fR)vS-OpG7Y5ry-jYgIcza}<N)<V)@P7uqBZO+x=CSltQa%GqPCD+l!VUraXk0?
zUq~y+I9HzLGZmV@wpV+*_rH^+%e=ZQb6<8|Cr_Z2|C^4SJE<%RYkE$Tn_aY;Dhnuw
z?W@+0wq{^~ULptsB;_UE);6syiME&>oHbcj=8LE=)2J)+8#y#v0s3uzeOR@=d6D!9
zZ$;$G^E*z0^3*i=pw^%Kbx1zyf9H$?BBkVqT1?TN6*>`E;|Ll|)iTPY$hbd`F8nE1
z0xx}0qD1p;4;aCZ4V1LyKDM{!4mZpo__pYM;1kxj2iDcEp}rUY_j66-{=Fjs^%9*T
zWB$aw6L-lpr9{p}Q4Zk{DdMXvdpZMoniJ+8x9e_4xYoTY8Fd_UrGT2&+GX=EKdgfd
zAN76?;yh?TN<e`)@f7gIM082j^WS=Hc7|U(SnG=`tVo<kXr^(0Ct|*h1VGh)!i?K6
zR~}sfFar>{Y@|d?9yx?DEI14!u8x=T05P3ReqKLu1iyX!xkxl+)3ag5R*Ne{_5LG=
zNEpd_ta9I>7E{P#*nI^yYjo4lNq67Ip`w`T-u(mAGUxfFq{KICTMLH{*BXnNuDS~H
z3ck$_7~A{@Sky}V&z;M5G5f+YiP?cb#<?Lv`|zrfdRp~3TMrN>yf&VIGSt2guld0Z
zp<@Om@zFl<cjd|I^s_%b?-#B!rG8F<dl4nwJJ=YAI^CpvKKDS4e3;QSCBqie?zh0U
z?94RQeICZXFI{%l3;a7_`sXwA|DM<I)|-h8*o6A^MXU-We+7JKBemfOv7)4G6=Fd#
zkLshFG+2I1Ky@P99K{lp@9n0|#o~WPY`Mom(-avbCVoEF#i)_9CXcs<wWRjE%jgX#
z;NQC2+BVV&DbOCst00CgJKb1Ly#%=Y->1NCWTc}=7iw~S$<q^RGCPBkftVCD4bl8v
z5I3uTuj{0HBQeX~Bul|i^&Yib&UqH@iXZ%-Q-o}UoQJQ|l+I1-`djzF8_(Km9kOe^
zD*UpC67er(SJijCrn`TVCm#n02dAeQECrE$ylM3(R%vL!j8Fsn3FltbM*B=k%6AM>
z1OEQ1%hFYqu7yTeHYwPJ<PT`I-)$JNM}{Jvv2$(OkfckOK1yw``f=q+Vx+Cz1s2>*
zL8Gi!n`-P8(2Rcu^fEU8k5zr)neM(pYH{bcy`TZFS~)?kvc>fM9}=iRbFJ9PuBw|<
zJkp;vz48;_gPAsQs-h|IfJTvAWFgyhFZGqAq#?NtB}3I5OybrCgX9`*iO<QLryw<Z
zs|!m<CQ}4EP!Q=dwwSs;pEVop5wCDlF<@kB5<F<Qp0IVDzWj$Rb)HL!WAotB#rhxX
z7qebfm>sbHeZUMM|A9F~MEyLP1xs1NTT&vE?7&~x^oXJft*pBVT!$UB?lqq#HfP0K
z+ZNI^3m!dV<e-2k`*#4jt{|SN!oq<C`E<anpWlDM1roh17pU2Vq^Cn%U3O&}>mdPb
zE&T?mbMZ;&-|^5y*8gK{Ze3bTHeGXJ%6;wIA>7HzZnry4zZ;h1CG_I2Kcdrs)}YpB
zu$42QFFd`jz5b$}qJrLI(<3(K?<_NiQg{3#T&t|D;vk2~Ep~?d=V8kB3;oZydKt^N
z3cli_XHWhaOZ?=2l7yg8f<lGPS4m~|ZU8X2n2~n@nJjhO_%pZYRD$X(6iA1LAI}|}
zIbkL(opa!t#m53wI$OcY_Tl@J3-;0H-w0W5R!~tR$UFmr&xD>WR-pq3)cO;_nLQyt
zDv4qi5BulhrKU#z87E)%KQXOLpi4FLmo0!v^dBZhX64H&-Hyn5pv#5)+UMz8Xc}sn
z`D;~VY)tR^pO1DF=V}p_1~PQRMDf8Eq+av>O?hkav5KAvbIMQ2SMRAoT?|WN)koa(
z><+(5VKhs$n7mTdP}!ah8V`A7{1(1{cgtu$_;CTS(uMV3KFUV1@?nXif6k}?Ak61~
zm!6*IPbuBpouYDII6=XB9}VP6qelxx#Gtgju#Qnlm4_Ak$q%m!g;eN*v|m%mMh|-k
zAawxD1Ij#_Kt)giwFsH+mr>2#cH_XZR}wyh`uc>w7ABznU*HLx7SXI{Or)qp?{scm
zG?>>x|HL0_q5m1|&d_!Zm{Hxik%)D=NY6mC-UFOV_UC80<6oB|GNe?Ynj-^z6MruZ
z=~%zPXZrU|jvW7HCL(8c9&vz=v&_`!Ie<QniY#!ibGmxx?;Dvr|M$~~dZ_2f066|+
z^tQETxVxkY5L<awuKvxl;~f8U52nLLM)u&2A_~n@i2WGYI@0AR(#0^k0%?x9r7_Z=
z<-+v`Z6!zkJrm&V|DY|Rw-jc{7g!GbxD3)<G9+R+!6RUsL!-afW2*B%Uo!Wu7Vy-+
zFZ}zjX~O><knCLWe+RzLz50JlywCZ6OuSe5KNIhNVgZz}<R_D+##x^OB$+xxa7n$S
zm=-8wmG$LPr0|#em^W_JMW<A+pCe0Lv1d=R8shx?FKF|9`Tx(!e=7<E%)Pazz?i>3
zxzyJ<YN1x22Y9+qCHX&CAc1Na-WcOT>|ii{5xF<9z7weEEB>jd*vbb;-NU3(Q_Qap
z-Y!PNR*w-y6VZ`#we_v=SEp5zTdyRQd9#6xEjES0*ytEFQ;HSo1J9^d<lh8ZCQ#yo
zx2|SVM+pQ2O(pGfSUo-xc_{w6NLzNbQL|3p3fbmvFx5<7tPS`7Y8ls1^^h%M^nLa)
zY28d8>?_af+63-foZkSnek7c-OCMY|O%!7(6hHk<^Plm@7XKSk^q1TVCA5#*!)ue2
zXG=x7Qhev}%XUBo4AxzkK;>KM0bf+A4WBsZnn=AJISOkT{*jGl_V3B<b>h=v<8*Fq
zOIc!OzE0(Cg}i*e+5*!JH0v0d{L=b=kIW)6t?T~?AG)er;m!RqF;^R&Y}E5H@*zc!
z)+9*m!9BA|?BgRv8@F9?0PyPK+fmVd0o+0jb;G1CF7ePv+J7co|JRnv?&~k5wEPm!
zGAtN%OwYY1l2T<TQkzUh%~ye)SzmjfUx^SrVJ;z6%%kzF7MBXV8w{j!^i+Wl#0~x#
z3atk|NG}}%ex-MddwUyT&YG(n#&nFxQdZ_$LKUbhcGLcOPUl}p$;5=Xn(#GRe)ir8
z08)S1ViJc7;G0F~*c4wfe+{UUAS>XFnen99$7t02T8Bh^y1v=Ny|~75h8qhD7L(uj
za{zMvzt{fvkGF#m)_{(({b)KM41GcS^6}M&^CoJ{=GoZz-o3^KaPlG;$P(D`rE{lk
z*Jc`2ac3Sj+O3@Yf6hpIEBaFQ!Qm44X^zz0VR1>Jgdnn*Y|;gOPcX9Q1c2yjsmgi7
zC$2Ej0ECas5U)`{qj!-u#tu3>I-aYhY=`Xe&vCa;t`7>r4X>Sj-x~0~9Sc!EK)gn7
zn8s&ntI-@H3+fv(*jMn!)B78<k3e^+opYP91M){IPfCW$ZlcsNd7L#V$?zndSt%h+
zL7YXI&8&vvtO4?T9n$AM3UY!Y-VL~Ett3or;*BShF2#6I%-JH*fR|=$;qEBq4X0s(
z!ZT$4J{CRBf5S+U&Vbc1Kze4um=_krfhSUt5-AU)7Q+H90KoROclO5U0AP~i)r5H!
zQ^$<mWtTtIE0^HH4Qz=IP~PPH0QGrHDu3TQC~x=z_%Z<J4-<YB<P86;RSMIAogM38
zX>zc&r>|ocSxAK{eBhcM5ZQ^5$GV&)K4W!jD#Kw7)XvWxVn^-9qehO_Zl)syE0cD+
z>SX+PE0fAYk>1(Z_PNieo>jYZwjY`sNIOiMIIpngCG`1e?Ug5AVbRq!l2Zg+w${60
zk}xkjdeg~;CDgv0Q9jHS0DvLAi-gyI@A@ZK(tm4OQyoy^d?;;hFVHan`1?Ftc&geB
z#bTRBH-OT7-ONrkOfk{-M5|m@_<KWvBNn^;0cc5cN%;&6>%T&tR(V<p#F<9CPhL{$
z_q$!u5cu)dsJ7Om?4#}iHMNy}mJn~GXtyL)iVIc-G;}bc)i5Kg#88w?!w=NXirLf-
z%>`>429wE#PxtV4+;rp_FTr6z+Ob7l5$DvV{}2FO=JEiM&wr0N-Zm~O8!wln(bX~f
z`TS>hCo=Sj`#bF~pNszq`@#bM&51vg-Az8smREpk&`9*jL-Xq~ZEafh(k}N@3|Z*1
zog2Fz``hvkUs83~s(7EH8YloDOSV^!2?{pSABDq;o-EL*W$WlK?eBAP0+A+uHs4Jz
z&rDbcjWSbD(+qG|Xk_(AQnEhLO#AwfS}Yhl1AWuZAY%p7)Y5^$27~B;eHNSP+i?H|
zLQ))KU`M2AvuWnJ>nkm9*q3`v(+?QvaQ<U@^WX3bsHjL4{Re;=+RffaDgViza_O9#
ziS#i^=c@btJ#)btYj7pgrv{PW14n#$pj#Ntn<E}$rbldadW13V>pztfO6b?gfdg>s
zuw20`clGmh$-*X@)PLfNQOBs0m`WYqXyLRko{;Wl<{;{Le^sS4`5=2(_#H`%eJ7)R
z0+@!wVP8zhSv7m#Z`-XQxuBajEgcpII&c(qR#oY2Q8L1;Rr=;bOOThHfu^imF-ZCb
z1Mfd4*qI*udn;w>B=W5C&{0P%p{@m<4?u}+N7*W-1y(n1y9R_b<@?^#s!RQHz>yj6
z(-lbGQ@B<s<a}J?v-%ddrA)%K`jDMN_`TwfkHn8Mw)CPs0ZgT4ls~&ETPu~<LH~#Q
zVQUf`tFJE^E=&F{3$wQI{Xp~?e^o1AgrwhfR)Z=%B}jp3Gf3fs{fw0K8~eh!kLl{F
zLYj=lo}D`HSG;z-8G+E2-a)}!=vJ9U6yqb4w$Aoyg_c2c^!8sfDwF&-9O!Ab+~Kd6
zXI0~%z$LDBw7th3SG(1a!>E9oDH_&`;Cyo%h!wvA?(aVb5Sn~g)`LWDq<xij#e~F~
z^|zp)YcWR{=sduLwMhuGb=CpImKg@vu&8MVkR192pL?DS4ZiQ$ILzVYd7x-k@}ci(
z-zg~DypSbw;f)G}M9sC6@>1G<R3<+4vmNuz+=hLiB5j@f$exija)3Yk^sz`ATfMH{
zqNi|6!Ne>w_Dli+^u~e3@t(#O=^Cmoar#mJV`^9bXML)3BN|s3$)IG&28+Zp%C1L@
zjlBp(lb6l;tw8{?cM@Sr0O+?aj`+M3%4P5dwNw)RbHsC9{6N6%6~V^k_?Y-))y$U#
zB21f}fnm+N#bdxi9L%j2H^g)|GBUUe4bF^2Q?A=N*!a5Q+dl$~6{_sgCw4*Bp6cVy
zST}<JB1MqwfN!8jk@d_!hYr5A68_h7nb5I~jJ|P%pQ6}>(xi+#{NKjpT$_q{DKB|Z
zUST3L+m^d;cp|$Ol-hB879FQw;!~HZo>zhNW)SJsr%Flg8*J%((ZMqESx~5jX>E+o
zUC3SplCm}UMe>`XfoViv7LU~1>~xbwQb|5k<v-^bzP<PVvGo>UQEu(`@SuQ5cXx<%
zNVjx|ba!`mm$Y=ZbayvMNlSM~cX$3D^qlv2eZOxmFJ{J>d1CK--?7%(d;gbBceul9
zmf){gr8iIc{0ObYxJ8(xk3eqvl#;<P&@l!8K)Q)fZhJ$wY$JMidPjnm8xy-DM7dQx
zy3CqGt-g>=n#DTReAy9Z_RifM{a6;duY|72b?w9ZN(>*>ZPxZ=-OS>Eq2LL?;%Ti*
zONWu-c<rUYT6bUGI345TOWq0p29UA+Yk%nd>AKadzuIOor%Z(CqU0UB>rdz;iumx!
z^W6oV<7rQw<go4S@9JCahHu7;5)dpgiU)(|mD`v(7pN547iyo(jTRUy<C?36M)uzH
zFl3wo?t^*Z&@o%45S0;FYRQE-reB}YUp#lrxPAesvoOLe-|@6;ztjHhrvT{ozq)6O
z{103tAp-D<0IvL!oD!$e&8TT`!m+rr<8dn|2(h8)qVL_>Q*x4eko*sB3N1eTqEn6f
zJ3q2*PMT8qt~Us~x)ZFmixy`(ENsvunET=?n9OozLJ`;3K0ca5{27lOu>aXEy;RcD
zlexbMGMnn=g2tdSfDWWIso=LNxwo_Dl^*eVBk~88<-fb+4F#}4Ax6DWEL}njVoK%(
zITO@8m4Ry8OZPffHxi(Z9lqMN@K$s{rNU!@VY+L8gx~9i-6hS&c}w&Kz=7@ztwz|J
zf5u~WFxWoo0KV)~{nd#)qnt65i2u3#vFg87Sk_-TZCe1THEN%~39z8r<2TG}ZnVQC
zHjS3?3>Tzu{I8Z3oQQy{_&#!_jzR#L1tW|mH=2*KYtNBEm#}}SxUZjm{?56iewiA;
zBFm-=^rMTRmXwQiKFB;bU@MttM5TnY7x>7qF_Ig7i^Auo`Wf#r5}(QIrT;qI#sv&F
z`@4#5?EsdYV|3C$5LwU%cfvm%xBBU?rn5?Vl+`-=2*lQl0@#-<W(0mw=$%vX&stA1
z7rk;ins3hZYmW4<ZxM)q+)Dn;Jnt*5F*jR?0B+xuX=UPYo@}9Do7_2{Z=Zity`-1A
zfIT;J#c+Xvgk7Cxm!Xd^-b}|~QWLzXHfp%|DyB<BJ}`A!tx;_Ipd=t{+riF31+Vr9
z4<5#-e~NM&3o1#RMhWl1i~j}aeqZ8o_Ydxn@&XCR0IT~xTTYRI5|-%XpY|0JK<cI|
z*St@x#k61zlb+k#&z(gQAlh0^V4B#=x8~a)a`TlmUV)4fs23^l@<Ljv<N)f(CXs4s
zzCB8eM^j+jbm|yykyi0dLWN-N<1z#_fELOy94%KQ4<EC<IbxikTshcx`c`<3^Zp&}
zGWCk_1x;8X!G(e?wp8TYu)pc`UGfWCo@ez8K-iQ4F9c<CTeAmAocOzh%B!uNk5s9r
z%iKg?h~}_guwiaE_%^XnO7j<wUbrs<m2Uu>-oQ{a_h8v<H8t}^Itu7YbIkeyfpMEA
ztz%7}y0>$v)%y~$b90){WY!YE)}TiZ?9{LIQ2unIVlU7rOP;%pS}G22J94k|=$>|X
z``erLD7^>wx2Z-S_Nv=ke!mgXeCj)XJZ&y6d)3gw^9<9>%6nz@`^{0y$<sE^)qGJg
zh_)O<D-z7!Q-FkxKd89Jyq3-8xWL|Mm4{A1wFF4X8t2m<`+V?AU0vvh^_1ndfkqub
zt8SfxD7<)_PwVT4FIBu<HLSFH{N$7)VAzhT5NfmXl*@VVf20>I00rP{HU$B?P%?#c
z1ZNC^0&~l|@E$u&!=H;45Qsk5?ey{a0Z^${`4WYGG|?ADK=bo~lhoGWqkriX5J#RP
zjY4q?4JR_~&Ab*}j?^$&=%EDw7_aNp4Ndj(q5HasGYw$IE5qkdfs`%rAs2-o^9kvt
z>4GAos0Q~6=u&g^@bljT(6c*Z+X|Oc*PkY&PyWw^p_8Z>Gf4<cM}CSghC|>LML&XN
z!UWjD7v%0RLPl9ToIg)?5Uy~uA$BrHHURgNI!)P{GNj@i0LrfWHFz9T!x=i>kA4aY
z?aO}`mrUuQ7ykR=za38wH!V|^g>#QKs^%W|ei9g!JhjJ^@+@~9r*N#BVendSxtpG~
z=>AAc<E>ZVl9~m;UbvKh+`RYsUv8cYV2TZCSmmHZfIN^YNoG*V5a#j)@q1;5${4rB
z2iYguIL}AX&#xB{^C8EF9jkhL`zz~W|AEx$JE6v9zw+f$x}u6&Y?k#)|8QJxW64@I
zc}hua@s~bnNBVvO74D;9*C_M4)#3PrDNPrj)l-$18;OKy!!a<mr?z_j{|iijAV~p?
ztD1!Y?<L)NseiwIUU6R3&Kyfm-9y?0VEpfTsDL7`nVtQXP|36|(173zfC0uxGv@g;
z{*V`*rndkIW{3r2$h@S5j&A47kSJ_h%Cj#;K{V@ClUXW{@rJO$5JiM-?N}x;imoky
zxXR4oU7G$=nWLpQ`tR~|hrPv8u>kd9wX?WhY^K|r|Csx;`{1fB!+-e$F4+fv@$4nb
zY4KcjR3Yw!I}sVwP4QgS>Lp8gXq4})pK%T>%j;`j?xGGS&LH<aDGb0n0dQdgqXkX>
zP|#z73w<92#p~^}=_C@gf1pJCsa+(4Xd-~G<8Fq1XFXJ3;egFTa!F<l%TAu{M8h=i
zVN`6VeJEDRlJ;@dDJ$X{&q>@>G5FBD{BDDW|IxB`e=P6ISJL^zD_IWHxQRmD#b<MQ
zY`mMQc}FnX)|Hzupz?b7iSYOncL9*;{sN$Zty#hZfRINJfD(o`TICG<aFW)QbUX43
z)_vQWrOle{wH8}y(lPGHYU&Dm8m~lv%d>&44&d(T=jgbvf!hDYQy%mXJd)o6B_V{h
z%cJ{dl9tC+!k;%k2A^+Qc|Dz`6hH|x+g>`?TYa7uzu%+*y5q$@O*IF!j?K$87=SDX
zU#;5HaAkJySK!Y0akaru?X08d<<+R#9Ic}6WTv0&qz9*=q9_8S)kLGcJ$}1>dI5z5
zuw^{37|iRx!5NVErOy9Y36Qo0DL^>nN33<Zl}=5aDFeh$k%V#-fbo?1TaPV=!Kr}i
z0Z2Hq2_Md_DS0@<<E<|D3d*%DalIS6tk$35tc>>;WNAhMU{9PLH67sLF9c)$KSa-d
z>2y%ItK9qVWU5fpZX1ni{&4*(%DNyCE6q;{rYuqAqdfL%m+rbZXI+RIQp1Y?NB!~`
z>i!Sp?7;3nNcfE5Ktt<8(%6c!2FNw0bb-|oUqT_V#*qrs3nzJ-B6N$RQw{I-5{Q)u
z1Eh!Z#;d9M%>g8DGH5E0NYIUyh3xqsXmj{KdIqE&_CgbrBcLaAM9mFf@}dFi?rwld
zQ!BaZwKA7S>}kB%03en}IF17{Nb3vuX-^yG6min=TieEEY~~QP-<dBta-sRN@_ux)
zLdDY`TGhMfKTiUx_{?~p8*@+D1C%blu7#7r<#;%hl+08(sDd{5mws!zwheG<PmT_S
ztp<R^p_a|{P#Xvd0O;t5FF3aQBfRfmQm^_<>QB?m8vj=rfu=>Ptf`cS6gd%!oc7~}
z_|dfuD2L}S0hu5}y(S0JCZIJ*k{rlnivYl}ey?p)oOljfgT|$>*0;*-H<Oe1=E_h1
z#{A*GTATw|CXBqe31DGMAsk`=wHiQ@s!(m(%FHY!;fH3uABYpLHEdpL-~u#D01z47
zUg_hM(bRMR{~=DbEV=aa@(DEy?eEs+lBqrTR|^2_^DJrfmkFlnQyw}R<YA2bgW{_%
zEb0T8SZk-z`oLN9kbH9P(?z^XkhtXFKfJLr$6p3lv8xEIa?4-!Gw~X``@NDGYP6lj
zKTJ&?ARhnISioPFg(<Yl6l$jAh1--lQ?9_?EZK<{I=JaUZ1dQ|4$7-)k5T!H)5h~9
zkj?;?nE79}2I_!pu4*ca0aVVw-H$j>RlEoMYAMejGy1#S26g1A{?=XsR4?H6N;b2S
z5{QIf94#v)`oDY%%il+Sr)mQd5dnmnO!J3^rA()u+Aj_gC|ZHij}pu5-#ZrgtFyJM
z)ioTzGyt^#k=%QEeyCDKDyEd-btTzCd8q!J%K%{59O2!1weBImyN?-I!MS@7Mr{Yr
z9(f@pOE~{-I-+R`?`#Riq?*-#to22p_rDe^ss6vHB?zP__J4PoO-TMfK?+3s|FhGh
zx@ih>$CzoYB9I>behh{EmkEG?snbxa#xL+aN0eyJppmAaGOsAGl&&C0TDe;AKU9nO
ze@Q;bJJc`}3^y$<uu#5EIZOcR^l=Ldr)Z@#04O}xO<CghN(55XfaNG&9`t`ZDQgBA
zO7Cod&_(H^LWscQW?QK0T)zzLTYH(oQSvX0V)A!CK{z^qS=8+Twv!$g7||1Wt?JKj
zJAHLGLx5keIGk4mMDxT}p+DzJO#kn$(Dz5{c`Emiw{{OBUY<dQjt0wcz=0-z0`BMw
zAT}9UUJJSmj8LP_8^66f2>YkblK=1I70$t-m#-^)ws1RSM)Rxm;$eWrd`gp2aO<Dw
zzaXxevgp7O|2c7}_Wumt+n~X=(i6n;Q@y=)@3yiS2webxjHN_ahmR{+slc2<JK9o%
z^yg>2e-+%aHRb9B1m2QNeOBfDgD<~&-1-z&cZAm~XMu@frRw`nKj8k?585ha>Lmo;
za7^`9<%RJC0f^!FLsg%w_bvlFk%5h~ULITh|JjwqL|b%&C~q-O8pwzt9~2G<(@_Tz
zvegS6w}ov39D!Lg%}eIf;*G{3r9kEB5zzY&Y6AYB<x)??{b`Z&L*`RqjhFFNj}yoL
zj(za13*b5A3SI`2#TsMxo&_t|A0&Tu@vjE`bV!A<_eL9KcypoHQyRDWi5@6I!^91*
z`~x3f<}xf%hiQDlBv`I~^{0ir&;LqkY{_V+5jr4~ML;4zwA{h4t8Jv_;fkVWny+2l
znJMM_;}up@|2m(((x0Dn)hhu6LBYM3P_7}F-4p(u0j1v2YjnT{+C}~}6fMGEcd5c@
zlL>U;815^!<ZF>q-k~MRq0?opB|u0<l0c|<JdH6y@#kj=zW-TVJdMroj=z1)riu_Q
z+!DQ@(XrGXP6mXxP_sYeskg*mVjl@7vfroq3H(4<b>|L{5gR&nLw5r<=I{7tH|LIA
z{r?OL!i1c55~1TAhZjiX+>`6#cE@#Z(IO%NND(}E+E&!9NB>-tI)H%uc?EDSP2D|r
zV6%MN2&<R|Mx2cmb)3-)oG?d!x)v1vmoe!m;wJ%m$cBqQEX}`1czO1gD5a@+r9Jh3
zSj3W8KvJmCY<q+*JCmI`O2BQL)_iyh+<gEKfrb=KgVCR((9-_ZxN842E^ybtr3hq1
zARN;;L+GA+w#)+iOTRoCzWmjN_QBo$|K5QMOodQ6U&%@-yq6AeGe124+(yIfU*9+X
zXUrmYFzlgghidrDQm)`vKG6fzkaGE-<Ey*>)j(B_Yy6KO^Q>Y!V-f&iC&EADS8q{i
z8+%W^>Zx9OHoR7W4=n1x495xr?AhDlG;ZagAMyK4`$JER(bIs|BM#9QtNU_KO9N&w
zP>0}O2WHGf0~NDG0JMHsYXK&L20}w3S&&{lfhr)L>F_Ptcy&hCeiIQtjnT0P=V|8%
z==odrQs6+|%zq_EAlf7saO<yXKA1oykC$wx?)RcNAxZdikcZMPhz+2eXOZ9DFODIe
zyncAOtj=vPAiMmF)+n40R!{<2?+yby82^^!nLBhB{}z!n)oT4nwX@JJ*5ql>C<#~h
zFn%NgfkIC%>AWf*YiAZ#oMRiRrd-~gN?O4XJ;b%&v7&o`x!jwT)0NVBo4^_(JG3Q0
zR{OM_rSP(EsTH69K7aYcmQ;8@T5=n(9eCnjQI9(A#a2?JfEf}11|yH>p;y+RI;4n@
zy;#b5pR|D$4^^jn>6cPpOmf8NNbF?h>4g0nq<u*J%y=*|wGK>?%OEQ=Ps;%oL7&%h
zd(4Uc%+z}W^J7#m$ph>w!wuQum`mRIUVpvZ_Pt%Qp@>K6@FXzis002QpZTo-HJ+#h
z>id%7?f)lKDZmj4HZJ~A49km#D~HdT*78@+xm#Z!{Tl0b2#nY?jT;(rcnzv}8=x7m
z^fbS(b-%u1;k7=N@pOg|fQY<uQP;R_V?WNy5q8}=34DH7-s*YwaM6gFtt%L?SHml(
zH5frLXY7%`PXo+c>s@#bEq{3O+OqGICfDKf8})h|)WUW59J*fUht#v>Fg%0@jdZkU
z_?-S_(p<saMO2`MgDqfzotL6qp^xsPJ9U}jx-?wR?Ij*HFF2@Sr-`oXNNVk-H@iEo
zG-jJq>1W!ClIJ=3f|eV%je!}x2Z_Cqo;~d=0y{7lCqKU)1dlC2@~7<V$0^?Xd4lI7
z*2mgooZ82mE6(z!uP5CF6AqiB&*kOqHdDL4JX*KI8MrYnHfNje$Oufs^=rYMVMhO6
z>F!^qCN_V@3R1KP2g3y<5yRs$MQlrtO<RL!0w8UbMSPdrf7+1fOE+sP&DdoH$`^(S
zg_U(q?%*rT?3hlZyyp29PT`#6{(hiF%w}!B>TBmTOvpTRN++<RxX)#FZfbkN@IvCB
zbzOgUZ2G3QO=XizvyB{5xp4$)7<#zJN(}j{zXd>VI*Nvw{|GsnGZOb$wCzyilH-AV
zFx2U3fduB5&98H?qWx);uCD#cbuDFiy54{sD%(!{d4gbMyD0JVT`z0foe=<rJ_%m~
zOHol--A)5nUK;brF!mEOiUNjh%=6+BxKIJ8!}6!O1g!^=Of#r3tM>C2swpWp4Qoi0
z_D&88QJncDteX{%4&ieI{+FpYFK%e{+35JlFT3HWW$fA2m0FL*8v_0#)v4&bFQ1O(
zC*@|;!PDxU!Y}-q>t}kcsMwLRO8iy&4L|Nk<Oj?{vZ@+nHxN|X*pQKT3Pqg-iGDS}
zR=!WAWUuU0oExla<Ll%<6pgAO?pAIgZGkntd|V&8NIlw}amg6FmP2Kx%O7Q5nwW@g
z8og!t;}SunuU<B)j;84a!xfZwT^$Il;j+;~3|(omZE{h`Qho7#Dzs}0Bq9XM41*L4
z@wrxCFe(VX=%CyiE6+njm3XM7XKK?2t|B|U{^<ORfw*YE{6alFDmKtPLb-OvO>^vm
z=%-hiVO?-Eb0B?gUSWzCF4w0<aBbX8bwPrVsMRfEZ-RXgXcxX6@f9Dq^>2+OlWGf`
zr`UMQ`w$f?w|SIt*EHVQ24#;O3(mm{=EKOYChCx^pSO+*Uz^{6D9YA<#+#&(BsLe8
zvG&q9SeeXNh_Njy<E-a_2v6_cq)16FDR}kuG|urZYpJ5&_m1E#cwA;Pv~LA5I_(xc
zHyVW<Ze36zZ5wVr$lV@=DBGieqK}?-PRHn^o)70Y7kn@jhqR)f)AF@OhDU=f1iqK{
z<B%LjsN=PI9NM1LqMavfcp0gTW#vzg1h*BDXHeZdO~7DuH!Oe<t@@ga^6Z}=DQDH&
z=^Q2HKbxr0zeAs7GrwC37gub(=^h*23`d9X-q_!t&vAR#E4PBt{EjjY57&QjYD<z}
zabt?t$v(%_jCAQGr3#9hePfMCL&Z(#B|6&r6Dy@TJf(9-0Qf{ScE>6L*T%VzI-iqn
z<ozLW9Tu*5si2@Q)xjoKw2{3gEBawRb@gPJbZTMMdL^1tI&#!OJIap(C~BdzRtjYB
z+oxf0DaD9erptZq^7j$hVy6`kZ1^fIj1(yML>{^FKR=DL-$CZCBYb>Negpaey75N@
z;eNlQ1I`V#XEtj{->n?|u;98cc%6A8ta`z0yd9=$JU?bu@r6`OzP?lmbt&i<2-pm|
z{(zOsSMvC{1bN9rvNE{0!19>MGPvl<DdZ@iD~!Pe8iahTZui*J?7CvJdo<avdMsk6
z<Zz&cQrsJUc$JPj;kAD>+Y?iZ!K>O-q!%G%_byJ<>2cj}`A$v+r)gjv5$_c!n0S%H
zbN6~FW`>DQtIgx^><}ftj9R7rK}50<Mt2NRa`~p=uoxzxYq`<=9_Zhh&l<P(LL&$%
zN?ZnMX*!Bi{)1W?J9%WlOycZ3|H>T}Q1=p_!sv{!*Xytgd--0<(Y3nj7=CyMh2C0M
zP*ApCifh<t_l2W;(2=JUH8IUbBHp{A7QA=c;gqhNf==s5sWkyAR0mI$AW)<+d>#yk
ztn2DU90E?rx;!V6j7^uClK$71!q)qXwX2){QV)Sji>#9;hMR`dM<|SpAUKZ?*H*~=
zg1(*u0jtEFLT6v8C95$$Z5<1-TS%oxiN=c|zS_m{<@03;5CDdqqT$gBY>w4?|2bZj
zcP@roK*t`N4=_0l^M<$%Rs<i(k*`?S2!~_$#7b1??QP(wJTLb;iauTba2)%jdkZJt
zSB~jf@*1QMdf<lwA$9%kVnVAlnY#e~<g$gWFf9bwbzWG^e5)^_<MvD8iAw;7C_9x&
znSBB0&!kcNEIkS5U^QnI+t$fCr82dQ=b#3PboL9Yz1zis-aYEMyX#c0aj=wX)v*>X
zXsisqTo_G4Wvg3iLXLhQH*0lEbwB{g2GGwxSp{}Yu(APt!~W1t7a_caC>c)RCP~NW
z=j3uWF*D-nKAlyhW<;l==x{ix6zIibGibvLMmW|)88*&}IzAJ}tua<#^Wl#2tBWgQ
zJ6L316Xn;J5A21kV8i(tpu5=`3TNJXkO(Fp6`bIImGlGciu@??0)b3^;I$Ve#No5i
zJlmg4WpqC&YWKhqJZ<q&^x}^Yca_o|;#DctR8&-aTfJQyFbcJ`Tc0O?Ev^J(Hoq{C
zeek{eq{GOF=T^Ua`<$X2LaCzVec3+7#O2T88J}nkf*RDGdbQIi%~lAI8gVms)x=5o
z0i2caZ2PB9L+W!CWM2t=xirc6=tn1IaeuD|7|ytALIf9KWOZVsw<_n#j{{d3jb1Su
zN!o{u1kOmy4b<FUS4@2P?jWyUalS=L7b3ntu9(75Cx%(oHp{0Zk&NMrsbK#CbpMi5
z94x$uTU#Cry0!N0C3q0)OzAW67+xFI$h-!VTUQE;*DIbXX9MiVgRTgNm(b90Nf(5t
z{`~#2{Kk=rXS}!`Wb<>TVt9)SywFl)C9{6r+N}+*l|gC+Dl%_<$_OO-c1i7Ofg4x8
z*6EsgN_cS`eK0XXT5dw&aB>FKMh>uo?PRljU_FT3!n?!mrmOPcyh`p%j6Mr|=_5_^
z#nwClR$jgO>kW-&n(!HT+<jWQYBgMtte^G%O?w)f!-8;sKML>tbiCFUZBR=o4nezP
z59yYl-TimE5yNNKLpk=QV`%%U@n2e>=8JaE;<^qaEX-7mn`3}M_gcZxjw)xs=}ovL
zZA4yaWrL8uvnxm5x&kYxd#Y2?aTH4eZQ1WJy>zUGtZ(goO+O>8Ad#e!bKwe#pD@_?
zS92CO0LVoq0<s&yoUi=m_fzd{_%S*`zYQZlcaZ)H5%w0)R*kYUO~y?7wY>{?6YlC3
z;Lc%XL!~Uc9(>IP6y;DsPgg&p{GEnal6T<zxk#d(!^f^<vX~|*(27jwxU6MGtiDi^
z6F=iU-(c`hb18mJp`$_U7ZkK5yy8#kCTb8(Cqm9j;-l3cjyd+wQxJvXh7O$~q_r1t
zF;$NE1o#NeV;2LPVYevznL1@jIU^vp{*dJ&Wc;NiD``7?(Cv7c1o2F7bhLQl^g$29
z;X*+CHmc9adPg`W@*}X@<1m?QiD!w?qZ<yA#`A_$a`J?e(o2z*$P9#JWVjh1b&(aA
zfIqCzW*A(W8$hPI!*={6@vty%F973E*INg(%ptO#qSo?j7sov##qrbq`tplSOdaZo
z(AVEvlSs_%ERUCQ`&#UN>k31vowN~kTSXG8m!26k-QQVeEIkUgFG${VB{6*o0kc=3
z$<(1O;|k4G$^*TP+zWK!=T<ANvy1I{rNraCj&97mIBSas1jlr6IeHfv3wGjYqeNjY
z*O5l`b@83O43{vvipzHf;u#T#ZFggs&oEKX8&`G)W19gWKK{2%>`(WWJRW*aF}zxr
zkF|E_h)8eiN5Y@)$EYx9->kH_Z(jFd#95@GsHkNhYH6%$DMy`m4LrQc{S)}zu|r-p
zKAJ2D`gkt)g+)y5;z&%>dEIR;KJcmbJ*pXrcs+;qOON*zmh;f6)>~Q#wnBP4FWer@
zB=h<%&Ph45R>g*ajW%l)KY>TLm2rIbz#*{=i{M9rG2OcsQjZz%&fhl6u>$23QkIXf
zXOJjR8?3LsE5Aw|&Ew%fc7ghk(N5G<ynjzzkX<fuyeQ+LwVHF;DLHQ^qM?8)BLLck
zyTt*4`nSXdNe0DJ5K`j}ZU+Lbm{BoAWCd)1XhWdcNIaO2pnX9!ySZ;-ik~RzNSB&|
zX=!0OU(9<a-NQOrU0>M9j5mV>(sm5yY0*=Z+0vtWR=2}9QbDF>6Hp>_=0^k{qtPEm
zsoF}CA6Z6I&oUcXKOiX9`jJe2e|*!|?tGFz`s{wvM><;3bn6(TA8nv8E>fY4OD;h*
z-2Xw4@vZL|wCy7U8I&OD{c++OweFBKlf8Gx9FnKjKR;OO0#WIp`lmbY0os*i)m^-L
zwJAoex?BttWnJ;f^N+oK0u{Sc^kRbj1GV+0ORX}s=Uf$1l6E=su(IoxtcSrChnHYF
z2Q|}9Nb!<R5(YErpzBU^PIZ#P5Et#aH{VIeQ{t+fx#=zV1z3z&(#?JgUO`+6ob<5t
zYQFo)Y1uXry|RF+KmbO_j3+Clj<~dd<GYI^DDT!#m^oU?<10xJQj$wEACG;P{PO>k
z%5KhX0vUOQS3i<pd3WPH1dYxhR<qi>`HG<n))-!UN-3GJJfpR{nbNP-F?C|O<fg-u
zdH4lUP9Ged*~b~48ocXTV>M>w6^rZ6EpBQWxaCPWq&VKujLYA<n9_@K#0Ev{Ts%@`
zeuExxIh`?^fk0*m64k6=aayADI$67u4mMlL&PHN!=yD2it9x=&^K4wBXe&+imtuZg
zkI&ftGNAr+9z<#6HwNx9<aM^q``%1;oHfXJ#J84e!1LaZpbQGFiQe47jvOQ!U50eI
z!Uw?|)_DO(OnGf-$(3CCgkoGobv7M-E@{w5>5&b+ivuz#k#KvL*Tj(6_EAtR!WsL%
zcH2FR4f>P;GYe7WeeXMq2%LNFmqQfIyXG`iYE5inW*EN_ePe6*_!=ZCawN!7S5kA!
zXL1fk+~j(>7f028KYC54nRC89x*w;}>dFjESJ^zK)e1)n?5J!ip6TvBF<LCGcIUo_
zOqVQP@Ro#SV!gB$K!AafNWYfnQH#Y<4-5_twwa=YzhnxN#$Mj1PP6)Hs3bG^hM<PD
z83h!s&ROyK`jV{NGxet@U!Hcg>1|`5c~-5AG(<|bB_mb}gubz6?4#2LmY8+04=g-4
zLb+Dc@;UT=N7sIW{xhDWH$<_^){`VL5Q+P%sFTz`N5(zhaC{w%8L#PYMi9O{PN0&Q
z=tLKg46q99=zSZn6Gbfi%I96L8nq&F5|3V+e>OF)Nj_tkjP9#Jn8Zz-y{H~4g5_2-
zVCzGY_#rR2AIm3s$jrL-Ev`)Mi9tZ-9gVjgM!A9YC>ZpuF^B#*B{y(}?(NfpAZmIl
zYQ(X8mFI+BJHi($sw-$W)?k9Q1EfuO?{-we^Ap}=nuzV{hilyL9|0Cc{n@wR*BwTg
zS1uCpQpF<B^vrOY6F=?kdTF`<&2p3Uf&b)B{<P9clXer4@@kZO8qHz8Edi_j<++NY
zb37}J_wxeLszJJwfHOxsEFwvCbJMVLPU>xH3l~txq(B^xhv!(t)ErqPx{S*I*dCAv
z>EuQ}Gf9-ikxv80hg|TTx{Af##s)JaWE>v5fZ*ayy13l%<D}xRhC;ms1z<2NrJtC?
z5od**!Fipe!)g}BJ-I~|$SaJ(9R(*YRT{zVDpxm~RaGjQj6L^xQrf))6?h(0y(a{(
zR^QM{>7jnByG{fV-scNOH;&uQ(}#yJJ6smhpBvmqiP7h}6W<{?dNv@;l-e3R&ji0I
z)t7s^JX0vBy^;*~>W6UQx;?MmpWdsxnUxzJEN?^qiot%FFqY?k3(pA$I=wF2D=+m;
zc1@H18g?lX;iM4Y*&J0AF1OMfl-xYbs${7=RF`(_7$#~%AO;Px*-*$hFD#pR+fz_<
ze7s+@`nlVDFPo$(KZK^=-A)ZIG-+2hgVS{KI5>ao@Nl(hQAnOZ#tyVQsOt)7_PaIs
zfhjEUG-`F`Dt0D?4q_k?Tf5K!wUB{OHLZD0K|#T`^z5!p8X>Wxc#W>o#htG2(Hfj;
zAvFRQwsv-QgTNQ50^eL6!l7PWGS$*P5?V|ngJh#_QB|&Jx2C*3evRS4x(PWIl`@WT
zq(&!qrbFtIJUA(DfM*aBH3&*p32R4bi@qrqNskJ8BnT>w3<f8Y5s3f3{qaqa(%$+L
zhNv}G><rS_7gLAHz}2i(;hz$7(rEDJxO4O$n6_*{$q}Bw=%|~^PO8FeM@1M6&vjPk
zCr;X4ub3gwY-e&f3sYeW<%#<E!AAFd#v4d^H90Frj6)b9jOeO$!;_~U@wz<iSQbKP
zV!|nooAA|)<+Xt*u*E#Ic)!)TlcemCAAO&ao@N1IgZ9VHtA$*+GT_Sal&=$ao4@#h
zydlfHKBlhTVSxB3ft6ar^rt0Ov{kv+>NIJGvm)MCRSucbEK8RJ?x5$I>^CAsb1(?<
zdib5{2e*npGVFf<<%As36s`Mc-a$?%lOG+!;&pCVW~mxUnnd#?!@^ua>!SHXE1_Ur
zL4#y_8n-HcqS%-;e!hd0lSy|P;KY9jS$XVYFZvnQN^iCwvABfLoeK#>20{5NDHU=7
z7XmL%;clmchu3B4?X2uZa>*nCcMq#vNC5CM?hE9%KOi>e?tJqBEHSY!y=dt60wd<2
zzjlJv(qd>Tbd;X?cE&8TuFWFaKT0cBAm7&_Y=unL-(&IaY*d|l<%8%7A#su|I*ppm
zrYa%ph7~Kh7@a0KDB9$k_$6y3a7WdZ%S#iPi|C)5xDi#{dq20;-Xh6&urDhx`yar0
z+dEmu+`Z3<Y3r1q3T?$(WY6%6>M=So3hVjt9gHcM2*1nM!&kX8jkAcDQ5`^bLEg^R
z^cskS9;mekly7<PIX<Kgm9fz;Vq-Fk36eF>FMJmvu(b^xSeS3DU0mRzM_K_mhat=G
zwb$c!S3iLAGm+!A_cC26*6RQ*|FS}CD}q|H@Kz-6d{i>;y1D^QmjXwp6NOq!fE9?5
zv-|Gafz*;Q)F}OXOmKEpo(wAR4<F)Me+?Hn!TNwJX<v4!>ZiP)EX1OSq|Z7`EI|z_
z`=Dyh>*D+vy>%XEL^|h1{&2Zb^=!jc<-AG#DfTCQf8Urk7Vha00$)>#!U{#oqtx{j
zxVQZmEhsP$AvfSFL42}>mWSkR7K$}7?O%!J8o_l6%Y8yu*!0nJk+MF+BEg~v1O*jn
zY;e%@_sVU75#kj_Xy>h*bjjd%CS%_%X4wjU%!A}JOxmcTes2_^GPdM3%I!kG+feva
zto|fC9U-{@xVth-sgn#F>L8SevTz|9P$U#<pA}U^1~sFYLW0R*eTICx#?$_=^HoSa
zsa4sXzZ*$@9M9<UufZ*`6T9@g%_fU>gXd}NzEBX~_@`4Zi$>NIw9BHW`Q(s-MxRP)
zx3%#wq$#8lN@@=Q>H-tSPzmpUwE)ym?G(769HvWlf^~WrVh&`#>ghL)SAs-UutrjP
zyiP|@N!XwOeWL6qFi;K&xHVtEfSM|%I~!v>I?4Gv$d=Auf&R>gd433fwxO!?X1@Eo
z`|QcIBs5@zQ?o6St&h!fM)ZTbRqkM9{rp!eNsV3;yifjVu)c|bPP2fP7+PApd^X~j
zcHDa#9PES36RaM`JmPBaV7?mgv7yJ5stf7<#K_#;q<dDQqm|S~R%Sabhp<>OC-8Ar
zWkUasry#%;A+g{2X?pa^Ac`Ue=_7-ur0jHN@OS}9=i|E8>>`=G<X{zmkxXF-vKH1a
zkc28hPWzW;SLOM9u7YnCMbu4i>->^m_TGfzfJBtSd?0sBq(ie$Um|^TjH0wU(D1{3
zJ7o7U2IP{E?A3LQ2zq&H<8@eC^oQ#<hC9vK^3IQ=u%tE;pVQpeJ9n@+$q2rte}oiC
z6>NN+bHIv3`0jupdu=Hu*53~VxS|(`%zMM+S4I$e4AQb<n1N+b(`kUkMCMJH)2Bcg
zoGNU%r2*c~Ln$yui8s&Pdq+8vMI^KFP|0Q`oQsn`>svFe+|pDSreKi-9Y8sLPwh4<
z3)G=sLb&HLvxAuo@V{lQxA0y#A5|GMGAr_8yHY=1f3xZA6tlcUR}>s01(A-oreqJS
z-_D`mqJ~1xP(_e?7^T(6pO=yitjl>}G1Ibf;D+F<#&U{0ZY>oKb_P<l$$5W*^S*FQ
z^&;L%!dNc3Go#8n=PYW8@Q9k!=5t>&mh_;e0{8Bg?om=sJ3lQ5=T9Lvk#b-mKRftZ
zbVrij`?LrAb6qGUDtf^zkw?-F;6bDl04ySNOisd+L3hHhqodP#o<3kZ_{+<@KXcXB
z-=8c!tqu|n8;fTi<v=hvf6-B>aWg$Sy(6qEU<9@v6%EVby81)LiCIy}ezj0zl)YO+
zt;;cDaY;p9>#%3U`p{|e=a-D|CP)pgb+k^l*4+8+W13;ZCE3+n0ZY`GX@41lYRJJ%
zZ!IswL(<a`MjH!g8h1FR(NTLTT8TGYNl8}iDganyx;$O77LJa3RCJ?VKwZ?w@FV$0
z*t&|~l3!C*O0&($)?hJhmGrMT!{*hI;yZPCm%MAAPVUHb*|7zEC1P`v_=q7xdm@L=
zoL}$a@Y84U_vd^&$pvr4=z@WI`P>BEO{Xe2b`%&ENa|V9M8qrYVuDYnbe~WdZbshs
zW4}_vyIr32<MkNfr@e77F!1${O9m%~Ce;<4SVRwsk8a93be-k`JX&PxTDx8w1{ql!
z5tm;VnIsBQuw}Q{dtAIUUC;M^4dQ9LRmm$1vbWbLA`=x3Fd|xh9{o$v!Rm59f#?}I
zup;Qaqk~ZYMG_noISo`uGwr}1Li`(1Xc^=B_cTF*nGmW>r|9s9JDZkN%1c->c#2U(
zsOvnh`NX3XMCgkPU@phL`4uqGrp;$H2^5>izX9by8D(O!M~qJ{t;Sg01*ZzC;x;3>
zQq!ugYxmO|+*k7!!F^VU%foic*q=qw{ME>PxoEG*;O4|6`HGL29j4lu%Zg{r;_9Zi
z+e4MJucamV6T6F4dAsczsvCz(*5iU&GCJ)~bQH_9R#a*ML3V#Lh2!y!)}IWt&y^KD
zV1%uUThH?Yb<h<*(09!Yz;4;bgae09T?lkpkPf+)nOSoCWzd%9=9a@Y>!^Simzw$9
zVBV;iJllF`GQi&nBfR;N9#%qVLHVf}GAXkN{wy^%+1e?cDGrvx5xgW}2arA~JD`08
zd?+K-sy8e%Pu^EvD&-<pcr)Dw2Q6!GjiAx(^V=)<OVs!SMqH_a`kYbLntGE0z(6fb
z^dIqfP+A@_;!yVyh?%72Rpi*ZA?fzQ`BE{Bmu67vvbxL7hDXqi^rmF(E_A4;<*_>Y
z=<puQ=S~EdklR`XxTDOWEhUP~?sN2&)IyiTwcGBw_e)$r)<L7I^dY0p`lsScx;`8J
zB`#IZ;n5<9;WYSNJPW^ltu%-5XW|dkF)+$|J_g*V>JW2YfnGA!!7FojNJtv{4_ZFn
z!K=3r@w7zx!zqB9D<UQf<VP9L%$U<jZ_A`13_2=lG5TSTQi&!LLP13~Dypc0X{$^a
z^etpodSF<T6pEB8^hDxI7V9gre&X_Op2119L0I_np2K6B4f-BG57}Qb%pYK$bVe2%
zD<2v8AwySa|Cylr&3|7@+{xy}`asB~{VJ!F#Kl-gjxPHR{oewY!(~bpNn=*6L87cv
zBOx=$Lg<(4d=75pydS#{PY7W^Vx57yvlWJ!&Zm|+?J}ueh^`Mid#l+`c5YABHV<V*
z8;iNcPuuCtc_xIpyy^SgSnA@~&B|I{%vO(8OK8*q>x-33O1#+T=iF({ZA{-iU-S7E
zc2S!vN8US$gGBVtv^3i`CgaQw`U}h9BeotY7~8cEQy~Ne|KWbzT`4EKqvs#oCDCHM
zTJu^`SXpmQo=U4(pB{HI=+xn9n{}UJM$z1#cl}{XWQ9P4Nsb&;nzYg)@`pAq1jLdG
zm-ai4b!3uK9vX8KzeX&XBBl@p$G8c`a6r4mI^MmCA!qMN&qat_<jS99M3eJEmf^>i
zoV!gr-sBG~WO6ud1h-In*Z4mGq5*w?>(tK*_StI9Tfeo3j4HehxMB(02_yfYF+MXf
zsVckH(k_UiN75Mx-$#JF$#d_n6CvdVo;7k)N8gs2{!aDI=g63~wG6(V1)|;(I*jLQ
zPjWZ{(=qd%wFkFksRVBU?uW;lYdEfbeYda9DS__7hj51$n%Hd)jbY_8JzJGQj=bJq
zZ0+A@Y%$bnm$)I5WvA5eVsdwtXXhp2e*F@8?2)bJXLzbyIj0sA{VtTvl;IBkea#Y-
zq1cacSeq182}C9zJVo$Udic7MPo#t6HPJwLHnRQ_m||?W%G=^;0u&Y#JOysZqH{lO
zhh0*G438u$8O0X^K9(}@kY(dgb@;wjo`iz=#-Ot3x$58ViTs!#N{IYyd-PGTuXqUk
zcE--pnch{HbgkECEc>>{9qgV6&<Ff}+*v<T0XUI!!_(&EjFuPP#9ayZX*bGSOQbac
zJkej~tJbJG=pJwjvHm?JPQ(NT=Gh30jI1EWxVwFgePb00mf)w5qdmvNFc~2yjS}Ke
z5MkMkG_1`Sw#Xn4o0Pmx`%R-&!+Q63Xkw0fqh=OYzeY~SQ~kppURsT*y1ohtgM028
z?%2+c(O&K+x{s?uOW*AzxGms;N!((+{wmT>dN8Tft`-lpW4M8b{jo7hS7sbns`P&A
zXB+wLi4T+CE~4iNVl5_FY|PfI2G)fd?Sj6j8TA49Bh~zARDzXO!^4!nrK;?GD}jaj
zj`0+T(56?xYcZBnRna#6P4~8ksN%M3UOq=rIO`SGmRYPDE9Vgil!JZAo>3YeAGVr`
z;HA0<SfZqp`CcfY-@0}j>jMBdW=xn8=B~~Ig5ewrAU}MUl$eP5d{A(WVkxhrBWW;e
zUHxnNCt=k~fl74_RHuH-umnrn27}c^Mnz+$D6jANyGEfU91lyVp{dy!xU44^K5XD%
zGh*W#QrdS_eb06Ij4`62%GVuFfR*y?XV_qXnt<M!d@EU2aPY6ua92OpA?!1FdGt9<
zj)bgi1ih-fw)9q>!!s0=*CN;c=x|>#Cf(06Ic;xU!{HMa@cP9zU?vsU3#pAV=7B7P
z#{gDe*`mm&w#X9lisy`d;qh}Sn@)QM1L`?6k$40<HNhtNP7S;FY?PSwTUwcT{GxTf
z;ZcIUw!zkd@JekLlm{wmkFG4E-bz$K(jcFe*-flq{Up(ME&(uLI=@J=BAFu`I)IFq
znw!)`IGrfIlW0`5Q@2wH90N&()PB~m&DS0BjejY%+BBC33BbDm1yM!D6DX4S2w3wS
zWG6NOxXWKwRC`u=<H!6Imv%|dcqE<e8EnWNE69zGKo~E4U^fJDrcez;fHJg&zxU;I
z;9I^;fyxSOeMRW_zi5sZtjpAh118Ct*LVXb<;LU@*&>*Ci^s{dbZF=PS_t$5)mIdC
zT{QiO=04w{7<CCQ5JztT9VmmRBBCwz1K?s1<fTTLXEd40cLBon?r^eph1xH>#H-?V
za#AB^JteU<v$$qRMG1^>c*EUHtfu}|3<Y`pL4BxjM1<!Yw~72zG#a}UYx$4SF_--<
zHr`Fm28V|gc&8M}@9oo%Q@<t#>mL{~uy8?1gi<+C@1N$eof=B?qdLJr)t_Ux-BX3x
zIQv&$IJ%uh8E^`G5Z1@=(?dl-XvF|noa^Z&Q|qK&yK{e8nvULZfSRQ6&&@6L+8Q(k
zNUQT)tWp+tc`Q%ms^Zq%AT^fyf|lH!xm<Ks^bg;cKqcCR=L{Y$5-_hx5IywP1v{Yw
zM!`k0cgsy&B*tW2gDv2ue1%&#jy&@;ytd6fV#%7#my)O*&5BIuS?gO2$iu)D7q&@S
zrP^{Wlqzh8p`>xd9cyjiw$S71Yh$XCnAgSNoZ5G;cmM>NFOBb`AY)VqH#1p-GFp>W
z(-zm$kF~gDjIP%dvD)xFp8UZ>gv6a>1JRx+9H&?Vl#w0yP%7t>ONg`4RZJ+<*`=|i
z2`*QGy_5Z{OlC$kN6AwFL^=Cxu5p7ss|F;F0N(w5rm%k<01vlnU+*6{p%LabSUzGP
z4O8E|dA?4GS^7s>EGpUXl>6p)o(G`f4#DB)r6kbR#;11zr#MF=ihXD`p}Kep9WwQP
z@U0fJT>+R<b#J^;$vE)qmL%nIBGM7ZSwZ7-*~Mvb36`lOjGy}G<1Qszh3<qqmO{gt
zXY3CcT<;gtc@y)j5Tlqb`<Au`t^}v0K%z}#E_P`&^sdL~GJ=KZ-+=(>o)0&{8niRN
z*?!yhuyq;+Be|cE!EMtI42IQzNxPkzY*3{1ntQ$kEMgod#R_xj#2yMn%HTBw)#3|{
z;*DTZWJW%945NhzVhlgG<9QQY$3d}@dyWye<LFj*1wmDE_SsC9DQCBgvIzLrPy8IK
zj)MR|F#RW5mV8gv;2iAg^QF+65_iN}Zl|nwHyS3F<mMTcOyuuc6nESWZ-b#pfQ~&v
zpZ<c-1^b?omdJ79eQ}f>fV=TuerwzaolA~C7*{D@G|HemIKAa3`<?)&U};2{FXZ=9
zx!|XDMk_XJS-LEgRVMJ)LskFk!O&u&vtH|dOH4^kZ7*GjmNAO408Qlj`Ft_^0X#c2
z;!Tk6Dls~wjg>f_-meDyPBKV-26J#_t`<+Rs}-GRvt`}#m)Gf0;0dSN8fo7UeS$=_
zDR2aJK4SXr#2e*k{|LfuBUFP3bz}#yzJd8uDt=cSLA2h?WqWGhFTN_I?r9}LlCW?1
zTwONGy`Z^}mkuP7Q(W62awkU(o}9jdKYOl;UU*1s%#%4u<ehCN9+^MsplynHOVM|M
z9|?p<P(ZmgkeL$)3CH%O!eG#L*KmT8AYV)>)Nfx4gf*<z*JGvMiZ}LtUGd{$@6wp`
zo`rGYG|btK*=(Md(?^gu=2*Kx5z~@NEdc{ANPA>7-*6+YY+eHi>Id62@6FgMs+V(A
zZ;RWHYR;DMr%qu9F%qXf+UXf7M~67!2|dsc>uZq(x^y;!q-2|W{%a+S&)Y3GgJ`$C
z20RMmtcpWkRkl;tGb1W|f&lv4I(ALx>9XgzV><jV4HujC#=FK*{al^8V2so(_ITlw
z;Oq4C^2P#qlMRimVVh$-kUoz^NWf0nG^v^S;W^@1r)2BHoT16Q#71Sw65xkKgy?hu
zsT7D%3|Jf`DX*d;2ao`Wr)35Tc1Q41v*7S!VB4c(m-WKg6*H;E<8Z~>3(*j`|I}4@
zZ+}kzp^)n?Qnx-2@?P+0+;=rv*e$Ge*tq`V()_`JXQodih{C+Tn1EA*v~??3<JiUj
z`#7R~WvovR^KnlL0uUr#IJMv(%-c5Ej6jypMFJ6x_l1uoFS@4#_~Oj>oCQE?1%NFv
z`bnFd1<5ga!TQ8L;658z(}{UP)3`82k+~+s$bzgHgB`ZGvC)9@xLKw=rrb0-^2p{w
zR|mqD)YZz{fREtKANfhKa635;G2oBW&zcOryph@Y0PUe)HN;Z8r{QqDffWq?(Yi5i
z2qt#{ps9%(a6YBEcYPj#xZmlzhUVexc4wa3=eK@N0#H7HJBdY*m)i!OKedUJ^`TQT
zS4%0H#L}Qi+KRa1*&zpbXl}h>RV4J#ll$sxh{@R;Kf3MqV4IK^y>l)0PmTr%5D?X$
zWjim+p}2tDjG+XM&=R}O|MOv}O8OO4ap4}S+x|`G$kTF@g|1?J{>zq<Gc^KN5>P9^
zaF5R8ES|@?OO9Sb0Et?xjVs`}O`nv<D#XSs{H$DBj?v4+=&R4jel(j54al_$n!}k+
zh*zi}-wuD_jT18lsZ)wpJj*21fT=o2NN=+bGcwrFT-)@8Mka=DZ`ti;o9o|MeQn`A
zD?}c(aXITAZL8f(XoMLu(|B&Bk%VbvWZZlqDEb}}tlS^wmBh6#{j*V4w9FLHK>AfL
zoC1Qgr5861Qj~K|-tsr~#`*v*ZoEub_(C&slPnGpD76@%5;z;g(&I^qk?)Jc!ui!A
z<R|i$;g2Fp0N*y#>%>MR=_K!rb6+aQMX2y=UR|F!brYD+Q0WJP7o`Jmy7_xYNmvBj
z%PUxjz&;npz2#IYWoL8J6Jz6Ji8^6Z3RWFx8r6!)4Rk3acp-^`;NZ8|1kVRUNm9ZA
z?(rUYob$MgErRme9M%(HX%UQ9xTPk)##;&`*w%75+?olH*+AZUtoXh#gVNE|H*Bv{
z@;<i_-5-O?Vj7Z?Mfq?W;6dvPNAd}ZJAWB3V61;3h*wgGqk)uZ@)v4Tt8S~pw%N{0
z{YqBt$BC!2-tr~d?qfH{=yZ6kacj+snPd59f-62hN%D}8ltg}R6JG6+{u)&DnM~W?
zKXHeKfcgL#;1~4MJ9W}qPJq}7(00j<tYj<gCY8xgoPx1&*I&~(f`8=bg_<p|S0mrx
z`L%vu7TWj-YhdLZ<6){F+4WshYK;XUVl;^G_Pag=&AABX!EgBs2!seQW+4#j&*R5P
zWnN%8xp=MW55f*cW~d6S8Z?&s3R-@V4=+~7VIUOMKn=eriwM#!D6?w5^9QK0qLfTW
z%Uxr(#l+dqV0_p~#gZ%1v4^KW9gq8)JGY(6@w1r?)P1-wi!Q<%bJOdf+kBY~+@@oZ
z@Z=P1D;vGSnlM42C_W>)!f{qpFX=iPcqKl~$m)~AJRzx8F3*Ens+n1{1VB%GL+rp%
zi#7;|5HjcA8;vz-(X}HL?TwPeG)ip>w&H^#ds*V$4$&?QL*p{Fd~+(-1##A=C_srq
zMRoX<N+&^TB~WR%Q5!si=fjteQ+pV9O_cK6;T$ZN%=n#RW{Db~-uVDHCO@I~>v2s>
z86ZWfmYAa{au;w*5&m=}!_2egY_5HRu04>o)z#auAVl`bxb)(lg@NH=ZtO$G*RSA~
zyGKD%+}|vWt<J~6zv6zCmNzrys9>5iKl0qAe?pUu*st^lkkv{pwz4#z#MSMLecfGH
zKm~;bhx*lA`#6AH*Z52OkGxH)O4ROg0%f(%cd@;<f$rUqD{BBy_L=b%rb<&~i&>g}
z^P|9zWDMsS((Uh&C7KZX)+o6dFS+$p7WG2dt4g|?e(bj6c0EmR8Agnu<g*nWwSKN9
ze+y{zVjUH7J$!g%55-q(J!gvcylteGCnq~_y)xJmv6=L{W`JAP;;f*ZYb7LBJ=^O`
z&cEW!Q(;-oI#wf|IU0Lddy{Q+fDxWoY7J!OBgitI;QdQppf$L@_Y!OxfaK4<;R%2`
zpkI_!bU(m9_d?V=?>F3p30R3NIBs(u0bp>v5><DJsmxy3DkC&VziGZ<a$b2x1W;4i
z$uZ<1AWik;(tt$4)`%Z(VBM-%4p`M%%)PEeY6rmOo(b*?KlP975MXY5w!f8FS*1En
zEx)GpKTgNa2^oKX2MaPGd}e5!g+aM~THnbzBL(OHBDv3+bED69BXqC)%N-)ZE!CsS
zwYDD!Shlg91oBc_@@KT(=6No*-EBVRf2!gF5iZ5{RUU=S>5k51@-|rZRZ@UN)zsFN
zo#F_ukP|wPKEho-+FQ%fc^wsm7Vxkvy%vRG1pERh8V(FM63cdQ!7;40#wjT12W!nW
z+5Dx?7-qL~s6LePQS~UsRRxlrfKh;2NU8(3y<I^CgMHCLwoq<GPM&swQBcqyzQ2(u
zs8UVNd+x_p4)&+~O26UH^;&G7m*~ZUA{*_$^8M1*^?k<;0HGxAQ?o;kfO?ABSRg)6
z<(n|!63Jx1Hxzs#C{$Pj?ajCk;6C}Z3{=NhT@D-k0+_>B8{45(^$+V>k<>6*YVr7I
z%r?I;e1#AL2*LS<8!S(b1FeY9Yr;DycPulcGqEiCUisj?dP^H^Ac<m59t()Dl3XeE
zhd=bo{hTz>jBtO>GvI;R&zFtce6l*TNUl|R+uy@t=sXXxu-d;gDU(;(O^NzWaxT?#
z=pFt7#jbG`t(FTDo<zmYeSAf&8$eh+u3Q{E!|Q2LdWBaD#u9)%$$Ae3=*5!Z0(ur%
zP?;A)!&BUWC&)WqpCs@5XDn_HI8R#7bv!Bk+5!cKwh5#3P9j4Er#jgQ_#?DXdRvfa
z(bWhg&KoQnx2|6m7>A#nBT-O$<=kdH3PMtZEB)WMbb%t1@l(keMXRI%2ravJi<YYT
zy$rxkR(-f>!@Kw6l@}jk5ns`EbTVM2{D?n3iJ!HEbcq4VIvehLzAuJR^iO}dZE`TH
zrPIV|m(WN1%m6TzNRFSr_2VefQAydQv%q|hG&vl=n300?`|!@gNNFR`Au-6f#|he&
z8heCpAXBcB8i8{k3MAuFpFQ47Ab6B4Y&9F|$WPmgg5jasCVznWah#HG{?t)^mVH}p
z$(i(^N1kI_Ku`^$(l5)xav#Y_3Bu{f7)9nL<h4HSh7|}zit~z%P!F}XX67r?Wd)M4
zfSX#VWFR5>boV}sBGRJ%Cx;dKJ`zN7p8M`qNbRmtNw>i781Jd6b9{Y$b9;J5w>CZM
z=o*H<qQxJ~&&kG0^ymq=$=N~amQ15+D~j`J3!08hhf@vSE}tf@Vt+;{=!6ku0Y^c3
z6$PiBF67+eFU#nT%#%@@DRxch5)fn{Vl1^!QHgZ+wL5;?Vhne1w}QI2JdMLcV1hw~
zH$uR=af|Eoo}}rdngI&Z^~V%<^iCSX^hct7-Z-Nm@Iw}ISN+q0H`y6jwS%rB<E&rR
zNQWsCBLQ`&A+&@xgUOhoe>@(&3&X|7QPDs3U{%8ww+M}Iqdcd5AAJ~EvNR^h`)e$~
zmlnoYon#sc->d){S0IJ`gZtOK0RLiw3~Qg@$aT<Ziim!W*DZ5!N{F>OEX4K}A++LY
zW&tXYtoSxa6V+;T-T%kfSAbRdG;bfk01-h#L^=d1=}tjPx}{ON8zc_~Qqs~T-5}i|
z2uOFwp}RW|-=e-hedB*!-(Cuc&)Il(_MX|<duJl5i%V(b&<qVA`cIyv`cU!u*bpvF
z{X8)5N;al4^)s@AV8jmv^`A^3N4ndWrp$pefzSP)D~tmc5#3;^_{%<s{yU^3bFfHt
z<&-;nQIpaot&WihBoelDr5~OYNd^4eFgAEQ<z6|~DckUK&NuFGtPkIs>m=@pmBo%r
zOS~FI>~K7FAgFcn|KV~UA|AyXLUykbW4X12_Q5TmcC0ZX2!-fc!hRpOcnMqRV)k0|
zbYO2KIfJ|ua}GlPy8tEN`}1J{PJ~-+^f0x331j5f`~5piAKN!8-GDnFP_$R1FcVFr
z`7!JFkG)(3Z#jW0K}B_LcEqVd881xr<;%UAjJ>g%NllA-Z+dUZ2N4^lsCO6XOvjW%
z{e&^}O46yjmbP&Fc2pr`WVU=|JAmgCKsmw}qeC$Ec^;^hkH7%w`kLip9+Mp3g)?H}
zN4F}-K`MklOoD6<=zibsCop{*79TKL6ah%~qKnSw?Yq-QMHE<Vub}W8nX#LaGPLf1
zc-MAp6;q)1rRmt}(M^ra;c+NnZiq)Tk+Ivi->RkxBl-G{&Yj#1YZvc9Rjf~^0O}Rt
zW5MSU>GUH)L4xlvh@N4NBHjF!X~Kg~pS<<FKbewKnA$wyV?#Z@HWGDp@|BDzk*Q~)
zbIs#{yMaJ*0H;samwp3hpO2B8g_Ed@fs!mt2kCRGr1oW$4<KP=XxLtB$!+K6BB3s_
zi0v3wBQ@y7b6ENF{P_4FPLFfPYfSqF{1_V-*;e+!N*<N*;Iv_DF4y)V&g;NS?=#8~
zZZB}oI&DwmT-nN$f7%nKD5{ZxK)(Juof@g{ccCY>E5Wi+d+7dLIlcLOZZh7dM8zSd
zHgC($OtTshf-;oDZRmhC6!Ty^(9TI|5KDya{1&85smw=a%}etk=drY(f$JCwf-j{$
zAV&zM$F)A-LdmFL+bxe`HK(m1_p`=;h~uC8&<OX?2|v9R$`SGaFx{fFf-=btI++*d
zbmOY^$dUo4tv`XskM#!<3H@VMN)Ug;9@s|Aex}ID(<GI>@nr-s<Bp}|CL{Z;=LW*^
z2GX+_>y?X}9h_K5+G(|+uRWsW#DvI^$s1wHbdSABTFj*tqid(EUYk?cXua$Fc3X6Z
z#k@&`y;w%#F*1}-F=J&XR>`TEx_cGhJmRzEmycu^4y4F*>O$bWCJgw36mmZ?6GrOb
zmc5H99T2tpYV51~@upeeLUixEb<u`B*f}L{)az9$k+=(E+=z_z(O*q!j*kUgUJ{7J
z^JhvD0Z9%a2E)pAwxnj$;Pid(GS-~!bhTa7o$(IGt60fXI>atKdr6F`Y==TJpRFAd
zoT5mJ`*>x~8j-|K9<UWxJG^K4tU4-M9F3D)pL4$bgGC{o$W-vv>1oln?#cR#7oRDk
z#mA5~l@geUwClec_IcM6!5<fK;g4M$-#NEI&*DAhI;)$~xPtmjRdd2S1w!6PF7`0J
z-+d+$p(BxU_Z~BUaKAaXjF4dyQ~P(Yja<>VYFc<08nUYoflvd_{Q>tokj)ALoLXHz
zWD?&LeM1+WUDO<gndlqo)KB>R$2Y7X;$|08H#5Y(init{I}wCtJofeZ{;`g-)gMW4
zO+8*j58g0;J4^+EAPl{N3YGS5$EsSfT+JI*E=x6U_e>vu0_<C!B^(^CH~S38$f_<z
zV!@+7xf-~O-M(f$pXh#E@+9kWn1A?~xgfCgg%Lv0=oXUx8K$0|`$85lGf?p72hzUm
zVoM4mc~<x=#RWNkw;$A>PmxpBQQeu8*Q=+ho4ST%A8_ZGq{;Xr2@&B_)gW<2%hcz>
z({J0!nh@;4t%Jhq7dXQ=w#gq<Nkq0`d=<Yw%ZYG(;ezN?)L-we7UerTOu2ZZ<HZ*4
z7AYj$H#m<ugk#-=YjFqdvs5^#&#w9aww}@P{MN18Gg5Jg<n}C&u-nbu!lFG)0;ClO
z<9$CznfT?J@XL{tJfwLwShyo~<EUznX_tko>835myC7lt``Hf+;ss?K5WN(9xk$|t
z+R~S*Mc&aWWz4*GCeVB<HxL95B9^n}>71;m286OjV8i39IExhfeyVDRK6pWiY%}&a
z-RuiB#^lGM&iZC($mVU?@79U->fx_MKN;4u872$>B3hZZ{#ue@gdOXIV!8E<Tw8N1
zqd9s>^i}AV;c3tMC!^5VFB;c71$|%6OrKXU_s1pTiu8eZA$XqzRCS-`-wy-Zwh!w%
z-;S*>C6$4EKrpkzjbrzoo06=tQrEl}$YO{Im%@|A1RUxTssbvAZXTfY%AblRU$$eJ
zp!U5AYrN(-s!8x^968x`H~}<?RF!S{`s6M|LS@%!?u)=qUU>Qu#wI&;O>C?VjbUu~
zpw`)TmIvL*ckYe%-F7C#LHWZB`!h9b)4dUg5J6iy&EpA&bpk+k+P%<)+?GD;&i|of
zrf~9c`|}=WMsIGam)X3NZ|)us<&}%{vU2pW$GJpIO+&&MGTQ9AS}4benixVuVOM|{
zVK_{DFoidS_qldh<2GRWFw<tW4h%36s#oo<81By@T|9TQscUM1kf`wnTn4tQ-an*~
zO_hEwCUwVES&Xze)P?=f(Hgp_cFTxhA<<_t*j<^=KG-<NS!B&l?12{ov;Q=q6<|Px
zXtr%Z1alW7wsYmTh<HoN9bWF|XHe|)g2O~^-XzuL@789z6eKwOSGmlM`I(#*E4|Z8
z2=R}pZp9pnGIi!<8mN9n*gIrp!#{XB+WK}WUCzvljVItEV&JD$)s#nd;&p{KuC)tg
z1qDWl(}v<P4+~8EW}O6TwcbydA&{kZejU0<@iEjJk4qE&##Jk`E^9M}Kivr&+<?bp
z%1-F&Rf|6|tX7?SErX8tDoK!^g5&Ibt{`t{g?Sb@tOh)Fr!HqVa1o-4a0yrc5LIpe
zWbjji-ZPT9o_W-_L@P^X$nQccgTs%dRHj)(+W-d%P<-^K2BrqbtC}=E8*I{Qs;Z&U
znF%Y}fd_!!;JkY{=)p)$D;&=~{RM|TKdRcj!EH<s&f$~xOWzr?XL9TXeYnGWX2xm_
z-6J|`)I26P+r86n7_B^Q)-LaS{3`q<`wKjm%(uo*vaqUnoORI;U~MNnRXux&BgJ9|
z0{2LyMTnJBb?Y~u)L%qhz>y%)A1+H50keGx8aZnWc90~%dH;4@rviR6_ze+Us;;Cd
z*Z93DMR6EuC?N4}D@~kcROOP`TGe*dPS52{4C7xcTQHkd$cERGww!Kxg9|VYK&VZ=
zcSraH=|2B%S@X!$#Km7^SGceA76e5lrpdjPN?Mrh=z`;Tp^u<dI<NZ<UqpHgefk$P
z8C$V1FU|IrHL*%lxNC2=vNE6OGvHNtz2`5~8im>fdx(p@vZ3?zNso%#x=jZCqMMoT
z-pQbjXr5otJz1Kdp>KG0#i~z&sOot=U><z_ws{Hy5R%!oQywzedNY;5l(TcH2@^@@
z?E11Qx)=55Y|wP0Rnd$*1i`9IRYH8J6nFOKy||zu`5!!a?XFAR9A@2w7DbtvsLO%6
zN1-h1#1KJlA#earVi}y&jqyGCg@@Z~MZ5Keoe#`0X0Knr`4Wyl3XC`JLmToN?tJDF
z2H$4i4lkt}Z`Ugg_IC$7BN(4JP$;f~KF7-SXJFezV|!2~sV2VFk64%EM)(5l{ykO6
zLQ)?a&4Izj%}QZcWixIjrq1nB$bDZ$-4F5A7GiQ1RpN83B1a}P#Yry>Z=5f?&{>ab
z15d=*ZGBcQ7L4>6%tbuv&#@TDmBGF~3wUJFHzG-@lns17w}L%E&rDG?YD#t1P_1IT
zHI^nB%9^5eyP^Zl?T&GuJz=G?hgWLI$r>0_?G(`B#hbvNyO^|dB@|8gjelR~<zw&W
z(D+l^hyf=y;!K0bMOWQGgFkRM+x~h*zkH!_3#xyVbc^hZL_Fh|S0f82v#pcgY>%1P
z1Qj`lVYcsP5TIv8#T!1jKx%$k03z)r0}%tz`5gR(v{7WYGUG3)m1J&HM15!USAqLd
zls<f5KA!Xloxbc*#0^i@w1f%k%<)W3Z_4{A!zRMD*%9KV+W-~~>Z?kBPGEA@9$m|=
zK)mpE;GL+0*<G0g%tz2f?Ai@&Cop$(k17f@PkD32qTixEr&10UOK~<v)<HM8?71_a
z3<-XrRQZtnipfN2YZ^+hh{=?oQPQBs!n{bQ-g46@&X@!JyuyC&^PWAtDC2bIQ|q0C
z*aF_lVLsEi#iT5%4C}mN86-h6V_PQ2N8vovY+Za<%aJ@5#))guN)D5$=BjDj8+wVa
zBp~9hF`9Rq!us4}DE#we@YfXi!q^Y-K9iHJWy^82L~ahm@$3<?W*I9MRFT~fYuaeE
z9EoANf<XF%C#sMI&!~ab&x}nF#8y&Xdh$MkY_V%{uS`pi;U?@5?Qnx|tuHU_YkItf
zR2Il0AKYWz))*7>E+OHE*t<n=R=J(*{+80mt&TKN{_5rX$80<r2bOT<YjP0w!_*c0
z@d^E*jF3VU8s-Y$moFJQoQB;RKWd!09HH^zvtKYJ%~hsTi<B;pt0_X3UvsWAe{`Do
z=tGT1aN++z4gFalKlX62g58t%@6n5OgKpyb2lTeBBE#kLI%i(<P@w9}oHyNhURhmq
z?}lsIB{(>cd|*HeaCWeBpJPLInD;cH%|Pu^<OkGS9J^?VKHs^g=5WW^hJB>3n+6bR
z)4<KdDV0Xaq9t$w30Q0<cyY5adN#!<9hcEoh~AHSdp8Q{M0}tkVL@Rb)pBtj+Qm=O
zUY4TD`?dX)y3!(!8s~xKSl3Gbxg94w2!zx78r}@<UMuXRON&CwSmtTfJ*$28+If=^
zY}4K0Q)Rt(JYjf-v-h(i*9*&i8@Z(hOHx#)4C0v~!CNk|-?Qb$y;y6n<PMLjw55}P
z19l+0>Dw$aCex#vHUeIzA?Z2e)#owx?|8E(c$81!$zL|nA*l&Pmt(|3G}Jiou*3I*
z%IfLVg94+RE9iVU#S>%_MK1@+4qjir+=*sY))A;6t)WlWo2Ydg^*%XH%Pyyl=pqDG
zVwPd&#N>A0W>7j&Of_8%aMXz~FD~%a!e20^%^hVFXn9<W(C*o>Hg5%$LE)<gizeLz
z_{+h1K{&h|SKtsFsYk}$fo}V@0rw#?E0@Oxt|rD0zisTyGjfK}mpl64m<v&sO7`ND
zw)wqnyO(_H_3_&^0}u?Sb9Vo5zNGpxQI{&m;XyMbhCnA}nJoDD+QdptT9>R3pN_YB
z%-Rlt)xp{zi`-DknO;Zjn9V3)>V@zJ0X8>M!q#SXcK*mmy#oj3%$`#MUy`=Q`3J5U
zl@Zs@-?bq-x%il8$zp^@@{fILv+cyLFGv!_3kK6zc`r?dsIA-<Tnfqa2fksq%FrnU
z&yyL;VG-7ap~>{@%*FRdF`+`6LUE0l@zth<A`^FU>FLIXRO(nmxAv|{!~DiDgYjBj
zVyDvb)j)oV*G)s6wA6H2V>neAW1|&eGU@$V4RclLS!tQg<~>hxv#6O|&>@D2;Mzb@
z?6g_>xZxt-e9V4DFk6p@$~(eq{&=SnZqX9T(~z4M829h{mxJpI!}c@8wB#Lh?(g|<
zG|AR04Ml(s&$t-C_POIft?Q1Qkm9o$A2*hPk?^7M2b{;eyI#$%sJjiz;919AR24f`
zryO>jn@29d1W<GCiHhvokOvm6*{dSQ7t0RotlMQ>xp?Acfpr;fue#t(fx4ood*_%!
zcjyYqKgNfSMxi{o{3+>ObLT(h8|+dl;LYC-89XU=^s}tGmtn&bUxo1wCPu!dsK(bl
zdmTX~?>;bMGW^QabeG+HeaP{2^Tr_>1T&{QkDo+pdwa5fn$T^$HVr6>x@XDNPm@2I
zqWiDmrmkjRdY9Qh>8Z|{Kv&+}9xO}8?MHxQWgN8HRT}k!3-ng|TXx@;mygVkhtabQ
zBfGT<WWIm3>Y=Hl9#)y(vTJ?EK(J11o_m;3(}CN5wo4>UZ#F&%?U^P#7#>02SVccn
zD<Ep0sHbeSL=wbYsIPI8SQ%UJy}p2b5v%bw1Dxq&fdiv`U7_3e`v4=(CzNjCCO&Gy
zLW0gZ{31S43m#N_ENme?O>A*h?xbC98$f%cZ?J`Eq3mC$!y|C8v3Fe=k=fOf3W<d<
zNxaL_)y|?<$iXAXTtS`S<;bwOtk_UB(zgtx1Sg)sReTD=5wSrz=ftyuyw$A0cBg7b
zB6NNM{u$su3`JcfA;V1E5gmL8Neyz@<Vu_ll%9V7E#o$aGY?IP(JL~ZI~IJ&{C6Y<
z2yv$KvwKnY-pAGu;?$0}1il;+zP=hhlHTjQ{$6=H9}w;=L|gz*3{U#^t?+9a^LJDz
zE@S+J(qcq;Nx14gaIEIRJhy-Pjsf?<g9p6ZS#%uL!b+CTk`u+@O=8E+;g7-nw0Dwd
zt*OVABJ>u-B~o|M-5i&&ly9q{Co*^pm-r)qUvGvtB@%*)Qnj1kErYXkPE8?U;pfkP
zy2btOn)a{0s?13Bwty*BpuQv+VJ%%)%QJoT%Gb!qK;Cl8I;U-F4x+2V9D~8;u!oky
zRo=dSUn>1I9zjyu+B-oyqh~P>o|p5Idl``-C&zZV8r?_VjYR?kXH81v!ob22UnNr9
z-HNbcOV69@p$6|nzgVB77Z;LffGkZ`a)V$Ew9Uxzv+)?b87CCIj4&{wJ2$5?zQydh
znAFrg-RMa3SVX-1)(z3oYY!vm<>$9P%{#3KskpHMOyP|-n?FNYeZP|rojp?}1!rg!
zisuvHz$4qHK9p4pM1`m?S=llla!--f@_n-sU$JOkbA7N@I`N8Zmr&NVrV0@vF+8y_
zU8LiBBWE0zrK4_7Y&i(%gj&9%<$QD*DOt}ZVOcR`&qA+R1+cGH$<^IE4L2av>E>72
z0>j!112kKKTDb+anK~|8*=2X8IJ(*P!+plIv}U;o&Bm*!@XeG5l+z0RsmvG>SXU+{
z`}c}=RFB`Vo2xFH0e|W@|HJpU+$m!;*2XLjh#;vX(YAXYO6SW>Q|0dqChK}H0lDt|
z9Bssv@jO8-#|MqptAHGe<z`qua>we+lC6;rtDVY+kyHYg%?lYV@fr%}?;Xq>&T4e$
z1?o8Yzc;^|taDt;Y07Anq&duWp=qtagNWz5$0lYfv}%iAM=&5fOKYY&V_&Od`jY>g
zR~?I8*uyTRIzul8DiSK1&P{_R`^8U;BQn;35W*~!*jk&V>e`)=JhJj+fnOo*yycn0
z<JVA!QTZPOPmUZjyHCaYSsj!fCR%NcNOvI@cvF~+rxIaK9lC$CLezf}A6tP45qynO
zG)-R$c7bSU<#pe0c|~lTeR+Jw6hS2uN0N-I<HocWBjm=f^D1d^bCU#16VMbky5{uk
z5OGIpjFddl8e0;uG-eg8otFk|#!5_6&?AT1C<|#3umEF$IVoONLoS))$vwrQ0>YQL
zE@Tx?qAFeipQP_f--XdUE4o^7SNJB_m)&*U(LMMC4}B+zPnbLnxNWVLMoVQ~V^i&<
z?_jSUdTjdeJl`Q+0mOa7^I2f>&XQ8T_=s2^sZL~y&AF6{yHAdZ1!wEKB*m%YP+&}H
z1)K>UuA8eTt}=I_+WX0;)GQM-QAi<$&J9cwO|ztUMiC7Nx5**`>lXz3MVvcdWjv-H
zHlhSC!J_%(OjY5cs>OI5J(bF8ioLDsvs$j`T|~}DCoj<o&w3{)4cNNyFpIO%Foj9F
zYKIliuWBAS=14@v-&5W!ci`>y9YX$kzv0>Q=Nct=1T6uT_;(xmNfa;YZ?j(}cHPeF
zc=iZyaN+gJK+urx#tmzX`|?-1^IrnWe3l|yz>_{cM}w}dn}pz6a6%wo*$biZlX}>c
zdeP;j{xRa5%tMRieY#I8Ev7e2e>Q>`Uu(aC8#ZD;-LGB*cQuek4S>}PjC<^U&zfkt
z`R;HK{zXcF^+tbVFFX?3KbW|oEqLLed^CDyCP}42-XIF`mSHL`y8~{akI~>2?N0^X
zPjlQwS#Z*S2P;t>WJRN>#>IeY;nbD|3UL=cLSRSawD*fHUCd{f>>5q&nqkb!>>~ly
zKa=J0h<Lst;O`N=@$`GYlQ%SC<0&xlmSh}(x3ppWyP%z7P|y#1^V#pC$gXk~tPIaC
z5g9*JTV!yuSN9vLynC90x9w*64kND3#ONb$v)y>c^&ZMsIu#vU8!EsGA!*rln%_}Y
zT>2iPXbAT%zM}U>;Ykb}kY$W1xB^~+v)1}xBMJp7N;$@6@{bnJ+7$vBQ;+ZS#3*!o
zBV-A^(tLeO_C?S@&-x0q9U>^Tb6dg=&+W_gWfh30(#Lh3;Cx>1=|aB}a3y?NCjS06
z<@g9fN~U#yNXgtcAZ>~x<yv81?u*)AN4bFY&u*d(ZN{r$F+V>(;+~I<MHFnO*<9ft
z(S+{i?>cG5RcGs6#XMwHmLHe5v6Q_O*+2PIqtIhH4;;zkQ;S?>6Rx(M$!r}i5E6&B
zQ|ZSVdrvRxCKw;-wVf>rn6_(SI0%<G8jKaxU)oHl;y#VVj=fswQO(mZONe2js$V58
z$JRQlRg3eE2NgUXB+e`JYw?jz5OLNC;HD%Gv;fqDtnGp_zZQ%s5Z*$bqvTMWP1%`;
zuKRS<cwO8f#eq)+kSjs4`EX5-D`8$B%%r2~L4)eO)+NSI7FYVJwCwJw|Fu(7+}aId
z{G7du`CW7zQyeKV8vjS(dconE=Q4b3FuFVY4DvBDkpUjv0vhWegi!D`cZ0f;hR#K}
z82KI>UeKm|6>*8YQgEO4{kr|s)$+~|&dZwalg(Sf4-^(g7_YkkfuI5xhe^iufzt!4
z_d5OB_884{eFM`mpS7bI*S|h-;a1FVf?x=gdY5B45R5bD*$G;sI@AU*nPrt{z0Ity
z(z<6wvjoQ6m}j!CE`0gED<hqiiQqa=#Fe6n8l4^;xN#4DzkGBtJ9naW$n!HuL^ZAX
zV|PH!0NjfSzMZ}(9Xln==G?f$MW1P~Kiqp4K-<o~6huUPxpx4|*Qz6Y<J^4l_!Q=R
zwkUGSRp?h3Ol>q<r4x}GTeD02D=HNI=kH9=wzGN-mCsk7r+(aZ{x*(dP~TPWaMV&y
ze7PUcQT|?nU_UbVq}m|-^rqSJ^Z17#YLVi!s<3@cj_HeKaOwE;@SMFRH_bg3-}e^p
zw#@3~E2^hcAcmBoG?>vbm(-Zd0<$nUSK?XSsEFvX^CRDTmV51?)AHgwqXC_|hf2-?
z-Gt3WJs_yf&s9g#hCrZNa8LVm>U>-0`j&lb{_xP1?U|s-2DqgqUiF*Cczxn7-KAIU
zsPbU(mKdIl)Lp0}AZKGg^mzGF(Zt%H(ep;*R2hqgJ_}cTcW@9XaKQ<J2o7?LAIT3?
zcCywBH?HYwrTA4+?5HU`O#CeA9ijruP~eL*S1p<J>UY<8i*nQkZqB_QJij|ZE5+}W
z`ux6!Hd5y!i3>JhEa(pZ3@OC}P9^)baxz#NALnqxe9Lybow5lTu$;XSNak6G2?MR$
z#+xSM!?!K`zCMPCBjKc3-2bSOYqDMz9`miVw8gOZV`UNFqCwvXA|%UP2Ht{m8X{#I
ziEFM|-SzDq@FBQmo6IWSXH8=>`Zhl4V0NGPPB0$(c)TM~dQCx#n)CJ&vIcXZ70WwJ
zgnmus%Mj$qNtf?q)XRBdy!XpevoaXi4<Y(Q&%_<kArQlA{hQXBgl>maSL4o0Qkpb6
zgpuk9kkBXPlZyStPCk6Q;IhOc<|i+U%rd4gzd=GFP7~d40u60Nd4>5Y!8QMW0r<`P
zfM<T1h>~j6>C8)AEPL9ur#rZ#I^P-l6u*h<igOS$zd|m$b^~39%=1&^q)ro9DTc(F
zAbXWHBG@qhTNzx?cOn_ms^vbZ*125o+cy#-xA*hMr@30pxYDdLS7YY61;L}h1JW>&
zpaFMr`$&Wc(a%keKzEeipPq5AI{^2Wz_Is-S?^$V>+(%0oH40xite3G;Cq`rlq|Zs
zcT<?-;kU<k)b>Mp%H!DFHygO8oW>}-YF3L1<M)Gv(Kx*ZQPA9cA<WYoTwg%^7z!+4
z_x;@8|64yHG(tr@!BzS9>dwM&xtqL~<G2Q#m%O@1{PCyHy51Uou!aa8U=lUF?ehja
zC>m3f^DnlqPl{8fWkuci4Jcm!zCsAXFEJfqD4=bX#qahn3J-4hwO9I-jN&NPxB_28
z9F)M6<h6Lb?#|<%A%1-ju3wrD_H~c`E}l5+eOyM|IyxTkIqwYg&y?A+kl-v#f(E7&
z8c`53@HsihHHSQtcWes0ax5oO5nWX&w(Ifs;Km89eQ2T6G`OzMa};dwAf7ANx6uEs
z+`r%U&LLWD#dJbeD8<eXpZ}suRx*+vwZ_3=M^(T%DXi8tecww}C*nTfzxPQPLt2N*
zpqo<qu0LxD9%o!84n=h(`T4Ks=>IwA;Ma48JE|`(M4P;0+yV?bUVxx4Iv&bz-UiI9
zUW}9_MzAwT+T$XSgjHaUK5+8<sp)>!1-^*80?hr_KmPR{gG2Rd-)%;|$$RYpO=R^g
zIzA$L55eWs&<A@|2(t<z66Y;g>%Y<v|N0P#-G7Z51U2xH5~95HZ)2}Kt4?`4G0$3)
zOI-i#xzbrtXoJn}3#sv&o@yQOUn3ZQhi?DfJ~PUH&Jb`jbbhBSCQG3+<EQ7)z}2YB
zj#`nrs0O}rhF3tCul8%4-_Ff_9LiAzkwdCs0aba69Vz+00z3b%wc)?!{kqm4ZCiq8
zRklpQb@DZLQp_PFG$^Pua`!Z6{a|N199?N0m_xs_ko>B~bLIMm_<xk|fx#R1QgeFU
z?Ny+-<c_~8VN|#n4j<^^{QS4D34R1&wf^JN5Njy_()~j-1={G7Vi~iG{R@saA^qB+
zNSvS(Xx9?>rwF$Ej|zM~T(_V0O;u+_KL~t=8pPrxc>xCU^7=h}T56tahqmkp|EUGc
zyV(Djcc$XAuy}UK1gM9Ort9lQ4=SuV(+Rtz6SPV8+mPhQuaH0km=+;F#OBb+hTv~8
z@+`Ub8vjSZV7b!Sd(E~{3~Rfh-I6C%y{)4KchNn#HwSvhu8Cr1a+*E&q4eLR55wwz
zbgPhDf4IJ)YW!$&50`sve(oc>EV%~~<TaTM0zn77lG*&m-04+K%-5s8-!$Yk%6~5u
z5|!(F?ac#2Wy=xu9v}IhTAu72yzQ?5!z&h_L*z78+plLe*5<Z5aC-PV$I<UWt^SYN
zAegIbGdY@PEgGu9bqJ8c&kc}!$^D6SX_Y9W(IZf{s2@K~tAAe0_j~_CuBCnmiC+5q
z=UDX6Lz7rQzRsOlAT=Fx0~nA_6%@3YF9u!pefOSixXo6wPu~0$9Qt>S{h0q<ALKi1
z=$oq7V!`c34nJ5G8SxX5vs;kw>aUSFeInQr;a{Ke-P-M3??^Pm|4#C}URVh5On-3!
z{^=(8L0tHMOaq4MPo*J5<*GUPvy;~h>o?-|klXjeiF)zgpETOeRxWE>;1kri;0XUt
zg!`*jlKB6D4$mclYlVsRWNmMx=c{+Kl@^!#7i>Vez@XatQ1*a8tKuS~{?vL%RvVdU
zHHyWEyQil2w93VI?)9OZh=$5l$0}0VeW%}f2YxlvKzR-D|6V*_2J6MChfwF2@g0M%
zzx=r<uY1n<LLbsPlhg6un`5Foj!REV+^lq81OkU1Z3+CyKI_JUA0KRflQ(2SNx!8+
zmzwQM9fn81p`6RekCW~g_Ey4h#}_HRP)<nl_Hh<s_9q#?1qNZRM>(He%8~=5d-H!F
zOTGwW#OxXjRwq}>_(YqpPHL|#jDnQ4d;5R|WAk8h|H5Wv;@AnM<8F2JrZ!u0o9{So
z``8Xj2x7B(z`#}slrQSt|LCfQ@2fE9)|JOL?0gLFNKwWXd0<!dZ$($sKgHy5;GuV2
zHJe>wtX59)t^={DTu(zM#lM3-P*oCX&}Apb(yCj!h9*`H|F&>Ioh<5;YlxJI^H#(r
zAmh?q^8ft!m$YORT*lSh4d3|$6<mbQC;#0U2t+-r_E&P+zdG4zY~!P@h1eiP_1{$j
z_FjoP3lir8==|xjDo0vMbJxl*aPY+I{uA8y5xFJkFDr{@N{XI=m>AaiFCg;tUHB95
z9wLR+b$8~kcCUXV(IF_ULiEHd04W^*<G&z+v}|0SDOYk(sGTcdTY}>0JO6imx24f$
zI0tLr49zvmW7BU>SMCRg`24Cx%Hq!?^<vogqz}mE=*W1gmv9v1z~f<K4y#KcS{gwc
z<5g@Q3_95JPKP<layN{%!MR5#qme5=Yz@<tBh4SoxaVrmU*i-pyeXJd%9fV_hfYFM
zb!S7!tIP7$2}C-o(b@8TWsiUuLze%s0F6Jtak+#@iEMVQWv-kw$#)F9=@}H?0dt$4
z?mQ)6<#3&EB)IEO3y>)Om<J6u7n}X_=p625l48YczeeOKYW*kpe2cHT7Np_R2*^pp
z9HCWT0Wyw37;8J%;inA^4ap+@>RQ+N2^d+p{z*Y3=rE^(3u(|to;lgb(7ZzZS*ho0
z)1Q{JaZ0Nop3(VvYJ_>3wOND0R`B$CHRWHe3+i$HQxya<AU1%=+5Sn8ckgI7UHP}5
z2F(7Mn*dC)0gzdM#uVfUe>t}A=1b~Rtpem;`}ZZ^UgSU118K1D`wj`jP&p!i+N<@6
zIp(QPyJyx%m;P?KHtrw%_(D3jdWgh`!*(4-I*^`$80w)|_{`_}BiB>lM6|Zef%!XE
zLDsiFLwiITqb&eYlye<$-2mO$+a2QBepz}bKKowG(L$@#Y(0&qBZ3};l<x63Dpq#1
zx#VyuvmaCiFZH@8wfw1R4W}F=O^(59u*q~ib(+q7eMv4BsDD{-CsyV71`@`vR}rQA
z_vl#rIw67TEttPRuwh~G&ufB6p?0Co{A}dU7ZOzhk-Vn}y6I+S|AhgN-WQwy3t}_?
zsr@=)^uHB#8dFdKOTvF9CGk}8yMF%vfCpc9`Z|w+bm^{@m9zh2u~Yb|Yauj9jev0l
z()002126o~Lirga7Wg#3Raa12bwfA(9O`Om2B7Lw+3ae*N+hJy*!${y?dK1&d`IPA
zj8T1OMAfB0c;!kf!20j){HLp)cnf_z3FbdPeVD~l4?n!j%SVi=B!aqi4U})1I^Z;P
zmTpB$LwB!uM?jjn==DTUcYy_*_~&DFSEI1$#R`mGZdu;CbkKlcz<*^+C{X}-3Nilw
zf5|B-su0HcQm{(?vlyoxfyMRjg)ya>)D$dl7A=jOU#sn}|MyfewR){{$x~f_?#RMj
z1;^*p;Ew70CAkom=s)=xSj7iHCk<;1Z&?#ST}Se@x>k$_))I6fWWQ>|>ii>{zd*R;
z|G$6oIkb?d-EjaV|5-{DT>q=nQGIh5U|z2g;(uW^%FKUOqAO|!v=~~E`B)K{^j`(-
z4}dpFVT$GDLv&lr3?Z#^>l_y2O0q~+b*F~NVu1wIN24(r+SC2=vhx6ul?t}bU*t&Y
z+B$FRsB3+@aXvXtTvi2if~Uz}00xPo1PimTFfAj4iT;py{B&yDd@7s`bN9z^a|g@=
zzL*dmu==|nEq|@Tfcw{L8oUNUX>JwERp--_{n!Wt2HqmGp}qLaJeq5W2E}yguZ3GO
zICysk8HA!QELvV%=2)cTE^@FrPI?gj+X8p6`3J0ZTl`Xtxi3r37H?#VfY+MY({9CE
zWj-=Yt|sZ|ZnVq++B8-Zfk3285;e}7mh2DK>IzEAAcipt#9W|VO2R+t1o8#vx}Wjf
zWp1>`WV*9IhPSJbfwg#D6t7p`ze+EnA|JrW<`IStjiPlPo^uU_-!VQ0|0o;JXsU#3
zT<jY}pIJE>G}^}{9J>k<CHaL^jBJ0j2LvKLG7YSc?;+(u-o1TmG{4&U!t@85qo(IF
zhj1XBZvd`*vAnEzF1DVHt+zq>?M6(e_<Izq;~H9mwfEee)<YW8o+8}(SzW{6A8VoE
zF1B|jSht<_iXb&Eba_v?T3zdx5ddNJ`7=<K&lhN(D%VKKwze%WSnGB=_geM0zVM`O
z{j)eA#z{w%M-IPGyEuScjOLMT#l~5)gX8a7=l@o(KGy&Vtd-0+C%RJ+d8avn>Gr>#
zYxwlH82rR<<A@%Q_7mY>uOuD)&p6Y|uThPW5Q#H_?<^oZFaNiueAN6;zdTJCeuDHg
z>)n~rgFO2D-&(ez<c~IOD8ZGSff$#Tso?e+$0lss{Oi-{Xn)!wfhBD{@k`D;m&-v#
zNDnc=X(%)w*l$Gx*(k1yZq=XR>3;Od80mS=Ggxan^8FaGlNG3xFY*x<F7T{`aKgN&
zzeV^_<G<R0K%OK+mu+|P8=haorGrM<YTt4YagYxC*)NmFbG6~GXOqBmHERTwRNc{a
zxrRCa`rV8D9Q@(Yx#pylrl{X9O8qab`iMLbENkpuXy$pKa&@W|1S>5Xfq;;ZHsIm!
z?vMZJJ|rgTs<(UF!zgD5O;?DG<v)#9Pq%AxwlQ_6-tKfRPp=QZgPrdxZ@s?QG>7rt
zm%jf2ci0XWn^MCqMPi8paNC~t<~+LFQFTB<f6eGWUmHF>0Dz?%dOdtAcfF4GK$Q^$
zDWm_=yHRlu3OwL%f>zLU=l%;7b#2#PRu%!tW&yWB|Ec=t51uM#^;fM+cSbH|Vf^e#
z;h%TI{rqo4`K1d%K=|UM`Bs+W2MpSieSLk6w!c34<>?<oBRI#a@KJb9jkr8vgA%JR
z66YEE=Eas%*%r6L!MA@*T-JY?N7R4K?hBvas*Nox5e^E?F#ljJuX;`P;7Sul<8Kp}
zUI$oZevZFH`0s`PB_#RUC_9wWhlpJz#ZPrag5;JWeRIwJxK!m44BOLcj=eKUTZ{gx
z;Y2P`^FKngavp%+nqd4(h5faUpcLh;^?;-&`=ZRO2$a|zU~1n)LSY%%qT#-~E9a!N
zI@476z{x_ZrsiBOulct@#rX$=fIhGQnKSM=FzD2nuD?k^4k8$<GTq(VntbNtU8TC2
zoi{*UtwQfa_#2ccJ^vsHgmH?E+cTw#_6^}!sp;-$TD2*$M-*$BLt%&KuR-^|{PVUM
zcYj%;r5r_t@c51#VxbD4E!KnTYtjH(Ij!M&M3o%I$HQJ$rlm|90#aYZnT`v|2kY6(
zLe$o7`1c#|t(Vy#RYX0PO*|q?Yag-2UDCTy)+n`pkFhw6#?bn)siGK&uKG?aZ}4CH
zj~FtrP!NFO-v9}O$2*zlos8>@1)A6DE$i^%9F3INcqqd7O45uBF3}Jmj1q%fr}SfP
zI~#o%LIfGY$nu}>kfs`Tb(@vF^}^x*EFM7-B!fdq+(@DHuohfYp$PJug?K(v+Ka7$
zeR#F>M}q;^w%^B_)`PqkAki9VZ9dQd?IHA`(CFc0QbPE(zK}w4knuRb^4PzGE2`?V
z_sGD0L|ys@K4Ru<p{LuI3snv&@Af)AM~##f0_{~4)l;t8yX*ug4Pcsv<^~*fEMY$-
zh8usi=7Pv2E$AFkU&iwO{kD5qfA3I7GQqf^x!Z`^Vu7Eh-Nw?T&GL0Fvt+$s7fY4p
z$J^t|b>!s03a3$cSkPPIJRjZ>d*Nsjao+32DU&F$ua2G8-YTdsP`$CbvYLbn{$!~Y
zhqd%vuID&Bh^YPfrW)g%CH*X;%oT*!#3VaM+q@O$CZ3s<GA1pp1hGf>j*3U1Ls{P(
zh~Y<GbZz-Q0UNm-iWdcgR+Z7Pg>xO3h5|3fu)p_8EaLAGIoAlGFB<{AKutT^J3W`f
zX;nW0!}-#?YWt<K%fq?Y`c-t=j7=p$i04_U_4J#?J<586WRB44>lyoQ;{s-IFn7$0
z?YRD+VaLm9$!)I+b^F+;tX({rqhA=;R6OK~04>A)*43;8*}m*HZ+h9fC`?$ONy?&?
zQ~VyL=~L~RH_bSzA5dg6CHgsN)D&;shA2Hgy;%{;W?<dWIrqn~k;eXQlbMOr&VLfn
z{EhVUoPqQ6tw93yCSyf+9%&t|>_(Pj<4cTQTR-zb5$1D0CR%BY2_d?o9i8|+=P1p|
zw_^%c7Ro8a$l1&sm#BPa#cf&Jtr93K7fK(ruEL;i;1|&&YjnAC*lBli+S%KRxw7Nk
zMl)r?Joc+(fF!b~rr1dDS+yQ8I!|wn>tmYwc|?XSqFp#*o7S7l#M43`A_3?AXB1b`
zM8=$dv>@n!iIq5QE9@yB+<IW&rl-BInh+E|hJuTh?nVFd2*Q$+6-ABqPX=3~dxI(G
zk|i^kx)4a3!BPOOJ8Kv-V_&4b%u<?BPpxcK_ZF+-0+8APVca{ytO~`3R2?jKQ^g!s
ziSE)512%uKU-aoBlLm@~R@)HZK2TY!qzeClajS<2xR<m{DIcePg;8g8PBtzCND4`T
zU-O-%1bA8LpfR`MdUbTT8RHv=y73tLMAT2Q<*+BE7bZsNqqE<qMQXUKsyrB@$sWb^
zH?ZNQF(D8<rguGfbF|XLM5v+d0pBzp%2&X*JX5Kh1t}DJcU8?IEcg^G?~+3x&gwi?
zyo$n4*!Aeu6i$pO_me+&rbNd`^0n%BaRv&D3vnd(;p>4;(Mz5Ez>erN3`$ngjulC{
z;TI9_oBw40)1$P|&tb{!ZsK>=gXZmjxSyMzj7oi6@Ne7Md_qS(cyLd~Hf(k?A69Xn
zRDw~?&JQ4D%pb!NLMaHR!);$3-LtJp8xnyG$O<EJo<a8wcW+-CQ~gcIs`=B{j^PjA
zUYlU^N8`I$Js+?4G=f9w{8lx_nj%FtgC!TqNe{<pa-H?{IX`gvjn7q_Jo60+nGHej
zxir0kiOQRH*E}Yr-5@a$@KbPZPW65Y{(RDS<<*B@KYmsIigjKOJkW-!U~!}A8edav
zNPSbkaqBT-wke7ySlRM~MafYrCw-C#ckA1;5=#z^i)kZ}k#9+LsWwv_x-AIS2t>-@
z8uAw>kL^d~xZ}$i!($G2suFG;mz1w3%)aImsYQ2GnNT|BvnkA;P(*iPuQLB4;&|-%
zV8W=x$$@WbC4t-3{|I|?s-iI4arm(1CR&<8ma&zp@+L(3xD$2aUVL@b`JuGdZsAqc
zx_>!qMaD?t6#q_fnDD8qAFW~6mA^)vyGC;t@BZ24mD|Q(^F?#kVD2exj>+w)P7kL&
zDk>fK)5bBql$7-RSQe{UVyNZeJFk8G$$?Xr?f5ra&#&6P6}O7wYH=7%HwtXM<aTNc
zh#Y|p8b#=w8zsiN8l-gGq8zo8bvt}d2Bp%fhUa_7^q<F|i}J|2yI-B%;oz~hg59R_
zQ*s|-Teq?h9TP<2Q=6;k{rHg>muC8We6~|N(-rpd1^d{P+MAb8J}l@CPxIivB!CEp
zKf-K~e=)Y}Z6HSCaioYNHKE~l$$SmX*?oJ?^KFyo3V64p3A{IW{cW&li^8+G`0QU@
zPc()QYgMj3qcK6yhup94rC$G-8B6h^HA;Tc;@(`8LMzJ8t7_QADbs4>2JbYp;kCW~
zN+8ykRuuN=g$N@>2Ux6goP?rw9DAmj^pxb-HEhnzRyr0*C;$_UPs>w`ag`SbrAX<a
z)k$1;n(S#%<Pjmd{dQotW>{0JHSy8^F3LU9)B)j%GxpT+V_r=QR#BU&<AcQ>R>E~^
z^U>kRn9{oAZ``qWRu8PJk_?dOi>iw=7N$)qGK?&;-F4oR(GMR+tWxK0u1b_TZRdT-
zvE58<8c#@2%GSDjTk1uDRrUO6^J^IFpxI8K)*aT9U$Ny{ZD}kn+hYkk9tdkDexXCa
z?O?5QoTLUl9i*QZ@%T)~`5{;aJ)XOw;9U3RPFm?}PF)(Ft1z9bf#aUIE)5s+osGh8
z?-_^^kG-0&4EEpA;`a|kAA8j82b2fm&<U`=IU1cA!wtHct2%|&Ee=T1v~@q38q{c=
zezSi5Y8n@QP;Z_v5modS#c?Lbb(;S0xXYI>s6axi6ndG5rO|NvvBwnk=5$@F?B@%@
z)v5H)Xsqp9-P8SIW-hM7>Ep*1QN4O})jX01I<}VzH@$-iI@Kre6*i|{b4Df26>>2^
zN<=ltA92AQEt(*>I`z3HNa320jsC|Cz$&6)mUuAlH*c?8iEp6a&lD=?harS5Vng7K
z;f*QwI60FZEPGBFbghSQ<DJQR_}VdX0<aTy8qY7IMg3?L_knOU5u>%(a0d>6@Asgr
zFHO6EH9;ZpiuE>y1k5K{X%7n|!94)arD=D5D;AsQjAspvVdL2kQ;GuLeA&g*y;Pq{
zz%s47vL7Xyx8c%9wq;SrD1^`z%o4^?)J%o1Kv&<wBZ)WxdPgVfOCXbuaVceHhnwUs
z&`V<46-81>GFv@?go!C57%B3hp=Cw^lTwgL*6K-kT}TMg^w&H%>Yc64aTR7}z1*Cn
zI<(^r3voqKKFw!}4XH!bWl9kuqAqVNWUZzMH@7zM$}@tAw$vu7t1r0I1~`hwCD4N=
z?)BO;OuD<aG5SA0n=9Pn(ax;TFyaW8g8f)CU&$!nuCDLA)5AN#r~S=W@%FKWYr5Z_
zXWPPpUj5m#S7)LT#SBDO-{|X!oD!MqSTa1`I2Hc*WZv3+F}J02c~NSk&V4mr&}y-^
z;cs;@vn>-}P$vFmHQ}WqjXOE{<yH)ztd!=>`1uFH+k(ppp{w|M{oDDgJaT68f&&Nn
z^tguig^)@4YRbyt^Foiqr3aqu=-<5QjUgnCer(ULW+u&8Dl)KsTlbD>;*XZr>Oj@j
z41S*bkvrVLag6cF^WJQ@7~G|Gc2`xS#8|%KXsMfxgTD4`jX8i${c5yZ!#|hNS)UJv
zGXbr!SUu5*<Ns=42NT4>m%5cU9g?*tqkK}>8FF|e6`m0tr9Onn+EZaSWC%^x1NPD#
z4gYz2?6lCsF#Au*iZR9SSwsgo8dspGs&xk|KYmzai^yX)?}{VKV+w?Pcu(Q{U<@dD
zqdn?S{5Eb0%VzKdGO1tk<WoQKN5mF1cSyU<!&h0IS^4afuF`IETfz+uCxBFlYx_}T
zq<n4E&(UXp!;Hl1y8np+mSpTKlK94JLczD)&a}Pey_8lPQSJ^)Fx@H;rG<3?U^bju
z2QX6$uuAXc<G^zw^{t~S(HOY`s{N0upJE7E!*hn>t{w&L>(I|!;C52DR+lMr`f=9I
zCe*qTZRJ*G<lE@+w^Tivww4{;w`Pe!q|BL$=L_G8A3oZe9#K9&l;LSb@A_fSR#V2Y
zu}%ZmeY}dVwmGQG(klnx=TcKd2X;`oC0%g5QI&L*U$%X|+WcaK&*8kd2R;*_;m5nz
zICa~bW_y3`oBKrLxR_;*8>%idCjcBmx57{xOrv}7DdH*kF=~EpZO*ZW$MG&R6S3MY
zSk;v(@zpd`ow@$|e8J_}?c>C(%rw1L=+5ZkmBs$#9myLp+YPG6N@vSmF?C4iBe3xk
z%rhe5d|Ur+zHOB%iL)s-_g7bEre=sShE7IS&8=QX8`I^l5CaQ0vJQSI0cYEE>Cgv<
zRBu=A#EofJqf)kS*s8W=D3C5@kp7wt-)ey@fR&LE<JK)lf-O{k74v$hiT+i_7i&9#
zViD5O_q`yDc1moj^nnA5RP|}zPrn1yn8)ZJbH<XG5kV#;I+L!L`$&O42Qa4R`^!&G
zoUb?ffJfyCPV4H%U{`7@^F)SW7ZKc;u{xCko{~p@?zC;6bvN*rMG_Qo->a}Buy7`5
zq5c|)68K_#b${HmeDbO7hYG7;s@(K?qAxCWF@@`d%|ynh7q+I96aJb}1R=cstGM?S
zhWvGee+Zopex&$fq`z<+?GvDHrFfENS#RDQT5ls!F)Z%*BW?Op-#IiXOkH<}s6oHa
zW}k!hlTa+!Wvw3L-|MC8hO&hL=#l!c<+|5;GsCd-m_I7jd}Fi9h%^4u+NEN>uJ>G5
zns4gZ9v<zZWU6ASCEGvl?ppoPmR3Swig5T4+hQ%Dj!c@5x5_po{j93-45bE>j1gfN
zN0XU#Xu8fxk2&@=m>2kTmYvh6#)Ojp)XjY^g{-5@^;Jo&%`JlNiL`@!R4rLtvkdD5
z5y!4|7d!my48t!;2~+lT3|M;KBJ<CWEF{`n){m{Y9cj*I{aefU?1b}PJpky&?DWiZ
z_+7X^vpZk4`-~BlcabBGu)x&88U0OX;$v6{F>a{(n<?mJ7`|ntlJYB_*|~2<iM4sg
z5w{s7;<*lGVPMM{wI6Mzb7jP#oseoC2i+V5!g!IVmx4~j1D6$uJ`ZQm(biC<B6gk6
z+c+QTy4q($E$BLMSgOV9i9cz6iTz`S!7-t#?9Ftmt3Z!5EDC7TM2%YA*|d^CSUhxd
zL0}4MNVhuQ%;8aI-@I&<n`f%~ZffG#a(Wwolepe=dyT?iUeE|uT?w74IUibX6v(wS
zN0H6VGMsZoZbsx$$TAy&9m?6BS0cp|FSOXD5^vwtfoZsUHQZ1y&lt5y;t>l%-K6Gx
zdZ2G?A0@IqZh9mn(k4Di_b)EMoRGriW?xIjw$#>!b~#lyJ$I_NB*ANtLzgp*#q<5}
zL(Uq#eP0vXG%{ffgfE%C2^`Z;+V1W*PB_nT1}0}-7IC#rQK*tR2ad&d!k@REvJ^Di
zXj)(bOY1wpLDCOT0_UIH*^KOx0O}SG?`ha8rgJR9_eMdKhj$bBxS5&b4tSWE2}C4#
zB=TFuvn)|9TgOmH@FY0JOR4XU6&Ph%&?dS?)yyd4Fq%k*T8SxJtx&YhJoz*&P9Yg8
zmI0#}+CzDkl<t(_W-3h9cI=y6rcaD1Vj=NecCS_j(X$a-3Kk`U@DyiDKo}W!dOPT;
z$CZHFni-ZBCIqJ)&Q@XkPz4Y=embQEn|R|{*jQ2x`9kBJ`^L_#k&pRmNzKv@dIv~>
zj?KuwNElwe;mkWahA&i+VI}OSdi-I>IwXc;;hy2sX(jEk=2%3^yc$M-C0D1T4<7<t
ztA$46K_;>TYb#_ewu3>nUPYNFEiD1A!pOCJX07Wq1@XbAu#xbA-T^Jr4*+ZR)}xIP
z!kSP8I=V;1-8UE#&t9v@HKW0%9faR3c~zqa8fa_s9tc%z`j^_5@=<Y;m+n+;Z~nM9
zQZXAr$T5kjU4Ip*m(jwF|Au`s7%OhKvpB={h@%+3ciXw`QS&&8u)qY7<(RzWkRWZG
z+t$fTIgzNVv~716))ul482L3e-<GZ7s#glH9Cn%|CA&SY7TOq~F0bgdYB_dZ57#nW
ztZv3G8+w~*azVF3yX*?I&en<b8`%OckG@CWlq_1SRI-x^ZjI}4J3s4zMa84oo`!RH
zoYD3?<vnw{TE|W4BekmCn+9!YNDejJaGiU;1(O)Dd>Bv&q(tL5AAPoxYS=CK(5U~m
z><4RAB-)ipMT-kj4_&eI*7`aQ9xVOCJXc|lu_kxY+=5C`-8pOm+_iSTrzNBxsPOGH
zGe48PePqBp77JC%q}}%HY02RDAQ@7mhdR*YDl{tR8UleRV>K{g(a~^Z?i{QRWE=^U
z%<l-=I(^wrLgFyUtF@m9g4X1ru(2nQ`h;nY1ue2NIn?tv9B|y<Z^3`c@PeFS=k-a-
zH>pq&<_TT^@j?%)f}I_gd+PY#92xg_r-_+4Q9mQUt0mqQLkMN%1VRlQZ*B008VoY%
z0Jl568GJ^4#5#OJabEa<jYP84h{P?iX_YpqbDIHHKN>8O&?f??j4VU-@?6*wFC~^N
z6*T%<sJOTFUbyi*N_!h}+h8ecsWvr?VkMO*oN7>s?OAy!dvS`__&8iwU;mAL0cPv9
z;lUeIC;uuV?3+EBHQw=n#*h<&r{N37K9S@8u>D9ZPdsq4M@hO|uNzZMWrS&LdDfD7
zdt|c>rC8i)3fC5qfo9KsYt{3xS>w%{%Usv9qAD89qds?8OUa<zZTj)q_v<YdccN-+
z)9Dw|tUis8mNHPiC!eb1S<p99f+>nS;94c+RM&FwaHZ%fv~JZn?)7uD_)SktRhJp%
zCcSY&Rvx*_ssvU+?&Lv$e7XCs>+IOpd*R5f^d1XWN?u!yAuYahD>Wwt;tpxpkBZp|
z*~_D?2^M9v%O0afdNp^x^B&<bGln0supgFG=GB)kDYYov*Ss)H+tCnW7<jPtx)Ybv
ztv*$`%`RPCEK{+(dSIk|+1BW#Q&(eMYZG@m<*Mhj1f}M1C!cS}7r0Dgh{NkTf4BiX
zbduaV<F;EH$|VT`@+;sDZ%Mw?s?+b7gI9~qi)dezhs7=rf5u0Y%B+j4|6x18A(~fb
z29=2tIm_wsg1*$M(5Bzobaks(leUFZn!$DsYehyh+($~?gMym<Q}N^Q5m{RjCg+7t
zUp*lGfU&hPeyqP`lSI`7_O%07ReXHPUXZ*Fd{mv184RJ(igH?x_bEFVSW;FF-5v;E
zJiytPr%LtNT27?p#Ode}P73O+v!nX{_IHn#ffbV8#)9Daakb(!7N;Ap1K{&jKhh-w
zpEYnMvIw+=`sUi_exmRv4IBWD$+n&yvG)>XA_T2t(b)3XUKqg7SG3&m*hv_fjnK=d
z$fHFa>~7KvHWGmW!b#*iZ%+_}ooc1e@i6&`vNRg0KBnlM&s}?&s(U69+C7{yM$=x5
zF(L993-3P@RRyyqs3jbPa%|fq8)c6iZAK-Bs@7Q>niQpGGrZ~V-jKEbd`C(m{p%7N
z=T<dW+k2EyS;?%d%IMLMNBmT2{2wG9$(s&uP-vzArDC^4Z(?R<WhSY>7UnKVAd&_4
zUv#IjQ<88)r&6tQIm(8|+f-6#vO@C8L%eRY@xD1XJk77qbNdo&hOQ0i)-Kq^rT4NQ
zoFb?!)dP}!eZGcPUQTO7VOMw%@csRIkt{;$KywD;=dj3wUF$X+R|PI3{HlFoSZudR
z*7Gw_5ov_5;jQjCE0-(@R8sSe_A#F3p+t!vl4no&7H3x{?UGJ?`E9g$420OK#*5Xn
z(+t0aXu}+BH)Kp-Wik3=SXLW68;(4l6R6h~HnQ^gJT&r-S?!@tT;kA(*j0u;`K;f1
zomPGmK2cl3P(s~_`AreTu!hsbxHq(R%cvtysYB~RD%{#)3R}DDE~ZZ(ar5kN8?~OD
z=||Ij@{%k*<91-$?%OYbv**MM(1+B44bv<5WZ*ip)6+N}MVl<aVuBOXZWu>%KL=*c
zc6*(pt;>#}-Whbux6t#`pmMlY!f68!7h9w_4505$Z(N+f)>d9V=NA+j{-^dvQ>|Wp
zF3hx~<U`tRqbD^@LIyzzR+F(4$-Tyg5x=NM;g`P6B~FL^Z0op$Tar(a9OvQ(Uagi?
zu>L>N-a8(ywF?`TND(6BB#3@Qgdm7ulrWA22|@Hkn?#Q`I)gEiDA8N=7DR8+`-~F3
z_iiG3Z^Ia4X5KC5Jm-1d@BQ=J|L`;R-uJ%untQEvUDv%<bb-$W+4Qg^RKm-6nl^I}
zi&*Z%?kI^vKc~QGuCC0O0a@jh0$rq|EA@F4aDiI=OHhUuOUH{#3P4`9Y-_vy;@-2U
zxrjtjJk5$1FL4i2o)}_<Uz85Y5D$y~GvwZfy8!kZ#TYsJT&HJA@gAR|x=3R4qjBhM
z`4<=kl|06*yW&jQd)ulq99_(Du)OGrS}z^><%uh(J#WNHXwQwK&W3PUpXypjM?R8O
z(=Z_?!B<gh`1#6m{<r$`+-s#!r)F#kk#{I8KD>_cqtQ%&w*ZO0BBJ6^2(oigCQi8I
z1@LYf@6}3K%ty7*^OBBLt|XAfWrUjxi$~F@IVqFBJ>_{h#p8u>vPQga?IwG90f74k
zWoyYlDF+z)(t2pVYn>dsQ~<6b?VU<vrRQQ)x%U~lqGj4$Ag+DOAD_rkEHt}#TJ2C+
zX3aKgPHp<^sQ6VLZKc%JVYhd-Lj<=|l=|21sO|Jd<a$Y|DdttabN=MqP#os{p^hLX
z)o)A;b1d-o{A;cMn~N7_<{IAvv>74_zI%i%cNNz?YolJYIpZ#G=}N!GLmB^SRh8P@
z4q?7N>sT;<(ly}wIBr-njDWADElwWnR_N2Wm*4q_PH~seJWK{d469b`aZxBp>9;K=
zS!;L%0jq*Pbu0kJCu)L-gb#k6r}1(rB#jb}s$F+rUT2}HJ;f3(wo{3%8}-K%2CReE
zYdieJzUWuZFUrYp?eqqrC#&C?SnSqZEqD3|)=*E-8mE<`Q0Mh!QkZSU(8`|+IkImx
zHFN2roewjRA3gwH*L_usLerh>SH{t1$@Zoci=5D<n?8Y>y!mhQdujWVK&jYQ$8-L@
z!vizl&vangbP6BNh^>blAzZ-JXA}nWm<_YYou+8<#`ua?R8|JiDkykE%t{?@SDM*b
zSlHi7F$YH;S{NyhcPU*7Y<=ymNWoDX8J6LA6ZK|u;w}({7$ltba2NR&zTbP-uuh1v
zC|pu$7<hL5-tOBKcyDpK1iGXf#d>SPbD3{qT-k(^ru`x3?hBRhuhA`rn#n{%N%;C}
z3s={dwt@0~2M3mv+D8is*R3e*6V{^6Re|T)0;Xl;mQ<dDz%tgJ*M^z7XO>P5S~(|O
zQ=FkPEb%G)P!VJSP7Jc5IUu`V-teAkS$9XtFSWE$(}KdRL|Cu4y*Nm$tMSfXQ=K9E
z_w{Ug=j%nyT`68=S_qiKt1aMFXWddVv2}#mtfP_D9h!w1?c?LAo|e4eIJ+FRM%oqz
z<-(Jegu*fz2mRxxpeF~P_O^~ckYHA8#*(|~#&-30UF<tY1>%jp-2|eKZ>@NFdE;VM
z1y1GMH`|tu6FyAHH14(YdM1$+59P#@X1({(s&=XM-+u&gLnI!LZ;r*S%DC@;GRXv$
z;1{hs@)+ZnQhSimdat~fkMh3fAOYv<vI7H+7~uWudFO0|@a{t$?D-{EzP6?IA|KB?
zq*Lb&&I~rbg6f<p=uW+we8h}(-3Ia&<d<EKRrdn)o7Pg7cmbAcmn|hOu!Pi|CR#t`
zskt=C7k+jW*2>}g-I8D;1!ZN8iGNdPj1RAVNWFmJyL!9}U$+<be&j13tTQ`XyE5)j
zd^gs{+4!w<!FS7{3~Rp0;)@{L8DjqeD>c){*J>%1Q2moA*ZKM)q;7V*F{PMtRXuD<
zh1xxbX&1PhH*6^@*+VgHO+~pxKx1#6Twe@J?OPb`UOM5e(q9$h{gZoR)9Dn$hvFyh
z^LU9h`SP%9zarh<yw8&iO6}~S?dnnN)NJJ^Y`wuQ)ActFjCw5)h+B)-Uc}1}!Pcju
z7gl@qg@LG(OP(HBJ1N%1!!iL#L5qcf5&8)DGfVh0b^S8AzkVp(mP`J^+QHHF^Ydb4
zem;-OZLSgB<5~PY*Kdi(9cp@ctJ{296?ylh?@7tt`;B$}ra$f*_NBK#JDwGY<(Pl8
zx`UlD)MkY7Rg0~CQ!j(0fs9UE&lXVr{`cv*c-6+Pg%WK<7V2NYe3W}o@qi}v+zI;?
z!0ona3iN7`b!3)zS5f!Bs;;6E0Y-im6@NRQs=TvN2o2|WBF-W2vTT61H=NI$lfX)H
zf4lf~`ByTq<mCoOpxmaVz0c%rC7q5{m+!v~0g2udDKdZSMY6EmI;IP@W3<=?Bl(k=
zsRt=`!M9p!a16C}Q%9v~X7cW{jfA7!?V9wrMH9QI#o0-LyGNxT<HMQlrW{(`v%<Px
z+_{kbd8372rs}OK#iy=qtSU)h)*XgBI_!h&)e5qJ#hDfy_Da>}k+7YZ%ho`aK9RZU
zOS>Z29gGt{tj7vhGZ-1wNC;(VF_`b~YuyR%{hUVi&V|(wD=qV6{AhwNO&$U5Z`9w+
zC5ykbApxY`mvIbcTkBn?`i1q0{NbT$&|LR9?gC2&vaijT?1eTkQrrqqBTp4N@;D=y
zX-`-yxkT`lD>rDi#LtF<wxyJ>pGgtAIcb)mw_5-qZc|fOD;xl#O56$!U)R~<gH8kf
z9M`nio}r&{eEV}RtS^9}2w}4-D*7Z`FZcoSP1#)20@A-hliK;TiIe7tE;fnqYyw?h
zbe7xDONQJkQiR}5-HnGt&jGFivtGy%B<+AbJSl&ZEw{Bs{c;gczSXh?zwVt%+KtYi
zJf|1^%l4_eR#pcZ=;iYcF5CjR?tlQIv~NUL50LsuA1u@=1@*@M(Es5OmGUU+oB6#r
z%n2?4C&<dkc!L9=S^I$qw2dh+>WbWRqQ2dqZ*FJ(5@(~{`dn58kkT)h<nQ$5Q?#`x
z3e(WVz<3TNbz>?t24-~g_Zz~>f_tA#-B<e#UGO5dUrG5g<vL537_lGOFw_+%0xOCB
z7BiiyYdyjUXd>R8v`9J0zzl~Jt)KH-)!zT6qD^E}1|Wzho)z|PAc@S9{L|Blf#P+w
zCuF>u@4hP7^7Z^Aix*BNS=pojm{-anq#yC<y>mE9sj8p$^nP~4BY>_kqbW%}vL{kX
zByc4B+`}KcW(ADYmMd%K9gnuAEM^6fMXP&b`o+ckg%;39E{)Pt^=ZeHf<Af)6;xok
z%_n%Mc9C~0w?@YXRmI6KM58fUP4V%QVeW7j^rP=@mouQ~*dO){L}0m1OJV8WSqO@O
z6y~5TNVWOCH{WV>-YCy)iqReDzhpmZzOd1DX!OKetoHC1Ggw@LG%&&6aNMpe!34Y(
zw_mllQ0@+e!7zw^jSy-D0ZxKBH`8T;5N^auED=jYBVBBqXEs;7%yE^7<Jwx?usGg}
z6~m2-B~jbsJWuAorl!4)7TpJ^I^L=$F@&o#$wT@BtZOV@ZsH{OA?=AmluLuw&95<f
zZU(t-w&O;YkmV0d+IqT%Efg;WZ}?uUNG7`>AVKJf6K3$QGyf|~KV`Chc~3U-CWOOR
z$lkpp);cJplULZM;B9Lk_Xg5ieeNmkW?uQRZI7MAU^On1j;}qh{#{$cLFbBc38a>~
z^|-FMWL!vI;ehUr_-wU=fo9#j=GCe#o+Bk{jIH-lB6|reXKUtg@Iz8QZgPC@=m9qS
zI>vVXKy{3r=BtOcJe^2&#kEI&m1Cldg6zcY!nIu(ELh$7qF<=%iA>8mUgEqJaPyb@
zJzq-z;V;$BJbN9=6BM-+7Ln9^v%{geE=!Exap~TJ2fw+*nxZ1<!yvPd%u#XGT2B;N
zax!^DfjDpr6Iv$N!qs(EY7~2HR;}X<%QFaP%v$)1TQuo216XQ2-77@Z*4ErwPx9&0
zKZ5h8Q^Yp1_@EkDDqQu*7aA|Q+Ka5N?92lDK)8}g-K%GDhmffmNz!HQPL;y7LX(-(
z?QJACTeq!<sI_79;QDOv4H-ZwGvlJ>jZwIp!lnw{+*AUvS%W!Eay@~5oCdVeYDOX8
z&6d#w^Xj&l#<t;dj)K;!r?Wp(cwQBNz)bZir%Tk@w)i2^wS}sD&jz!J>F@9PKXHBw
zP-T5G4&Qq{6jO3CJgu0&d$)%QTrcjr8idN-xpuKAiq8Bdy^&IL?%<Rr!_%6E4YpOm
zlWuGa%`RKtPjS~325cHmKG0<qz70k!_uveZ!O~(MZFk1v0%-eDt{Z7PU~e&Kqsb>I
zD}b2p2EZWr5z@hOt+ur2SD5PZS*@eB_Dr_*0J-Gem~=vz=7e2MLB4G{i6U4hfcA#-
z@v<WH-tfAc+7|4rAg_S<4(_5xZJ9{V$}l}QB~`^>wQFU0Y!x=YkbD-pk@_c@Y$H%>
z0f?tD9|DA1^XccjG~1r$Ub%8b4!B<7u7Q)aBUkQ735z?=<ImP#p(MWQl^a-(<;c<t
z2EUR1zK;bZ$kDm=k&-QhB~<t>ALEWTy0faz|L+CY97WTWtt@B@JB=a7V~<(c#Jy_H
z;+y=XBS)?tv-R6l5$8ogW@1!1YN&&r-Xt&g-4CM5o6LOEWF=38#ZHA*U8r8v;;~^y
zOoh&k<Z!7g2i0KsX%jJ^U93^WNN6{OvHq@ZB#oEEm^WONee(R2oL=R$Frio59pt%s
zc@M@{iAQ5CT-t_U*f6h`RhEwR=%I=SOLe6<8zXMdz|WRQY;~fOi8N+y<M_ex3`njx
ze1W7r_MY2}U{W+&9W6jtWhqs@is-x9VRki$O9kMrj?1<J6cqdy-Nl14UWO^$y=Dh2
znj-JPVT*u7B4E}&;vSr{q@z1pN4S)g!|m98xCPUhl5mgd4`&UfW%Oqy<TSNSQH!*!
z_Wg{$19)h?YJzchMMUEW^MiVQ;H&U`?ZDqz3%8on+wbK_EY}~>#P7{|k{ky-Ny7VM
zT0d!sdnV097C>{?jw|u%M@W$9l1U!eIXC9%`IF+hjr{xA&ZP(?Q5m;PYyF`qS`JZb
z`0gja;KTlBUuK_hrm)T156nvoi@((i*5D9v->m?M^dIwijh@ibgqsqg-#nl4NT~tQ
z=l0xKKlcT_<?=W`O@<CgQG>o~QR8v~Llvu}TWK$n@2C{zwGmuBIx$3WOYYftH%`*l
z1!w_SA1M>Yv7~6IHP`vE<cSn^Nlg`bN_d==nCLclHeidxwBY2~B;9x|x(T?nvyH20
zLC`hQ78ZTd+!G8X;V%OtlYAXP$-O}g-U8^4j{8W&S#d!A%;~3DB9J@{)Gx9Mdbp{e
zKM>BJ!LrjK-Nn<td=b>GSTdgZc=hWm-a4li59168(AB2fTs|iLQt(qu9%KiN*BE#C
ztmCXBk!aI)obu+=w_+oY&HfI#72y}$^AfDFVU^f!&P?)LdO9PAjo(mGt#jYnQ9hGU
zS<wo|=cCa#9aSY;10i$<rG!%3H-ZK=9_vLjvGdM-;u8l}v$p_SG}ys=zOOF?!G2~z
zQjwMZ*!N$dhY)ON^>9#kvEXrcWbfOq47I~?Tji2#h?Iug<h#j~>87<XRwMXP(aa7t
zw?N|tIeGe|RJrZy@lg9AcyF=b*QVy_qWWSEC(9G}?@?n{lnk>Ab*P~cjK)!3j)1}}
zUNvXN0^+o9Joo8RJku|L$gecpQ3O&%XO;l)Nh@FUgKqi5n~u5yQ8$yBiz7_{wa18y
z5f{g^gr%^Lufl1S%zymI6g@Y6AW$jK(T>YQy)rkuRnDFG_10J1@VYa_yFGN6AML#R
z!VHHmX)mv&vgW>0(mfR0<K>ejk--p;M;4>eSs%O{Eg!6+f}DswxDw?V)y+f5;O)F$
zA7;(AsNZF1JXHDyT`Hk#S4P*|lWtf@{l2<T*K<*iLwgDH4tDE~xjMjm`|de^(MyVH
zDh2pcKE~$K`toK64p|d}WA3q{G-CC{)0*_`VgSQW>&b@t(hjc@SC1#4Q|kJ70HE2*
zhEK(npnfS=k20=asAy@ptgBVs+=Wh|Nw3lhJ}dzsddDtIC@rAcVYp%h-!*}x>Q@mY
z*ujL&gLURO7AfMV4gnH0X;nv`dr&J+V$9CrPu~sAhpC>%{4Ja=W8lK#x)1z$?hTkw
z<#Q^Y3O_M1!XUJe=xngoHNSNd+UTY;C+j!9z1|gwj_>?N6aVd3K5C3iWYYJFeFHA(
zX+U@JUJ}T9_WRk&Ql8~EUE`tCj{Js=@t|!#W&^nOY7o*mvr%~QhmOqvJ6CMwe&Z%&
zdKUERX!%P}zAc5Zw99zu_M}i9-ApwdbDLO|q2K+z%c#@w6mTlPM0P?>ttFu`D|Kbq
zj8D!HRw!(8DHxvEItH3+$c4UrYZKaH+rcCVZYa&Tp$vz&<J}G5oBi9nXGfD<Ek1#d
z0C-R*PWDh<T|_KOENsL9SRX67B`Bt-qGdyPl-k}*d|~eU*%?0bIipILktG2W`f)6Z
zCzQu`5D?8`2K@j{B%qL#HDY-Rpp#1q9&2(&^#D8>IVa@jUhZRV1?cu1O@Mf_m0`YX
z_A=R5K>Zu-dmcYZkbvzVy+RH{j5pT-iN;KO=)&{W>pH1p^Z1`xvJpG@R5qq=Tk*LG
zx^e9I8-AfrUnZrfv4z_YZnl;_QaUm<go@6M)Ye<-b?o$~cDH=;S>H%t*pYEBjv_R*
zMR(Ngp4bj(LJG6uw(;JH@SCZ8#hy!5dWjdzgU~5nTic%m!;%soos47A?wiXv$SGb;
zkiW5CP<Wkrq7T=bxJlY-?-9K03ZUX$tMlGRZOx@2;SA^#+%R1ls<e3c!~36sqU<3m
zUMN?1iekA_7(*rEEWfcpP}F_bNiF&0&t3PS#gv6j&&^Gm_*0Jw^A4y@e6GR7cW!LN
z;IsSiCe!;L$iKGT7x7u_9M>*Ltmim-U13aUoXKyD8VNF!?<ja!W3Endd2y-Tzm*QX
z5mkBtq76~a`^@y)yRy!hQ)lXHTFn$Tzo)b8r1A#h#_vd{J|L$qK1_)hSFR?Ubj-Bt
z=bcyxD+b1`Z&(+z&XtW)z4>cV+xc}xrIHMr2{{MfVU{2~?~L*)4cm1Ne!Ys^1F;TK
zA)&#^D+$-4da_ib)cSui=h3#%=`M9F)1+9-i;BE>0#G+{pEG!TX*(PoIZckaCHvE0
z525c^a>`Al<UR{0Fm=s|_al=sfk?34Z%A};$W<=8<mI432B6^(Xh5f>T-V68d<)2Y
z0X`HEdh$VFa={n=G$r{vwYhNk0_K#XCjFR^*!496T$tPZ?X)%j{X9PInvK^=L;SJC
z=+}ODzdp2aY|!!k$3uu_@F65LcLY~LZ%^EevKl}%zl>qqn6V!?m8J==Hp+(&w5Ci^
zd9WEU$gLrB1f~2q;REv#^-Qw1!>!WRp7mIl<%jb%ps{kXlzKh2{KiIrSHA^vE!hkW
z7-<2$_2+wzKIa;Bz4v34*YA{ofAi$!-hMCAHs>)O1I5hw8;etbfq4k$1Y7Bona4+m
z6G{NnnoIfZ@7~nd6PLT@4lU-T_GbejroSY@QY22?5YBkCU`?4<Cgfyf1~kz3#)7y^
z1tR766`=9BIjCt^9IBEunl>kHq}hZSueI26OB?Oj^~hD<&Mux#OIqF^+^CnMz#dC`
z-`%V(T|knj&v(+iwc6b3o2IICla_weSlFYCrpSZeio;xpe$zE|pG>3&Sc-(`fS0^B
zZ4EWjSv)C|pWx5+S{~U=u5sSn3quuKe$k4F)8tK`{>R;!Bmqfx!m2^rt7?zb=f}@G
zKD5H6ZeaarTQcPN?2rJYQ#~%lE8<H_UxX-J((Ynf0f@te(L4y7#KCZ2EBRuh)gmIX
z;ItvUByoio?+z3aC^{@9ozymPy(nusSK9|UGG#`I1UxI#8@a!B-_L9(`?o^mMP8>`
zGlhfxXljZ^BoKzlbrz<T+etV8y5Fbg5~50o827GOV8-aVq{wd}{7NRE7jd9)$o@8j
z%%GI4yBByy@oozye_8FGv@^$>s0JpOpfY1Lz{Q1lyLST`bz}1{fO8DjKIqmx8SD2t
zqnzorTLqNn^oBWFyD?jG+!=jL_spIlPS`KkJ68Qi3qW-%(TXn<uwuELRj^cf_St6g
z-19uF8ITq&DVTRe6-L|0IFwi|0PRx&QEYEQ^|#Uc0Epl2P-owvk9ZXQOnId>4rjBJ
zphYqTDsH$WlX!og=>o!|anXEX+yKJRWlTHN`g)@ET&kiY!z>o#pFA0YEXI#xWnR~P
z_=pJ&Ne!4-HMRe?xI^o)kxw5lQFrXJDlnM>7agS?2KqPGV^>tZtZk?sVU`TlVD2>*
zwxxz@yfJV?^Tz^jvuj9U$(T?+WW{mhOc^`sZ&JLaJn5K#`L$N|w=QJIYI(_qS&jl=
z#q$Km4TOYlH5_jOXO$3EOT2IgTOB3IV7IQ7T4kNI(}lT?f&0N>tfZW@-ah?@9WKF#
zBOu|G0m2!U9t9d7g{mdL`EKeBSGy430!WVfXI0?@b{$ae)cKPHj3cCh_d#fG<PhOM
zB9C`GHU;|Ay6iSO$w@wViMHJBrg1gfaIm}K^FWEL<K^I&NFzG&Bh|6E=T<+sn&c{f
zr^I2lkCS5AV`E(;^7luv9|h;C;BA4d0Zxf`?ub7Tvacx-1Q5+O(+2)YAVoQIBt^!l
zCfA6B7x9{1^{&$MV-IH%l)n=3@UfKSwJLGMN+m#|`3^fdC#8iC0z#ki*1w|OFHH?G
ztAD9;<~#~G2Ta~Y{-f2PuBDKF`UucUb`5n>_J3X!KWk<;db5#rr|;-FuoI`_ok$hu
zh<kQRF0cxUS|9I%pXeS3-oMHAVoi~Ex74)XD~dacwfE7E*2b2^Syt>9bVG}1M+}Ue
zZ{0b6fFMAZ9fAF2!TsmUE82|*T}}^9{l)f79={ho^VGt`40uKG!cvs|c+8;89qXjD
z#(J8lf`;A8qzaF_7rQBKww^t6yr$hp5m#u?;LF>V*%KSnIB`r7{aKDtf>Y<ULhl{4
z&5L|tSFcCaR)X1OJ_Yx*CHxr`Vu$pEHNFl?el3MtKeb{F&E)N{W_bg$tkbdli($D<
zY&_Kdl?aCDVD53cjM;R?CZyuL@_Tk<2+2;~bMVv8>#Gvz9FgvvQ!2~nHW&p}=lg=8
z3pJSXmts;0N?8mo>Z9J&XVHLW-6@19NPryC-@+jJ9-X3IKmTPSXudgAbhTxHwcg^&
zJvAel+b!e5MI*sI-=NR@QOWa2PjltyG@5idAjMFtxZnG0o>j4h_@SYP>wZay|0}Q3
z9}DTlA2O7Ae|@X>*LV{sJGei*OLz<CdvbWT3FH18g$ZSJ3$u}M8HhAktZql%l2O<E
ztnl)*9kkO>j^zE&Uv;o?G8tUvjXM|uInpSNfXrvV|0)h3;?&qvt-KbjSNrGoi81>s
zam1MvKHeCI)*yYb@S~JdhewB>AoB(sJ>+X5`=t!~VbWu1ZVUJ!+rE4bpeq(8w9{l}
zm~hpcSmn9>nxc=ANEW3#uqy{)pNn#%mn>_<jrjrrM}S@#-`XY7K0Ku?-?g7_-{rc|
zf0rnPKkXggp2t4hP3^eR(VrqJ`fxIznr0{I*6uPc8ra$3wZ~<$>qD>1oBvfhs5nej
z(C7~{`)3bmh7Kkf#)w|N2Ch%F^m$|YY0^8}G`@Ksfesn683Y53Mq{c4&C92>ctumf
zB6)Ei?#A>XAH%_<ye;5^Sfi0tqlTpv1Wx|;q<Vaf`ggkuAtiL~@gGzhd4tF`@_hC8
zi<4AGd?mu3I7XIb<BO8ga85Djw+9DV!3;+I=G>WsA7)pR(tHLtbW@~~M3+;-_1afL
zd~0t%7`D<CXrwQ$&6E4|#yi{!k;a!Zv!og|lrl@7VHrM7Sk&H$2u4z*_u@4ZWq)k-
z9i(0NxpBaU@-VENtiQsHu3N5E2)^i(Jxm%SBS2(tz0_$mYN!(H6B4RB89eTJ%pkW#
z0etV?NosGO9u=+I3kiP6lsLNJ!pu6a(aInsE#h2bx?oCeQdVwz>WcQPv36EAy3`>{
z!F3t!>28Twz8>a0F}UE_I^J-oC#j7T9aMmlY$?qCc;u+sICpZKMhfnEOaqi7RYre)
zg2A|?)#8C(Qe8_)Gy!|BV){sH3!X->#-Cf9aPpOqG~#j!nE0Ynf!*D*H30g+$bI3c
zS}vDfb-x`9v8<^8c_+fSus{9v`L4d3Js^f7mO|HmhJgBsdjvjZ${QTMN#2X?%=ufc
z%XQ2xZaL*otYD!{g)ClWD1UHbT@^xdMX)S*OxLTi8P%CeXiXa^oT?-s+pGTIBoue#
zSJag7JxNH~ERLO^=<WV6?aeSfw1W0Krdc9*N6)J#mrY;CUsVY9<^(#-e~dd)=8+w~
zWFO)4Ec>4ITZ&&c8#Z&+HgAne+JoO9?XXCf`pm$}8wf%3d%vdQ*rssEC5w2e+ErrS
zKvh$|c0o%hX#NB_KTlKidY{m-1Ge<KJIE6FjIKg)eQZd*Q~t}rtJ~!23%rEM&DKn8
zQ6R&rAk59}i*bZfmQ!4$Vvthbwlp?KE<WTySP7J*^e~8sko?3+C-*!{gF|Ve(>(at
zW!)ly4+^=?!CcvUt9)NRIGcEg@18O9yc;V}k7IzI2_^Lj=#9qosrs*DB8Fqv+mv~2
zauje6?!;8`P;s6*8&g@T)5<0}-R*`KWdO=s<)Bon_ZQ`m*{DM*Zp%z9jC9>ec!`#+
z1b$b2yPzYxU@1-Q;IX)z(RD?`qfy6-XWR*n590*tC6?Okzu3Xp6d@M|uZXr=opqNO
zG04@p^f-aV9d*arf#%RWilhmfHBwMIiE1U*xo-bukh8L;Is07Vd<sMI{uC{eWlkmQ
zX(u1->qaakWN|@H=fWSl80`{e>sJ1qxkxs)R0*I*2(6P#dt*R=+$0L|r>9xj2~%)!
zD^^wI$8_-^?bo16UB3ymmHk|ScLDMkx4^ni$6hpkmI?Y6z9e7)Uu>pjdw@Pp*!5cR
z15vGNA$LsF5L%~KoO7AIZQoFYnpyGdJE*|<mIB<omKymAL<ea4PrAC&yiZ~`Gw~Be
zubEmf?j<q05PcUEu4I2!5OGuPNZY6CP2`Odu``CL;lw01JdCJX%;H{#imkk#wgrxu
znaT;b3*KquV<kR-S|=6QW$%f~==bo-&Qz?srI#?Ai9ME(Y{&xsJD_?=SbOBDVwLWs
zw|CIyC?zmwK1<hreuX<~>ZZWAk7%dC#zm$S#t9HMbIn!$aa-cdNW((`2-5Cw$0f-P
zz)91<xKnJ$;Z@F5X?M)h$}@`#QnlQmQ~R-e>S}5X!m{mPxvd0iN+1WVMlpeVulVO&
zJbLjY|EmkH0c%RV9AG&Nf=2EV;A<Hb@b?veQM?1zE}jRRbx6r-zuY$!N{vw6QDOk%
zI8ZD9?W9%3e(v|`dEPsC2fzxG5C{ZDiOPvjL;f9Oo|dG?n0lS!{6R#Ys(E=dZg&h=
zK~)u9e%SCraUmHL;glMA&Q(X8-3K&VlYGZddv2^-CfO?g_$)2cx^<(iyPl7rN5}Q&
zOC#hL^!+%W2;6*H1%_pK#Z9JqKb!kOg>h2oI&zx|OzeLe=oPNihUCU|3F4F(6qFNV
z#IzStNwwu}L;7E&lvb@VbU@`9nl+NuSpGwF2C?MFIw_<4dO!q7uEt7Se*Kc~!+YZ7
z^kq*umJAQSgMB^u#L(zpj+D2JE@>dEJa>!ShxrR@4>rm5*88#k(6ppFFfJTSx$ucD
ziL_>+!nFBH*<G<{;m?65U`)&MN3MxkktZ5BKQgCjYF!9KO}tI8e?ro{pe(UP%KYYy
zi_Re(a-Dd2q?LGEdFdzR3ru2h2!?#^_ooVVo1Ly;vI}HRK4dqhEpT`{z<vA0wX5wI
z#Wj7d%;5OE3-EMSVlMs0b4<F@5^Rr0v`2#Tz7H*L#zKv%0TVcUTyDp81XE3bDT_U+
z5gdw)c&IfK03%lF&glNZZ*Wb4u0qwXAk)c#>>clWW8v-_x%M&M>q*gwF*m!pSxs%h
zg5p2HcA5EIJ#iVjOmxvoz~-<#(jD*B6hxnBk9m1sLrqA=^|0K(1Vk2w2ZUkk%o6I?
zJ6_&V<c-;=vwDxYqOi{9ckR!8$9jRdW_8a9#fKSjw><Avh*bSq!N0?y<0rUlbE8e8
z6jCsG$rregL644wz3&^;Vd%13#oZV0cD_BncOspy?)_P#(BrX+)svun<e4KrrDI5!
ztfJ2a1+%7D=p!9}q8wNu7S#eP3x_W|OD$iO1BRdM_7G_-MEeL4Gf0ZbBn3-wf|rfU
zHkSYJ18;%#u=D;h2+M<mB~uEH#_guKO(SVv+7?-gtA}jk%)K0K<dV<4yX6}#a!yOk
zChHI)o?Mbr)T~qL*p1N1G>8*xbos?TtbTB7?K&o?nuOHbL?NVho9n?tiBTIGfTG6;
zHloWOiM+tN13<aATlfH>o)K;pRB#ypxK4Y0iZ<gQbHm-;wXV1*+U%ZO4v}w7-36PL
z*DG3woljspf@NvC)`Y_|RHsIgo~8y}lIB8JLxA?vr$M?<M~No>z7@B0wW^G-k}uFc
zlx5A(bh2x=n*L&5zW<0)kkz{iF}LD4awlc>&~vD>N_!@CQnyO`DJ9S%?8zOp<!^Iw
zX=H9u7W8}N@c9kvB8`P+FV4bk9<gUb!C~XRr@wD<8e6i7uG+}@)TJga`gqOmDRgQm
zU^PwJpo1C2Fy`o<{6zAq49#ny6(ZL-b?z(JPTm?X)=^hnapHO03dOSo)}hvn!Vd;b
zip*tK<c!Y#n$T80j6)p?LRWSK0pHWZn9$&rCvtXIG-beHzolol!HKK7W>1wjM#(Rt
z+q|5^w0!D3!1^6|X&-H1Y31_mC3CwUIz71OV)z(RanIwBxkq#Dc4>!&YVeJ=$+=~A
z@}YM+bviQ_Ihiy)a-U^z4V7m>P57d2uV_iOlD?m67U&cumZLL0I6WOZ7WhZ9HzeZ=
z%<X%V)(TAj4xXYA?ZF?#ctQPh3>6n-;TVLrfi6kc1cO*!BN^|wi}=pfUX7NLyHhlH
zQjSCI=kQgAl$&6i8;%ODTD+vk@brqXsRV4-=H%0$1SVA_fttUA8dCmYqhrK*T7M^+
z`&Z?}4bt~5o2T+Ldc?_FTVb2JDcUaM@7V$4sG?tivRK*h$!dE2O2kmb<shS@p)xEj
zPwAMzCo8?xAh=;x>{tV#nnSB+SNzB|=1LX-y?J@{q}I|>UA$dOY#o4}*t2ezAnlf&
z$rRlew`kC)gILN?gvW7xEgn>uKUx+<Rf3|Et2_$>5pi-`2m15)4}0xu8F*1&|0Sr!
zRFL(?gIktORch=-1%zzc8%u0j_~5Odg~Yqov90Ta;6qRKvB8Kv4<g!B8E8O~XzZDS
zqJj&P+%~K%(#f5E@rz^vY|fmh^GWnII&!jCN@Q1UE4&Vy{67snuP(p3i#|khyO!{=
z{^3SkO3l0Q@!gm$yVV9zzCmt*_6Pl_M3T#)f4p?rA^eev-~hP3+;LgOlLL0B62n^O
zPQ;^=n6|4n;Cvgo`7TEAZLiNTzlF0dY3x8quM~cIlS72mWL10CYN7@O91r%diw{Ah
z3`%o^i*vjxh{x8zqsvYGiej;u@?&x1Oc7q5zMs#tDBBZNjgXjrDP6hG?FHG6c`PFZ
z7MV4nT2}WzagN5yF*Sw`?UYS*SxRLIa^GGNUm7p+%5q*FsJj(D&omZ4M4RZw)4o$|
za1oeHiPUCw4(jw4hsmIOezmdvLTN}vCth1CRl3LE-ru45uaK*<J?o_WhGJsLn_*HG
zDY(|eT5?x1uxr47Z@Emx?RJQ1v1d@3v!o!E$$?VL_i+2%x6t*Lt1#y2eVov#Wzo#F
z7`=;5-C@r0U2X9!)bA9hvdPP(r09A%gkV|S@~raj*>RpTX$ZJ{owcPR?sX`^565gS
z>t_nb_+uhx>EfYJ?zC{4CCQw5%zrBI&YYJnYUFJoj*k+$CQ!Fp6#2wR{mih{ARtg$
z!U0Jj9Xt9U0z)0oty5gj;e`!dBdVzqF|E}Gkg9bW(mpp}R)lGnqn&ksh(3Lb7Ja)L
zX@8b5QBpX0HroYq#8Q2FTnw8~d7RDxJ~@ox+s*y59EnQVjMG&nkWKF(i}y_J=&Sjz
zKAqI5dcOcM-OV@QPTDDWL~r`D>E$C1!g^+D8^20G+Q8X+Y(Pf~2&6r^q^Wzr_Kmml
z(jV$4^@AU1tKw{@Pj=<ffOd%dolTZT?$#b_&vC}zpt3GwCSw;YOS3Q8W2n_$J{b(5
z7JX423|%Qady#46u|(7b*)Xq>2J4oGp64H~s(AzX$R|OS@Y^#&gR@g8M|mywkK##^
z7NvGZW)?K_sKZKIazav$e?kfvXG8Q5MU5*2C%kc}i<-6K;;Y*Dls^QpRp&O2oBoux
zIiRg~Mk%E~AAlR?#G>lnt;D2)5eI7~;l0su*HTPn42|cWl`>Q<uRePgNuOmc>e{Ct
z%4c-?=6+I2FPG()6`iGEy_Fo7XEd2yQ#KgQcGKGh>UNy!rtQv)TBBtmikZo?^<R{-
z%BT3IY^pFSJ&i#I2g;$0x3*v2jknG>%0lm@*L;1Zs8<VmtY;+A@r~{U<uK#Hn9IF5
zDX#v!iaZJJ-X}lPuD@NdZznfbAu&{*n`6N`2C+^uR^<&T7Z<r*b)%yS@1=gXUA9q*
zNj0Z>VC%t4?_>KD$Enb*NyD1zv!5Y~n(K2gh91Z{{J>PL`RIT=UZ2wGkwUe<FpM4_
z{Kun>8|C6(l%UsO7CA!%z*<AEUyno2G;;JN%W{|FpYoI^6lLlzL7S2_B7z-uaI38;
zZtYc`E4Z_S;S$^&&3JvL6vA!4+9{GT8tT2$yf;Sqjvd>OP+i?d(sU?mcazbKK$f1E
zpm{XK1`{Y99~r@6BS#=K?aWS`AbX$kkJku7rOWQx=}yvpftOb;8uos8l3EFLEi%L%
zzugr!{%xcvSbTAVc|wp-pevw2BjB|?TPrmW5$nl(D+Ma7%bL&lFs={8@%?A=$D^w)
z@bI=ODV+$@NTa&iF~69~b;EUXl8ENPKpN!&($V(njoW8tx{rKsK36CL7Djj#ib_;l
z9SGvA75&_GUVq=X@#@OguEal@Zpu|ZLX4W@R^T+qW6zOHQa{ppFEZ{%yO=@ccHrEm
zp0@lE$LI1qC$Y`q8NyG0-5iCew&Gf=nJQ))wnBJDjz5rT;+$rXAW^7}-<aIw-rG%?
z-07w8vL|@J;B-CRx>atAt<rhLS2k$)90xV#7m0{vv>f<<qRH>KWZf1J!)Jcvap0^I
zk=A>p7p0Vy9W5%?3by&Cv2Jsm;yAWa9H1s>6a~i{3rOCp7W!fMHVpxLx6T};j}#K*
zFWg*<YF7U%H150q0V7M%bB&08+Jbf`6Emlcxg^atXGs9Ar3!e=CZvCg%8ewJbZ>5u
ztbB-q@xi~bDJj8zV{Cf5G#4XF8n}Ys_%5m~Gu_Uwnep90_wfLm3~gNVBh1t%zw%4j
zl$(rqRM=-ztJAmZX&wp`NtjESIZ66RtPO@?VjF5zORk68vGUehzC-N$Q3o*B^#7XI
z@~HW_Rbl2#imbUFR+QN)Z;{jDL1|&p_y@wFb9`u_CCV`HLRY=hQXR?BO<?=iLG><9
z)%`TH?H&EV?JGv#N@g^ZKdXOIE38}b)6FiIX<4e}tWTTzqdsNzYWPwB+}B!%ZcOv`
z00)3tq%rHvo<;_)e+ddk93`OYcM5hPZ<3q$dz{qHjv$S~<^4r&2CN%8w8=M-_{CM7
zvwp+U8Q9{{DqHF`JOzl*9_n{w;Moz%eli`75S}~r1K(grztMB*a+NulFOy)E{m3>G
zyOZ`JFW&2@Ni%K6D5}|EE037Qa0WZ<Q`ycF7+-;Ar)>@3M~X>JnxNQmmX_5UG70QV
ze+-{b(<RpF)Oq1V5?{~S_^?oLN!*zLrmO0<ZhKQMhC27JsXXHgmT~eg7@Atj?XcCU
zlU47u{pbtCY&?x2*4#0A)%a#=?IaN+Ag45}_JmAn>)N{8qgSxnyppP!ZT#ql-9XzI
z9jbR@{#W(IRV&PUK)asbe^FgOzk(1pFb&8}#cq!=ZMuTh7qW_{tBR)IQ467+(Z_z>
zadC=(SJmP*RPLdbGpcD1!TjO^PzH@WOoqInq3%P&id?=JCHZycJguqgVQYB?pKQeR
zZ)d6ZN4dGJ0?O_@XOI1;tRB1mx!8t!DXgAS30+8P9RQsUJ^W#4PZ65u!LMyQ2%ll7
zgW_BIP$;llaBOdOzK%`#)3#J*r!E<JDid(w{ZxUC=`SN%NjPru*W$_hWR_P65W0ym
zk~WwC(ZnDgRofUE?D3J7mglazd5M9rs|&1Y(y!x@H;03nmCNqjdAOk(-%tj}rn9CE
z09sAV)YBO^cK&(&_s=`x>M5&5U@&a+cAE4P5rX^LVd>2KT&q7HNl~IUqILZPt}0;F
zz_VZ>Ndxy7`R8uimX?ZGHUSAKTShk4&TZZ;w4`L6nR|^Xw2gx`fRB!)YKNWm^VIXF
zT;wXpLkNV*d}bF21;<9t(GV?9_ruxPo(N-Fk}iL^1829jRAiskbWgzP3$ifQt>OaG
z5~d~&2fS7vyRqYr^)L77gm``bx&L@zwZo|p*sgdu(yb!hW6svN-(o>r5hMZCljFV5
zg4vNkx0kbtk!!4g(0NOf4|wF7<1l7$1a_}wf_SP%I@vWME(ls2AGHe(NbiFXrwM^n
zus)eV)UWAZ4-H9RZ>#m*OkF?;@fdM9dkajE7qI0tmr^EeXKym+Gg;Q+?fB2gSI5ZJ
zhzrdYIqWi0_Iq6|n9vAs-QgGXGxM4l&!K{=saxQat<`LH#LlE9;cBW^<@XKd>dHLH
zuc}-qs!*VlRwv$YO}5;d!78eRp&j}h?c-Q>xOrtTur5l#=kj~p${bz{o91=Awy@ji
zalAFP-Ueg=b^BExh<&yRDS4;-Z_5ju=r;w>4W8(LQ#CgDmc3n{c7fQ-5|^^<xQS%r
zSm->qhz9Wbgp8x?ZX6IWIGM8*xzuB&hY$yNMH<9zR?CFjqHW`d;NFt|-mo&DuNXe*
zn%j(G2TjE<I|#Y}I172&CZyU6*CoAdZbhDJ*;QYkX+2<NGdpa&q1_WQuI?|d>B?YO
zNDKyf$Dh0P1ovIQ;spXZ!9$!$Dak;l6w4~l-VNkK?Fs>3lJcZH;sNkuX?m$D)-(3_
zGeIv4K&qw5jKc?Nj}3!Iy^C|JAN6VP%!<Qs8av`4Drn2099JML?hdHP%pw4+0zXT?
zJ6T2eIkBLUO;RQ?I1TQ_)w^y9+$nk@0W-tz2rIs?v&IZfa^GOEy9a-(_CmEIt~v`a
z?>qFY_qhexXOZ<bw(D*B0Veb`Av+QkaJQk@VuN|Oy(fI;!b-(csk$rv!2_!L{yknv
zBZdLN<d5#+sX$RO!di~f?P2MdJ^THs_3pjc1MV<q-C00NK9i?qPqEIyVA5pZu-d6Y
z!p@Z}<rp=%?8KzkLjs}jJ{3=yoA@8$uB*^{@VUJaE(z9f8{FjEcgXgo<q3$p*_xj)
zo^l~E916>|?`k!PJ7h<CV<XdFWwx<Lk6%bx^rs{3Xd+0VC5Fk~Ui;h-{Jt7eCwpH-
znzR#EV3LMwOCzbwE7+Z!9PSEst=AA+A=|k3-U97>(6g~t=$STdI2OL)CgY{tdtkkL
zvK@{hY8df%wo#!8`yeFqh@*#;wH-43%+AMDx@ub*;r=s?lGKFDA(E@LSF55{lah?n
z+^g(4*E47I<3>txNi%61;8&RHIdG%L&n}`c-#4@i+T>xumH4cuGxb*#l)V|49+K(@
zxlPh0AvKYhNY;m4>vwowk|(eL`21%6y1;u@GFhoC<D8Q#U&()G`nl88G~NZ*oXQ*b
zT7Wvi;fS-v^7-MAa*KgtJ#Cn|1P?v6kLyx}mLFNL*RmY@T|rVBHr8(8mtWk|7d5iZ
z3)o>ZP@LV_3L4G6q3iKrZZzyZ_sBX1Sacx`dydX2FdJzOUj!izbjU8ChvA#rGNnl0
z%+JD!Dl&I18=IC#PuaK*WfZTe+K_c%5H=Q>eJiVi4Tt@xc4V@Au(#QIS&|@u|CS*)
zWJkN+;JW~jx))Bd1mTQPIzt^y2RB+KS$w<EmHvdEorbmeYCBkQX=nv0W$RC;ZIqMJ
zKxlNFpjmk|03*I+_bt7^q5?k7D0w=DFiu3G!+xTiEPD;476xx7+grvv-Jp(Nmf${p
zBP@cCdqF0t6OBjxQHo71;WoUeq8(*%Q&2$1M2Y$W3y^rb-nOv+f}Z+iYt|wI)Cut-
z0^EqGZC<zQ+N=6h;!+M6F0#&R**NsDRWv2V*jO}slokl!DBgV)|CD*wU+3TL3;(hM
zjCAl?s#h6lkXs?D_+$2Izr~Js`6(78_V0bzkgJ`Wl_?IxeAQk@k^UPY+&(63Q>VQ@
zcZox#-wM7S^-D*lS*WYp6t;h1x+o3|<!o;^)$=5js#J|m?>G|SG(nh%LlEOuQ^g4_
zZi2B;p~e(7LcgJz+|#v7i2WFqv^vrNaPx7#uJX^i96FiTF0hbaO-)YF764MI8Qv<H
zV=dBvHNJ7pFiqeVlN|TS^eu#a8rqNk<Iuy*N`iov*x%=5!oZAK0bxIUWhY)+@Lym_
zS<fq5SJm2IB54R{f9Hgsl1{XSxyVLs)rT*Wy_y4nBchS18t%o<e$=s*v&9%*ry_^{
zEt28kGn@Zt0oDLqw$ebAo%jC5A9m>m-p{rUn5=om02iGR_o~9Lm}@tfvVXf+E>#}M
z^NQ0~2w1nA`Sd_BN;$y7CSV+1Y_6820pvO_s|+iIDxIhLl;`IffgeLfau<d!AFAVr
zC+Hpq8o4|a40yFVG`5!>FL-QSximMIb@d$(ffr{D#J%B_2Cb+XygPjvg(d<}<wK?B
z&)uJ^Ag4B)$5cS6%3g<4)S8w)(7IAd*Dh*@Hw~#CFeE)guREm)AcXE}m7^hw*12y0
zVxWOl(LX;er8@^Yy0G_ZOd8ntYWOtf)8id!1NdWY(8YVdG9iR^l?ko0i#a8=d0wXq
ztZW`0JssVK57I2!I+RC1a_%cd{w13iB?r6lpjkp5b8HOs`xd9PTbvp?xAoyJYO{c}
zLJrcltH6-c5LDaFwRmHEyQJ-U_I$0wQCA?p<*wOXJI0;^G}6?jRuzNU_%J(!nr@|z
z;V|gy<1O*BOn<cB#=xj$SqXOoV(s1(&V_vKnKsGj;#&dO`d&F<Q%q10{_Ctfob?Iw
zT;o3c(Ez#DDDXmr(fH`K;k99YpV@<aT}4F(b%uw42YSFDN$dCKhaDnT$+|~nl*RCb
zRbMSn%Sv60nUIhc6#FdHNjH48D%Pms9{#w^Ml4B5p0C~9{$&>Qn?~Lwh+Aj0w{lHI
z_lVQ7J3amuzA`8*?Hp5Hi_O4<PNS?}WoDQ>IBW!88%bcu3QdZV(Ny)j9u~WEUa@+O
z40to*<e8pwCuixbaooWr8!vMha;k5D1lx^IdS>Lrf!NDty?A3*SL<+rHCp#qkZP!m
zq!8Zf!?!t4cOG9tCb<IyQ#=~>v`@KAyut+di@sucR(-w&B!*Ho$My5PkHpp|&s5@D
z_Z6x7j(L8i+UF3bITUFwd>$8=y}jPAST)Poa!4aXg7g~-Na}tY35KrKAErCvPiJ~m
z_%5zppP1h#*6+NA?F<YMe*Va!Dltgb^B4gXhKR*+m5Gvubl4+$kO66d1aA#MpRD(E
z81*IByD!Ad-cpW(o4&x-#emQno5<qEWrCYD>VN@!woY=w>>^IP+o)={XDhzdBbGy}
z4dAvTL09zb4nfrc_FH-2GgvRXI7pyzx3u<-Nhxmcm^MsW%6>ns_{;6NqhdGK`3i~1
zG9Uk0Pj^KMh~T+5iRQ_FuJiKls6dSK1u`n2?7Gj5?`Kb|izMBT0q8eb+MU!t!FFaq
zs#_k(ygmR>M}*7(8l;Hmnf+K5(dAmqHRpyYfEG#<eO1w%F&LWh;{$-$bKh7hdu;qv
z))9!zZ|ADt9*$(sq_yQfMFT8kFF<kAJ&nK4v1E2`3<lUF6sy$@x0+>|)Vxg>_29|H
z$2uc4ZNnwn#GB{g%6fHnpeM90Lp&)zRA28p_2(PA2G<qsU<FQysOj?#I6`p%S2H4A
z=Tgxu3PjPP6W%91)H`{TndG74*PX*W-B|3L`&yqKRf0#qyw_wFyz_jTTc$3j`D-Z6
z4zB&`rrtuoA|PX!PkB;iRx_{d1N*rV<JN#!W(_TbZL^{X>>GOf{+m!F=`@LXN;|Tk
zq|vM`0QzRg_u+-E8D}(z*}w^y{sGt3fNDft<d~T4u_9wHds=eau3R}W1=vDw1qNES
zN(#@UG<}RLn!n^T{zub=!E3Qlwn=*+$<MfI0RiBrxNeczy`QVUR<4rg+8ygEu3E4x
z2{a6t83ArIPpggx&N)$&r3+!8i|zTz6EyjGN%!>*yt7dmSoh!nTn_kNv82#pkQ_zc
zsMf5dSQ0fyrKP&AT~P)oz%@u(#!1har`@|K!-PL>A<BC8uNRlp_hXnVitaE)jJ_Mc
z82|z%a!}^<-xd(@CP3+VXMdM^)Y02XTqouU#6ic_eb1?Wg3%kG#J5Vq!h|m$2%^G=
z1st)##l289CVB4@z(+N@cP%X|<NSc67zgh}>wWp<u)H1oj@Q9Ni8qFl;|P!=*b=Gu
z;ZBz2l<a?P-e|+!mVAZL&i^{rED4N(te5=XK^&MBfG&KkJRe915MRR!NM~KU?9ykA
zx&x_wq0bB|t;Rxs6P9z1CNv*@$WP2zAhFyWzGoG|JV$??q74@HG6WI}a(!s!#H8Mg
zX19ABm{2$-^>KOpJ*S)bTp$vWnDLOIQKCY#HvB6gz*vz_@+3I5?;+Ovyio5eD#CxN
zlfyJ#Fpzy81WF^X@$|%%%rtn;t$gbt%eAyKa);bTMgYs49un{`P+~Q^NL3$?2PfF-
z9x0E~%$L^NeyO<KwVS`MZ?SywHtU(ZS|#dtKOf0>;y*2%fFV*$eWN0iB?Ia=Qa!gh
zku!O7M?R3Pc3J(@=gZZVM_0k;lk=ZLh(z%d*v{MD#q_q>aqpkf*XJL~-ZImCp-cAt
z=R>9cx8VPC4PC%L|KA6{<-fQ6-w$M?fLXt%dOn)}J!H%No`wIvSI_u;nV<hpo$2Qq
zV(S;lXq)~nUHtEzBko;bAq&+10<2K~pCjU1F)(ZY_wjnLj3dVR8X47G#{csfna@2J
zE<;&q;e^t_z=Qy4Y53e+m-assSf2S$hmrH^t@wvg{OmMsXG2*mcDm0xHGb+y^KO!T
z_V>LcKmX^=k&#iF{%6XacN2Oi2TzXxFGxVXoDu_+hnO#n@H=vM*rz5(%DpWMw@&{)
ze^B6GD+hXEx_Zx%Nva!PU`Z%VKgvr@4dq;1_NVu!4V9VI&il*M%hY7xG`)01`=7UQ
zH1t1RO;Y$G$->U>+VY~YT}|yr1v7KL8D;@<UqMVKUh(?hx6@M~o%_w-M|o5V1mvSS
zN<2U9)xuAJZK25N{}RI5yM-Tq4o_l3kC`z2Q)+q|NCf{qFNQh~=d66XeaZ4!J-N6r
zerM|S%=ebk4(l|6b$h1*P`~N-zgn{Y`<*$b)GI(foaMWz!kLsEYHGJD3*Amj-g82x
zb%hiD=>Smi@xN0^7SAxv)TI7iqeRo<Ab{H3=*e^w1`@fSDaiCs6D?T)qyM{vHn5lN
zL0KN`Jc!x8{_~EM&gM6xv3B<3-$$(HIK@A&YWlDLJ_A|rEnA>Iz7X+ywcB2Bwc9G2
z4huru-RrLBuOET8xK&NL!)V&Y9$mco`$>GpNB*5?*=W<%3F@!6Dg}KwomRBYZhh(i
zS{dKjNn%9Yxvv_>ikPg&en^O&gSX!DT)cmg87w_ypYr$A{m&voCZhZ+k`buAKF!c{
z6P*9uZTDI8WGgMN8r?X;^WEak41Q7J&s0eh%75lxPwBs7?~6_Z-dq1j6<OHHwI-!2
zw_|M?-_+ODz!Gl$(>`cW<A01QaqFzE@|n2}`rICqk2V92LlefYSf*B8w*R3@MlVI&
zYs+-d^z#Es9?XvFmKAy)O%MAo=5TKm()>N_4M2hO_itoAD=+#=^rOw%{l{}W#CDh%
z2bcpN%gB^jo2Xe4B5vRKDpB{xoaef)k(mOog2x1=qJrh`Gvn$0YYQTGlbkO&O-;TU
zXpbx#2n)B_y4F$Nn)P^rY}5kC8*Vd8=^QUHfFS>vMeqCnKFn7Klc$f@nHxRB)l4~W
z-RijNnG+Xo!@{JTE%{a^!5aA9JkIHs`JZ#m{(Ek&fE3EEFPZ`M$o+|@u>~6QIcaAH
zxp01!9JVm8_mJ!a{AKzK3Hqi;*#4S(?dlcKQCsmJAk$Je6}7{lzirLu%KCq;?n<Dd
zdG8QeSgFb5edgAzOxrrVe-5X3yPLGk_zzZcDs|m@*U|E9kngW%znp$kImmp4QA|5~
zETiV@_khYIcZ4QvwzBE>+*)P-caHBXaTzule2GKd1W&0@z-hL&S*JwV<nBm_iOuR}
z)5%iPbJbyiZ-V*3xzSD{Jl%SKm3Fwb6_vTV7?}J$-<A*lcfNrKSBa8vvf413It!((
z(=AD<bK7{CZNTZnRPoTqWaURp$=gfDbBi>O@#9dI!1o4rRe;_3T;3jkI4}o@1-}&l
zSB(c@{%ssKD02SjU1VW_Bbu#H-WzG_Z$_Bs7y(<aFa(}^21ruT+0F4;i}fD<$Nz_|
zw~njg`@V-SpdivE9TI{dARr+vAtl}2ARyg&QKTCLq$MP!yQMoXU6<}|F7*uR`=j62
z@40^<>YbT0=j^lg+H3EbZOj_~&IBqayJMvD^z+w4|6kB($M7Yd@ee2^i2XJ0a=Ep>
z++24HRaF9hHn!-2T(<b5eo)?~IkeDi{KOBF&A`S;b+D_jC;v$p9SN~yyX-aI3=s%K
zUk6lpNW=jW5g_GZwgA<1QA|Md1#k3!jMqd0m{^imwr=J&GW!WBlDl@xE&x(~*2_d2
z$`Re2Q9j?jX$UP)f0bTrD&4+I0KV@+1-Me4e~?DvkA^s1U9^|1OERl$yI1JqqdTkK
z8PEdL`0YOgGya(=lDnCWIrhqHS)Rt>5lQ-sL3VuPp}=WEKq-JCxVwuvg-%q+?9U^?
z9>+Np=@1LWhkmL2-B83^<uAMovR!AB`kPxNn(LF|jp?j_$>yd<BtaVb^6t(P4aaSV
z&2B0ib5{<)*5k8q+kTJsYviDU^nc8%YoLf&kk;HDB=#uH+Qq;+{?La8F?@pM8-PA-
z$gLQEP@C<q<@Bli0V#4|wC4lBR>?;f+L_~tB;5-pm78pEvwUZf@#i@4&wtVTeQgT6
zqOct>w@T)mwr;$&kh-kwF?P6Pb?kd#Wv?-rwiKuH9Q~{#M0kgE*4=|r(~MVtPb=QS
z>g4I~kt2`%4{<^L>}&=7zYKxZR7QGEvu^q;vEGwf>{(}L=)huQGVs2QcP3jQ6LZ=V
zuM3<HE+7~uV^x2;Q<VGt9TrivG>vw%)~_vjh}m+n8Hm|~mKV>(M>e;q4S^K)c)LCa
zY?{z9j0bhDN`TMSy@BEPqzESdV*+FVQTiGU?4nS3t`=ZeL-+!<&+OMmM;j{gL!01C
z3;_^(l{K>={QFMUu>MXcik7+8TtbF5M9;z0V`ACxP`ZN*%1$;%JUcq^;|3l@(I3&~
z_fh}**@pwf=N1WC@L~j0T(4rEdvUXIl-n9@HD06XRRY}X|9+^SA9jVjCi;m5Bb%Jz
zadf^X4E315XhBN+Md)(t5vRBM*8fWh1BUO72$GxoE%Y!ll#hxfw9OjSKO_$zYTsmA
zTVHERO5@(%O5X43@Sgf>nV>Nsi%x=Z{(*!r+_Un2_Ln%lO*R5RVed?K4(&|a=9`yg
z1@tH=e-D&Y<nLa8SSo@4vdZxsEwzjPW##@MGbt%yDw!i>hG+7w9m(m8&<x>Kce==V
zN_Plafo`hg_woJSFP3H2p0ctvyy=)7VjO2DKgfv~>iB=3yFB=xwT_hT&)1Z4INv4V
zx1~(9Uur8pLHib88#Qff@*!t7Q|GeatYCjixj_Ea<HON~L!4I6{^x&eW-Id!#ajW8
z_>2H83uP`^_6!gQi@#ZVn!Mb^22hc?vGn<`Uo2#=-`sy?TG0z1enW6-D1G!X$UH_I
zBnGtlIE3Il{nCaKiZ;dyvvI?rBAzIU=PFSY|4k`CmH^^=yWAa<lS%>I#@<{tHI{CM
z`-M?!KP<NKG<NB2IKNXLQ3jz*+|u5DyJH)qx&IH+n=S1+q*Mu@H_hTS+YxdT3LG*F
z(8@EVK;DRq=2v8FhQfT`6~W2e(a+7Q_ob*#E!{j+2ZttKNdAT<@7llqORw@bmQ9oX
zjGS^F?3nh$3**zd+R<fu%R%(%5B6RCHGI7@?C?H-k`K)Tgt5`RCxo6vfKPt3bbhlb
zD-`gb-9`q#*J0@KU#|<&(|K<J68kREAd}op&hgdt@I3+qEZPy?`2N8kld$~u7o@WO
zfRrhybFxH(_B4+JU}(QSvt|Evb{aze_w*_64MAeQhyD!>KDzY{Jm<wr7p=|9I8y+L
zOX9Mn<gxX5SsC~rF)B~_Yat2LiIV-<nj_Y<f>Zzti+pWQe;WuKNh^bT-X9`>ij&<N
zCa^Pi_48dXKobHRY;it@rl!2Fmic{?Illj#iU#nopyV7+IFiGR$L{AcxFLL~X-7)>
zH&yr?dyNeORu7hP*|I>t*Wlatf84_hz>xwW+<Q%s*!2TF8idEal_5X0wP)-5{-dq`
zx^mD<FyOGMoZyQm5B-_RLKEp<0HuiwS&dP~bNLgo9QO-i>iK(sau|Pc=auyD$O3T8
z1~D=l$|jixnE<Pi>ix5y)&JSgAktihP57WtQsU?5O0ypNzt8qE`>)}A%liO?k1r?S
zxd1$l@no`U{KHbe_Wr#yIc<*2D4-8u+I_U<%gRhf%6vWxjC4|Y$?wHk;ij36{2yzf
z_&ZL}5c2espTlKYRSt^)vVQ9Phc4ZJ`~AN(QksLk=zi82h8x1kJz2`^P0vqSgO)L_
ztoOszdCx8itChzK!z@rig2y+*2lJUy-)waWr|$zoARHO|YuDef?jv74d4Q||9<6T#
zpjHEYa4P2)t$mHB#9{`Tor6z)dt^}mKmIhU`?o&@R&HbF6+xN9{YGni_bj@5hj=S6
zxsTy9Cfn`HXhLEQT`OPKGT%R*<oNG}LEf$z`JQ!K0Cy7V3uuni=A){kFVf&jxA{Pa
zCvd3x!Q%G}y+{5#L$4(61LJ#}zb7ko9lbI~%5R1PTI{F@mo*VULmaw<d3O5U@F!xe
z?uK!+7dkNv_xGEV_>Z+d`=#*BrE~1%yRYLT5vyAt1oLbY$gAo8cfl|Je4FaGZ6^bs
z`-BhJjPs6e=lnye)tNw`!2KWN$^Uy--w6N0#L&H0sWpMy{*RjUpc;l){{N8^ab8wo
z5Rzab0+8MSLML`v1P+(_FgV`^cUl<&uRmav_}2yj@cI+-VJtjKb6*b(TvkSUdbXF`
zg?`?SPPe1ez>!sP-|7ASLoM{z>6AtPih93@+|!^s>-E5$w!;RW29{+^ccgQ&)Giew
z*-60ua%%o}u=wADK@O#b56<W2s_d(+*MV28M|Zzt+$yvn(cfP<%qHK+w<i64!S`tY
ztQh^T>DG>dBVpSjaVDs>2D>etxvBmMC^-IZJwZOeA%cF+K3d30Z2mL!K3drSq_`)q
z?$wX&q5r4zma7mjbxZ#*8IZXD=O=*7fnvlO5BP{*AMdQcS9}RI{uRvjXWzT)`w&qn
zZ~ZVZQ|dol_#5{Y>t!k&Zy^qS1k&de!0BGH3@=19vzdk^^OLsZDpD;k_cNmWz5>xd
zyTr{nnhE5c2|z`BN$37$l)lgJ@AP5(y|4GcdwjgxXs~`p5)#*%E(fIjcklmv=PT*`
zy<30*zi=NnR_+4<)cr?8_}|s3cS5oWK_m41Z7)Bs{~`ZVz0ZJ-XuD_&FRw9w9qE1f
z4;AUB@J67p1>}RSovWzD|E^E`&ltl#h|>UFhJ~tfJnMSuU4T^2)#fh18pCdWoefl6
z{MR@^-k0h5mhR?<7p?R24L7>=H)jBQda?%uB=?`Jzu)|(e!t{`SMXebc0djWI9qik
zxIb$wzct>F*g5nIFQ3bGaAMY8>T9PR2QEi*qx$z2%gbU{?oLF1Zu0+7Q}uhXuF&#V
z9K4iDz{}4v8%NAFiI3<UiW@FSf{|8KWS#fM!IX=waz7=s?nvxu*!~uwMl$K1g~{0O
z!C}h&olxm}!VLHGQx)Cf4F=gDF(6<)@=QdO(!A#vI-_${$8Bx%^DVvrx7lYf`_pLB
z@~?R24f}0g6+n==bmG!C(kxAb<J2B1NFo3;n5}NvOP0mQctZkp41XH%`|AJ4K|ed}
zLIqac5@jJ^3<w|`^miV$L2yclH|8(fmpYc~$jB{(%zhsy$NA5Oqa;>1-i!s_NI4W>
zgupdkxEH2A4Qy$nsa!VWwbLNNLarm4Xz$-99`WGsfuNztD71xu2~T%$_oFsDMep{o
z`;n&q#~%9d{@O#%i(jz#o(a+B^WK<+RBR{-alR+Qsz1Z;eJwci&myO{QLzCl1Gw%C
zQ|ay>I!}|haRd}qwU+HH4adg*=p0-BNB~*D%K`4;P&M719GEJT^?|*pRlvjhz0}d1
zDKVsCA?D}95;D^SdAp&3uQZ>2w!gD^T4l6a#EkM3)pO^3ZPiOcE*rd&taI10q6pBA
z_i%r2^VL!^;P1WxR$^=TSeO8MdNBz^QBmW-%C>fiqgk?DHR7_}VuJL}B#WfNMWoVA
zHEl!P6t1)%t$i5(xg&6*-dLVhq+@+ko+^t5`le%lhVhs}{_oVhPq?QCLx4Or1c)Cd
z4-5#Vs`*L4Zq-F~e&bZc{v6UP?EO2M(6W4oXrMDEwze9cwt{l`^{jwam+t&5N{Km?
znce(_#kb^%3FVui<6zhAqnQ%o8k0k1QQ{v&8ACsC5s@ruk8<Haqv9@vav%S3Ry2u#
zcmnLwiAGQuAfmZDhjyaklK`SC+sI7a`6doeU48}G{oi=FsyRN}m>XqYPqxmp^GQD^
z0;%QrHQf|Eo=6qz9lUAv-QdpO@cI(?@^XM@I7;yjIt7%V-5oA;_2}xVT=}Y1>^nE>
z%#PX6y)_#bJjvCLPr9D5zq5X{2qvA1(<$FbXjfgGe&yt;U1M=fU!Q&lgVk6((LEhs
zt-9OZUQhRx|Iz`Rsq$zdF8RtIW%wvZx{Ifea_R#|4Dg6AV-VTegsxOb%L_#UlJexP
zY($V8^P;if>wjiB_SY-}_jyExDI5?_bT8ahBex1YT8?>3CIbk9bs9JzZ@Y9Uh=my;
zdoU{&Wep}k7FT%oug}4k#=H0>6BW>hF8Us=(K2x+Xe;TQM~TKj-Eed?KKN3_>+xId
ziHR}d1}veAa3rz$+Uqmm$Q5!?eb~1iOZvEKEZ`B@H&1~#+1xTW^_i4uZ#tA;7xY?-
z^-@)|(e<(VHout(^e_j39I<o|mHy%|pr;3O-0zut@ub2wvm^Yn@6&&^0KZI7ydX6A
zD+m?zwIPp<py>M4SYAQ+i|4^zne$zVe3EP9e$(y|K`{l%I2J35UJ`;&RB}j5V8Fxp
z2l4qoguXs~Q+3$9P)wLHxyJhFN51H=%N<EL<r+o^f$r?uf)Q!wz*End&!m3uGc3rD
ze7bN}xrkUrUZ@ixkjcr^lL#68+;)NrkU7`X0-rpv^VLlho};S3DX*-mvg$_R`sEil
zJ$g2NTU}$D%t~!lyeCl)thg<Eh~&bXns}@rL66s5^mKBKeb0`!)7@RXR!ZkIv5{q@
zH$M`+rI!=t(|PGUpX~0eJnPrFj1|dqz84hWmUqOXgtm-xp1~;^RR*_QIPk{eId_Dp
zWe!PvY_|@ucWpi$u!7KSZvJ@J-+-5k)#HQ&0>z^)U{0N2>~7bgtIantz*4k1e3#!}
zcizkbGYFb2NhR3#a8tH1r^kJ~IfXQ=#L)>{9y|v*ozV`pWpJ1TSNWm{MsX&OXFTYj
z-9yUGGN@QObP-D1Bxg{v3z8Q9nEU#QOykJYW^16KE_Ra`o8Lu=sOzH5v%wICpiPH6
z!p8MUQPFvC?q<tekb$0c948+`hMSk)7BF(EroBzp!lkX&^u81y4j5`i4$aosi-A6)
zA^>`;MO_HJxvFpfO+$Gr0bUqXT1Z#cUod+oySrVWQ(S90(bHcMYrL!my~RkxRs*<W
zA9^E(OzyV_kQKTLk%q1M$_={D@98!5J(e2t6ZBWkvLIPBFvFNe{jv$6w202-Ve))b
z^xmi}WS0T=4Rbs`29b6fwGn7o&q72_t3tzeQBim`AN-iVP=VDuMv0wP)|s4S&6?`j
z6Iu}sV6s>6T6#fTt*3tybib!RQ!uCa7DPdA%;ha?w0bMctRC_vq!I4xB&uq8>n_09
z1W9re2#Q+JsYy!og@K7#!>3@8vEnA(rop4ea+Z}aauc2P#)H7&RQV5=Q*MJJ!ewX4
zQ>*u>7e$eMh0^52Nloa5Gs}&_ZfX5qV5pA1Da6Aw6BW%A=gLf1K0z~1j}@bB|238)
z;)(w;#QT4UD9>$yA$WTzCHZow?%kb?goj_>93u_9eh%wINm;K;QfX+si3*<q#9VK+
zlfc8+qOB5m^k|~lQsftMiii8Nbqticy5x~IIb!f1(Wk!82OK8nRSliY#v0{~F(_f2
zY~Ykk%MH&=sm<$W`gShM(}ZqwZ>HIsi_M5runM=kiWKF<<_{tHU$tDk37CJhd0V`~
z>=<RMFfz9HYrrdTh4(^oWH@(2bn`c}w5<NQq1n`>SPw^2;q&waY-tE@|E+0M{#C|6
zLG)rmMZ><8cp(RfaE+$~vL&K4TVKc7&gi(mQ~sg*SVd7-0ngdVQgzI{-I|kD70<ix
z>N+`YFg#R@2_o6!CEsKz_Do~HkWYYR?!CJG9SXWC@s|4?^BpgTKKJ(+pq*bGejeMf
zCFuD{>lg&TiOczFt-WOPVSaB<9J9MYZ{Dhu0(5ehnDZ_0&JAnTPAeT$W(}Z$ButF#
zj#=u-v$0tWuod5NZr<@!>aLlj@C00@l(qAX^Al(^{W*Jm?2_L#^IR=E;c0I~fXcOA
zYvSnu!%>sNt6H*xi)N(xALq};ghA?&&P^znwa$L&+w#i0Uir`C$>3d?s8c>T!+}id
zK0L4q=<1Qa5<s)$w_kB?dOhSQBQh>fORq94-6#eA(gbc~*n8i88VbIqg+AN6-%$oO
z{46>oolF4%ozbJ5nuZ3xDoUARv^nTGtFbI3HSzse4ZdU8rySAUL<FhTMN4GqalyNO
zBdZ!RpQk4}k_(AbtI3lRMuLJ~faTx$J@&Lw=(`mA>DoQZ`yTRikEcLOD@E|VR9>~Y
z=h;pLs<@)ay{6?&(sI8?OxRMLf0~c`1U&FLN93+FZwFSd-(|73jP%?<Ti_9*&q#@*
zU}7E1Vi!g#XU7#N;<(n?sf`y)+R4FERy#%1(^PdQ*2rBfPG$D%21y{!%xh%0OY@D@
zy+zx6{7*QB`;N}H_T34$u&om(iloLxE$CJ9eyDoWD}Ih^o$~{MxPnda(}Fp~2>L2n
zFv|hV_&DHeupuXNSn};b?~y4vbh6xBKaZ%+c6n!CpvH#9{Ib0S1z(`%p!SQvRavOC
z6t>0Q*`|Oekv5xm(t43~6W+G_i1^hZx261D;};Q)HY;%vYOc$<!&;~3UI#JNEF?u_
zeEirMozc*_JPUzBz6;5s;1?L6sTg2U+xHF}l(lDgr~^~Kdh@Pt6qbR!9q4-uW*9hk
z&pnAAYr_oP-FHnAKl(6mUh+3MY=(z!8^)?ZCvQ8YQd#fYQ@5=f8PvOBI)8F=uo5yj
zIS3eudaWTJ8So{xPu>hGb{J7TscK(jb;%Brnwc|bZjlpec#SiI8qm|e+P#z8t#<(6
z@Rs$;Oycv|3v=E4M#8J25l)4%ph5szW~d(;8U-4A1Ri5Zg^`y(U)f4FaV{&KVC_C{
zevSukJe_AA=*F(;3HYEhts8~}zzsdV$C|I9+*KWt(4mL4Rp<`Spi$Z|zqAr}1kyn1
z%>Y6wc#4VSNh?mXHzFdRqGxzysORL)2g20KUW7<HTJG=2d^<}fP5SK3dn%(Ha)PF9
zV^Cdrm3(r#pjVGBLdZQy%H}DIPeNR{ODb8d3SwQ&v;M}+?b*7p=C6C^+EkqoRB43)
zKHTHaKM4(3<KMJsaNzkk?ccQF&NrWnEV#E>SccM$i50m}P<tlQNF%!8I1~{O@XpJ*
zzHwsB@lI;DWCWn3O*}fzqx<RZpweehFZ`p|8#dK>$^7ec2#>{o<2s&c9Rek=NYJL|
zR%tPb9;tpFNBux@@$R$%1%GrzjNJ3EGm+ESY<6*Ik<rrUFj!exxjQJ^d9*OUFsC>-
z(|n?YN+%e;ruD$fLPNHK>(L3>={^*KV^bo;6i^{{^j@}~fb<CLw6`2y^Z*{tQ(akk
zEC^4wDBM<Xt@`u1P^FWu4jG4FLXK;DhkB>EiUP!5lXVJhJ-S;;YHjiX_eObodbgfj
zI~~_^k_Nf{&hhp4%1lf^5r5WvIT@q=i;TFwgK_&VyT<sG=EbW8C4)JYcwCgVfijP=
z>NJC~1&$F)+|rALBBxNYuaxL(6I0}l8uIdH-AMo^dFwp5Xltxkc01|tQf%-p+zUTT
z5Hrz_#7m9ZA7ffQXS0TJYXzr@+MUriBpV%K*Bky61nPQjdk2<wAq1gl-C=k8r*_ud
z#xT2^BTh6$>99z+x=w=3){oJ-l9<S5-Nw)N{HU!a*MU-|*-1guxsEljxKp-6(k|J#
zFnfkiGdy+CHPGDM1_^0tX{-G&Rf*cl^X=-7x@3i!ox@*WU5AaGp(p>4ZJpRW=P9bz
zTGl%+hA2Pt+nz7l12-EZfLPbhHa|khxseO1J!(%sl3M0Jc?Ow0!1uI_BaNxQFiy3;
zP06CMGH<dS*IxzM1vWL#ZE}{@X58L-2Kr_N0Uc_bnHt$10!9M8lC*AUmtwAfCOByH
z7Sv{pG!43zfmoUMz3{n9=A<7x*pymi7abN*d#NBDVW;7NA{<|1{#@+okxo3b4&u>b
zi33&_`N1@zduCqNsh*VM(~vhrA9(8dt>5Af$YconV^p9azR*E_0RK@?&!$)`&)kXh
zgT5SzcX@Ewyxb`Eskzm_9us4H#?njb4i7VAF^o28r=ex!cc*GH-kc^jIa<gpMm8Ha
z-Tk3YR1FA8wzzDCEUUy~RfzC8_^MWaf_b($yoFVNn8Cv2q(fw9nHl2^r=4N*EUX>5
zB9IIQz>F3~NAD>qnOKXq@(WefPWu+e>7u=r)~F^V%QRLKo8pO){x20qQ6BdwFP^{B
z+G|H0g)LaP3{z!VfT<2F!|=Vel%2<U)zlc6&IB_25U%P?X?;54e4^*W+wR^ONIP1w
zV@Qpc8-&lR%uNk7bam_{p9Tikhi|~(R{UWy!8pU3(&ikZ=L0gwf$y_rv6IRdc=R2~
zrssG($+qTS6HY}h1iDRGIzv_o=a{l=KwT`J4Xw+OQYyTaYeO2Rc<vi|`v)<|trj?~
z%wV0Hj_FaS#bj<;oc+*rc%Jg8Qr%JOnY0uSeuo1O5neHyAIINbkxSP5gGM$wzJHtu
zrTthcrkU#D(eDvcYU_S65_ikYtcBX-a@hM3B2{UXl4@g^<sAaJ4)8r6fIKA@m?dlE
z8gg=7OCR!2P4cvx@ne^<Xxh435lZJnh+r=wYcPf}DPAw6b8aj5_}zWasG@A#dSE?v
z9%M}SR*PA*)8l<GJ-(}MVqyQ?*6_aAr$}gb<F=PFkq%qI*o>tNbf~;X6b*8E0@H;t
zv8GVHS4Q?uy5|V$ZNBf@+f)2D-7@PM*<P7(CrWeQ1tQbo0^)+;_$r`8^P`|v?$#(!
zoWmJBGGI!qIZY_|sW+I?9FDRQ@E~I5!!s9RTI$Y(s1xUW@?JTyf*dut-dq*fmrY>b
z=ZPR*<r&}=Cq)&c6Q1n&EsTEYSqM<(XI6Ww#FdSQH)XvJI3N1fmLnQZcQ@V?xh~lH
zqr|un<-n(#t*?v+UUkmv^9Pr#1F?)Uc~sz{+#51Cf!$@c52b8LKdkaP*IGY)k@9**
z>&}QkGB6S2LP+usc#S1RFg7}KFH~F-WaGQ7+l2~n!!GUQbIP~y7WSMR=YzbKN7k5^
zIrjPb3$`|;ifynCux(N(N!$;-CSt6@;nAiyA|CT_hK6a}kzTfKJU{!KDP)ral4if*
z?X$~>zjw)3Z^O4W9bQ;Lj+{d-FNe%3tjvie63t^r4ixSd4!UMqgIsr$!5nKA;#0wB
ziv%cGkz2y2V@H=fj_aK#;g=Ve&PKbHxMvG`=U;p616~&!wj1{Qhs0A@<YN>R#17V3
z3=Y;;S?mYbM~E=PW%iCK>Jl-v^VVN(;4E^XVnzB|s936%W%f9)o%lnftQx%zx7KK5
zX~$JW1m8|&6v$es8TK80ZL;ad(Y(pOn{``**sOH=3d~Iwn*?&PT>H*TfLt@(sgK+h
zR^bd)e9!wmDI4Y-p@fakJSP}O7HUcD&HZwpVjVVC%8FbwR}e3=eWDT*Jrn)nzFF5r
zCU7AsLLuI8E{*ltt$tu=bwd)g65?^C-IcVYzf~OY>V!jTG+Z?B?g#oxegy3dAIMuh
z92B;qJcYJML`vG=Ww)mD?qiDZt)t(_7-P0fh0(Y0+)m6kEp`VNlM{Vq<$ef{`2=G@
z&Jf0HN0Bo$M5TDa!IakC$;i#9CUP6xtRL@YmD#=+^e>GgyzrWX7!&q?i>>G6XQE_O
zR^Hq;ePO=bHs=7ui9s>VY3uj_V-7#m6iZJTGbSh~n{#tpzFH&;B6fCFafaGr+lpGJ
z1FQxJP2^cftTd};3z~DED~|YYR!(kxeD29vt4al(SJGCf$gOMyV)3%rhHkx~<--7d
z&Gt|~tbANVyt_0c*S6Jz)}`WiI`(+4qHVfFB~9$Op%&sP(|PM+Oqz`_wbfP*v9M{e
zVW1?m>8PU9{N!93347b}$l#PdSR4N6Q7Q#)NDI$b2K=~G@|q8q0w<mH<Y5)#zz_{@
zP@rp9tm%03SPxea3c*j~Y;G9*SN4%e#Nz_wP!(ZsC)xV<-x}FGXdn-~Bb<;P6Xk_a
zLO30yZ$pzcP19d}x;u;v@TuGycnfvH6inF&;ldJ?)&RIn)||CiU&W(m4rP!oWNeKL
zN3{OX_GE3ptfhBbIs3f&V0ApfLU$;EmppcSBUYu8W)nf{iw>Pzt$8(KZ6++whGbQ(
zB5bV5lJvC82;-UxG-jP059ZnJg-IUD+&(n82qI2+wlq9rSyAAH!2Lj#(q?YH8(qgL
zvrLgrqpO!gjf=E4tOYaT@cniNn#GIW{=3a0vb!qJMYRZwJGTlmiR=1s;<$Os#Ky<U
zjkP;#?JQSIAiIv?0Ba-pGy+OW1?UiYWHZ)}{g{A06eOJnyu*gv$<n7-So?`c*QY?>
z@kVi8%qYVDEXYyztU<x`<l<drKn*tNH5EY4@-ReI45#$5BK@k)-kR97T>*IaDiGQa
zxLV2wHks<8z^6CTa2sRaj2IUyQgE>Qsr8$X^LzLsIz2~|5hy1k%n$3@ntpl{L9%l^
zp<<8!^4mA#?XbFzm+<cQSjF{eHTkpke4}!sjl5*-m&vw^YGN^E_x1TX02C912)`T#
zUWLCXaHxw6)f9Ggj~(>q>`-p&;VwZAk?S1;0|7JfW8aXXyrG~CUx1n$n{<AvIZlsD
zWaUzV3-2_{8$XiH6~U;0G&S<4*K+gQLU+~JPT%74UDDLbFK52##YN=9Z0r6k=uaM7
zo)T=wC`!4f)$<_g{R@j?Ud!5qZbA#3)#+#by5__{i#xf|e2hMZ3f`AH6BTLCa!p($
z^Q0*PjB=#}$rf=wf0B<tDvQ2LOuXRdGHx0_ckUbv*E;jG2?85TF?3NqrhA}uC-m0#
z21OA4*p<qUQ}equ-QX(=1U6*+<_OugQ_%Wdzy}7WW)tZU%hx%ui~)DpV~vt0Nk;U%
z5dyW&Z<9z20E;x6E^u@pltJG!m5n!XXyeB?bt~UxRjk%&t5t2p;pv_S9nBB*`aNIL
zQ1=mq`IT1cw_nC*h8zY9mvZ1i8wK{q5m+;g`hx-&8902i(QbO%`fv(fyoLD%1nj73
zAI-nB>PqI1{A!IU=rqfJbNcY6yDa>Al+sf;{!VefHPsJ?3F!1=>wcW)Yc&H)oiOop
z<h?2**d(T;EH&x|2*7IpW_|d~-IbqO`zL~%gLSAs0*{J7<M5TIE-@U5px42Kn1^f-
zUeHqosNAfVmTcpr`J-SL?TS}K!$awuyFG6170Zr(M@2z>4t-w;N|{X<ix@{x)1F4g
zfV%K5&+FvqN)xJf5<a*{ZBJX#kss4O`$pm0hSP)zqDTfxg#&`NrnQxQAm_%lc9u$c
zC<fK|!kG308>Zy4(3|`$_A}2-k<9ek)yYDJEggMdHo7N*ACLzjGG6eN(Gm*)>^-z!
zSC9RnrNz(B?{$1ufGZ;*dD-T|=k%tXZlhOQl;G!6JSly`pfsIQtzO*FG(5a`L_1Ma
zBNyHj;K<xgSPGi25dESMdu9~tJ9_*$u&1ZULvyw9ANLPaM7;*YIy69$<U^H_Zg*oo
z^D3?HSyN6loH@C)Blxy_)*6-9_bB6oF>1R3_v8m9CMMY*hIzMYiBgO}bYaqx-LEy1
zS(Y@t0BE-RT;ZAT)^x2v5uSiqnE(o<AX6|2V|{aBj&5|po&t-4aIk<7WIY*qc4&TX
zXoYk`+#(K?snli072c(p@IiJ>WOVkTa=&drPEOlg+O0*{SqbK+Qtw<vupjl90z->E
z%u?;_llHK%u|obJZE7`r)U|VETAb+!OOMRBD)|msSkuTlm@w5<;Rt1hzjxX%+5%^f
z%p0}Z)AL!*(lD*1BGWkUNsd;E%y#v~jU`qb9Em_k5@d$n-8|Scz2iEy(()>IInO%v
z7LpAyBgL?`(yXw#KylVah~R?~H#OpRqL*Pta<a!Gw?FdQFMZl{$dFV?r{mH|Dcu-+
z*l!MHxM`SUo=XVRrBOuavUdecqv|klEf61mX;|bU`B9njS|Jr2$e_4df@dVnU=Ff4
z74gwxqHH?GNO^Z!8OMuv(#eG`(eq>XsRd*iBR`fgRJF3Dj;fxjNQ+q7FdJkP$1lsH
zQZEX`_v|yYZlV>1wmm}^Z8v~q>EXz3v(QA!2fS4Nwl-S|kFD@%JB8`=d~EC|AkSVA
ziqV$~wb&w_(L|?yo;{g;;?H)SEE_Vx`Nb^{4=;D}VL^_#D8~lzM~TsHMX?V+<Oi5W
z$0`u;cQ^V0xHn(ptSZ7w79KQGoSFRrh#XW`MMW6s>BnDr;sY`<Ao_NMdlM)iD-pVH
zP0CqI7LA`YE~zZVwi@!RlE;1?6qVViA<Gi5CC7@wHs6jPNag0Uq>ev9qos*{ZMg=(
znKKkh8m`-w_jTV#5%9gVQ?)mw%Zk1Fn(;f>5n>k)()4gwnv*6IB`=_$m-P48^%5|$
zXZJ>ie${m$auw;hh?cM@F*GMFBPy~YhK1+Q!9-xoEadu3reYpW_`2im49hggE@JS4
zT4zPZzRlIh6uzGwkk@JscN;>j_vb8CUl^SEmpv!cIGk}egaVbd1F-9`MFicWYR^Ud
z9@Ol#5+4apuVQz#J5NW^R*vla8%KjR+ivKdcr7ugbkqONQyHvPNjIREO`|-x70!%A
z^59IQ)_&XIVsXnj=IoNV5EVRp4s$5)1bE&HS?zYv2=fw8$P?O358XD-(zw}*m0iB#
z90iCHUJA^meHUevv%Y3O<`$voguFNz=Ebm-NGeY%$kfvW^(C60*3=^5Pn1xceWHDi
zMMskMN93JbYr0(P`wqE-&HJJpjj8H>|1q9=y`0@B{XCyhls>iqGqCpI<S&(il2W-Y
ztbevxLRSOgmiJx5$e{~dy1(2k=<-4R7n5tS-J{G>sl|R&L<8s<3?x6+f7Zs1TlnPp
zcS9e(i1&KdXm#>z&bFAQhA~>tG+jk;o3Fu3`|+6)iBz0&t`DotZf!tXh5VhDqy5X9
z>hfv5iZQ$nN4o_&X{)N*-H^r@b3hAG3P&g9ALX2LxR*S#F)CdcMo!_#Nl9USu;jP<
z+<i_G^R}v~-i-2Wg+7Xk^e+S%?;K_0j4J@P$LG~z!U~?v`E=wPd^>?*1xYDj=c^Kh
zz8KaXiQ_@Vjp$-j)L`xWN}Y~}w+Z*S0~Ni&j$IV0Ke!9QOy|kySzplpLfXOoRVC2N
z0(n$`-&SC+dF?*k5s~}#a6k&iPx?#vRFHU_C8<Od@R`b_kJBnBy@|*1I%gxcTkoaS
z%g~Y;O6<qJ*gTbnqyn&5+1u8rtUB;^4H+P#yFX=c#;xjJ5MNMSs6<rF#?R22&vfk>
z`=X(^xB}t)qI)?FhB{EtC{I-_<=G(CK3Ku3O8s_swZCKCfoLTC=60W=gmZQ4ZaEj0
z>b3YBu=z5iiKgn?_EY9Uu<cFow(eJ&5UN0SS#Z)wr1SO_o!2Wri91Bg?Vo*okE>En
z8|t3Afl3E?Fl{G<n>@dp@|Z|kxG&=(M)*+5iX6hf0Dd+)NE=QP+LDzfM?IxXH*2qh
z8VEg%$e#68f00UM2%{Oc7Qk0e6oKqJ>mUn^_9S4u)+yQ_=rgo@gAKibTY+~@BS49Z
z4SXl2eSM`1j$l95e&x9A51t@rPHh@$Pq5tr9*b?1JJ}iJ_#vU*0Ag(+kWb^oWKLg#
z>n|3{;naEO7k^&i8P=OcId||qz}l%&o)lC{!wkkZ-*8NU-woD%K&Q4I#`+EyRQf^_
zjs%)y2D^efB=f|@_JE^eHKJd+?MhxO1{pK?Jwzj&hjI`EsNp_BMVv+yC^X~4He7&i
z3@r+;#fSiM3dIn|q@pq;VMmQ8*UGFw!t0R%>&-h4?R<8GkIj{QKEPfeRge97C0@=p
zo}YUP@52|klN<lkmygzeyj`=V#eL<APu<pr;dF>6s`X2D-r-V_Z<fWhXD_Se?-qsm
za(N`9X9^_1Mt<n~@wdG1is)o{wyUl%s&9C-*MF$@aU=IZDwDhmN<kr;2(Vh1REKgS
zKmg-LCs=J$m?`9+Qgqw!u*|cj2!hSEM4?336>pS+W8Wa2U3J}tik<gY$38!zs*9TS
z>Ha|HC@xT1uFxW8LiY+MBX@)Wclihi&JQ<@UaI{dECY)&&?|_2VhQ;<HdAlEzX!x3
zzd5PWng8KfbvJ3yR!q>YIK@dbD`{&vHO+JE=Yt{Qx`kp#N}8ME9kvyLvtAo^uF>~}
z3xX*NusB7nJIT{pt-BH1n_eqRARy0GSub>n78kQa;}{&{-s=~!d1pnsF&e~oYdR`u
zru$3Ou|Ux%xlvmiDDb-ob>^g=-~ml7Sk=b93Q&V;QTcB1)P+O~><kDeWVr?a{4kNF
zujS)bJF5o`J7bd~XO;ldwlf>es+6!vSCW5AI6pbJi|s{^h)DNHDLeRF=)MmaXvzq+
z-9~Rw808d%K4Ilm6rTHBx0d4iz<;Kdv++c?&W1&6cdJxvLg45EtX}$3hbX{q?7Z7|
z(;x8bc9rPfvxj+gOFH2sva(`*I*xIu2Fk6&%h7)H^vR}RP}jHXTkS3h%RTJ3z%(e(
z14I!@;%Ywj@dTBxPdg*k^9NMr{ggtsh4$u55p=op#mmV`aD4j$Q*eo%Ji}g1)XCh(
zD6&sq-wEL8PB``7N5Smuwy@;V0TFcd;EMF}+UZif?UceTdH+7WnG^>0f~*;b8t1vD
zCdTm@l#`v;^*c(t>@${bT}@1G0&1vrYz6CE_qh4&CbDLv3VLmlm!h-&Z5`R;=Pr68
zaQd4GE%Kjq4kc^yn)6!c*ImYt0%OKtPq6ds#iUxt$!NLCEDwykX0<3k5&G<{BbAkK
z$Gpaw2P;9j$Hef=YZWvZ6^GJTBcr<4N;TEI9|^Cd2Q6*CqUfnNnau%I00@kN+55ks
zWGK8GHN-Q~4{F7X1hNT078p>GA^8dz2EJH1BFM|x&k`g*3=0dMAPpC2jzGHrDUtlE
zYF=ebHL|Dph1HBzIFmmbjZIGWKF$}rT$nfzZ3kAwlt}k5_S_fX;NspRo75kSGuWwJ
z`L7mWwfVGYzwwIc`)T|6f)9pCwW!e$Smd*Ib>bT`2M5BRKftZ?wPk*&=*Wn)#@pVG
zh5jn_6L%rf#{B!<I*}aChIYbfa7GezOCl?KVQqoQ0S4QJ`>4_hhTcw{DMZ5uPiV;~
z?g(37lt0{S6o%eewDhAl%<?_qpZiijS=>KX3nfZ<p~FBajO5-wOx$xZdEh?RdZ$pi
z7K)E8vGE`R&QA_iogbW*uUrJhir=rl-ffy6C%@|#J{Da$JNW*wlDVI1nkh`X`l8!+
z;`?=+z)ZDkpl6&_DjC;Jnw#3$=rC<8&pw(r2@BET!>%{1<!=sDB<rx#+}pSeJtOnO
zE_L}}HvyZ8b3SOuA}*jVZx6sc1976UE{XT%DeN9itbstT^viH0rU@x-jTV)m{9okn
zo_cymmqkiQQea|6<|SzJm^W-8iew1i3d3BAe+9-X2jyw4Y{I#D<{B8|Oa+g^1&B#W
zxm|N)L!A|<I<adl<Ct=v26UgW9Xf_yy(u6fOXlQW5ty1_BLK=!$Z@`wa<Xpi6OXFS
zI2Jo4HztIHp?hZSuQc<6Eh~ac0OiA{jvw9mIj}WM<wWxG%u~GFZfYI(Iw=v|rr}|l
zi8`JRqhN(5=LtQRN6BXP1GAYuM|A8V{2^Ml4jY;J(xzf&c6ICz$>TQ%Hw(vowhXzs
zxv9fADrPV)a=>?e8?6DJtT%6I7WP25t>m{0nQ9c{bloD~>xCwPydotlzZK^txp)CU
zj(JPVN6O`V-%!UkrV+4lDJ6ZLfsJVZ`VdKKwRO;Zgw10=Cfvt(78<`QBjT}H?kPaA
z<W#X4r%LR{JYA9UbV|I-#X}e1Pe{JZEI@|ANe6EcS&I&Cl_RcsBq1E#&C<y}+C7w&
zh}b{%ecD#0Bcm;DBm&#!mBOrpNcK~hbYZXE`)c`sq-Y_SW=6IUJhcGBpJ?*_bVwyl
zDMmq?4sO-C%z^Nj=|%(AYTsnZ=V)sBmf51+_Avq-0hu)QCXRgOUNUKPtg?A~sC7;6
zGkCcRGG?sa928ZUoBwH=-1l(y9h<7Nv4E%KtS5Ho#Wz3m4oeoNbzZB>A|lB>>LI7Y
z4+P`TF;o?Wi}LdE4V}Z(ezZ{99iPg=2KX8?ES39E!-(S$TiCzc9d)NNN#tmPj--5_
zQ9ZwXYqI#=$&h$@#Eq8Wa(l*F9bD#q5Thf;4(dX<JgplV2n#5S_t%XF9EKmBab*pg
zbeYA;;>N~C#NC0{@{O;#Q+55Ziu875>}InBrGoq}pWO#l<%@QPiMOG~{^<Vxw932O
zPFvzqj=uT?N}{5Uy^--TNLSZb+t0q$ephaq8`7IueA>zI3^GVPPRgCT@1HPkD5inB
zd`O?yz}SJ7xW~zAFIJr}PQ|FRY_5P+5{*C(Ub7NVcFC9*r+7IDfmE6f#xiPFewb>y
zD+a&sl1^qX$POXSPj8vAle2z`iJ9L&WoNWD8Q}UwGA7r&qF<(l=SS3Z!Sk!G(Q>k3
zfh;_{u>pY<CjNyIsM)7ZW>97?;8vx<1~CbdAU=^+<r_C{KS?NNB%loePzm_k{5#QX
zZ4u@}7hd`hVP9h)Kg+^l9^H%;taE4dlHjo~U`eW*#CdE$$%C6BfT35rg>SJAJ#Bi<
zn{#!~9X7${fsN^hED_qtcgsElcZbdHt-||gJa?@3FbR4cu&oPt<;qm8yC9$YU*V6u
z*KF(jn1{1G=H+Hd!WY0v6$N*IVB>dzcE=WY_E3<?7EHM%BgTn40zC5ben5Dy2Wg+$
zA?D(E@BHLZcoe(!59v&Fpr~Q_yY;(7f<kwJct4Pl;N{4}J(X%$>+XX@dd-sw<{b!U
zmiBe&pfgp6g~0XYxvmTbr{>K?NB=R-4vP0|8lzpEp8)Bc%jw3UM62R%v)@EiPQlj`
zGzu{8<i5~pvNqNsaeJM6^foq7Az;xS|6%7H<XAiY1G6wZpL^vA7lTH{_SfDztFiL8
zMRJRC@{kNa>4$~d_`{vtXvxH=MUK?tFFWITMAT>FyDQrE;A7tjM60RlDw}I){&*W8
zI?LA^_{!?w@qVPDPA6%Jn(N%q;z55t<wJO%kCn}2f%V5qA5{FmDgAJ-&7;Nq`W#DJ
z;uJ<iSk5E)-F)KBcE9|b-G_;oO;Q#YtDNuCRv9!Enu--j`A%cc&Lr%P;lj^l80y6|
ziY()6M86^k++7atqFgUxfB|^TQ$n{yeE^?ZfO%FP&W`n*k+I)N>}7C{&10$<g<+st
zhT1}fJ?G$&;k0i)3<zX^^r{1Gs2P5HLKZfQfpWS|5Z<LcP-WywAXbu<i$pKoS}Jfh
z_shwCbp75mH%8C^^a>YgNVpCIDTqT}VO{D9cra7m*&WueTs}>R0<ULo>Po)8z$hl!
zQBZjYee$&c?wsqxf{Im&ja`Aj^T^4>o>l2cF@d!eDFIU->cQ9G%U)fD>Ew66U?^&j
zcA|--u<w+)p$M=hlcI3$Dh;UI&<ld3S?;0By$+Ln{Lq;9_M9(8$pl%2Vokp<7)_j^
z8*qU#XMxO!=@O^T581t|rgsDo&R==B>CMMi{NVM}!EnMs>*kUHpAiqbnllh?K#A<>
zN$8kyxZ8b-?b>&1GnBkA1qQ(<59<$^haE(h$y&`6cWAXML*r$c^ZbovZ_LM99GnS%
zey}~c5O4c@DjK%T)>w6I!uh|V3(mkrkRBsn6N;mg^xF>WgDb$ex;~*I+IZNnaQbpx
z50wDk4@pyN{zl4tTvLZat@ei`wYlI}pC0{oMOJ?ZX&oTUY0c(!DZmS3MsJ#8Z*S#K
zqds9ny|JU9A@(S?Y}%QYuPdyj?B>(As;FHRcMR?VpsmcF9Q@%mu5EF)<-u~oxKHfz
zSysD?HD0gEXN+dGef|M?b9Xl0*02CJc$VC`gwpW7AmhMYK2CBQ+I<I;QCiF=438E{
zdi<rO-ZnO<*HmjsyWaw#EkVmeATpq&{7z$^zD<}~EkKC+#3I4X5vXSixOZeqWg3M9
zx?{TZjPW=Oy@Dbgot=PM{oO2KzEvqOM&%{sq|FmDBmZv3$YaUNtefPf;j1}jz9l6k
zU_={v`R3&9yMC*&@4mHfpnCvP;CEFKa#ScZw3<9%9%szT96}zI`(_tCrV$Vl33U@f
zJhQoI)d_~v&*Q!WL55D#bUkvpvAODxmN&qtb89ohXqKfg`q)fQzJ`7OLLz|98E~kD
z8@Fj>tXsh4g;MjHQ#pJ8%@t>1ob73(1mxM+l!K0dh_>>tlKj^nf5BKI0AqGR+q$t8
z?P~Gcx-7gG3qH?(pmmjRoS<1++1lT+ZIu-5eb~@x!wX*+FKcwpIAX6yU^6aXTDvWG
zzypblIs^}NR56x$gMK#n%(0~r+`B??`3jAxgj;;d#^@cjI?GOtP4}kbhLU(+RX2a?
zsV)rA!&RV2>%9hcWO9ko?NL@}Z`CYU-F`I6qA~t;9GNZ-VF}B1`zA&<dF(HL*e$3F
z*9TveFv2Oa4UMy*bo$)AA2vbi`5mk)NssU{nEI?vuHm_-aCucCvBDIi5!r7}W`7^H
z^1Ct_hOUO?<>lcqr0AqBT$JV+OC{rzH8>f=hV^Eih#5vu?)QD6tNLSq13%1p)cV{;
zlfE0x2iS^h=@pC3MN0igp{hAo?K0A*w+qhPG)0UY95!I4>o7_rDn!JYGjWT)<R^ES
z4vrB-92!ZdCQhGsxu!CmEPD7^_bT5QY`=ETrnaH*t5!^~#HLda@>R?kMNrhd&u2`!
z^*}gr6LS{;Mc6ULPY;z)qGqG*L82A)Yac%DFKudxDUJH?N&}uGB`EN*Fuk)x6*Ewp
z7A?tYX9Pr(HhQ<#<-@lA<@at@!N=Nw!_3{mP%bd$64xRNU$3eo+7)IRwM*nBja~`=
z{{FcDetQo=kn*JR>g(Q1xY(a&Yz2v<-A}oXcBP?6*`<y__g>Ds&E{U^0E&!0Z|fdT
zEW_fEfqJ8=kY+;4biDmj0Yj~!`+RTa-iJ=U20fdRt4I-Co*DfB=u6<?1?0(_v-z%v
zjW{!*$JC#+c{~H>_i;fX^ZWJKUO6%BO(h(nW3gX>d>fC>``Gz`^JguCw6`O)W4V$y
zpoo1H0V^Qy&Zu6xxUqcKxZk_U!i8THI;&<ZK2>d7xqEcdY6`A&JidPAjQev3Gr3XT
z_2-z!soyO^B#K+zb|aft+3#u*&n3PGa^DoL>*O3@Zx83^x=(>3E(-bXd83&cDg4Kd
z*ZO0HhS>((b}Qmi!JK2x&ZQ1F31znSh^mF5S?Q}=vEvB`%gvRo!e`(A111buo?2y#
zS%0_X0r1~M$DU(=fRU0?$#+k0O7H2-D;jsxw*ZT(s9NocRLqtk<uUm%XOd0hW|4of
zG`eFWMh&>Nvq%Z3O2XnnGLyCmf8NU(Ksjq)d|<ZBYpH6Vr!4c)vp*)d7pN}!%KC8h
zC~Lj)+DSDd;mGp!wmVQg#%T?u*R7^oJuW%TU5MSUva_vGQNm41kaXq_w6{GEj#hd|
zd|#DoiA|csa<tnIaPXX@0Dv6P%h0I6-qPVd_7wsIyb8-uaMW!%zLXopwu}w;^odBO
z8ibcjwj^s43^>RihV#`vqhieC@&ycl&|K6WbX}|^bqATSBfG@1=p>|ZeIY8fqqQP8
z%bCKHx};Ph@N@LYNGN!uL?W3yo!+O_@}_17NZ2mn2gy&4)#h#qw~*y~a7p`qobr~%
zeB6O>ED2HK3dX)_a%H_k^e&_br^BT#1b+uAXg^G>-m`Oq!W|NTqyMmdO?Z<bv;1>8
z2WWV8LwrV5*5wnO+#y*>w8S|-w@`=zdMx8+MLu1AwcF4hNt=#iX7)%h1@dH4O#3aY
zzRu|#SYiB5D9=s)$e=-+(`hu0Q06_(-H~hF^uq|x-gV7RQ!a(G(IS<pf<~h;oWU*G
zj@w=EEvDCXYpSL|&C8pMMW6v?9|QbX>H6HfE>_Rw3TlP|>kWZox~0T9=Q#-#WQxv}
z*?l3PB)cJZ8+iYLH&1)K`9>o6rswSRbekA_gyVYBjt@@q(9Bn%(M8g-b;mCzSzCId
zCy{6M+U~>6qOd0x=gQ;6Mn*MdfTufr=nV?milxdnzqvX;c_0Fmf?QXO&7-15_s90_
zs!*K20(92YGZH>H9bGa$S<QavlgQsHL5772igXE4$=Ysy%@PMri{&%mWsF)zHIl4J
zvJZiZ<|6CI$*UZJ6=^T30`tG%sM^mlUM=K3?OksBI{l4|*twx+?IkNIK@i#J`#QBA
z9A#W~YiM=*Xd2BTBMt-qJ;pq4dL$zqKOdGGrN+}GfhgusLEqT6NDt&%ZI|n~3zsC<
z)PiCUYP$53&!QwUy3{$4)LO)ZYB|kj!#2q$;g7Gf5P)1h&-2)WTC03%B**I{@*0w=
z6~h<|{CP2K@ezR81DXZwX6~?>2KeP3lhkk25ZxY9cU(3neH_S?G`QoGU4w##@K&&V
zKee6~ed;(ZdH|&NeSiY?%+{%}f;9`WC*`5;H}WVx=n6mYTHmaejQUa@aT)*#*4xnO
zakvJLA%-P<tV36{<+zH_{=joYL70P+rL!I7^vs3Is1Pk}x*RqwB`485J;(h1t7=8+
z+E$R=@Kv&3Qt(%mSFgx}G+geA4pp!1C_nNCb`ABen~rM@7c-V<y!djx8^%Rb$rwGr
zut+O2PvxCC%e6gr{M^8;+z}{E0(uwSOBCd${fW?Q6SW>7HF22E7AtUB?7_wNS!jP|
z7XpTzEIE7xilEd|q6R?X-!nS5Nj%Kx`44&$fg7q?D&k6jSsUHPO<PO9eIX$V>v(0o
z?*kN}^m~-KOHji?5yZbwUDe%}d3GJ30d0tYVkF}i-c)C3KZ)*Kx~t5LvtTU5N<hR%
zy=>TxO>nQ0LC(o=su`4)l5=ycUMtK|W@HGbChuHL8c-YP6<gl1W>;X4&ZnPwL(r97
zoJUV`l;I}|gK-Z)B3E+>G2y5Y`h{y|nleAh0KuM=IxQARP=)izL%oD55_~NPlPM^{
zah03Cw%MOFv&1(5%d<^eHIZ1HJyMG#kKKX(m|iaF0aBa5LpfbWFS!U&-FV5a`j<qw
zS-H}~$r~6?1s&&IePdloO43=<TOWmpRYY%Y*d|uS?R|+=5_QkyWuqjmA#-iMEsyWJ
zl7>8UbPRZ6gxL84p#4v1At5OHRLG#Anjnazr56dv2xFp&)qh*J8z`20vjoGBK=8_x
zP!g|EOb8BsLYZ}d&S>TIw(J=x8$!}x&<D&*Is*?!n{S5s(SAOpnn)*3=Z$sE-P|D7
z&Wu2#Es+%&h*d#`1&}#?F0_Lo7eK}w$$)HP4ydLi_&3+g9be`d=8`MWRUU^qj{8|d
z3S-#Mo4%1T0ph-!<@BeE!-$S^f_gUUsH)8*!Kd`#Z(7tP`p;IS@T=*YY|P@txzgWX
zUi%&rzdYI9!l)8`(t|;@ZXy)MRNjEYX)YqN(kdI~N#J$8GkEr@UFsGTA%T|=LVX?j
z;xZBEj0nP#WdT%*0BQVDPTVT;2?YLU;Xyl6`n#YgWeT9eJ;aRk@g_0g4v~xi>+tIk
z@bgG4o%I(=gD%ra8Fr`pQFK&yMBWE@oTrJjj_fK*O75CKyidu>&FyegX5Y~q@G&J?
zSVznmE^%|HGd0wW6fLKBDYT?ei#TR$(vzgYsa=9mV)ye0u#uMS<kSPiAp;e0y_8Bb
z_%mRShq9fGEd!vHhj^px8Th>#r_!*2sJK?qAS&q{GGv)C9r8H6D$s2|%NN@)F_Q!R
zCo57L-CknofM&_!OJYhDIp6A}sP!OEBX8iLhhz)@l2}Ea8KQ=bL48?VMQ&6rkE`_`
znAh|>oB)0=&k%TqW7TKlCPY*gdpIaa2}Ao~aCJ;ea!g}iz?XHmmEe=_1bx%e3H)0F
zXC(uOXXg!7x)b*vimj;$a9U~xNEo@QaBWK~I3ki*N1tHei2mlJa5208pvcxuq;UzR
zx@C0L=>Z5#A=|pc)wjlg+Jx)vg(d9gE1TI}S<MW{EQG|#@3<a$^Nui-dakKgiThX{
zTFs~`zOcg<pKXNelHz&N<-XXxg5L7nuTdHt2kN7jTDuE<z0xuONSXtJ*`PkkqQKF9
z-s4~w%it?*Xk!0c2bf=F4mi~l2^))$75khIzAH!QLYt1;Q<%Z4E}lFR008tZJCN9N
za0vgq_#k73tj;L36k_w-0im8svqwv%+OglG?!zu_Eys;nJw$FpMKe)H->TZ_8CGrw
z7g-lg?jK%R!UJLyCJAZbmCGA%1+KVwBdpq46FH+st@>}pIv$!PJ91g!lqtg>mrLUm
z0|E>~<R*eln>1aQ?XW&t5LuP}3!tsNBvkw|t2pF3K$+Dl<1D%y;5Qy%XDt#_J~gXX
z8KW{`3WSs;^Pe=lfJe2qJm_#syOD<RED{19WM?V80%(guH@^8Cu$(1F{7H&WneREu
zg9x<~YOGU=y>S@@bAhJk0(hCs6Ncjk(OI&r(H}A<PJxG|fTDXhtuN`fNCl}0K-9De
zc(!NfM4cEByy@l<6CwIxXGr!bUJ~zi?$#byQuh%o5S*WgE0NftO*ARcT?wq&-!Qmp
z-6;SI+5s#m{4&Lo-t24Kw18f?!OZVsp#P7yw+@S{>;8vFQ9+Rs5hMge1wjNwLRwN<
zq#FeR=^h%CZjkPhF6j>G?oR3M7#LvYw@1D2$NRaz&tLELp36(l%$&2&K6|gd_FA7+
zI&s>P0>_!J<kuak`PS#tMovn*<HRm?-NOlM8*4gT@Xp@5%$7&&S^h*lrc`}AQ~gP!
zLW2~kn}<N(oyFs_Cz^MZzv-2<3F6^ZoH(9C0yebZ3w5@7wb|i!xQpLBsIGMV7KqWg
zQblu<BH_t^04O2sM4o>LYcbJ|hd-E*c<Q(?JhQzqn`vU5f4C!-42<DVXPpV}qXGpM
z3;nK|=Ga-*w7neUYave?>Y9%IDN8(SO#l|VWT8?N^7o)5W$3;K#qr_S+60jtm6}?i
zYv_W8=%y7!&XMA}Lce7YU@cafoz70Kv{o3_49uQ<*<Ibt+e?Y8JmG+@YL|bCKDt}U
zs6B?oLekBMV3h0y_}$ChV6&cfj=1q8URnXE#9<#F@q(-H7erivjnCyMrB>=QY`JNz
zAA|!oV0Y<g#=j`>$nTm~+;v#f^26bb3CtG5+n&MXd4-I!^Zcrkx1PQgO!%QIm1>fa
zbK__-;hDHVvk{Ri<6&9i0ITfXlD^2iFI`_>Il?j@$<iI;QNy$CPTxyd-1E}nd~emV
z&??Xzz1ZB;R=uk2^e$wX&1#+LA(v8z!!=LSbIA7ixqWY;sWJ!JgZ9Ypc#h@E_5xIe
z4UbHu8G;IWAoAsiFsGNhTHK>`u~r@%8en0Ul_T6qi{I^3<#q5-Jx}DhZYOxo;@(Rv
ze9OT4;et0$sPZDul)E!p+ud$|lPcQT2Lnkw*SBxmcBYx?HhH(@eGQ8vQ}Z}ln2{TP
zk?mT6&l_)(bow=W-+)kWrL}<^R^WQy)9^+ej}F_j(D5+w^q*cg^?~HX_cbKDIO7lA
zRN7tDocMx;Fd|V6O+Y4ud!n%|nrohTFBur#n8k3&;BhMN08e4Y;i8RT$Y{lG;9gvY
z!IymZQ!!Opu2HBQ*9`1rIn=zaLOeM6(iX$w9w_8MRYdoZf!2{91uwUn-D4;L`BH6{
z2|?V?GvfZxK|dVG^=P>|Ou<wWJroEYcoo2*uOm=2cD^T-j81VsmA0Bg-)<tUHDL?b
z&=ke24&P+<Cp8YDT}fSEv&tkF{XqIT(!=RdQQkX<!X)Cu9kg~pUt9L^g{<$(V`REM
znJ{VXW`)yRjZrlw0`iK#7_jRz=ub68^C*ic*z*idc>p6-*I`DOukOa*^IX6IR^OL<
z!Dd}2k~sN++u2DeytGlv0^_rNwY$bpjf(Pfu=s(h6En^YP+lg<NLh;{xzi!z8!_Af
z69({3{$~R5O!Ca=1b_gYXJ=pXTqC&`D7>Vq{dNjkskvPZoNv~j;nP~(XJbyC2^AfA
zvv;T8L?*E2O%+FpXYqui-LdC|P^p7l4a_Y)HWQnuNb4{bZy9!9E>bc6ky%+-Ha!l-
zM&!EyzGMWs>}x##wi{aUzFcqBg5&^zTo-%t3{)JDWC!t#czMnhz_sAEi;e{Q;j@PC
zMfVGIeNbM3lg(x-@Zd*GFXz8Y;+=8av{o)Y7B6bF!P}|tin(R9Q)Qf1cd4y>I@<Cj
zC)cL<wtzNJQm#-hh_46Xx0hU_&@fba6bBeiUJb)OUsp1E>rE-q@G3oRqsI1<fE)KV
z<JbJ;N2W~1qdY~m#c_s->`yIh8}s&|d3WQVSdU;q?iQc7doDG0uJ8TwgBIyZZgj3m
z<WaxijXk-EiFqd%R7p9RKU`mmS(yy2g{OC-*w^c@Sma=-5s%ft&@N%?L_;<R=i-tm
zn7!>{Xh0Jl#p&Cga{ui89mmJ>90O?AS%|`FrO$uRnXpR&6#s^VkWMYAqK;2|f7+uM
zyZ=sw9tN+_;?@oya1V#LOhr^w1jncw`Y9g|*Exk09P7HHT>R!*U5)II$&YdT13U<U
zE$kF$XO@X+*-+Td_=D{9fO1HS)1!jCB;8N<O44fQfoTN6mA4ZgYO1Rve-kF4A)z5%
z_cFqlOm2>gZ7_vbYBHWa%bb5}aCm3-J<;N*z9V7bA>%y?S0j~U@_;EFPnb#6U7?k)
zx*}U6u_tY8A67RtC+{tI01Mu8>)Wr>fyxe)v?YMN9PyMiwP(L<9`?%z4uRfY(VRNx
zGDjIA+-cBUB<sVNWDvo_LMHD>1L8`4MwQytyUmozZe8Y*(RxlKdkHJthnc={jU-Pe
z8RZ5x`yO8U_LNv3Ei0Xtzv(1hG|eF|aVF}$#Gn>8DR2nGlfLv!-#bJCuLdfv!mk8E
zoT<d<ERL!CI0<05AKq73<dqVT+-Y+>M$Rz3_jn0p8Qrh&vOF`}j=Ha!EN@!h=5P|0
z#9B}l25e~D4Z=t5ex0v$%Mq!j7oAvn&PkB-*G!z(tw)+^VZo2_8js%0J4p$Msq{_|
z9AgU<AE)R}m0qJr5Qx7g*%WKD#MJHJWsxwTSA$0{@~vnxZMP@jfs{RDzh&VeQ=qHX
zv|~Ys<BVhTZjPnZwYtKnoQuh~jYWG0J7b|-V_Z6HabaYl3JD;I#`&kH4QX<xixrZB
zj*DGgr1*WIQ~1MeG`3}hJchx9DTj-BxtPL6jabAR=Y7**;!*!gZ6pMEvu%|c{&IPv
zq1tm28O$=kozb=)B@$PcqBQqK={kwIhzY7_mGwe;$laiXI?Sj6*5dh^Ut}f|Tam_6
zGe)DPhUeQ8$9d6>xS(Si_57?b$2DLlL0(5dwE2-_v*3wh$%Z1lCAPT^QxUkd=uD$f
zBRV73Pmu=lTN5S8>8+=1d2Spsb6n*c3u{ScEZG(0q>6J~=HfS9w+-pa)|IZwmXS_c
zJ)CN;2PCwCwOy{;zBw`li0XF#_!tSci-qL&Gj-+X!VFrmTPb&R3_Sw(^6L5><vZn6
z`L3}qpj@W<fK{tb&oHWP!)j@MAK*Q!{ghc~z&jrgPwTh|xFBAvf&UyUI&xnR-eDor
z0w5zn@Ee*3_nV^upa^11{gSA&m;mN2i!-dv<Oy?k(*Bt01vGO*RlRjc(dfx~{nNjE
z0fNk@DoG*p=KDKbl3&^wNb_D(#@r0(a<01a1SsKX<_jnhxOwL$%2{gZ-e_GA63(ip
zqC4>PGYCQ5U}1s9+;AMKU^eG0;Q1urrwek)cF|vZI5u~om;V7+T{-TXkgYz0q4U05
zGa4y{5^eKNm?b;yY^%#q9vBIN+^NhjY>TlMy!FiObzzg;cko$&JBv8=eG6yE?I5zF
z4@l0Cx6k&}VH^YBj65^@R(6Ks5*Rrd?6r(MDs@xg{)KZB&o0&!jbI!MLNKMT`DF(_
zbD8p`^cRk3KPJA&NpYTiVB31XIq0Zyo}ME0MB_c}VdGD0$hs7{)BA(-E78j3r@_J+
z$i;@kqRw=c4+r_llea3&)}~n+$8vPZmE(D}Ap6Y*XT_CKp*}xe+rzTaK`ar$Iy1WN
z;iFJ(;JlIHXWZruM=w5B@i4J2TMvMr!IZ9iuC_ZYfA&J482#2qY~JXlg&n>$n}GRM
zY?K`LS$rnJVUVXdpOu~+`U9%_e9ToTsd8zkK#f_B<sm9gjpNIV%H5Y0n2NV9#NPuN
zq<W^1oAr@vUQK88Bb~Q(8t|YRds;Ot0Gti*?;gqn`pX%v6|-wq9(1Hp%-(I8!MdM3
zQ|TKIY;G3wR130JodqLfj=2*hbTjBU^)Jd`L<-Kpgpq;#u$0!F)vG@cyeIV~#%$3P
z<(~pDW?&`*J1hVSfs?%k+lZK?&X05vVE@R?pc%=@MPZMQ{L>Lt=4PjWA05?!n^;U}
z<%Vh%8KLp=vWNB61L(_PK3+h}RDQ1eeaaS|u<pHX1R6N9zE#U$c_Z*){k5{E=n(3G
z!yPV2o54C}IXqfNH7{IyY2zg>2v)iq{F7PHIRR$X{1Xtuw9U6QUW`L-V_Dkt!W*O<
zXNt&z*GQ5HSgD8JgzzL&UYE2sBwIandNC+JEgYwU&b}?ElVLzoZbk8C*P2i`-s=%Z
zrFGAgN3p%G?x-~ghlBEAF;^4V-v9Dr@i+jV<LDsBaugarHq%dFHDTeXAP2GiFv%e5
zC)GN$vF&hrk+eIQ<XRI~veN2D?J>|=7^l705Ix|vf(n}+so@zSbee&I$foJIox4o@
zUl$1(v@@U(Rpr_xv&O4#o#`;FST4r}>yy*1IWb+8UN?Mp0f7$Z3if^hzlVXyLc`_p
zX4FO18&zg!LN@Kdk_W#Quyr=T_#$mHlfEC1401!1|HC3{Ves+7k?_S_IwB(AUZD6a
z?$&p=+WSY=Bnyl__+qoTczegub>kpjddKOb4!jCEoHA{-)hc}7UhN~nz$FUw#mODM
zT6|zS{>c|(<zm}g<hPOO$06C+u}n`l>=2p442~UY^>4F)J-<nX1Nf-G_QH569O!5=
zv$B-;4g-`;tR7u^i{g;xDC)dn(iv=3!cUEJv}PbLH|#b_>{Ii;1H~fE^`}?+bkmN~
zs)~TYZ;ne&ddn#DxYQ5(l!pVh*Q(H@XkV_;K9q2IgoJ*tEFNf|?ty+X1jF>g0mImi
z*1&)E)q?e;*|*Zl<oB!JrnhDw0&P|He%bLL1mj`w-8)$a?JyY$$Fj$aAZy44EI2hR
zEutr|OW+uIQ5Pq(;hBt2e`~>T0j4&$+32xl5A84ONp`A$_dawLAS~L<4LRdu1>3qp
z?Xluj^OZ>9SMwuF5KXFS<P4fib<0QGm#AIy3npsU1g4W?pf*r|*rVAa6B@8P-lWC8
zh6>VFj2Q8`=vA!!;hQ6157m;DZ1t$I=#AcI{`_6Udpz6=xK5F8Fhj0>EvvB6^*j|Q
zu1V>vES4u-U`d}24%0K)@sAGdpUSWjm}!>%{JlMI>sc><G67TIWf}*L_0e-N*lzi|
z_LDJFBiJ{qHI!>XiNm|qB`?tqNtUk+=_zo3k8G=HyaI~v1#_@U&spCBzBMf@j)4&p
zh=_8b{sKe(Q4RMUCvNMxBdMp=_fmG(TUm|cO-?(u2GKl3@6BvF-)B$Pnp^gd$*<G$
zIv}b#ouhUn7Yf6qLoS8zJak5*vB}?3k)3s-+b0<Zf!_-wsU?CnN8t=7RCN7t90749
z=f}^RwF)n6;s&wx!_tZa9P<G{mEcE%AA7!S6W3fmU`l}VC>C~!Vrzc`niLm&O8Pwj
zyya*>mR~W*3K}F5|Ar=X{7&x++jv$aow1|V84scnA>qe&#YOujP-|N*MvCJ_e{7rS
zIG;3<3oNeiAZ-VjyD_d@mXq7tkwPcR{PcZ|%{-Z!mAE_0JAl@=$?3SNHYwTnE*hH;
zvxI=C*s}JMktUIqBkYWkBQ@uh88ti<Y82@%@i6NEaKT7k9Qwhh0*k<WoD%bf>m1mP
zXUI#YBEP@eG_q`s?V4b=UPFbd*~nad;<!RebCt7zGX&Wq@?$9ClRD+4Sc@Q{T1D{J
zG>Eb1<!jzAE2Kw7U%$d5+7CheF666-PrmbqX5?PBV$sPPUiLEVD+>{Ye#4Q=6fXP%
z0@Y|_lFGO6`i}t=`L2IV;(_vY^EM>xjL>=baJFa~^XshlR)8dq{2+DO%Sb)S`}aAz
zE$H8~%1;q#y%iQeit`{z{h6M_(m-!9boLCfvCXxkT}Al14k~VY=h;}3{#Oc@1x@Hm
z*sfM?P6hJn<Um5540J{$A@#6enJ9^}4V^l`DEp!gF>46ByapZC?P~UFW|!wTyHyV!
z#-n<4v3BUGHNrp~iZmY_A$mt{cGKA`hn|yS2bGoMej`<W+_^KNH@u|`)maUwT2+dP
zcA6p<+8G4f^GJnNANvOgvy14uOIk4Myc_oOiJn!Z2UUGJK@4@6FE@4l;2^L^xDIwP
zib~4ia)m&*x*25O*wTnHMzhY_s6S9c*LKv>+Dc(S*c@cZfk-6thg7#N@d^n11~Uv>
z?*yn<CEt^!SBXw?x^(Jq@3$fQ$o(GVo!v0e&37{5()}dBKdo4t#q7BP=d&I(fu?U3
zalM)mFD$w3@ReBRN~e{qR_~LWd8Gwq%aEgtr7uu3SB?URLYJYe(YIbC>w#n&rZ2Ik
zeViA7ZzLIP-v8~04-B3lQp*gn)rL*lOE5i5{z}t3pIhUP_v~%~vDVuwj8%pkCX5gD
zB0oMffNxklmNnTbMU?QPS9#`-g)8qV9#8`RA<l2rA14D+HEvcry&U27{ws~fjp*pQ
z>W>Qn)OAF9zT@Gqf5o$gyn$bHIUnkCt`AMb6I2<h>y6Po`gme~wERgyoy*SQ!3--$
z=Pov+f?5jXUU)bQviI_)X><Qhq|+{w_RQ=4{gukf9+8L#jyT&tVd5{jS0LsCnhABK
z6DoKY4H?KF;$)h__fgFokoTIj<vs6AJtUqzTeN(g_FlrTfZ|woHaTLY;imQAu{6o}
za3w@1Y*ES)Wnz9CxS;_FVgzah<BL;PABUl}xSP?f_Y5Xfr7WD+=NaLwn*_m?MRBM^
zH2uZ%I?1dq<5_dLHTOYaTdt%)bNqD8<JXs>M<%Ce2GnNn`WVwL8C_~4A2xX*+qtq+
z?cmgPyD=#&^y2+dDH*7|Bkh=xr2N~P*=;TL3qjz;R#HhtDA8q#0QZ=Lu}a&cKhn;m
z{*`v-3rIV2*R$YPP_PNk?8$)h)+|GSyQfKLSibs6LQOW^XqS}W0J;n3!WI|Bz@!>7
zu9>wEM2+VUQ%jc_6$&?~Pr`h>z{WZosOVTER@w`BPEYPccq4Vm5acoqx-*XP1NU@!
zz&)oJ6JnbMLX+gUKDl_kPSov1804ZGWYaaI+!Pk}<mLd*pKM&`-}DveM=v1dX=%Ex
zSUp~HU9wUh)b=FQ$-?Q<tDPzKVbgav0hiL#pdkG63bL5v_G`2UfZYS~1K9K51jT<m
zrY_eBaZafsfU~}!9+Txn_}dT4GP#KBO<c%-Hejh<+8&(ki}5&kV`35On*lMOo-i_b
zKvXP!A;e#ZT;~}I>hu0+#QchccP_Hd;Q+srmAY`bp_>F&u40C0V9sXAQ3`vlbv(jZ
z(25jT&%PRT{Rpe(dfwVriHP|uOw{@FaEl3uJy^>k=Do8BDnp|{G7XTkkb<KSx85vm
zEpw$R8I?Wfn>5M|ZIC4+{t9rfxY<IZIB3_mj?jes@M^T3NP7=ViRtMa+rWH;O1JUh
zk-LPSpmC38ukmvLvVR%N!~s;s%Z^IdrHWdhe;nrGX=rS<FB?E+@Ye!XAntyA_D*eY
zL|7K^C88g`$#aE~Bop@raEEvh8=aJKx?)B}Yi@eY-uRI#NUtFh|FdGqS6tu`LGcW@
zZ2i$NuVgzYtQVCghfV~tLS9E%c(bf6SBpX%$u+KDISlqfMf1eZpj?_h7kEN+Dw=d+
zR`zcmlZs!v6l8!gfoiJo=fb%1Gd0uSO;bR%pSvu?P}Qw<Xj(wtc+sNVFNbFgp=iMV
zjSXIUN&8naO<cf+Ag_s0*<t>31d%)^syADPPW?<>X6l~!EAgxOe_aNF(A*ZhYp~X3
zKZW>Djx%@Q=J=ncP-nOO0PN>*kgxF}>+#>eP;C7(hZq?9Ihr0>{)nfsz8CZGmg5OR
zQ0!1HTR!z`ZWx`++uylo!Tq@S|LiU32I*apYVw2Q0y(n1%p1@3?}yOb{O5QBqCwvN
zP<DE><;VRyrzh&Uz_IS{TYjFL(r8^G&Az8bUSsOGr6WZC_nc9XgBaX@T<g*hTmT6O
z<3WfVHYA*bDhYr8+9vkj4^K1&_OnL#9nHV5;f)7|#lJuOJh@5*9Fgc)fx|8D`MO43
zyW-j3zw^@m>*=3^D<IfP6-aUt2^8|s{a5m7aL*$3KWPFKL!trmN?F(r6CT8EvOsr&
zJ__ObyK%@{)SCataofZ1nyDhzt4$!cQ%zm{UG)K6a_arBOF!R4H}x{ySJT%ZFIb%V
zBp06^-NSeTm0#-DVGA5#|2qHmcd>51i?zZWnI5s2DTm&3H#}~`&Pgt!4SwX$bK0K5
zP<x_<7YqpC;Y346SUVK|@4HQblhNNFex58=i31<PgYGD?h@nPuq?Y}H3ASY%Qy3dG
zZ)57^+PY|A(y!ep2WZH@53hgIsTXKP6+V5?OU+UGYk8oqZ6o;CJo4WDQ1zdd{%Y~}
zUtSix`k%hP|MflUql=RNe<uO{&sAOe|H=gX_m7(LR3OFP|2T?>LAKTZaYPPd{P-U~
zLF#klW51WOobpj>T$}t)>jVFuU9V(A=>KX3()I|{!}%WvWA^_qgF)A0jV@mM(qj40
zen;(OJv(2ZCys-)#_|Gwbbmi2VTc<(zuX-opVy`HyUGQsdn^CqIQQB>=_Qlemj_jK
zNQnsU`91+GK~Cs)tR)#6#r+MEN#q{+vB~?tN+3WWNFI9_IhBId>UD&y!sH%qwGVmq
z6KPL8u7aBAo39UainS{4vM4gX-}HO4dxB16Xz+QfG~!ihf9*McU*Y|Do?F+o@EE2`
zUlm@zMsjB+k<gwKZGCZ1ud`}mQDj&BMQM}}WzF24xL?}$UQ4Le7Wt^-*~pH{^sw6S
zbV4BRIta;lxiXVgnO6ym-MOk9^2ah@jCsN1vM|54ooj$CM`Ww`3Yo-ZfW38HjYE=b
zi}@xVB=N)f*^F-Vw7XJcEnsIZFM$W~-*XmeH4#02(mXg_rW-!Y0XOACCNXzDFC#A>
z>y!LN%JmN_{z&;)By(_??diR8#~WHW)Q3CIFu*!rKN{TT;@O%R+*Akc)=&M(*t(P@
zZv0@Ic04QOk|B3s%X991uJu>3DFFNZ-49I`0~5X&LY%1?FBNBk_1U5vd)r#5m237@
zNTNrA_0C6Y!VVH5qSeh!Lu*v;T~5FM`4&qPI@RDM{HtLK14+!rXup!b*Na>H!LxO0
zqOd_%YL@#BhiB?^v;Rt9+(q$cKd|)@`bVDLJCWX+^HfvgFrMD6*`dej45zOuU2Sm3
zq>M!t@o=0Xk`U`qoQ>I{qk8D<bWQQ4xL;W@U5NkS7%ng038Aq`BDJN$=3Lb{;!o!i
zw$;^)H-8MIit#Jf*d1;6ku)?9qC4%UaCy=W4>|9z5JE|Q7t5+Y{Il=N9;J5;Hx3iz
zsxK5qnbMzDPpUW_zAH%0hQey?=Z2Wn2;z5U4?+e*HYk1#>Ciuhhf0UhB?VMjAOm;C
z7;nzGfg9_<eF6JY_ux!azb7Qp_>XO)UiBH+jR<?|G^|epWNNIixk<|`txVi0y>=4n
z*fNQK5eEhRpJUe(2{uGu9-eUxojq_*cKRCwaQy|fv+-YrBL>_5oHY)eCoU<~WV10H
zyP%Sp;zi9q>!n}J$R%p!kKM6Li0Mwf*_=to6eMS3vNKI}uh0Y<_H+{&gDI2uI~OzU
zKZYv+hy^qoBTHiGs;vV{Wh{x*zlMd;>d#lq8JhYxPu443zGhnOjF=DID7Qge47jV!
zs<SNl{CeL+$$#Ac0tgG8wbOEy-;j&F>Fk&H=6j`%r3!BRdRRuxpUV{5r3gXwC0uEu
zZiPBMU3P0NVe>I7Vi(m*S^8hM$o}KC$-rw%BkpKYk}3~!ol<(GSB;X=mN~S-<fSTq
z*R**3XO`1QrU~7g)YRGOzLNC1uB}S!+;%q^LOX>~+ul6i^s0z3awnq7+8V=KWWPip
z7k0oGc_mDchxXmuml<a-ilJIXRgf-&UY@eraW%+tNeFVK&d8e3#nuR!bQUvxRa4zU
z9k_)*dbfO;jW&??=3HO&hn*cZs~1NUP2dS-c`s#4vvB!FJ$6Q3M*bE7)|bIlmYi=m
zBZYo8j$BKJZB5uFkJmt!^H=Ox;Otjk^gF}X9Fe2&mFQ_qO$59we;eNT1)1{m5}wNQ
zI-cr{VIA1z$hxnEml^F#x;0hAph#2UX6iGXT&3{#sTs!~^=rO_bfFbCdxxLU05^Ny
zNgK$0^Ia<rl;n3&sQ+3~UdMNhjE)kx2d(s^mI_>xkbUt8e*0qq9FRtMeraikR9<M5
z$+^<*2G{SwI{fF-)<I)y7?|dnD$wmtvUDE|WRLxE{98ErCr5dz7qfOv0Tw^;%qVP!
zZFuAFs~-Hhc8+8AAOX1YFm;<W1LE~9yXDV+=d@@0^SV+4K2Qch3|cq2`pgUP_?jyF
zkNnxcyJ-1m7sqoSg8~r5t~UC@uGc87xh&??SbysviQa!+sFb6Gs_~SXZ9iLMx*nD3
z&$l_T*X3SqkIMN9jy_<j`E>?9)Ef>rfsXPiR@y;#s!lP}tu2SSZWd%4v}R(SLhZ4B
zOV^1X{@iklq(jiy?u#kFYR`<hodLgM-a#K39xhpkk|BJ$c(7ADSv4`n|MOuG<)A-T
zO;s^#720(%g`JriJCNc6x1rgWu+bfpuq8pM)KCdSEP~kNcjbU5|JZhJ<=(Cj{${!M
z77hDX*CUz!d}aBOkxc*+{`25vy8h#z-#|bu9Q=QYboJFEGpM^6&!&!tNV>QAM4g8>
zvAZ&0@#Xno)%<z5C}rl3ll}P(>fA*~|23p_ZChMvI$egqpgh_7H+OmZ=Rlti1pS*k
zMdnV#OoOX+%6>g9(z=Z6NN9X1_uEz&{MWAc*VDd^{(pMMNB@#KU2G-(yQnAxauEIR
zPA-=m3&|E%!W!Ip=!+cI24tIw^D=%9Oa_3w|LP%xvF^nK7b>^ZwU+bLC%=2&4ZOPl
zzMzZRsC)bGPk)`f2>Cyr$^36gxXa+*;%VgQzeR1v{Qqm>4NR=j-<AV@;*PnO@&QNy
zYL&&_sIdo!%i}-wF=cT~bW>GM#~;)@6BZeDn;}4OU^(BFg66Z_`o#oYlz=zTU)_i3
z$r&*X^#Do+Zu4+Z?<(qp0j)pxP0B+U`9zV9((>_SmDH&8@9SRxSE9e~Wb_mVb(Vn4
zP-#-f719ohdZ0Syi>SOf-bRx{$BN&sbsUfG;K#uPZ7CORf^V-;{?cI?xBgrJFZJz!
z92d{c=!BX7<&)c9O;rarKj(Cu2*-(OatptG`1FjUUE%*N)fj<__L9)qTEp&m<rIvw
zGgyw2wLNr>qG@p209zujAHi>63?*m$)#)@4KlGpOk215nq||cQ9D&Xm&X1^tfE`J$
zR5xnU<=43ODgL<vIQ#m5Jb`YU?~E~bl3N9&vVK_)av`gWHnsX+ZR!#I$s3*YG{NAY
zsldM0!W7;j<8#rUs-K^>i-7FM!%|aMerd5T9)IR#U-tF5YASAb@(zRYaA&--1aAR{
zAHx&zcPXc|o<HU-l82a}{@E~z%@IsNy~dz5KsWxS)%sATZv1#bO18YjU*)d;DP?dD
z%tY#^u_L~H&eN47o3622J82xwR`8szsUehtny6W1ytfB3715)S-?Rn)Q{X-jQQ_Cm
ze1`aD&<f-})wY#`QwIbPL{WUVI#2;4X(I}AXZtDZ7t^)%87Kp%Oo0!gt|BPe`quzu
zwEfwC1!8=Z97P+*Gi64LG~U@B7R&d4pjv5(x;Gn(!q1Y+!V(%bX+uBlSo|If^h&YD
z_;*7c|J{(Z^e7V$E+XvTuWdu6wlWIm1IXn`oIt@5vAU`wzh}z;bAgBmyppBtc5M39
zNTS)FG+2>A`Afn?c;{iw=5qR(9>IW6u?asejLv<k=48<Q{MT)7c>XcVA3?RYhjOjt
z@?@WEHui~O$QzVyA7-sFe^Z>!KQ<6n<_9hd4YP+*`ZYdO7=u%gF8caktwuijlh)1$
zxZ}b`D+5P?h~~jqQQ$j#i0pR9eLM9E)HSdUjAtX^voYB<uU;$-cu47{%bI7XoB;AF
z<2wvrfkZd`IW{qA%r063`u?g{Aj;}%4sEZCAJ~mdy^elwEin?OR&&)QYA*W`gpuRV
zS*FirMq`7xR8T}i?z&&SP45gIwYGolixRI=kwM}1<~lv3>)Ruz;VY3i^RrYHs!wAs
zBzB7`UnP@GBEN&O+5H6$7)dqs_<XDIkx|#_{{U5ehJXdSwAddujnf|7OIg~%Qg;I9
z9=-IzO1xIx(K#fJG#+^Y0!%r47+XeB&AxZ0Mhd=mkQ5T6-SI8Dq?efNs{b@X#c9E6
z^kSBrBV#1hTZ8%1G8#`dFx(#sJ%Dp48^itv$sjKx0B278p8PoYqp0LyQ=5F-lf<hM
z#)ZZ&E;2>ql13IY23GysRuvKZVRHx^?uvsonbc>bYxXLsYmcArU?+Z}C6n#dsQl9H
zw3ZmHK3J!JxeL8LVOSwgwPL$+V}8j=UP|y>O7V6Z5d(Q8`z^w6_P^E)YDS}9gWZXd
zUcZN^B3DbKqdjD;WN52TY-L=Cj1jZ=aoTgGvJRie)WU*mB!Wyi^GkcO+3iOx5(Ogo
zx@dj8sp!LW3VCk0+~wz~uKG36`*Qf2+<KV_!xkd@3hMnhjW!~OlYQI&DSn`rT&e?*
zkw}B}=#8;@@8F|HPhMnA7jk*C%M2yxi5zYomPitOe6-=Q2W^X!N`S8*=Xem!?sfE7
zZLh=g5!QBK8o&@-r+iXj{Eo9-e|5dBZ#cM9!9%!ppVBROYD`1*WRfF+o4H~}GQ?&^
zZCpNiY4Ae&awHR}EJ=nEY8G5aPL;S?1vpsX4;5>)Wn2F2UHCDocZtFCwz>T|NsEi&
zGzd-FGXB6v^$!#Rqv>jxjMnJ47nY38W^b?Q24ks_-6iX6b^fZFy@}<Ma_A2T%3kG>
z@~s+5v23!=736Mn1eDv6*3Lx0K(L*Ra&JUpFv=qXl*Wi*LX(|KWEV?C&Z9H7e)YZ_
zYWm{3FCje<m>y~-$>d*a?dSi4DDi<&xaq)IG==qk{}&p?q?vZUR3^p{Bf(ywqGt;#
zQxUn;%EoqEU%TsN$7sl?qNVTjI>HZq8R)|DU&3-~=tE+835Lmq&E0VM+WZ2R4o=UP
z=o+Qv1h13rsjzk#-t@GELa-JS*?qtM?9R{;L$Gt{K(=Lca?UC=1FQ?ivYP24rxy`)
z%8}hqEA3bIu9Xc{t>vB%e@WwGiRs>SILf=mdac-pO8DLrXwb*m+AUoy<F!+CoCu@G
z2Lq*^8@cOAji>uoM$YB^)7O;iPw`{WpE^pHKbdq|sDT<yP5lV(sU1TiV)z>cTO$n5
z8e&Xl5QqDq%`T-ngfnu~5slK#`EUnuggl<!h?yoyWOi6NNOzKfQ$2DE>Yk}Q@z#*c
z+)g(>$q{dU{9!WT{A)bbFw`U7!MVD`UBiAK7Gg}Kht-PB^XB+w=g5{n_o-^_5mw5A
zzb-U>D%!k+DkealjRAxaiCdCD5}yE}sGjt;gu&w)qfJ0wg-q%LP3*EhJSbikTmDSU
z#Egf|c$XVZ=3PpYKN~UC(qe`-t0-YWz)}doBl2*o(+*QW2ecs$Xu>AbcdPIVe-AK8
z>ghZvW&(LrHglva)Y+TZwv-qDvYcfS0iT}6`uK5B1wCIrIy<;?so{HxZlUk?N1H|T
zKGj~HF_}FIrinp&=hBO?ts^0=P*~sGOb5ODnY~@n)-;1r%zHV?2ofx;EtbB^J;vMz
z*$f?A!6Y?@Em5u}rN|<M9sMrFFv2sO+Gnze%&FB~Pc-yV>?hSb@mu}@=aA+2z`z|6
zJ!^88@cmO6R#`Y6K406?405xbgztDHU*egEmXHHt=q#ic4pHt3Mpw^pC|!1s#~MEl
zQ5lTOiLyk`A|<06vn(EH&k5@E_lR0$f$&~MvCwPT8;C8egj%8xG_oC!y}~tWpOfm+
zef75Tx+lM`U(9)9==IagWGiDco+>b&xnMk@J+k^<&Gtut)fG56J%|WQX3feE!=YR}
zz$`SPbW0u!&=bL!!_ja!B_u>xw_9O0o;cC09{^sV|9$s1L@#j-k#%<|6Abi<h6)(!
zL~bsF-XUV=SQ0h+w3Q=!YsrFC{WTKr7SzBa;)T`Y-fQt@vza?fZMk+`a0#oAzibx4
z4AbJi5xri6H0}Q3CoAVZwR+*15b^zzMT7T|?o&DB(29_A$#Iv7MxE}$Z56w{Emsr2
zktfXOV7TjQSAx^i$A}oyFRXSn5Fd)pAjLsDH@Tvo7Tkv+bfp~Ahnd*29(hBGI_Iz$
zhPWAu=iD2q>*^e%?cO{?q$GEEQ;JmdV+jYg{Q%6F7lLjvS#bf|-pE$VDJo)`GH*@8
zqEoQ6l+FBV(P@D`685++`C4y{bty6q%1ufX>7~BCxmmpw&QVEg_|Z4!5VJo=!L+p2
zB2ilZ+I?~1%Ytx-JK$2BMum1c{P2-%UFT5xgxdWzCKNp%Wqx~1z427NI5Cf~pg19E
zijIRlym#dQlpTH~W8J(uUXCUW7kX|-rL+&G=eSfHiJlP3O`^=qc@;uz5M8S;r1A2t
zHSaY*;@u#oc|$`+$;?M*g9AM}AL`hB=x<<dFuU$}vAx~S8b<4=Rb9Zu6|+5V>#*;C
ziGKbR-5||GH6-SJq+nK?pMiw`j5AyHTU%vewCu||u0=5hsV06^#m8Gk`LoRKzKj<i
z2Uzn?ce~~YcgAaNK?@e+8n9!sbB6BiD}8(eX(j^EU%m%lw2r3J!~J`|nt+1Yi=U#`
zNJbaa=(^_hCYWRj*e`RKr;r4Ags5z6;X~?&#0?;=XDd<(dCEV|&q7oS;OlSlKYjYd
z$oMUPM{!m4=mh6DFLH~&`C9K5a<8`b)AQb&KaayBGR(N_(sRxHS7!CaHr5Lrp%#QD
z8;Z9Gu-!u*q^yM=N$ORD&Goj(8zwGFQC5t#_hiZJRkcSTX?fY}L?OGPrE5D_u{Tsc
z*rycx9^QQ_7meY~+WY!xLc2By6l$yhQpyV_D$zc#w=ec7RlySC0|G5r4rb*-=K)xh
zOzh!-_k@z}33}h#%N}8$du$<|5}*JPSFX`urtTN$cIMYzAmnj*vCdw5+8nH2td&yT
z9POTBL2PZ(37o0Mfs*AiQm~dg|GjDa_*h}b=PirfXdQF37<U|r8q(V4D6gG)xTw@-
z%$#L1-g|Q5P#&TGPD^2S9ZRPD45D1m@7YPJtv+!l)sPeUV{YEH%T>5REF?zXZ1Hu4
zR#$0)U<sKm;(LEX{9*v-?b3@JLpe|TPT3O2YSXAJJMx>;3=&zRymU5uTN+?XSPbuR
zK0e%E73~>sok0wHwzv89YZ!6Y?Kdxl7_2wJlae1|nyghb8*eltV6&velBDlEW^K+k
zmopa~Ar^xx&eLZ|(sPCcmMRinOxtyjFPZb0lP1be7d-}eBndDFD&(H&+)gCS)2?AY
z&Wh!*DP2QM?f4+w`1QG^awPYT)@u74d)3wAED7PxSVBA7&XN#j98!92hQs=T-Tm+u
z>$zPIy9HKJG>*dY<J`%7K|ykkecnCWq<Hj(o(Sq$euxWP0u%y9#xon}SPMpGeX4b=
z;*OmzGaz;_g5_x~T_e9kyvk9Uy%RTHlgW!ICh7Mi?!^~Xb!&un@q_49#8_SVW5+5p
z%~u3_mJvYI$-Fy^HyXA4bc0_(*uz3ABL60}1+TJoPghRW_p{sS4QUBERW|*fGjc;i
zQszg*&?N3KDC1t$ciUU_ci`q>zcYWuUr$7p`zbybM|dwoS7}%&fv%|=?z~=La*tbP
zjp_1q7_I+|*MmR>31lSWbw0VsOW0Xh5&PN854hW`n)Lx6TF!b==d7E<WFbO&rnsgm
z(0#f1)@WTMNQQe^uP{FQs;hgEFKQjnzk_8%J$J2pCz`zGr}U&Sj(U19DBX7D=gOB1
z&o-{k5l^@7U{fCE@q`#YseA0Ua4E7T;C9dJw$dGO9d*w0fn(eN(u%qt6c!KM+8>y5
zDo(wHz6xqB_0+gQNxEIbGd?tjTN<%;aMA3UZ7xS(RplU8pJO`Ttz9crzE^9{VjOQ%
zzQ0x_?{Ppqf}`rgbD(iE7Ne6@GA$D!){Z7_y~kr;LTB5qx3ruW7SlN1+H=0%XfwWg
z#^4d=jet1Jz)qa)j_O#d2zkX$d*1X1M#gZ{i;k2#t?cld-4|+eJom?S;!%sEHKaS!
zUiav515A8)ZoY_lb~c`EylQJHmy(pD`d#sHd$frC4|hA9$IYL}#*gE!Mp3Ag){u%;
zI=*nr4cVrG@Lm9A2TfU6w<Orw)z!6*<E2?=&L=G%C!-I^hdwnXwH457@G;z#V<|uG
z%}rG+6wj|4@5eI{E6!{^yO#VMt<`#~?zFpxF1mlvR#U#lLeoxQP@=P^%bWT@HtMC4
zMUPZ4P`1ud(sNq-53jUB0VWlFmj0RWOm*Gj>wuDcd<2gaCh9}e>~&6Z%D>yEq&3JV
z<dD=jh~=pe@etIcIiJjT0<K@wOuHqNL?Vcja>ZX8E`+KCTU_=*sI@_nn|%MY%O-!m
z+`rr-Se2VrV<L*}g|_9BS)f`0p?94yP{Q0e&=x94Y*U>~fFbwJ<7EbmYIJ@yWay!z
z>D;1ufKHWYT_2U4Qw25D@F=#Gt;6Epwz_5^$oy1}U=Gn4=0H~#dPO574R0y0vaVGR
z`W~$ElN(jWOvvuis6^J{Z8#mp$Cv0nnjp62aDz`?R&JjA>?-~a++Z}R{uraEjmzdi
zwJh{XYjDGLvgsv*v6*S7a@Ky?sBeLu*n}g+MQ5pnSjLm`MYNjMG#Jf_ijlR?j89t>
z8s|#T_HxG<jJV0wRn?&~3<$^let-8hl7cX)*@MX=-FbB^<jTQ2iwmPm76a6EwKj&c
z8W&|cP&7SeuDZwP1TwL;cvT2*DNH6d4^KWtPI^api*Ea&e+;+8q-gIb4}RSxi`M>z
z1HP+kBnzd6JW9Aa!S+8CX)vlIKT8PPuXBAhWLLHkxIpwobcRnB)?DaGC69hly5th|
z&;^YT<R4odv;R?%cTshG<c`Zn#AdxeRXNV~#ALT*!EqGvglF;j#1H{VPxIsi^t=R%
zZtAMv>3V7U3{BN}4fS@+Tzr>EziKe;DPI3#i0B~IPX!mz&yaV%+v6Ns*bnsCn`&g<
zIz^!~j4bN+Bu2&Yhoy^+k91R3-N3FtWjPcjT9%i~&0H+a)JRuzWzf+TwibXLRZJUV
zwGiuTSH?P%DwzLqXk>|QZ#O8YT|2oCr5bT31m1el(xSzdtnEJl`><aVjgl?2`4oeU
z=(a@m{fR%cg2zS)1h@dId3z6(n=-TJ-t@0ntoh!V;I#K)Jl)9*vA~Fs`%Z|KMq~TN
ze@6Ks;w-My<i^2r;skDgK-Qt7(|T1w104Y$M!)aXE^A2)*;~jWxfABKx>o1u-QoPM
zM(1zI8gi*fc3C*=zH<CIefSflH99?Q6y*_DAOMQmci(Q5@aCtd5w&;~2D-;x-@WIW
zPG@jCT6ATZ-j2yHedZ{l9G9=zg5)loqGNKOEDf;UpDg`gZ--BQzQS))UYq9gg_~B&
z_C)ECy>leZ`5O%P8ar-{H!|*r{<3nHiW<hd`SEcIeaj+5X1$)wI(626$7uy+miouP
zRd36DsJkaoT2f&C-JjLdyEISz#w!T=?pfg3%z1Y<a%r$BnFO&^cf(of+m*<4MBXl}
z4}Z8r9of{-8y`!=LTtyE0hZO5TBcCL`}PK@yg|?DoJdW?O1WZ42wE-`jR&LO2#`c|
zSfx}~V^t)UQ2$NaT0vuq75UqM+wp{1LA7{%7;`p{swmCW^<*#zSaBhVb5~63%L52V
zN-dAosvJi=$+0{Lii~AvevCu&rY{VJ7s5<z?(nTD7Qia#$%SAyVoFlz--*(87mIKk
zj7adyM!)!CZVaHv^He$>c(b#8Z4r}WN;^@&K>ENae&@<TidGWRRmx}Gd!At4%fGzY
zMx?r0Z{4ZGD9|UY=QMDiN`Gde#?DRn=yh<Wo~&h?rPuNocU%<ul6lAM2H;3O*<R{{
ztMBYpbRI;I(}j4ONgVQVR+%jUmEgwG@z%T`W-^6ehEGN+pg=^*+ivL&p`n@#kHw-W
z@w8PnLe>ZeA|st>s95HrC9se-Fq%B_Naay){}Cw&2(Ywc6Mnyzh|#LoSk-n?#zfY3
zR8Z=r%U4#%*%h9P`OfuL&((2~T6ma<EOzl_>MQzLPDDAjv9TeK%G%q-%R-k|wy&ne
z0WeoWfi3oATVGJ4+oPoWb#t4JPR`@Rxf?jo<#y2EZ&{{z!d$zKw-o|NZ5b0fm-r$@
zafQBf`s%;BN9p#>Xp%0(7(<Vst|r>X)Iy;oS&9MMV4*mmf7XMMU!nKTG#e8j$IXM<
z1jJG{lDXiAGdx{tTZI_JKtxDv-0EZJmf~(hENmFu)`-X2kuu3^mbrSQn{e`><A_@-
zC(?00WFw(+uBgTOW~7q)ihHR%9o(G7P;a+=g$n1vJgcPf^l?2VSBV0<yp4K>enMmT
z7B%%W2Y~Z$)nCGAm!IjkHPobV!@b50SHmbWhRpA9SF#|!9W0xS&`uT@4`ylgtJUkQ
z=!t1cX577Ds?z$U4!)IIN$Xd5tz((ucFEX`yglPPS*g%L#(VL&^U-HKCl5OGydoF?
z#6uHdBpv--nhRijyBZ&n0-b~6M+~A;UeL}}MeYuljNAE)f--*CQIFcSgJ@D8x-qh|
zNK=R^S2%pId6Hn#sh#Xed7~lYjczr4*ky_NuYs>)%cV;jK=JzC72jwBlg<ZP%*i(f
ztL+ztQ>2;}KT{@Kb+$;o*swATf6<y9`jRX+<;7F#iLRIAmx8!uuSS{}Zz#SMX7A-Z
zOUKom@u_7&>(U2!=!?<GcJy&#kv*txW0d!+z72kxlAe(}+Wf}JpJ`bYw;o4cHCKr1
z1}x+C{;OQJxUKP~bRL){2E*{=RH;#>{Nu<A>086njJ@}nPAPCSOWBaK_~d8q&Tnf>
zNBzf{B3e2%?4G!#Zy0V#O*DF@j3Rol;xOK6xqqhS7UeuySh52BFk?OMbcXSmz_*Vm
z-K6xxuP(c9B%B-`#G5Q4joBhFDTl<XPCj6;it9pa_msYeo=Fj=tL|iy+_G4b$-i2h
zr`3*L(t~N#?Lqjc2a^l8p;N}-{1f3V&(IwO97PElP6kT{eMf-Y30-ERNLpA{;iIXz
zItElzA+#g*GEP=Aja-|xtk&77s{3x=E%M_y5t6g@C%E-Ja!|vGZ+R`{wR)*I=q;l+
z;^N}UzAENZ*4kR=7MC=q(4%Wil1-C7nfWo`KY8eiMaTYSqqu%KVq<A$NN#vC%7!7=
z?i`3ZKeets0|}_L^^AEdP{lSP^PoLHepD1Kl@%YRrc{TP-K*KkWL-YxGD(+HV@d#s
z=v$Me;T8UoVY}G)pqsBSyfxkRY-8(iB^rRfB-N~hLm4bC&E!ZWF`M<3_O_tS?T24b
z!j1W<0|pSY%gR0w6Ey~JIv~;U2`NG@(Hwca8VWNAT22kDG6M)tF(>)Tdp{)J)lhI#
z^34DnTLS5)Lj7kC#5a${L+#&IzK<=(<C+X_cXQx&1DjyW{%P=(qKMp8|B+2a?OS(A
zrRQv8`f;;2yq+~HdfuI!OvMsRe%N1RarO`az`G5#X0XjQz<OJvYcHa>9kp(G;w2KW
zQH*L664(M1^bpDZqT_tLroPMUf^I!*L*p?!aKjb%gty@97E+4m>{?Iv6SeO2<VgtF
z<Tz%-j;gA}M33@#WEdq<W+HJC-G>Nuq-}y;-fAfQxHjZ^wKZ~{<sPlg;coFEMDKBg
z2Q|XD&dS1n0BI^+e%O2hq*QXgxPGFToyxs#0k>8-qA1@wOdZaBHYN~-+u(oH;sCK2
z<AEz6Estu%knOC|(>RGss&xmj%uXj=;xoL46{&}{>h1uIyT4@DJ{E9ja4FZ2mO3HO
z{o98d666vBoVw7o-u*GzT|K)V@k@{wZV>F(MCYrGdpsr;6#@M!E5|!CqpWYf>1igK
zXll%IWZ&cY1{FLl?b$+_?TJ$URK+(pI7TYg_t;x{{9}mT+H5XXgq4i!C_;a9Tb0oH
zw-!INA1^aem~Gl^DtvhD_=qf<^QQ4`$psJhNO?Cy87sA&?ZJbQylf`$ET5L!0r^DR
zuSa*_voD9sET8s%F-i<&5}9v!#M$)}YzI-#YwibVX1smW(XVJe-N2Sc`Ef7#Dj*zS
zY)Z<eGV2qHMc}#{H<?{e(K-bP^+rcUhEk5`6T$Jgpsp+4_k`N9Q?&477}t85wjS}b
zimC+KGkY_7JzmZdaP(D2seX6c{g+xp80OD?ztUV`^C`>tQj2%A-jA;<Y_1h0{>UyV
zB^}@clx{#?Yr>-kk~^EFl{Opeol$bWHtpyAjiBxz)4jl3tid>xr<Wfds19V)HGOTH
zl6<E}#+2YCz2<c|^J6KRylmWfK!(QgnWnh|(lBRjGQnz?G|iOpQgrjBibqu$vPA*M
z2ima9Osz-x=44hEokt9jk7M)+Mb@$Behd*$RnXY)t_xggU7cX!8b9vJh6+9-t%;fG
z(CB`s+u~_>){|L+@nVp_#P<OgDc{barFH|qr+r8W&xDw<i!vLl_IW;4FHa;nJtw!j
zVJTP9#hh~Heco*Qy`2r(+Wo-B{Mrn)93OvJDY=8z;e_i{U;aqEEi$}EB|-h(MJe0a
zd*qo3HWqattH#z8Tk5q>74sWv$irY~20Ju8m-yM+=bIllqaxz-10NaGJvMsM))#_L
zl~g2^hJ$}eSuu_}q0GH+Ffkm*<aLut?38kA^inS-8PMp*5BOBdhD8Z3K_XdD*7V|_
z??pRJQ0L0in2)3IU7L2I!ZF_2+Ub0(@Q-UH(^a&2mPXBLA4YSsw3f)?MiHBnWd<;W
zh3nDCn@wn{7N!<TXX8;YK73c$7T9UjRB=%S4GZ8F&r=v(R+x@K{koHIIPsDFU~X^P
zyWf-24VQgKG3x!{;)9u&E5{1c1Jl$zS<_Ft90BRw{eTtfPvS)#-g+=bmwHQJjfBz-
zkQ^{Y0rC9#8>){FSgiqv0_oI%-L1^(iFXL$`~H2P5$16F*5)O;(Yh}n1Sa23B>db6
zCnujvr}3w9h1W8PNL#cc8*FY0*WjO}jJ}X)U!;q%Tp3J<gDu&I_&TLufSTCQYXpx>
zxhS`r1BTY=x#H5He@1g%`EHpru#xhMNVbBFS^t^SB*3XSzmOPoYR|({O72JP(FD<S
zQr$kL&7!u8d8pA!J4X?$JttBZ-37{jrL5NVt&TqJSoNl3xJ4iM)|_~Z24J^+Px73@
zKyW-DI1#F=t{znz%B%_t#JskxJ5JitrHRo~tAOaxM*)(1+@&W1Z=uScVUCh_xf9HA
zBF>{aS{cI8CVPC&?-@8XJE+1)kmIN#;?Qz%HL%njH{N;Ibmt<ax#$U><@x%W<L>T3
zvlOJLBVU*r(SvY>-#X4yj~4fh*|A%!{Tzdj?aa2Dy_e5lXP3g}xDZSjA(cj;*;9c?
za8Q6>>%-+Ir*^(`v9OVzzyHnDtkN>DH=&YJo!)*h-IrwIVdBRdy_iPX2GfsY3EB7S
zG4VY^k9Q)p^X1jbe)NaqHUy&w92;g>s;|1oqX*o3>~tBTsZ%`q^zuQ|4+EY6^np|o
z-aB`Rh%hKE^Ck*EHdd~3OS`V9pY7G|RbC5-<1EYDRhI-{!Oigv6iCV9P?WNuY~kv<
zB8fP$ndP-4$yYlwaVl!SYU<cfgt^l1_$hGr0XC?CkyK<cOa`^gRjo))nc?ukBAsN#
zh3-$RySnDb`Rz;(Lmi;ACEW=$(xxF<g)s?4{0imsKPp1JnFo!W_O0l)TMeiK$u~$=
zH#V}%t<ltd4NP3<1CbPV7wG7{?y^;tGhl-@@D2{Rn%`c6g@q<pdRv)qE!T2z_I2qD
z$rpuuB;g8=q`czjg&|6o9&iUn`ay=;93?4PiltI<$CcZ|Z%S7v-R7ng4`st7XePA6
z1SkmspHn;@XLF_Uxs^fp=Et}{T}qqFo2PYDO2a_p_dy1T3I&q_mU~=kV0I@piZ}}Q
zIgs?HWIp-{?%j<N$DlNR=%;JiB$oWm@3^$<bzXTC8wfzT9p=}_EM%?)U{%4>#~&6n
zMKzX^BdpBHQ*GDseeOxY_Of*NAM<#~7_Rm#u{Jvk-z6PJa|J}|BEJGxSWBxwFxTZf
zt5v?pyz3xiy5P^+W9r*wj#-#wnPFBFFnjUZrb^c_>Q$%6xFBrP63H8B@_ZaO%^~RV
z;kheU%7^T(QfQs5&bsj2R8U8Dt9eB~M8>ZAFJFK&93EuHF??RoWqjkO1X$uLS9Y2w
z%#y%UVEbc$a9DKOW|oVDRjTAM#O1PghA%GPaiH^fTZetE`kPaUI;o10rrx~G+{a8D
zE-y1ESKL|1BuL**l#SYYRIFC_p)5L%>Uxd+RMcf7gbaKJW=7a9N?E|qAlsfn^<+>;
zc8EeOf~5+*i}oj6Wxs#I(f_&#mk%SGB{$n7XC|u^=Yf!qRSwu_D{yCtHW~x3HTrnN
zd$!2Q1xG5DD$oS6tSocJ>@}M`_Nw%c$efKDeJbGt(tVZy9{lvF7W~s{O=jUxm@ddP
zSn?srq*FVx`CeZY5MX8RjLTg;(ooNf@ospbcS~YqwBY^Dr`E=rQI?|lWzU|5X=|g6
z+06-_TK(DD3$+ptig=E@CI%mQMz+)n_Lpg)GzMb^1?zig8}i~g3MH-0bZL$&vNHn{
zB+4@h$kT-M$1IF*G)kB><}72Z$UJmVqE&x^>>7czxqx)605Ra6w!rH5(YzHH0#O@n
z|Gk3MIh1L-kx)bEIw4iXzM9{2hKUv##^;nx&Ce;Hw3TA5v~*an0o3Buftwbek7fc*
z`mhDC#k=8nMm9sK+RDOHKVBO@329$ExJpT8_V9w-;meejbbFzrt4O)|lIUS;S-8Uz
zP?&)Qe{V-Kc%7qZJDIXNnAkuWri-ufWC#CID@XcMYXu4G{^Dq~Dk!$$=M9Z6^E~I_
z36{~i<<%nwJ)`k~?90rD9(`PyX?CM42Fo^!LNn`FP^mnpM1o=GAX{#?HPMocpjv5F
z+9A=|@<*(8$hlonHK^=W<$+{&-TFlY!Vn{*O1Ejel{+gQ00ZTR31_{=+vZX?Ac^nn
zY7t7W_=H>spL%dhj?^-$EzOGLgc&&}Cw-#e-N--!qQYGV^`NcU?$+(?hgZ_QZu3w^
zSzm-$NwKKQBva`DP7a(*fE!!OdH*~JuCIhCRwl=(#Ak2hsZi$FRa?3TSjEVb`3jX`
zIUH8uKJo}Yg8Apman92h?4S5ocTDkSrv+G{&#2lOBTu@sZ(!X#!~){qof|uxa!M6O
zm}GQ2tDG$Ar&+DA@JwyN(X2j@xQ>4%oIG^)8-abo<a6%&f_&uE48bsqb8_we1x@v1
z<<zkB6J<-|@i((#FY@A{lDYI-fTZ^Vx8uqMiFiXo#X@OtAAS%|UIts~O-fF?c|#(p
z5ifyEQt56$Cy$l00g{pY#t-1`U$TQ`gF#MVVC&1jWdf8{Ct#}d1fWHCb3fRufoYR1
zA~2iYFt2Odr;rb^nyqmP=t=UBGJ*VV{vxoh0ko`r$T97<KKuyDdYkWLy}|Hs7m$R+
z7B&dMDyl#TW?NirfC#gW+)X%Pmxc)}N8+XNOV-L`@&Cu(TSrCxuJ6LAUj;=%K{`Z0
zN;;(_1nCB8L>i<!m2Qcl<I5mj(m9|YAnlNnLwDB@1M@z(_x|nuo^#f_*1OL6@BC&h
zm#pdeJoj_Qbzk@O3~eDB_Lw-HoL9+Y%vkPSiSU{VJb-?x+Sy^ZYbL@v{1r!=D;-*;
zzZaNunaAWtl)^BWts~~ZG=R)u5Apj~U$e&jK1K>&>%k;Aq}7Qm!|fpY)0F*%Z<^~(
zSN&Z%N-<j>@eB9X(ZB*u&slWY$(lPMBj}PGCfPjsAn4Yx+@dmc$7Bsuk$9j7i}PSU
z;3d>~EQxx%>}c1>L{obbZh>fN<PM5=v?nl~>#cAd_<`RBCAZ8Qz1iPSyd_D5kyDe`
z?sH5+gq3^0TU*_=M6Ls?mQE$c4BcjY`p)V!r((#`jZWBX!Z|&7Uq9_S(ImZ-QdrjB
ztM1GCp|Nq7xKOv*%-tZ3HIw8db+GN6AD!>@Q4D<~hDYP#=9XVkoe!FOU0MQ6{+ikw
zMC<$dYKTD#w>Kg0MtKC208i#{J-k|4mCQugGB<vhB!=;wDOke3q}W4IC@NCnnO1J;
z=38<IAQQfKk^$h8%L}o1auz)k(VI*RP&s7PohPrCy8|VlXzYc*E1|~8EVQNv@h>m-
z=9keT#Z{8JOWuHL{*U_ZE%JNLTlS<oxCtB_9QYI_6rXLXBjtE^G`dRXVCurFs1qZ!
zt7P+?zb4w`063mp;eRWK$CtLb(0A>i8QSK}4}-Mna(sPRe_A#-vA3I5d420h?n4do
zx?{G&wnCF_`nvZtGy&VM&Q1?&A*Tb@BdNQ)f0Xl;MEqsm208!SM5E);;rd=R`)M0`
z^_K(mTVAbww(nKBwW%fshV|xomjyK*1M&N8TZgGmP{(Z<H3%p|H`VzSe71V#(yX5Z
z$WeLFILcZf{905`%%_!!%snX(RU$!OH!PLEd*<ZIv}@3cfteSY8}$004QjqpeK=?M
z$M>21{@lI`-(y8vPPGxO-s(6{SXw@6!34ER_3GcvHo<6uzJGUa6PosKao|`4DYwDG
z0YA&EY3XU*SnWypKt<);50aMRI~)v~C4%?9qKtorz=%$E0(;7mKgdUme_4ScfA~tv
zFQS4OJ#+rvKRfKc?o>3Xa^@=@HL5Fnf7}6ld2!>rnq9z&0SR=uy)R5Iwc(J*AY$_&
z%}r&G3Og%eY@pGtx2<0q?6#49gRd$pMY#SesYePDivPO%P?d!y$P&^JUiFH^@vyVN
z$R|~w<VnnCsg=u%P5FGrYFaN6UR9<LT2k1f6Hk{H)b6b=h_1WOq)kX6%M9eUkM?@t
zl^j|mpC65p=8*#g{OxBua>_m#UzS;WkPrc1P>2HE7T9~oA#R9w1IKidPr{Gl7Cl`#
z_?8}>qEz3gW*S4%st0dlYo74b51gyTO5Oy^JX>eLT2u1dPk;_7_2Ti~q|*^VjUHsv
z6;2;o4!XKD@3C55rGA>spFEW@gt0XKeR(zuz{ZWTK<%YZjqnru(Drv28+P7#mJcT4
zXpzJUvFpD3vPgN&Rhh&u_CNgUV9iehX_NgKgR&Ki>c7n(f2V#rj;?SMJzP9oK4Rf_
zFPougkF*D6n~TrndIev$(OaE|7=Po67@@b9+SR%EeVZD-+xakL+tOTK5|;QZRflU5
z`R#-o@ioYNLpjwjJZh^vT4BX0uA}p%b@SO8{PZwJ-%lM2NB+wAc_Sa@w-@vQfu7JC
z8pI~7dcx&wRwfUeH)rBhkHInquY{V;Mm*U%4dD>**j<iAo)~P&-6RN4LY^v}j%(G9
z6~1(gth8R7ee3%IoL5J7s;hu7O#AF4--lNlMC1-YO5{EE<V^c)b3w1(=flS|1MO8?
zr~dnINR%fTz3PT*P@BL)D0-r!V}e0rj4>*eQP25>oZBBLpwHCh#UIpXXf2l{>^+xs
z)kzh(AwB5=$O&#pk?yYRo3b14ufGU9__9sK(##Oq8`ORlAaDw`9kF`MpAUtxa>uib
z{Ajn;Lm@p@Wy6+~*o;Wak0tb3$bu7tfa(2EeF0Kjg=b$)>G-t>nbhd9@=r$ipJ__4
zS-0}%Pd}unFjqTVQfKz->%1=n)?R!IoN;zi5y4UbW@Tz>b}u5@X4N?eKrv(5e$Y()
z)_>Lb*{58FzmVd=o`pW!of>x2r*a<G-<MHga9SD-x%r=U+Omi+l&$yab=D7sNM_z!
z&Z?)>qn?!);G7VGSh->T=eDg|aYorMNr<oC%}eRDR_<Qk!qAXKSe|(%fjT_n#1p#u
z072^I<N-w`Slt#1IM2)Ce$kElkIH+hQ1tyY%NKj882f<_z?T@V>tIDcxW>RJe1n16
zFG)$sa!Tdl++ow~PDu`u#@3|@J019gCGLVqM^Cz&Eq=r3zi1zsB89%0&FjI1Yn-l6
ze*74bjo4%fd0_Pjq7jiKOK|VefUz<+tC(3b=#uY;vbR$7tx5hQRE!1w-sul(D}PJR
zh+T+9bm|6AAz^s8d<0NCDOt(qH!&NmsuY#f%m;|y^Mpq?&8Xpm#ivos>3lA(bkq`h
z&(-;upT&$Z3az05NE2xo_g>F&OWN!(T#QT8OUrUNhf>x2)xf)nXR2KzJW$PYdV4xS
z{9u<Ks#uQ4ufG7aCegcFOYWo9tGo((wG(95Ls{pRv?q%sahud*vfg_6;g>!aqbFuh
zH$Vl#+rr(C+3L?Xe!Meu(|=HIJF6<s`ojLO<~HVX+BQ_}<a+BD;|*@ZleT5_?V;z^
z{FMHV+N3v?XYG5FM&Z48v`owu@6{JFRao-ur3_&V&8Pi8ey^O+B)XAUe(HcB3ts-*
z-u5ecZ;}?M;P(JC)k_$qb{Tg`@Q4C0L}h9La0$)*v~r8>CV!=;Siw`SZ>bOaJM5_J
zsMXIL{T!jkt}uL%Bk}C@_C#*NmKGLG%Ng8Z5grWz-6v&)^U?)CwPQJwsjzcuIxsLa
zeXmDxa)ZUH^XYwQgND#-aD~(HPToGT6F-n@6!7-lDO<1{3gKA`Zf!eSi4Pv}G+h6A
zCtT9t^m3X(Tb^{X|A36X&O$QwNCHFL8EZaZ)e^<p1B89ujf_Mz-AkuyHg#%xzsa;=
zo8aP!m-<y#Xf=vZa5Jv>JXWVP@26XJO&<IhD<u!LC=f6}-ISC90crHPPowTB^~FB#
zvfxT(w8?Ra_epG5sk3*?2IQV$(B=kMU+hZc5U+G>_u;t$76|$epXf4`)p>DA0A>{M
z<QsrjycB*V4X2?A0tg@=p+66zpI|K=)@Mwv{(QKCwZbJ3&sE#bJH?f1x%Q*e`kr-E
zn-)PtJeORgbSjzjicGge>(`B+wO3l@khpf|aPWyp;Dg{<6;#hgo(;FP4h#gQ#JJ^3
zMI8r|4+DhV>f(0wEFG7?hqF!X{GxxG$Z>TkC4q60lX!U#8#C__DaDaB^0qUOLf{6o
zg(>tmt%|CV+3JnJPDw231<J`ZXAw8^j_=`>efGf)fV_39>}djyNj@&u$z&btbOP3(
z07ZN4B&oE4UxT?~FpF@u`IEo%P;b1?9I8^uB0rQ?_;9t*HhY+sEt1>gVv14t@2mP_
z8JzH9f;l<h0}ijC2-A_o)<^B`@ztx<{;0mD-q0DGY?0J_$ih%oiwSD)4{dYf5|gL<
zVJ0!?Jjr_ISV*STI*E`QhLKNuI@ug+t0vcvmF7DH(^_fW!mxaJi-IF}QBRfE?<YcP
zY%haSKiOK`fWSWLWI-m-M0!OO7Yu)<>E_o66&%!Bwj^eji4H$rO%pJOx_CnNgPZ8M
z1$y}iS!16s^IPlsWf(r9=aQ;MB)q}tIp2r8!-|JR560x(_c3U(d^LHcl@_IP%pvC!
zrp63y%XJBt%c079^*HZJneS0>`mz6h&$>=4V3uPlB<|R$e(k(zLWZlU7E6m7#mlJG
zg>Ibc*#5I1;3S1d&^Fq5l7i&pPm|mz1;|emUHQ}TC*kF=CWLDhkG`8(sRI12h+YN3
zyZLXe{hYRXi&G4&sT#;?7vOO_>})8z0e1}-g)qR}vGS^m#5=oj`Hv>r${nvtM9cj-
zl_~H{c7Ei};Ww2K?b&HGTXiWb(^LFq`A7_V;W}J~w}UAwdv%u(KjW4ZA3AWq>sM`?
zv%f!I&?d3L(6vF}lp?_#C&U3C40f?7B7W8Zy@8XFQQaZ?_;d19tDZ9q^PD9bx+)Iq
z5Os>w)m;oQE^u4G__Zs@(`#w`eXrI|%XK(eH4CUe3XDAVEftLUu)i&|bY}yk>AfLe
zW-`|<E}yqK*~m^zCNA;arwMXcJ4~0rY7;;^LM*J_s!~birEbCxK|q!%!co^YyY`fP
zC)yj4KvvA7S|@+dz8=d+@+`_iL~8L?GSKvr$vckLX)%>Fc`ExJ#Mwk1PQj!K4X1yA
z^I$839jH*fwMfiaZ0+(6yIk`Upx})0xh}KJlhnMl0?^y4_hj_SN<TMf_v{8Ae2v9#
zmt~`#a0V#>z~sZ71fd==3OBU>!hR`+?hb+Y3XM^zv>!qE?d()~>M$MbVwOddn3Y#_
z)qce`8W=yS$+fSK(EHSYp$>>Q@u{M5>r3iKDIAzm$S2Cqfk@!JpKOpT7@X8+kk3xo
zT%1~*7(Gz^`0|wfQ9r*D*61>#^vimN_<%oUuh(Aq<svr;(?q6yGjR&g%W%4~QPW4i
zYn4?J9~$dNvo$0eH`wTU{ngK;{_^LjfA_D@7>b_QRVSJtTO`Z$mZ1pvmL0(+k|+0^
zWe)=0u6G5SSe{KGa@)$Dd^?yf3CHwr@tN|vi@{+9$^vY_qrP&<vk%%oN6_o_DH%zA
zv*m^WTs%Kdrq&iahvDF>xuxvJCI$w#G>Flq;prd#0qsO|{5AB)^#a85gXgYz)At~j
zCavJs)*GtI>C%MOxZ&1JkLg6ndD&)K&VoCEZ)x3m_(3SIS5jfcVMEK|5?;Xx*rGgo
zf(M`C2fYK6bI)t?`bvU?ez`Ce6op<CtS9@zb2r9&KytP4MxiSve^-M^B~eEeRwt(k
z<M()(TKY&6a964O)*`tm0jj6^?6s8`ORZX(-6Q_E7e=QYPb4hu`$kXBBW$+N^9i4S
zGk~Rmj>c}5AjT_~lh?#z`1q7Y!U))UdqY3pUx)<(mNEV5RQbXK3;JaJ`u&@voI!Pg
zaf^=IP;*ot(jNz1j;T{H2YdPk$;6i+9Dc+Q0&)j<rHXSxYIHm)1w>Q`2QMIOPTp+9
z+M3E|HQqA9bi+}!r~Wosp{3T&H-K>ER99EOnN2+*6Ow;<E>gDo&`0K8%2)=zEUjo4
z?!$*?1<YQi8s`%^+sUWPq)}cIcW)?o|1^Q5rpvgAEoOghEQ>|V#uJ(r-MoFhyCY1+
zkkY(Uhb{cqPPKg)+;-kb7FP&+O5JB~j-cdXLQ#E(&qZ2Q8JK{+vNeG_$e1z4sd|=Y
z+aTu0+zDqkawu3!>-@Mep}97Zvx1g%n;$-tIUhEq5jMzkK<E7kz3#H6w`ufZ3er8N
zY@^DQxP$Qxr4ip62RpG*Mmg!Gzc_P+TwT@F-Q^h@&t%wQZlez(244?T!lO}cA@^=~
zyfuBnNZ;hw@18Um##;E_ZgH2G&gN@&pBj)qZ?HiR4DA@G3PSDnXRduGy82z`D!ER=
z>i=+2q$Pg=&}fR13jMu_3ldv@FtvNYXZ`_YvcTAL19v;yTlIl1o?wy_%H(@Pi*K_%
z8@DJ&W2SWe5wNV~q}@H{9UExwC=tT9X1n^2a}o1eEKS8@Em;jJTt@Yz5foa#cI823
zs#MFH75tcBc^T{{Ob$Kc(R`l3B=@7RiXW;UTR6o?11)T1^n=QYxCYCBkY7}l3;l4z
z(S?-86Mua=&6h#m7y5Z<jhp5@)HqUL94a;12~q)cvKiVGHLTrq?ws-PF>(_aRR38L
zQLv5F<$GX6xfA?;FAtYs*7kj1U<2SrbhHux#U7m$m<ViS7Ic7IZS(jLSsv1=r(JU6
zF=X-yf4H9l%fB9bWND^^rBd-;$Iun>7%+YVZ*fH4(3&Qefpn*EB{w;)QWT>fOHeZl
zk%1A%UroCux3Q~e_~B~GSVAB)r0Uv;NB&O2A!sCN?4Yr&&^i8I<hz}~uZhzLCo`d5
zknJ3Qz~Z$p>rphdBagl5n5&+W$u~!V>nCtMpL>np#7yLC%Xskj74<lfwyO!n>*nTl
zl4Owhty6mz9{U)NXjaGG2AIf5&~Y?=tLpyNW;jm?%IP3sW7BUlWazoI(OE4+p7*yG
z)gtKv1L;C!?c)*B)G5cFIjqy_U7v*aB^EQB-PU@lGMR+`Nyuf^B>K`5VoiSwgRBe;
z4F`BybcV1Vu+m<h?o7eZwV(~MgnbWl2v+BhEDz85n7-M#q|x0(tuQL$V2ddC&Yi*}
zFoP1gA<{{fkshh40VC#}bKTm}=-K+cx!FtXdO!58H$w<KZ{sbr^`j|#E~-QbIdwhT
zTs%u?LuM4&CGS#ak0@V`A*T>d-e#=*lEM3^t6)aK-M>&9mY}AlM4MeHG$7jSXx7Wd
zYFJ}D>^$GM>iz=!^$Hwh(*wX_)}M8?W?Xiu3X@V%Ro=96h{8#i=Lz8~OvOi_c?WXJ
z!uUkhn)}ilPs|_#jf{`$NeXl|a1jkeL2*6rll1F36vcA+-;e$P@U$f{nC$^M#Ar!j
zDDY=DO}n{|a6iDWyUeZdqo*6@hgvKRFzu}AEX4CPhfOZtB)UtIlS4|!ws+Dab;FtW
zZW30soDlx1I+aXKw*SHJw6nI5J<q=(dFU)YQzOccT}~1KJ>l0M3kj2mSu2i(1btb!
z1m@8ed)KC41owPiwMjpr`YX^ZRz%5-)5a?^@-fp~zlVji_yqT>Q-jys1auES_$>(p
zd=z-p?w#;35Zc4|jEceC?(O}Y9C9J6hrfUSb$=)(q`C5ESNyak_l%B|O=NoEfx!B6
z(iXRMd7}yg=^g4&Krgk>*3@=&br*^KJm$JthP=vBFNkSY&O|z=-19-NTvmjfT}R?_
za%Xyc-g>KDG}|W~PBOhoxo%l({N>VLqwL^L0ru;3PkUJItpw=>nKpaLJU3tPob)Zq
z^~q0h66Hu0NY_>*u2s)`6l*ZO#ijLLf%P&bhv${!qSnX8oVgCjl)*YI#e>JfZx+#I
zn~~izzI1q}LdyoVzR?#h*^h(XYWI@4gCVF0MkX$D_i99aC4-n%G#B0^%|$R|UwX4-
z7p3Z?M+{=0XjwI%ZhR_^tu3U(2&m#R<-e6?Ggn-2+Cph$&hL@$4~|{0!fEE)sXz{Y
z^DN1lqJB&~tGNMN;Ow%2mtUo`#Uu)NnD50L5)Y2)PeKj6lMH9PpEn)7Z9LhTs<C?I
z??0)kdv3^M@TwBRPr>Vigb&8GXxnXEQn6?MWJ(;9k9hXvLzb_kDfe8-UX!!MfWkdJ
z!aV!YD&DTBp|Fe6(KkB;bI|g5avJ8=nk-9`V?;yCS*T@$-YRPD*Cn;z=jJ3IJ3{YA
z!TM5qbaHgzm!*ysI!<eL2jY&Pq337r^;-75GvO1h8ny%PxYW2_!@dHwB1S)2UN!5<
zFMy*?T_Ir`VNO5lI+_XH(uusEHE6JMy5|^ppXHu0S4vQu;BCdVAdKju`0t$2Xo*{e
zv3qSk=rMeRKb$mXF1gj8X9o<{Y-G9i)lsurA3F_QO@4|DH?n;CXlYL4;&^@bj^qc#
z9d<Fbx@=dmDQ%DXaScb2Z=(m><Hu<~U`o{@zT|cGo^^$)Mu@cb!2Y;<t8Z=J<ax(q
zrf#RrNe<bu5-eBdAjRRY^1o&c$P8nS(+TR1j3hIaH7bbhq*L6E8tU@47vW8FJNVu+
zG8Af7!cWB8w2!(INKHqD@M-`$+`-uy+#b7|KXYe(wMG+DF-;FGBqI$JL<fDU8mKS=
zBoGaGjJ~G3r}s&bM%W=&SNBN?Qn+G7Q+E9X0SvM+=bDYs^{J|Ft8d<K`BkY%pl<M1
zFDB2siC`_bDSLMyah#&PiTW7uH>aj9fW+Nn^O|Oib=gR_%&9gjD{Tmofjs|?e{qhG
z(tzH{=(Wh0P;BV$k$I9O9{Vx%G3_=RD+37YlxAaFd7gKcE^pgrq#n|cXBKL3iPR{s
zwHe9JgJjWupOq0(=OKNb#+-m6p5j-|S9_C6S_IXP;uK<fB4e8f0SkLGWG(8wbJ4b`
zN2{wt=BKJf$f0^w&ArO7h01ehDU+Dws+u@-9PDiFpnhMm5;}+GXjF3_X!hJKO`Yg?
z@fdLRf=G)-aW+W(FYHCDmItSyU8WV*IhKj^?sKOF=v7*Ojy0rtzD0i~FZbk7Xz&3w
zO;BUC`5f=wD)cPscmHve1{PdM(+!fF5tD9#P`TJ1)$D8K<b)`A!xdX>tgNzL{`&ZQ
zp&~|HU4TZsqqW-0@tDO0j>fv7UF{U(7&{$YCGWM7Encz<wx7Q3Nxul5xhz16?C0L0
z#lsWc&-DcqaGFdj+k4%0#1HO0XxwhSR7MK#AFk_c8SXZsk6V%0udz2S>v(x6INsRy
z5K_Ib62TltQjvLjSo-dL<n6>v?TUZFN6jSsGf3Y}Q5kX3dUlY51M{`<!X=E=A{(yc
zRrhJweEBsW`70!@oBX2pbVOzB+n29ip5JG0vu6v#?3YTwcfq9c^+a3+CMQr|yfi2w
zC$l_QXuKO+fxFv`cG=4%e%&qTvv7{7v?r9xYFRgLXmN%SP<(A$Xnc4fb(78;CEyL?
z&?Vj)nHgU^l2AsLUcK_E%D3Oa^J-$4CQZVW2L!;`k*mN~SK8Q#6QrBith)Y5Ab|!Y
z#y8j0%5;n}4*Rk40xj2U--9!_HXAN(%uFSVi>oVZgM|UISbC{(&TbB>*-TV6<+`yQ
zd8eZFu($K`x)_zo0n~G+<x4qg)7_QT?rV)_iJM4m!{zpYJ!N!rq@Z4{r6HCDIaB`2
zJ?ZKJy_&6}t62-o7;4`Y1*hh+OF1Oe^Jq}??T(xMuSTd9hmn`xQ^LWVi2Q*Xvv>ta
zymiKkX3<-h>C<e~8bZU%gz)~s?mnh!zf6WR=j!-R!#-{0dqwm)w4U1Mn_S##(o7`q
zE+!#4^ViICF7RUM3dQo#<Sc5BljOE&HIDc=BGtz-&ZaNo^ODRxfAp0M&1VYPtOCoG
zIw0@ap@NwKTZhO=vs^D>w<+O@NU!;hwc0Dpx@OqMPq(<(M<q{Bi5H~acrT7OXKfhb
zihrWgYY>%`9FAN2LHJ=K8C<DCX3uH2x(Xh(!_2Y;RDQp?2<?+w^_oHcETw+3VzcZt
zsp43QKU^<;uJA7|K&p&kp_UNGUEyZl>ehm@r`7pu;^*TV{c(rq<e?9Q-iX!1zRZr>
zZDidECKYYeKmC@L6;YeVt(V1G?Qs;h157Q#tJmBcx3s8&+PL?($KEXZ;JC3QoL)P0
zaMu4q|9m{P8810oU+`<?-wAsD4W)DiraiPQhHSkPXY+vlaf!(i^`t$cTqX;A4Gi(K
zt|9E9x|zed2^(Lli`-5tO7sbNJt6Ag<ive7B;xSy9kdZR48<iT7tefL)WFBE@N(k;
zgUxWNX?m|HTpRr2vo*csk;;|dYu?S-Y%kf5U6kE}iNA2`#nzRQVFVniL3SH)zG*n0
z_N;ghf%0`&q`-}p%-tQmkB7#WI4fzpC~-vcSUWg;EwS0Y9H||($U5s*na(W9-Eb50
zJ7(84B+!&5zSr%o#ADFX<?axnL|I`*SyC~3^St-C(rT&~lF@oh4^Cv=&IZTO?0%<7
zFWZPuC|reBEeeueqBY3~@9Ih6CC_+Hr+D7V_FSaSeU`vP)l(P)<%%zHnBZ4R+BpI*
z(LT+L@Cw5O;mrt~S@=awHSCSkuNzYYRbc7{>48mdSTt}-_%M{0r7ZM_v<Q$YHt9&e
zIbPea^p3yOR0t6cgPIN=YYD=Hdo+%`6hjcsDNcf$DUt*)=YL^fX3PhP=TbP6jo-K6
zU?yGTWDsN}P4hK>9Kcl4gYd8I%=WKDk(vuC4P6yDo7OXvmpB^W+z1rmQwhK-lHw$=
zXs&+Z?UhtAsgH~Kx_8Fx4AnP3(5;{wg0?tSl1+@^e*@fq6JN)BGArM?rkR6$!L<1L
zmR>?xQ32<LN&GaTda6QFABnnCz&^2G$xi(5_ZNZfy|d%ZTN6EOkDp58lD-;o*Us=T
z_HQmR-n;<h%+;--7J@63nCS5u`kLLp4R*qDQtD&UjrKl5)HXlpvRmJ(AN4BF`*C(#
z;M#V5D|eDQ&mek&|AVJ}O+5LBiwl1%>pg{g{`<R#4|6nr^>-4N-zI9)Yvta*MPZYz
zl#wB4FLrG_Hhu-0tp!2QkcKCyX#UqV=9Iy03?9p_--;NP_44ey`c$4(hgbb<FkjA#
z9GmMM0(GJq<-Co>TK<&&n^2G_uIk1LT`?W(@9CE{`xBqOt{cfpX^*x-ZXTa8*{D_>
zax9c#T|A40`sajfIH+D7kb!4bPaM{^oR)j?%Sf{|9@U*8430jqT^3JG<PGb<9i&Ma
z;qk+-a(_iXQ(Xd<z&*mPMvI+G7~B8z6BDc?nK#Qf(9752WN|$}y~Zy0AE5{XBR_P*
z_ud_ht{$Y3OU90ve$pFl`}jc(B}8Edb{|zLZ<j3Ga+u#(aPn~Gv=WaCGQz*<@cy!{
z^7R9nFWl4Jas$8BBnQ9~bl+aJS2eF`*u@Rb7`*h&#Z`Xc{azYE-V>|yiyoY0R`@-;
z)_xuPZuLlddh0C(TD@3*zwILlocGo>zw*hflUi?K)Trjq-o;Rf<(4tpHmtEW+#pCc
zL}{UokC#53n`K!@CEBQXFC`I4-nqF<F4&WtdP*nh-C8~KB+f&qgIxcEyfsY2s}=jT
z!L-z<{;>9^@$418ES(0BfqBM=)IviOo<-n!v4RPrq1v58T5pl-MGf)&Q5uJ}W6_@I
zh3A955ttE|FDNfoC;3~1P_N+a{g(4xmh-Y$COw(;ys)2d@{)W5^Q%p61>i0OVGxu$
z4;o{L698X3Sh*8yM$<BLB+jSpW7Fp1rDc5y_fF><hF?<MrzyQu9lE5K4F0TSQP=jX
zr83}+De#aq$0ud%?w@<U)Co_wkVIP4f52Y&6*A|uq2e0zWX#9XH#KW`Lt*o0(!f#$
zQHlZ)wPBrrzK7q^{7MNpjF7vYExXNYsXafs#<7Q%Tv7mKy;qcj7lF&M;n0C9FH`2R
ze}Q3~)~w>?cGtLS(rNn$2pE;K`bU^;!FxzdAC1k_*R^}*n5Mt4x{a{|L`;nzq|~2f
zm0y4V4$!l0MCH+rh|z{2kNq-phkWuNE``{X|LVm`s6l!Yacxn5(fel>^CK=jq%ZRJ
zOj~^9h73skAHHgVtPZA4CYBKnXp(!OJD@7D_!3TGQ({6(_xl*h%DbLe!DkAZVi?TG
zEzG}w9BtC(yox=x-YxNms_t(Cguq+VBK6M1>;2mp96c{-zDaEcag$(hP_?2#lePIO
z_?}TCxXdT8`E}n9Q4caL`jTjITAhyKE@DX2v+c(1)az|O)R-Gv>o#0C1TGLq;<1<P
z=}&dh>+$pP5sHIu-t@K82J*UURvN<&Y4qX3i@W*qssc7ro)hW2lbNnNDZ;V9XD2c~
z&r(luiSob{J3S9klChr}wj98~D5C?{h&iR+dUp64W&y<vv^d=rP|sNy7BOOS_vd)s
zc)4ElXysL+FH6BdW<+-zCZ8iqXLja#uvqK<mYf0aO7#dd)_P=3bj$;9r*NpshVW$n
zCU1BsBbu$8r$!Uk{7529LYX2F$eIaf;QwtPxbW~E2G39Ae+e`xZ~>`<{1EA8rJegi
z$e5Y6k_?baw^rH2_2G4|?&gk&AMg5XoPSjDa9>=<Ihm7AYlRdmYuj8{6u454X)VW(
zs=d$a0*+TSJ3BQiZkr}z4h!g;9PtV9VPHfh5ZqP}w&1|{TM+SIdU6~^So0VuTtquw
zn;OWx>O0-Btxk>(9IcNlOG<3Mi)=L!v!|_1bu=|xuJahD$+3bD1S|ljH>j^qvNQAj
zHS{=<Mdi=>_^5y}fPIElgx@~%mODXtAd`3G6_5%~`+5J?^DMB&D#3egpq23FmLd!^
zz!aE{R#|mxZQ#Xj$($#RmyA5mueVWW?G*`qddHJivof{edE=Vk=g9B2uwGBzyqRwI
zeQ=UM&wpMpTyVkG^;gHlO@YTa00u@XcH<}Dn+v*A{_3Ne%ZWO)VAP}wr3A=M|GFin
z!FT_yne)$k|IYlMg~@)w|NFlGz6OBr@_*b8xOUh8;kRg~+&;tB{l48u&(?p~C$Pwo
zo+!6}%kcfnaCZ`|z*!m?-#jwWQxR^-|GN7h|EvH0|9FXl$XZ(+be2BK?k2r0DVYUq
ztpB{JfBRp)7<k10d=~gSAm)D|O!(jTf$Ng=XKns*xkg4nW-Q$-k0CgFoi(<@c5B_z
zkMj1E%?h><$R&Tlujgmm?A-5*%Lvt=KmJ)r_1OQ?R18RJK*v$nHUIT@?v(@L%INpC
zPjd-|-~`f33w=(kT-W^IZ|9rL7==Jy?d_D`Ue`zW#a`Eo=*8CDJhZc8fIg>W|MRy9
z>m95BCY#sI|5haX=XwUXfsxDqVQs}_{%5hx|NHLi1E3d3n(Jh=%E1!VjI<JT-{k-Q
zd5fC(-`XwzuoMOAOC<^$_xm_r?^mX(`U-Z*YBll({w;8J^MUH4|63pQ-&^&+-PQje
zm;DdJ1bOta@aYbIyI$LwS<7-v>@2Co@7BvcyAqIt{l`r(@Wy3rs;3cfMxTHe+?V}P
zSo7wl?0onCK9%5a*PAp-rGzt=-;3(<^!;EsW0`-ycr+AZXYlOxsI!LhFnL9`|I!K2
zhnh{VpU~m{k52(FWCtp0;tbYL!*W=CW7JuIBtW<Tdc-+8SwNLpI+X4XZ!fdr505S4
zHXrcsg@I_+RIgh~DPM9QjuB;+h=rS_8P-2UivRVtE;qVt;ub~pVL5Xufm+8el~{EL
zI0G5xE;`x0!8cr12xg4N4Np$~P{<6WN~_I>Yn2^u?~g6uw{O3!&ZUme<>&wFFN}bL
z^q@GOf4wRAe^d!5?EP{q#^mH7vi=nz+FjI0uQ?RXxanVI=AaRa8=jPn-e~k{W;GZI
zDbP>yO%JKS9m?Vs@=EES$KnkNaQV4Bv;j*S@!kQ2UF!ev-tsr^ZBFpArzPoWh}RKq
z*Q5oDR@;-3R3E#Od%UYI3Mo^$lEq<OvagK^>EH1PuT<IEfsWC};c#$mP{628&nOp6
z+u5sv5<@{dfg>L8-`_i+2^0YsQ~616R{O?Ftrs15?{SV&DRE|r>hW~9J}Dt(UY6p?
zbBz3v?tZrEz?LjuFapV!($Xe_JqmB~1<Q88XM2&@H|cENExG2@2pTaG9^hX1p^&JX
zMP5S!ZYt{Kl~%Q70S-++%<%)Ges@u`8wN*oEgSkjvYvqOn{30fU|cbWo!p+&$MuXL
zyQ$;?n<bbmamQ!9SA{Tg|HxSko94^Nh5$uQlBiz&g=uSD(pUyGV(x<R=hG;<w=D-3
z=gWs+;z2I2*qsE%;boB!QvOgnlgGE$jOQH%O!aiq<8H#9X{f?=hF?kQjMdW56qPeW
znJwNY6Sx@p^+YR({|Jg42RC-Rqq;<Rk)VvMI78)odE?UGOkL!#Nai@LO2^41tS=7c
zh98{UOaNF667f#Jsd45PB*#V2z%5bt)8j?bGn%06?90hP-gM}X&r@p~V^2yr>A3{t
z5Wg&&H7B3<>`>Ex=m#dy%NsUVH2PDINFvOi(FR=R8+FeJ*|1A$XKAt&CsM$yydy0p
za|f6-8*8aSf|$Gg92Hu4Rgc?m#5KmoDG$n+y)N*1z`O#rE5{7+0@uICC%dQ?T<0BE
zr|TOH>^JO`>w`&jHU%TqN%lg+sPDMtVi5nfe#_L<KI(GMhN?_(Wk(NI@c46H8!w)m
z0_@^xuW-xC<nve;%d78Iw0SYMv-;|opvh3xSyAyKa}s{6i7~a@engJRZp5meG&s|Q
znj0fe7asn$pzZHxy&kxTaOL9f&u%*j4_2glXj%P{AAL0_PZM-EmhE~mt+HadcyXZH
za93z@-lil*)W>r^lNxz-I<bamCaG`R%8$Kno)$f~U-L(u-jn(3d9vh?@L4*SKf~<7
zSm>$JT&2>sC#qM^%&Y1Aw4&VBZ&F;Ia@Rob)OH}hOork8@$Q+1ve)H-6Sw%d=!p7d
z&)@Gh`VSO=nguV0xbMx<+NoR;UKq7qm4M|HyNGTpD?wMC6Wh1=i6-|+J5MeB6)o$e
z>m8kGluZRJO0DR<sKl`|WHdy@<Rk%3Ua8MUJkS&?!6;}~<(u~cV)<s$PEF+ZN|VpE
z*ZOPbCl5LV$P!pys=qBBPVAHGOK4Vq;YZT#5EQ>18=j3Kg*<xVbDfP=$6AQQE5-Y>
zGc4cI$()QuX*sO-{sJB`_VCs7r=(vv;Z)>zkwHIN30FFU@YRYmuxPS}81RgCN{BHS
z`K}K4z}i~=`o5d+WHoNhX)hZ!*Xyz>b@YsNJo{qq@=0sP&y}jisgRPHH*Cq+EXscD
znS)^q0Zc2uF0P*7Nljj1j(_1E64BS=>C)k&5LWU0^3@#u(Zzf^uFUi#-rlGEE{7S8
zt+4N|bt8k*WFLd_o)Dz*GT57#-|S0A#jemxAOxv@8W<UPxl7$*(pgnD0E8f?Rx*j*
z4XxdA#`LK?jv6zQ*^*wgd0fLH*AU)2;n>((FZ8RuugeIGNPqD#P@z_vA=gS?Tj{>!
z!~9~GV~_;9Qju6BO!I(nq<Mgi+_I>(c1S~1??^$}y>Eo&Vto8j%>xXMjR8wg4vDW7
z(=+`&C%S|RX08}8ua?w)(E6UbRZtjskTN4w)kRdp`vytE%0~myvB1Dx1StnnV;5f*
zBeir*G79jo(!G~vz@42eHNM4$27A(X?Dgp#ljPuSDj(7{?z@q;3T6zLOJe2F+0gp$
z=x($_C$zkZ<@)8`LV8(rv-(yR6+Y{;CHJjVQTwH|xv@^U=Fal^@m+gXDO~rj9Zi%2
zCdG1h@mi(5%txjrZP=k@g9kyUCy^q%DGK^-&f^LNK^XL`9ZlK-{nOP7gUD+$F%(Dk
zP=O2`9&a>k$xH?T^Ylj9tJ(cIX()hc^;@sp1~5Lim7Y5TBwNPFo`?<dyB(Q=IxFYQ
zkEn8eqsH!Yq#l<Q0a9$A*~Io9IZ`}ye6J$0HI8EQTUuafas4F6u3A9R>NzGe7{73u
zbVg?T`Jrej{Tkn0yl%$?E9~up3cQu@Ej{*XyVh+&rE_KN`h2f1&z{ae*mGfzzwLbW
z2jO2l2ySv_`gcE^_CB1gawk}o5@cVRcOFN7DyI@t^>!3bKX%iWLxinV%?=&cm4Avj
zpau!81^|FnlB<>35Z?LM$*y)&=+R=&0rrBG#gj}58;6bIJkO&kL+Ks1Okc<OF>S*}
z0RaIL_N~oL-LLam^s|w~ig~<^<qKCi213Jh(;rRZi)`l3DiGJ2(X1uQ>^~OopXYFh
z(4<uw(JRyJ3a8B(`Gys~N)%N|d#$Sbe$-+T9$JCxzo_mmZFiH3oX#4RbGej23^Jkj
z2FO@Q-PM=}=NV_8OS7kfydk4LzRk_<I9!AI0q@UHt|AJs<CB#N<dcba?u@yIGuQ0_
zw9zkJsUWxE^ZUm#`VKXC-L&K6x4HcCIB`@L6+ooSQ`=Cu5tcP~Zk(ARUz<Cq6w2L?
zeBM>1L5w*mCH7-U$$_-?Pq1tzU9(xncJK5+%7<JA{Sf5iuslA6h&<>VszVBJbIR3A
z*4VqWH-UOwL9`WZ%>p_#se5h4x&ps_tGeVPomrGAg)31D2M@<Z%!j+zEhixb?cpnC
zSw(;c5AX4?xu?4aHhisL@dSnN@k~cLIhEB_3MX~0q}c?;!y!#!KZ@S;q+6M0#!DN0
zwZ6OT2=w7Mygh@NGZnDMQ_E_1I!9ok$#%29&y2+{VduZakeRmbk0xf1#jd8~`#kT>
zp)MKbN;byU*iY=TjwV7^5mU!a0po%P^ZYhfk+vO63|Uv3y*(V`Ypv%W%Gx+r=4SUj
zJWfk^>+U@_L`+#&i*UIhk(7v<ymR_vq{26=ESKWDkUJO}&jRjPJ8o<zAf~*ZD?s{w
zLY*-Cgt>uP2ncwrD)_JKHf^vp=l__(7%l7l>k`ENp7#<a4`t}G#;FFn%i}NcdBdi<
zs~1gsJP)OF9zrfCxZ$*uPW=yK8rpS7sa~t_F+kO_m$JVbF6rbrq-%J-xt^-!h<D_u
zV$80{Xnqtq_p9oFv3jk)b;%Uy-CUf@1qV-kdKo{>=OqI*SoiWg_mRz0<2I*r7I$jY
zZ&~8axPqq)*FRT6$ZC<$%RMw8jo@Z~{`6J8+-Jo?vMH_u=zxIt;5JKJ^JDfGb2Rd9
zzn0<8aWAaI^KxhC`Wdb6KB5EuoTD9}!itsG8VCBvF3+7&)$pjJz}bSTy9he4pV)(!
zvh7^Hsv}sgLs&AiYJH~T*4L78-7K&pdweT|d=@<5@`i#cX{~Xko>xWNmx`sNmM|R5
zK8mmy+h{dD+%hp8pQpMA@)%n?%P(`S!?`NYYgZ-<>*pl5B4Z_d$=j1G`jC)*gsg>|
zhV?xKf>0w-_gaUZPvy<oIwX0v4}~ud8zXz}|2^qU&i+EQ<!fb5mcMbKWe*N^lPGDd
z>)BX#Hrd5fN34ta<-R-Q-3iEBH6#1!u`sxMRh>s<ny(KMhBk38yNh~%^sOkaq+9fA
z6pbpVC+787>&#4!k0mbE@j!IxiTZiF_>Szae02AlabKuCyQn@Q&hp+(VlPi!zvR5N
zRJnHG9(k?O>$AMNXCD*KuH>_O;EsK*&<neq&)E-rE%_*&Z+ywObAx>2n9|7C(*F1;
z2lyyO6JeS1J!g|XBzY;BWtr=h*&sIMG>Mt!aoG+#kkLJL5Ex&6*gm(Lw!|MJQUhE4
za5*lTpi_H%YMzXPp)z(;k|!W08It!H{a+%!xNLN0)+P!8dR6yfCc7b73FWs$HG2OA
z5PP`NnyH42*$3vsfEh3gm5J=O98V6HN3=h(?ujo<B*QPU(L$xrj#CXkng#`yYLg6P
zf)DK1$!P&{DZAGRFUTMdg-BWjDe{Ch-e`o!KaFsMF!JW!#1d|l{8p5aH*BhKUV~Ya
zkn*vFV5U;R2;3{Wajix-u~oH(19vG*(ym?hOggf+kpE}CwAtu0IfU#k&AFT-$GOjG
z2UJSc{$E@GJb7Y%RKAQP0ywC?T^RhZ&OZA-zRYLBD{LG}@>}Z;u2)H-JAvgus;l$H
zGSvvp<FaZOv1#8B(FBFYvZLF>6$B@rC5!yp*I~g{?k?aq5_nW#5A@tG_Ya=XcVsf}
z<_YP|^o6&l#l8SfesUMKT!lH#ugmkx%Qu*M%f-CAE)GyR+gEAv?7hp+#78H7%Wg%=
zOr6ej(B6oWl%>aR|31}+t6;JmkEL%jf`(&$^P2LL{`nTy&MKMLk2xi<BC{Kianx?J
zX~KT`ls>KBa|~e_!WwyNb5|W^dI_d>`c)^vX722=t#p@19XPo6Y9{uT`S5r|d~@0C
zMfuWNA8DcB#lsxgAN_q<57tMNS+5s&&;NY-WIbC}+vcxDR+iV(`Y16(<4l_M$8nF<
zvi7!w-LrI7+Hew@AP54znPefRc4jebpJj#A8M0**ewldwGs$A<&o2Xyr!SUQ#q#U&
z{?dhYTG&uci2@xSiQi7PBd1H9Yr*jhW@*Md+`QfPqZ4*c1H;|?weRpT0(t>h9MI*W
z2+>pJgb#`~=K3!$c9fUrAxpOmeJ)!(4_ci-*fgR<kHfhGRnxy;Qz{RN)!L{EG4Egi
z?v`L}322ACQGF(!HxvC6)UToC4+pZio4W|ZLB*D=NX@88p9<|74fvUo&`}RU8mRfa
z@FcZEhBuA@0<f|>FrDiez*~NpLJCL#mZyGIiUyKFf@|MmCj}sK5M}<tSI4WXN9XVF
z$#7jiowA59d{X;D?oHT;jvRyGYrW@lAEd{+Oo*g4rN2`+YQpu@m%Je3Dof2;R$;@>
z<YjE-YF?fAMbvp(sYqA%k`#PW_7$b>E=`<O1tBvfy`E?+_Td{aM?uix@;p}H)0gdP
z(BFcUpnvw(nkDZEIhKBWW-_%8FN_=*QzQ85nTu}A05x{i(A)`wyK)FZ9K)D-|BCp@
z*|p8%gDYY}Ax)0&&f^hM6tB0dr<Gb<j91pj+8cM1+Yg4uMMf{wgSFKv&e{*Ec0`-)
z#*6l%yFM-B0^Pjh$z)tYgIM2p=N&mwp|KSLUk{fp*{NFzC9Zhy5G)vf|HxQvDxb;`
zS*Dk@wWF6HfM2$UBU6`DS5jWFk5N_TgNJ9|M<M)~q?kR3)hul#A(Ph&X<=Lpv_cy|
z&OBzetJmUif}MlB{*L#W%4p7mhQd$>sS=9TJA{|9zhKJ5p_b4={iyQyoOjV5!$cnd
zCNx?{oZNuQ<FYZ6n60ip>?8lBQVhyxqP=a-SKcb2&)$I2_v>P;7Au?gQ062x;z}l4
z^t^gYm2%Dh5{<IH&Yqi_(fyd@*)q>wC3OA#k?Esp)UWE$tjYa!xL5*}FrnYCZ`rRD
z$!wHgiP0SK>$MuWH4Na*9(W)umt}TS@u`a2EdTDjz#hUDsX4K<N{`lyrQTIs=2{!4
z!tL%wX^w}vUWu@VBN*5DGCd1-M|3bSmVoAHeu6wZm5B3I_aSl7U^>73TzLTz;D0%F
z&lRqu=CZr7JMpq`ENrITLBr1#&jguprNR!m4$pmfz%hS{FoM|xQVO%c$Iu>+p2iv1
zGC!+N-%`L#M33hxn0(Z*o^h@b?lmeDU29=vfZC>4C7=)<S3X!KYiWuHWMkw#oK&Dx
zC-but%#7g{kT5wggPlYg;ewjlcVbb5!wcD`v=^JatzXSFsQ+`t$&A2|`EIYqec~{3
zLe`?=knsl7C{po+hv}B%qa(v(FCJ7CW!VrUt2XWWpCQohG(j?L2la|Y)fekDgkS76
z)Y5uYwXK!oHk)Nqcu48Xennu#%W*d)45bMxV;=t7WD&3-^szpeHNuxT9yRcHLe})x
z-^Q>M0d3lZ-`{sX^`(Q<TS~N<)%Valmsns8Oil8gLZ@`7iJ~-Vz?37<{n79;Be8>t
z2Ep4On70&I0CCwc+jIyO&Y@+Yx$$iZU8DWV8y(WgCVCiJd{Nw`aQAm4^h<1SXpn{a
z25)%ouMm8>lBv(zB~#jF-*`B_ayn`!9M6<JS@UF_X)<;|bjHbOBdD;j`+#`qQFc{N
zrIz-G1}{Gn%B`zD8ZaeUw5Yv@THfzeXF$lmXxbQIX}7s6|NQT$`nKjHHT`dgZR!3t
zjatc=G9?$s7QJ*XQ4K2z;5idZ!gH59(|_J!;YZ+Ie+G#Bk1$(5#-oAOGi2h-3IR>2
z?T8M8pXQnNGLi+a5jpPitdDm4e$L|G;@-z)tY=cm^5Cb=r1|ADI2sm87+S+{kuhCf
zU${5)-J;}sZFc!Q7ez9QTENkmWq<>bA18mOLci^DVpv4e{t>_ZAKEj}8YSNr^K*lw
z2i)3$i>ryNgE|qEV?%5543#ogWYH(2nDkQBD@z(Na$5qcn=&E~jVNP=StPJj86~4f
z+H>C?_t=pn^B}t#bLi4d`@P;t=MpR0ym)XOcAE(RLMMb#8rA^763p}iunb5g%jr*Y
z-OK{-bhQx=06?rp+Xqx65IcO#3Mm%7qHHj~`{Cv4B}tx#6y<YCHs}J>B3A$+%jJ>R
z1eHorIV`77o@g~yh?>&oZu~B<89{k|GTru|b_E{Qy$&?=y_TAMT6q>3QghUEn?$ik
zcmru&-2S?c!CK{uOfF&Y{W0+TZxW0EE6_*(_FgN+ys|}iWNG6om*Yq0SxHzT1&c_@
zrp^NJjv5~M9GrBToSbw)Dul~%eZ+rxVgQc&$3knR!8jKk?yNiTj_DhB=YJJh!j(#P
zB9K@pl<7pEp0awqa`hwdT)vps_uYpEvDSyrF4<{DDZjx}-kPlq+)04%D_@^l6bv#D
z&D8E!;0@pnT^-pF=yj)uvg=1tm+I^n4D|OkY}B1AZsS-MvGlb4<T5H7Kfc`fYI4k8
z#bQUkrBaoBOsRQ!Kb@Kkk?kj5@&xd;G|DzAT^5%e?yLp(uIr_TAp}~j`lc+r(%^PF
zdY_l#M7217fb&=(H?-!zJ^fBjPqy%yS7SY&yU!K7%2%L)aJ;-``t6!;H)Wxt99m`#
z(Ykn~sf<ZKI=lt~r^Qferd%GFp4s>ZviWkyaV-xAI*~uJ;o;@O3;}$pk&Q<F!C8>Y
zNjvXYzr7N8kt@rV99cyU4G55fzS=wcop~Ux8tQDHAq0%2XXRi@+Ig_VT_~TiGzYAD
zgwUITl(;n2{XPpWy)WTuLe8S;1GomHz2?;b9+1loyM}=RWE}wzMoY}F5&nqN&kH3T
z9onG@CiuL&C{y$AiF{#m$Iy9*PDXy*3{5Ua<gmdrwpu2k*Up+o1H(9mZzzP~QAqfC
zedztD?68<8s@BR+90NUeHWd+1Yo31klGHd}%O7afA(P9-r!BAed-X*0Hbza-UR%JM
z?Y!l{<NnR)<N^Xbr`eMj&ChL|uN__w6c3K#(fIJ4KSi22-@BVu%zyq=<n`fY$6ZY5
z3C_cM*S@&WdXak5X=kD0Ir>`TQ<=%am`_B{UB2_U)4c3C3K1)x^XMrHmmFJ>*`nTw
zB_O`D5|ULhvwvPsfOqX+az%S<$9<%L+5kn3J76yvH25vumthI>sBXJ8Ppw0*`9p-i
z-%j_Q#_mC8cC4GgTAg{9Cq|hI<M~qqdQrh3{8rt4;>oos-tm=XrhIbbxZXXEA_^r@
zlC73s`JZSz&MMyNUMf8ETgq9@D>+3H{;=n%w7>{xdf)Y-z2;w3QHhECD>HinSCHw$
zVL3O%(Ov}E=I|tOET9zk&!23~N@!0Edyl(T=xSWCpUYbqQFxp{0H3+50z2p5L>m6C
zw*gbu?3RSdMu%1M7TKtAcmt~NBL4B~`VQB?E1@K0=%z?=sq;xRsol?Iq749gI1Gu%
z6UEkYcJG~@8_+xq!M_Kf4y<taV%idf*{iL&DXIKY&i(r-vz_WRu4=>S1Ty5_1WX^&
zY`t4N_inMQ#$~9;ATr)tLtA^|(+!hotmV*K=HU7*I=|UvV_W>3m(_R6Sna0zfcK+S
z*kYd=f^+4Hy}V<F|0?DIrALNRflX?*QQc8P0`&eqyaS3AA}2|w?SGmvDgEYw<Hc(Y
z3n$NGhkn;6nQ*_tsv>#qub@1MMeokQ9kl)2>7~BkD%xT=Q7igX?XCJjdLvA=MIzd?
zlLazUT*+zrq8K0~fd9OFL<1pfdW^832RrCY<N>X{B44^<p4vf}YTzF<AE5YuOn)cx
zL3!0DI2$<Frg!x#7_%<`dq5WA5CNt9+r?S)ahswmjMtG)k%h3#yydM#V`9cLusuqA
z!bAnpr<2uvd#Q?qJ0$QJ9<pz7a=}$56z4uD8nUDP_8)~dk56b>CU!wa>Du{ic_CMo
zIK3qAx$Th`^)2JW{Tn*J1Xe4W%!0EO{mE;LI3i9il;h%4a>OEQmx_S0U+H|*RFtc<
z8;QlhHz`J8{AkE~)w7&koE%<tl5yhx1V-YqxPF{B1rk`~UYCcHQ1;g&H{8GR0aix*
z6rE%-lB<_TJ(V8}P-6WLk}G0_&nAG-E@^b-!JM9QX3QGI&~JY1BXQCTHY?u!^^N73
z5xwj}d;Gtzi2l^d=hXJfDxLev7q0`2#I8oW#S9xa%4~>yF2b{6L?Qy`)x=`e9%GBc
zeQ6;GAKNdASSFp+S{wK&`kUC*4V;v7(MF%e2qg^scX_NzMeYu0jjP|{I0u6uf`YHN
zfVh09B+1rLwRJov>Bi1LGg+~{-FK_J=cf-bR2XhijP49iGPKnq`;8aczklZr^d35P
zXPrIL2xcKN@;TJI>a2(`d?`Qza~aWgi}vgsh0@D+q;dl>9_@Q<Wy}cFc8LX)I*F}F
zMGo#79KN~LU}!hf`{Cyn`TKnZf7+EOL$%-jsO`l{Lt!vxQed4jYS;v8pPx^K2@OUW
zX6e**&d+gQLD&OX+9acPaa_y28>3(9uAuH28fbG`WY1J06%K<JvZgjgFP8_o%<y(a
zj^y%U)lPU>uZr>W1EvygTdVT<)2I<Tse(#yTitO@bwCrGt1_6IG-k2aZLs3h2--c4
zy01Wj_TCfc%blpkH(R}0g>xOCAqcVr1w`1hd*y3y30EN+y2+f*ujNWu$+zRvIB(!$
zQyW4UlWG+Z7vE&yHGM7H^6EfW#`8fAhZn73iauf(BUyIH6oE@4^@bX3gUf|b!18@V
z^al$z=Sm$`f)0?JB+Q7y#~GVDqMW<Ks%7Kj<>RH3Vl=gjDu>6SdSzc{XsCHVmyhHU
z*agu=Obl6lW&$dtFLdF#F8(*{oiq^V6~Q11!vFj_GN#IJ`}rHZ(0y1)_r=$FwavI}
zL&4t$AKEYaSMXvX_lCpt8!@c@hVhL~_&8%>>Z<=N#KFX)i`7Z3ihC7)PP6gZs0}TC
zP#;Yb^tsQhBkc-vIMq&OW{WVtssmkCz8Z<<$Y>UHt7g_>AKDg5g<U`8ciFuj5i|OA
z(!)UX+F$P>h2P0qL5Qd<mD{RI8P!C<w;xSA2-cZH_njNEVo=`@%yAR-Gq9P+&pe8o
zxZY<Ko^I&92wVSdBZTP(d$-_AZhwBZ17FeTNXDEycSHnFEaG-3_VLdRqo5zhe{;Jf
ze1R1GnmDawqemMzyDk?<GKjjL&)Oa^<)_=DH+w7NMFZS)wT+HgX^wfUmgjbV6lE@@
zr;XZRZCzB~Er6s=aASzi{^g*O&6S)+kCQdo?CiHeZ(&wm#e6h;O=;p6Odb<W2D@Fk
zUapxNsM&VyFy9lw!f`V<7fIl{^rc;XoEbvEJo~Av#WQkmV>9*xw(HGf{||fb8P-&^
ztqWrXMT*jtDxiRLq)LlYrGr$dB2_vPdI>7M3P^7P(jlRR9wO2NDTZEy^xkVABzNIH
zXP>k8_x-!~=MFp%YprC>Oy(MMj4|Kw&M{O@C}#>v1CTzhE9i)+Qb<?AfrdNM#47pz
zCVn(BQblw%MMBQsS$+iZz5;j?$dY)rZWz(v1-9Ai0nb#|hxG6<&OGy>Kn%Y41^L3c
zBLWx<j}{Y0HS<@TnKF<i#B9ev7O?U?MIMzLmek|9b}pdG11k_8Xis$60u5U!5A9>#
z7AW0Y8)8|OsxfFHX<S(|8fQP9a^zve89^By5L3m;0tlh<1CAbh4idSd`21Jj9D0<V
zj_0e^ypYOWO&Wd~^0G$UHRL7zA^8P)KIf&^+HsdB|H=bE$L9G(1Jk5Q)LDa?o{$Sr
z;~RUAi*<nVLSs-X#d($7W33`Z-ekHKacsE8jIN<abn0FAB{8|^ygIf9KKUoi6YuY`
zqki~!P8PXZNSGhS6<g`U0?1!d3--s#QepgBHa#M<xC}pE?6&2??aOK>$x6zN_KrDc
zu6kZ{&o~wX6_^9N+fK~-=FegHY=m!g+=lq(b%?Sq=_cFHpZ1-%e$PkTxb;zzUq3wa
zs~YoD882%oChmSshdPcub9+qE9ipw*zj83<6-*rR??<m+i5yt|?G*84n7M<YhbcGp
zK7-tc)&4T8Utkv4Lr1L~+Av2uo`g4w6-_$la~`uTM*(1WebJ4RYcGZurrr9bPDZ!3
z8&`0!12bYgb}W%gf6~y*|5q{}%q*;9y-wang46V^scwQfwPIw}44sUaxJ!$Wk44Ec
z))F_}3IyLQ9W!}_e}~HCQwQGocB(x&Jd-e$RpRD1gc?-+3dH6F6C>E%WP9>(DvRu~
z4w7u2u2EMqsr5^$L%uZxAx>aHE{a1uPPl2-xpP4Pw!tpH+8#N7pQw?@*Dd|N3d0uI
zzt-R;8|=&2m*<K15LaUj+D&!jTE~qb{n`rglGt6og5*Z<c65&0NDsRCI9LNx>o&pi
zyt*?Im-c#g7-#`m;Gk_}lK>$qs0|_S55RN-l_iVw_mM!rHHR-BbT$JM9i!)WjzH2(
zjt*B<pPJgPexPt|TkDl|#P%#3pyoIb7vS@ff`p9pYj5K2eSaa?IUp%4sLj=6wPSIi
zeIl<somk}8Gck@!?Y5qXf9ATdU_qmxt}%Pd?(i`6LyNdu|NMinJXjRTaiT_lsT7`H
zl)n6V71M%+jdkPaP23?vRp&7*n$2GqnXOyhIgpW+rI2J&y;S*O1hJPA6&q{fW#=<!
z#>yH$JcS=mN-iuMT2eUKVkBX<thV27SlJ#aP0x-@p+oCS8!y->WTj2K)z)5(5OLUN
z(`}fGtHjMKed%V%wQSw1jBAf?IvLB3Gi=nSK4G1#%F7!m^oe=Q_RV^1)8FAwl9x8{
zb<vXU;)Y1W%x-yBatLE*SC@>etQ7oD8V?z%s(*ZESAY9di{C0R2I7L}74y6qrIiI#
zemlOi_TLPklib~UaD<E#LG~=a8|ViTJBviMX7T%K?aEOa3Sg#YLv%PX+d5GlgiP(T
z6Y`H`^B;u)a9oC4oCAlB(kus=hf@A4?FaP5I^NJ1j#?=%0r*3Lp7U`?`#uJ{0W-Oi
zm@67jOf7x-@R=B@UY<B$9;%D_^jX>JX=*V+97hxYF~|(3W<8b~gPH<`Hq?S`YdWA(
zW=&do4_4JFDDtyc{iN}PTu$$H_;><t<c2t*r5e3wj6T!z%8q#)h<wYO@i?Y8v>3<{
zKBlLpSOE~a_qbBOFVtHYK7(=TFQiAjg0O-|L<_6m8zS@-dVj94k7a{=YPmn<6_H_B
znK#RHe_`SkVnosfuW}0&gNYo~N1wIqX)dw864w?M&YJ);J)r-si{qf+RU75gn_3=_
zs=`FCh^|15VY+sh#(`O+cdgo$nWY}L<N$KeYxk~fVQQ|LOG)3?n7wF9Nbb+}g(=6S
z>qCti%m=!RIfz7S(O@sYUsf-zafhG6tId3{`bRBq&qfN^2c%g)U2`IG#F(zO&UMRS
z6(j_oGSc%#%>HzEXEs7Wz(fAM^5s&9Cyc?V%=VnEFSs(!t7J57UPAlmn1oTR=Iy$I
z^5cVo`3D~o8waPR#z5nBEUJCGr9MWmph-dUjcL4n$wBDNb8Ex3tqrD4BdM8lWx?f1
zC+v@34Qnm*9Pdi&>R|ZqK$tUy&+Pt=v-SBA8+BIvuju9Xd2J}tgg$7)N^6t;$&s+J
z&06jD7Ly;haT;T}F$dl6-&3MqZ*~qvNRoMDL))7y&QJP!5!|oO(dd@eYb!fhS-TH_
zDS&K3Q?}<Ks%n_zjcNyfxY}^;WPBn#yIJ5&uIKi2H4Q;{c5a<iI<YW#gM!PTW=ka$
zdX_e7rWJ^xxS^_QoSl-9afdlb#+(iFRy>f^N2+HnENi5`6jfd~dwiVktlOKyp|SKb
z;?uKv%2<1Sn7ovoT{IgzyI`s+CFMO&v3`;#m0%#}EV3o-&L)Q_ysB77h!$noe6ehV
zdZk67Rj;5U4ds0O5Tm!{N|1KMmpr=$zNAVwCLNfyVqiTPc;(|ZK(M|J7N1!~!0D~y
zNRI|Z3l;&HBC{EzMtE>0nt}wo#m!6~PPBc%G_qZ{ja(*^>yjN8r;)lPJR~Y}zr+Qs
zWb;&_R$6eTQcaOvB?gS?&QUX;5?ms_8rc(?-drV#htbH`Zmu7{qG|DI@Br#r&GkTh
zmwqfa8I>}vs_)>o>a6bxNmo0{w*ea&vM(<9Ag_c0sBRaXFABkE&`9r3j8HjYu5;gB
zE1&@Phx~q9BA-5~OtYTvtlEvY8%mf<g^5ma@WayZrnVeGj%uQA`vy+~7A8E{Xc@WA
zcsXYJZOMR)(L{f~x(W*38rv-Ae^OS5ZxwDsfuW1d#a_jmNnIs<=!g}{Bz3(~lg1x+
z+HWp1K_?^(a)QygYrcFMeZHmdohvay{s)ywG^SFukh5jo4j^6hbPH-u({H7Cozr<^
z<is-6ts(W^Qev;e(ERKCZcI_EZkExwsF&@p!|Pp*Q@LKSvWlnjqVlD+yJ-kro3(y!
zGk-ffk6+L2&sFkKYJQ(1LZIk0HqW`6{&U1;MJ9iIXU|QEnA7`2wBAWI7()?MSBcdq
z@5K$Q$SU>A=a9>!_&fnhI@U#S$Sv$r%-&%Jl4<4s3>SNwcbDqmDQ%o`KA^JTrDHvU
z!X+cGPyc+%vQptXW@gMz<&Gyu2ON}#ZvZ0F*?ts$`YuI%<E6EW+_Q&TLt6J0JHYq8
zx=KmOC(r<a4L}x0hwm2L1+-Oe@zhn%y=_$A2Ku8{J=R>9X4&zj1bmC$#WvM~lqY3L
zV1)_4$22v`Q1lA}!z^Hv56sNB2Oux>f6~kn@(aG-(fQofNfX;_=F@uC`eAqX<$j{C
z_L^Jx6#X(8Y4hpsPsCC=J%cU(^0ifu;E1jW>|z{AXi2*ZJ(qV@|0b`}==fQ04Ec|}
z=+Jn=T&(4<Zu#eR-U}GMWvI}8dhd}v41F>PLhfSDSDID6N6-D>gOGzvFLl`cKL`Cy
zHn@XWh)bnuW#qyN69c}x5rpl@iiJ<V2(+g3@A!hbhz+pW(0SQ~USLP#yu0;PWz9&Z
zS_GX<u2C0RTmusB^7O5E4)Ysh;U`HRqa*~sZ(pGo;ugRc_bj%wv>vPw()m;l^Kex$
z3;7-7t^@jM08}QR>tNk8E!qE^e&)n>eE-k)^%nH;biSjZ^X#7wEv@8$H8b{FT1w_p
zm#1eyi=OSLA!2ni>O}K<o&)*2o*`OGNZgkk)5aqi%K|K_+v%Idr&t7#@bwV2VLhuk
z+PX;M?zN<m<u0f(YI3x6(5WNG-crrR!`z@ZOZwcQo6E6Tox9&P><96F_)bjT=BIcr
z=r1>>c*CE8w!gE=o955#BhQZ8_uBTjLbFe;58rmrqkvTxiMjYQ7Q>}|@pQyKHtvY*
zQMNjXLq^qIb~Vb|IhrXxY066n(fyMdOAmJ@58z$~Y)}nVg&A7|$^e09&n8}mIwLM>
zBSBcLj=J2#=y<z?yNuoC6iZ;a4j?+w{3b^;3`oJZWV0|j<!mpB%$t;MSdRr!6`XYL
z3zwWKh1b6PB;if0MFf{MvV*j<v9bLCv&do;%Miy7GwxgKaB>;uN)3fMZq$n>+kSLp
z!kOY#JoF<m3FbZq%Phd|N`k;(K>Qm1VD&rJGx3J^SCC13&S%G;E8gS_0&2}9JYXH2
z$s82DLe`Sq{N@6S($sP2dU^KeESGDN3<^rdxmd3h7vGP`qZ^<>JD&=XJ}1Ed+X`d0
z{h*J8EzWCKVi;x(Qlhrycc-`Q^z9v_;1Uj!;74%}gle9Ak#i^b!lf~5(8BABUzW)2
zqET<zA37KdNZ@c;YS}UKnsXyDRUiXbha>4|W3&CNry@FA%)L%}{XtpbTSTHT<GS~4
z%dtF2tCD^$8uZ0fj_ABsjLJ#?c58DT=3v{_%pa%#_>8?{W9(+*bw@emKhW%<XmxDA
z5N2^g)A-J+2wBEU)t~_f&`f`0C?iMDtgZ7BIL*zCfcw=|F={~B$?EPH2q|{*nWZ`T
zb?|HzyZT8sAM^+7n|fnj<zrlh4@n%lV@OwetP6$t@w-_gBcbIDba8zO@0E?6z+#Q8
z`aiA5rric+N;lA7N{f@u&d$%z&Ix6S0IzcgkY(_mW&8AM%Z&LBrl)S+&eBM<PE@Fw
zrmwfPw1LNZz4RGj#lGdw)L_Mva!1=q)4jglx8(sWX|JJ`!inFnkc9C<aqT9hoGqs1
zejSlDbWvohVI|#HhaS*NsrbadxMl96md719X4X`1vjH82w(<GxzF3xJy3MSruAbCF
zBk!zANJ4o#cz^Z7v-sXWVG;?-`L{`eT!9VY`M;7(FOEiTP67PHV7Ydsj@x+-!5p>=
zTcWcOoUsb5%F|)W3X#c+9l-6D74i#v4}^zjQtj&oWprL>&p4Sg*n1Z9RI=JxVvGKp
z3*fj1OdVk2EN{RP(eo?-pFrvkOG&whLN^KDBq8VaP2%J40B`p<!(TV9H7NssyB~7J
z3gDz^+cWX@{P{^E65_)7`t4TRnj4{tlf@ZsvGr($6P}=sCWXI`eC|1QJiR?n6GP&d
zqS}e2pYU>6Dc<U^s&$s4qJ7Vqpgx?$M^9|-TfR?V=3fltGf_H@OYnR&YlgsVkEF|r
zVJ->>It~ok%x18i>+^33tkB*k?1I{UKA;H$r-6Yi4&_;c^B;vEqr(xUf&CGwi&A~;
zgc}GO>5QO3fY*#*_?he^%4;ctmp%43o7MZ$8i#BY4e~!26?l{LKsqCAcKm2_xmGWr
zO7jb{GQS)kA=M}l5cb(0c_{u}ga9%%nE%<{sl~kEG)q-pkeseQNBfgM>?x`&`t0Tj
zw4uhi#+Dv1rGRL8i3NQ&8Ny9{JUTu;IzBl%A`9V{mKNF9nLb->x$x9a{W!i}3x%oo
zg&ReU?DsDfKZURC>q2=<>w%1gvi&n0eYP%Y=}pOu;B*d1&Sc_<Yf|eRd&>i}OtcP0
z#O@Kp!~S0{3x|Abw{IWK-_I3SkrB-;M6v7{X^F^TeZYwU#2aE%&Ua-vC<}5$Wr{L9
zv^auv?B>tpemM%S0;(!qNt}Z+jRa>~;`ap>N59Jv_b5?0e`t+d2H=ozP%4WkAHL=2
z^@~mEeH(VWPT=8J86hPq$Jh_6^{+c)YEbOkhPn$05{l92H%8S@wjv&SpKVshmO;Mm
z{pdToM_sN9!4^m2PfDX{vWEJ({eC&&W=eN18uxy<O{UTfG;c)*un>HYZo|MtTyTJT
zD-grWkEN@n$e*?80_j}~M+?TX(;Fn`>-a)koV7m0xu<?`OTbj!P)YBW#OC~HZR`k?
zx8KMZZuFmWn>+244Ihu{&LOXb8ya!?b3D2yW<4YIMoS6ut&(`!!kND5i<e$fi-5(!
z3L^%kqR|x$Tgf8xoSp^a1D)P4?Lb_g1XOLGY&%${dh+7?RDYX$pLIqp(McnCPA+na
zf9zy}FxxgKeq~G0tC4a`<B=FgE`R&P!|0zautHHO=>x(EiLRCPotIxJ(lF5F#mWzI
z*v^+t*Pb$R_s{YpZW9=gUw7eW0*qr+1|y^6LUWNaAUY`!gRbU&^g5lS%j))pMkdtj
zq`G+OBS74g(%f1$u0~A(PWvD1_f$nsvK1h&!r#TOcT4_s^HV-D>b%oIVC}^u1BgJM
zDXqyXqfcXlim9|YbM`wc9G_TqGzh~EE~DoiW2L&lCWDvOKtu#;f}qY<`Sz-~D8Byr
z%;%5yiqoCa)(kjxztn3z`JTJmN{9kLcl?<>jf(iFLbq|^K9`BM=nUn_w-TAT$r1Z*
znp}?P_dwYwbuA^QI3R3g*abLUsFV}#weXT?`^a2zTK$#}Y0A`q^Yula3K_t<B+Okj
z9_JDK;&)kXl;fvlkI4I9r-bfvGOM$^tzYh+a&8$p-R*zsRoZ%obd8R{Wx*9!dUXZ4
z9Cu*GPt(t2Jo|i7&S$2L)rwE#z1pes6qQw+dS;3CTN`ouGH=9LOq{!-y^YH^L9*Cs
zI*}|5kts{i2Won$s+_oMsO`+9t`7>Na>3p`!xKK7f;G%{TPp>Ed<dZ>y7_~OA2}Wx
zXqtMTwb&o2lw+~1vz)$r7hKFbO6_46MGem=<kd}aYV#BOK&AIMXa4{UjalPU|G<X*
z5%M01JN%CG*|>Axp;%|rriL7F)O>GD;s_wOp4*tVfsFG#vRw9+08F(MlRd}6HCq^a
zsEF|HnBp?sREzfUTb%O6jjpM~Lba($Wxgr>!gkqIeoDN@=*y<>Wp|huZ5MrWLOl`(
zUZnln2@YX==@*v~VC!oSA1(q-RDuh(jsUGB2NZR8cEE4@=PmKsR8LgAU+SUqoJE`^
z8SzRA3sXV0mITRdK==w&6E|G#AHLb@sBNJcl%GOCv~dST6-p{3BJARTpR~=WHlj89
z_Ky_=iyP!ub$^L}HB5o<P72@1K3}eWuBTCxOQX|gIr_BUh8U0HbW(CFaXBs0ZgmCk
z$0M(pSpY)@I85yr29W<zSMZu~6d)KAa+Fjkho1o4z<Z0Z_&(z9JW#omGa>aX{}7NP
zKxf&htiOQEeKK-JZWY3CVnG$a!q0NQ!!nT|Mka^f)t|*MSSE)&`jwC^08I7>g)0dA
z<ZE)s!IC#YFr<foD%m?BoED>^cVw|n3As#;X1&MJ<e-&`eVGMDeQ>{veKM76+g<H(
z1vfR^LJ_yLFQK(M(7a^UTfjWWkU8j3lfWaB97b9ZZ8<dkW=)-d0L;3AECKSUw20wv
z5PuIHWS(6Qrceuc{cVuh8-s~yB&Iysvdx$Gh$<tV>&Uwn+#L^pY4eozL_N{j$d-s-
zR5^xJEp<vKb8BiaG^Y2Fi;g5I{0B+0`5q=uwSL&HpffaCRA2f5I?mJwf|x&n5owsE
z1-!>3l5aafpM8yNKG@daRR@_K?Lmgd>m2moyIL!r_GXXtZJv&^j|DSjL^8(W4}WBO
zLY-=pzx!>dh1MD2BFnmG-6n@etH7taxoJmgwZ64+@`2uh8neWi+_$%XW{XMw{`m^i
zc8z_$(<9p9T?Gzo=hnt>jUYw)s+cW0|JOneY;Mia;(TEubYe5jmyzDUU&YyZZO}q^
zB~({O{90utCRF%bIso&k8etm=U<J5V;FtWz>FKxxjj58lLN7^d)R8Edrs9-Wzag0~
z>O@p12l`30v!t3vxSq#)v?2Jxmuf?=7UA)UNkJZS0VPu^kbuBlnboya<&PW#!nVyn
z8aSq9Fb2`TmP_-m$+?oD7UF?oV2c1WRhEMLTm~vR2INH>MHeiyUXY$E7R%3oxO+p%
z7;t<z9?c((@rrmId41uydyx&K$sdj|BnFh5j@|qBPgRqf&(JVF+^%A^l3_7}jO)f$
zoo*OW@kA@2G+!03x;EQsJ=O|RY}Wb#j_C%Ll_kLvKn*n4B*h}i*Y~>o<klFw>y)#S
z%n-)qGqg<EMIqilUBlADZq5N;s!*zZE-U4E>mikeVy<H)$6Li^UT9cn`PqU-FnNg;
ziQG=VybcAo<sM8xfA-wH<0@eNOaL{q%C7ZEF<ySJ`rVbZx@$*6HrXW1l2y15v^!S`
zh%*n&9G9{B&sLDiV)zb{KRPI3LU%r1ka*Y3yKaPpf*rzSdc}98hU+<3)-Ch4-s;9V
z%C41q3Hb=VtMBw(LJmm`{6rv+GMYq=#4a(4BL5edV6)od8HFDb))b^y!HHZDe5A@y
ztUj(OR@i&Qk7#Z=7FJ|t>wUaTjxIGLHQ+xEu6Vcndd4N9T9|!^_iJrn$2^!QoI46I
ztRvjgLO@4QpE@vfg%V>F1c5A9#frYpKPo^|aTP9?=w@FL4p>GN(|MOq_)4b<(F_pZ
zu|H<CwPiHPi?Df!ekx^Netdzn>qn<GTp5=tof2UvZ+KviK=>mr{QWOJEeM{dMeexj
zIZveW4Da4)#{k*3C}oHiXiT#hYZJNb%6_W)3oE{}=5sKg3(l$R_pl@7snG>DSi^r2
zFJQxNwmqaSQ{YV}vsxnn666yxlm6lutNykE2oJ>=JMDwcq96IlCT=K5wU8`G?xx(K
z@~1X*^*8p^2Lco;d#!|2%7w3(m9IRxwZ=giaJC)`>=yZT>ndE9BUfEv;(<HW?O;_l
zipNet<hf;^fg{~}^6%P%lX;D39pScIk3_X~qyuQJydM#_s|Bs8`2mYMK+47vj3rH5
z7ytC`10@EsizCkcYCLk5kJFlXbt4%tpQ!4BIueXfGU(e|@zajE8+Gh#Y`OJClfgME
zav<_jgz!M->p!`y9^gmCdT*r9-DFDynztU?UL&9uyTnn`So<OH9t@w;_3p@D2ktO?
zjz*0)Kv+RSMl8k4P;@iTZ!f9YK<ES`%QJelH-CaYtX6bpvcN(%xO+9@sDH&!sz#`^
ze&{<(4c7Aqd{$K|3%lZTg&vO$nD5@15wJmAwf)f!x;RNSjsn%g_ovCf((!17W_Lm2
zOv?4o?Dk{Rb*k4sZoRYjFIl~fZG6G_IF0j8FJi)vuf|*5XSg|M-f5(t94LLLcvY;?
zF!9-&L>v3ymZq^F6pFlLATTXeILXBI*RaZ7P!TP~wXT5-==fWPy*uqH$z5H18p0;N
zICG-B!ht)7x*6%otMYx%oOOi+Q>fVNlGeBA#^V_noYp<m1&ugFvBRk&gzM@r=;>0O
zzM?DOYH#^U8Xre2GmWmjzCb+HnaQINeO@|}O8z$}@X5N4z64a__AO30bK+}-t^D4V
z=r7Gzo{FuTikR^*yjiQtr=p1h++Z=eJcSKFb^v@b?@P1+{jS!&a*n1gKa{0$QbbvI
z_4ObHyP_)fZTTm3#lt$^759dU$^iq_%Ea@CNa()I-2%cAO^z?Ivf&(HR@#UnOX+N+
z$B`j`6=W$0%cZkDBLWb0t~FJmO>^-dtUdz$`$XXcU%rWTJ(?7v^2`dx?>$-b5RyYR
zCI=-xN(_psG4Wn%n4Uj%oqX&n|D8p5Yb<4Tb+Mju3cg}bg1er*!4xL?jTWfbL_F{g
zV6k`unk3Ow3@>bxfwxOMFynr1&m=vigYlIWqXBi31@XVi+r=zC2n3hMCq{j1JFEYA
z=2qP19Jx(9ZfP+jLOtuds}ngw&^KOr&`{9hOg5*9eS{PUGGrRblPAoz1`X4ZF(`3E
zB!k40BdN@!1+?pnvd&&Gb|q*SN`Jkw1wLWr6Vk=SSglPbu94vqVIjNaw83T#@nT-w
zFx{#wAWWsnhg?uhsJ10oWx=0CM(!=3HAegImTV%-j7jH`#Dk%W=u+*gDfVWw$Ew;(
zbs@W6*hly!@)qs9-2{0^+z98hWJiB?29Z9$4js{-tWoSpdPCDXF@6an58)HU-ZHs*
zyT(3Qkmkndi#B=1kQ;6ZIk|kF2aUu>0rd2hsC(lz@OJ>*(vDTo$bP`*@9BVV;jF%L
zzHKUqm~OeSSlVGPsFpI2R@oR_swx1sct*u;%}NQ{b9!PjX{Gp3O=npSM9hn(EiO5&
zm{8Po^KGdIpo3vBUzO2K+U)4D!8ZWj_1Yq*m^5)<xvbRr%`dR?`?e+~DM7vb?V=BY
zv7MReb(qj@1czvI)1VWj%=_4)sp>^jjO`N7xh-6(5pc+I`*rA{d#Tam{Ss}$u+sy5
z^w{1ylZgR%-4t6p4vUH8ONZ67pf*re!1ieq-<DeYfw3&HfYb1WP~W6OZj;)Y$&Y*U
zuFLXUPKVrUKp@C<RDqHpPQ5T;jm}&=ik0)hNeQ%`423M#*DgLad%DY_ixmoZ;xq1*
zmy=2!!9o8hQ++H3O*9lvhav)+6FLL-KoJjlvuNP~{n%X)e)DosaOsY1>ptzs`v!7-
zco|<xiWX5k6q^)BEfnS0&T<snm<RZAZXcjpXl>v?kl>|LQ~V=ZWuf%Cr(+jaia0R#
zs)w+7G8k_i(EZ%m3r@!Ii{}}LMvwp<*%|=>we+R718{~LE66wKE!8aD?rfMY_^U0W
z>w*5~!iY3luS$<KR{3PaY`PbAKJI9RR2?r-RjKz_F;ZJ{-I#?8{IfO}h>sY`i}r6n
z*YTrIEAu9}#e7HCzPvtb5s}RGO#C&NW74!xtm5Gel4Wmkp9>xlAnfgy(u<#VP=S1@
zXE(=99vnr0AkX#|+ju!{lr-%AVrq{ayFh1>@YPBBx?xWVA?Hqr`5hdXZEo+<s%qUc
z5MxwE-}T<_Hk3yiABqlE#lYdss~S~&R91{h5^=p;DkI_M*OzCHb7K!(%5%wTu!!Hw
zTdZVTc8g|v!bIV9cI|fDzYWL<2(Hl2yt6(t`bh$UL#)##PseeJW1}f;Us7g88l4^A
z2TYxgr?$t3D_ICS0>LLxBgDba8a)j7XsH*m&>5*nC2c^h&dn*TFsYw>t3NnrQd9LV
z<@bH>)1A!srT^T{k>5t)eWPHwPgkUbP9Dh0=bR~0RO=_n&J;qVL~>%I=uB~pZPd$o
zsp)fHbW1B@h+^%>Y71OmqQk4ZNO-VX&fmS(AwSs0V~xKNs0O3z7?yg`7|tU=6Cn+e
z&o0>8tF%-!7qDNSuPeAMN+qr-G50npaiGoWpbi~iM7gAS{<MEF64<XEExu3@-iU~B
zqs|SATr}z)4*yXz8Lfa}=p&V1V5ZEjw;usKd>~j9{bLR6GqJ;azF3IcrSL3|(Bm;|
zQe~^&^i1yXre9K<_BnTa>@0tmb8jJDK7N886#=A`UF`s(m|BE(#t2DIrtH@xh=H!x
zJiyw5lq<Z4@3<(_yMq5fkHGc?xIHF-hz);(@ZfdW4>j5+Pv)-^INk4uJi{+x$A*9O
zLt2~9G*Q=A*1f6u7@uUJr~Oho5Rdg+n2NPL{upc7%X_M+hPVb{ol6P59VE@bsPoY6
za>IM8W$}`cY9l=iTuSY>XFrN<A`EY1^5D>+Nw1=MxSWf^9!dTIUF=AawXLkHaNkT6
ztpR_8n?OuvncjA^Me@CYq%wmU2ibIetor;!BHt+Kx!0QV{QUz7{02i`?M-kbm$O+c
z6RbDj+un_2{92cOr$+z9X>YJ$Fyy#B*5<9Wo<DY0h}M)CY~YMCvt&Qk<MM3g=w?Cq
z`FBlDHkK8oWoH+m=98KD&Js2yNWk|}Y&vF%8GI`Y8xp!AkZ&hMJWcGBwL*=4mfs{u
zk$^)*EU)|`*)cT|OoO(kAyy~^8_sUlS}GP6SS=}po<x%-Q=;jPbz9tY@fJp=Sy{gm
z>CkmeGE7xQq>tI7sp{a)o``3!E~Akxo$S-)E$k8*k?biW*^rdW9h1iS@GKuBu&2jJ
z&r<eh@aOv{saG5w`4d=Q!v#_MeQKneTOb#XM{F}AWs*Uh_L6mHH_03uYMQpyQ7!}j
z4aA+jv95D?{Y=uhGVyQRt4o{FK1z?UvshKjh7Kci#j58Mdk=AMf!_jcgF`^l-CJ7Q
zvK`7LaMLUnfppWvntcEVktG5MD4HDYO7z`bOSyGDpE&JoJ$`#^VGJqB1YG^!?tdcV
zOe!A#0Kg#8!%VT`6kTMBF#3S`^h>RnJ&-*1S-lX@#emui!)fT<bBD)K)09Q{n)p?^
za;Sf35YY2h=NWHKgg#A|WqJPUkd-fjPKD6`3+TVoWQ69LCOy*W^OZNuz{0ddwjYwH
zyLlOds(6!Wm~IM6@Q8!pr{khGJhC{S90Hmzp(Ff$dC=?4`$|m4AZ@jPiw0IEUT3b$
z$N+Udw9u0pi4<h8zJ19Xxs%A{jV$^_)k8qsE+{cbNBUTLG&DqH3#~`1<LY25yZ2#r
z`1p6tR`FrZXE9PWCovyv0|V7ZNf8FASAGvUN&Cm+l$g@E#mO0_UKfM908J~?_{!*e
zZUx5e*TBFx6YTP|%0MU(esB2pn{EIiaFX7Bu*7ckvb}yGLFGNHlcr|{CWXL6D{_vy
z?VoFyDN7IdRn>q9h5)l4$%Jn$SL?MRuJhZ3yr$wc$?!S;qzF0PqmjOtQ0}JH&pMy`
z-H=+yg`eq4svFe)dhOU>3h|CDp9O!sz<vx6di@^<$7GKm9^@i_woDB&i8|WZUw975
zrr(21I{IMS%a@LVDo(m{EI%v01Pv<}et72LqvG>NZr&(~qKnbb_Ixc~hCP;q62f=5
z`4MwosChUuAQOJ$3IV`UOBbKHBA#1^|Be=(h;SZBwHcX${LzHV=jD@qnSNZXU^GNr
zK2hbl%%+|^TdZ4K2GTwuAZ8Zwz{7kQODFN0X$UI>pYx;lYw!KCi>L{!ftGT3gu5PF
zj%_+x4H`M)Kgjyl4Kp4TpVjHSUT7G-Y`u?BZ-zSYh(r&@*JQo(0Y@cFBW>0s$lO%m
zhur?$(mI)q<gqV75SqXyrpyLrwyIWEsxW}I1ArMPAaWHC4>Jwd<9~EdbpFq$Nq3RK
z($$E={indYJwu`;y78(@VYwX3yf5C2Wx=FevAZ8uO4P^rm^xpJoP;;&$-k!S9$Qd5
zzC6T30D6c@GS3B|c=^$$kvcPuG|?_KEkIL%-{DmWduo!u&B#k|!&dWfVFz$S79ar+
zYgIBZy8C=YFRupDeLgTtJfFjkNdhpRZa<`QbqnjF_1znH;YLW<&y*S91O9_;1)!qO
z{a+rv6>@Ma5L1!D;~G9~3^2`zUHC0TDvoAzk@gd<k->P%t-tAowgKVSJ3^YbDn2`K
zNsBMClQB$q9iAQzEzrd>PH6f(wA=CtDKAuZTKk+m?P6=YP<e{kQ&S8Tp%;HOY52}X
z;hogqP#CFW#KtoZzXcR49ciK->n$Sc>hm$`8J^vqhU6?XwkLS4kf!DAqHyPziBiCN
zvgM9*an1;PGPM*k%lE7$ImZ7eI<;24We-bZ=5uC@TKZ<bw4zrxO#hpNEnv?Yd%7Gu
zb|zC`JYZAz80qi5gY}=qA5$blj@A8~3&Oc}ntiP|PVn<aJC7rda8o3G?S2Fd$G^)v
zAaRI`)8ENik`fn(o5|Vqoa*gFYqQSJx^vYW{V(z!LlQ3E2Z$-#nE@>)r{QX`WC#=E
zp>i{4_}KU9Nz*gL@9;LVJ7j7g2A?;lyU*A3xp~^F`r~pI6Tl-)K356AG6ey5omBll
z5t5Qb0X!ui@Yzj6016^xqTAdcVII!7HvSI`dOS65h_CBch=WI;6dCEEev7C6@Mv)E
z4H=PilZO4Qro!Y0MidH)vH(h*SlQarX*2V9`^xA~Y+Esosqwt`=h_@9RTzfGkmq9F
za~-pHat8g{2_jE+cIEdyC(AIClJ+oPJp7gXUQk!#$j05w`Ik1@=lqNi`0e_yh%#TW
z<rY95fM!@-IR2gjd3^;<M&IL_Ju5jmCSx>^;uqN;UaN0x_h7Kk7aU^MwHR0pL@P5D
z0v-MoOLBO09JA%T+#mScF%5BXbhQ~s(&C8nn!9;U{n&U|ll4iS;n30pdUvCD<>m66
z%Co0H$j<JID@vN%rJIQY-giJoOc0;Tx_cRaiGUy@0@^(NZU)i;DD|adpxYz}xCO>y
za1lWlsO@bxX;1^e<?G2GN{#d7YH_Wpk3XDnUwXzMp)ktu8Uhv5lLBEbg}U^RCv)QU
zoX-)QtZ;zAp{9-b+n8?2b%U2#6LAw>MTgI8=6=UrHlv*gs}?ggRyu2#u>zX$C?I2a
z^6btMpvT2@;LdBkwKhFwJTUg`>^s8iN+A8T`77K{(?KW2_~XISF$!SU)A71LAHuK!
zysXW>Q*_!H+$W#N4VgOt8uJ8IZJ$5J&>|bY_#U+wC)Uas#QQHA!Ouviz1VP08^H%O
zh1uojZOG|^%JpFU5shni#o|Jp(8Tb%1Rh&DmVlv6?nelKp`BMnVg7Ud66p@tSIhj(
z@T<v;c-=iSOXTF?spit)T5LewK9i_TB$pECAo}*igTFc%>H2nP;W1?DVtuPCR74lb
z@eP8{@z+IP72iHg>b=IUjHICzbFF{B240@%qnf?X34J8B|66hqOc@a)mnd<Q;5TzI
zZ$nQU-Pkw|ey#wBUmyBF*EyMrro<)kX}zn^p{womNy`a>{0B|TZt&|dA=czyR@Fy4
z8j52PP)lB=rj(QvAnDa1WpJZnM;gKuifwHAwr{_i#xolxW#k#$=>%thpSSO&IwS3_
zfBeE{;ORF%3839T{FNNocMUao4`7W6-E!~lE+KCJ#qv(~E@3@gURS^KzPjA|F+~ll
zb-#5=ILc}@IRJ1HES@>AvVfCwrFtB#CzL*?29+Ny0pWBok$=}w21MpSgl-@{D`HEJ
zzA~c82lG*FYcYCw!my7j(fQEMx>`$|rSZP`2YihoIM3&r3S%k&8^wdl#qW)@=T>vX
zJmY?jtH`}UDpu1g5K7&<pF-i!^I-C#?-931KjcMO-wINEzD3b=6-<-jYt?W+Q8#KN
zOz7Ta-tM^$)(v*LcU(X)MgP`0hcgteZ1Bj`>#QZzQ)+h2l>a0|gx+q|!h~7RLiyU?
ztl%O^)-F2E+||{}<BHyH!le@?PdYwz%>S7?$6x5%nb@k*c;h1ZOy2$ZWcK8rH?^D%
z4kgoWX#1?OLfkzq<X?ZoolR}vx{wzji5cSM$;6d_`=Q2V*mZy&uv+}OIh7#BlNWN{
zBFZZtt3R{1(xuPs+~N!Joez~=0Z>ccX>97X>}|6dziZr<kbP17yB&{)B$Kb)1G!Z_
zm9)%3H?N6J4g<9#gBI9DE}AQ0DFG0p#D`?mpZvQ-aL!Hmk|D&KZ3a3q)bF?)pE3oy
zt742|u3iNphtN2lBWf{khc^X%;N_gH!gR-YAd@Zq+1o-OOCV2U@eo7c!~<U|A)nuW
zE|70rX+a08lRt6mCxUcHzb8JQO;8S<BHSD!SNUl9H2VO~<Rwybx#H>8-9T}nnyR0G
z#&Y|GUR4!uAOH^$2+~L7H~_Uy0C`GO?ogs2iMYA)*Mg#k08<Wlp+IzQDN@cK*!-1d
z$2<B5C?Zuqk4vKe%%S_OI3!mMb4jo(lk`ErBac_M7ur5PrD+o9iBvcR_yA|gzjoxZ
zXHSP5cY!2Vwt(6K6RsfZb2cTLc(tFcN+$K|Xw2GnB3g(0>HeDwKvJU%60_)fd7(VW
zWXkS;jXKV8U<(R)2-I;$1!*^Vj)Lw4gRMQIFE_Z4hO8^uWF@p*5sf*=uODO^zN(fJ
zKPX;bqf<K_U154%R})Z~_<D`*xl}UDEP`nkJtT9<o9ma##vNQkaM{xsjx=rESfi`(
zk~-`al{VbrtBM^}YEN0It(CsKkXAfqpm27Rn{-N%u>_CMBu=GuE7b&d1RjzQyfH66
zKe@+kxnQHNctBQdbh8)qO(H5wTJX32nzYA)O%`pxJefY!C0l~J7{Ew8PQ#r=?vejC
z6nQb-laU1Di7>HkcJa2EXXb>61|Z_3*-LvYpC%+EP+9Qh%0*ckRX1cw8eRaLMDfHE
zE&6kShBn`|25z)b8UUDP9P7Eo3VDYSlQ+^fJgHnye3A}z94b9zg`bKU-e1IEPHWyI
zol6S$PC?a4P=#xMX>C{}et>Ro167c^lEDVB7o8~kAt?V0{CsDHg{27*ALD}*Zi%^g
zb%+@JKu02hJGVB5l#)|8Nhx}|DQ`}h(^D*S@`=wi>YVaIk5Cr}TZ<)&wpyD>&$tdj
z36BsLfGB=DG^b!PVtvA^Ml+!5GC71t7&?${U_Bw*i~;Qx;I`d!G%q@sx@50Ejl{&%
z!fP<wiXx#d`-v7Su#194@&r<3HA~Zl(Uv{&SLjmd4z>q0?Y;D43CVEQfi(8pd4=sP
zNe8)|cYzwbJ2F2inxl;HGtNOS*!JWA)473kqhwGGEK`7b>(cE8>T~W@o3-qhc2`9_
zMItvPjK=D6s=rH49B;}pXGk7DU)l+-%|=hhnO-5_HoPik(A?Q|tT9m6gRY+R<8rb^
zl<stst^m~sKhvQvR+kW$#$0{p=pUSLvRHLBkPuUHM|t!W3gQWFzI11O7uokU33D_`
zfS($@k#W(1MTPi++mgDwyZJL)?*g&!qrW;BC^VDNa`zrHO&?UtF?vVdH9$`3b~g}^
zzNBOo`3i?13}w3`j}sWO`sB}vF$=7mxd>_pwDMTQ&1aW3WJ3J}$k-X!(U0M43_CEJ
z{=7Z&Y;Zmvu<^oEqcM6@%6fzELq7!Do!4sf@HQL`Z!h<NhF(r;ux`~BGQQcXU%(~s
zgqW2{V=yCXJQbMjE$GuvP8FBT<oB`*9lp}OShsA)icP@E?lWyW6l+s)n;~)`M-d()
zH@W4En9iIcRiOB7(dF{;c%M;7A<q#3aqIHdKyCeEtqoI}<HCS+O=-i8+RgnfA%Y5*
zTYui59!w!UW^GYi6FjlHjZF8ZUjjgnv7jVkW95?lUjE9Z&orA52qy%Fud0gz*|Gvk
z6Qt7N6J{c_$A3a>YuKS0a@{oDbv+4ulkWxa0RMw3Uq6WacCbvwg;rJh0;$l*s&SB}
zj*4V>Bkhggn$h3Jk63HCvhz*k<a^_SGk}fl1O!3@z;MMM{4?)4H$U*i!A`H^NSH+r
zrVnKs<!1k25OXO1yoPo8D*Iz;k@1y-@3~{gf#JagJjs1^_Vn?27r?@NfhbB0lUr;m
zJhBS#{w@i7>gv?DgUMw$6C?Jyfq8hj<ksdfw6BR~rWE%0SvTuvf2F`ADEMUWE|q{M
z^(5nEV#pdGfYSKoe|3ig=#HOWS#O9jZxDIagfLl}lRV8Dz|aVI<G1<a=UOt~()&Ht
zb_bEB{VYO4y_2IeeiznozK+=AV+WBb)OOA6$3}-1LpxdTkFs>~R4us44xo@D_U;Ab
zFVT;(6J?m^c2`ErqBuCo12s`#Z=iXOjW#!5E(dx;Q0|J0gf3zKnf{qLlz@~v^Djl4
zFx{xYdDpKx^83}1ztr)u*)=DE>S}(z^z0SW$LT7aE^;WkON@-0F0;;@((QLxaKpgt
z{U}DD-D&a+d~VY-{C}sI{r=-2kZ|#<Aoa4<!NGKC@H3*`yXk=D|DRXjjg10z&;L39
zuOo4wkoLb{G*SD{wf~F(0m1*(81M|O0mMR=1ux5sYnj_J=`&vTz`m~k_4?8TI9&hh
z{O@6)dJp(AFWUFq!@~(13828p|9;uVUV=6M`^x3<>NnLw3E-Cb<G;5A_W-%OMAEmw
zeE;&nK>0uak^lcE&-5R|I-PR4l2G@&vg$!dHaok<OJK4ELn5F<CIpR>+1NhFI+!L9
zTDvpedHEST{L*ilzmL<v?|<qr^oE}UG0eZWVm@tnZqu%T`2x4y7k9rS@jVUdCs$WZ
zAs1Wjc=wCezbE*)%{F}3sEzmx@mJrAm4)<{nseCS6My}i<OGqUYMI;r^3X3&!14dg
z#{SRu8q@xBt<kXHa{j$>^}^M5oRltrfQiRbk1y{Ap4UGI>hhqm|Nm-O|Lbe^8vZt&
zBb9|L-`{CFX#P0T_^;p7`~N??9|8HD|Lp}9{?E1li~%sU|6j`i{I4bXoo|ZXW0EDp
z&)VZz1C}6-8yYf1C-fIc@zFxgTxb%qilE6*A5$^#Zi&ZX2tPR{k>2qyq8vzlnS{G2
zZ%gUH$kT_p`qVnL9W3j1!h4+9<#k`?pdX9Q+Nos0<N-Dleeh&EyT}35bDmsRQ=;X$
zJkERnV)One0C~0u4C*g?1mHqY{3bg{sW4A}8MA)U5Sblj7p7OZcm(W{K97xFXg-j{
z^LQi-MzKo<v#LoBjVeU<Vyp!|5m7u}&Ev;@;-}vsk&JkZatIT8rwHN@Fnz9Qk<;-(
zL{CD*)c|jv8R-6UFZ7<a9<e1y_b6>EX9()P%fk;`8P{&y`n+fQyXU9s7tNeo;#qgo
zXUfhO)}YW5;0O1@LlhIx%0E7%AvK^4GI*;mHBd;sgYE`U)hH-#J&!C()aZD-v=utJ
zxFanEUw3;wr4Hdg&k4i+nU!XnSR|dwDY_~h1txQ=d@K5;eDTu+neN7pqRjGD%jYS&
zSbcX2xq*LE0Dn&-WG>y*BKgkE6R@~r0A#@isW17&qITz5?2VE0NY9<#V=4WkwF~!Q
zUD3{MZIFZgL_S)Bm+x_MvW`Lst3`P<(M>9}`J(;((A+a~E7upqBDC2hYz_?kB42o>
zUnGJzno*4?^QRB#flk0bQZ09Rf6l#1Fg}0Ub@$)<qxtFbi4zDb_jUPh^0n}?Ku$v;
zwgJeRwZjg>*wb6Au$-?Yghs<!`V6U_v-2|&wV#O-E$5Q_%=03OcS(<`waQCgTe&TE
zd5^!91_sX?bnofl^QfnGuCWE_X`c5@ghdfokT=|O=8+}Y%^v$GJodFbr;s&Gsg&*J
zhFKS+=my=$o~At#aD>WAv8yD(VFGW6St9@4w}*#Z8a_>{7`NT{Un;S6Qamj9eLnnK
zE;GY@`CAB1=71k7hdCb`Zq-JGaF*mj)1Ga&U<0E*_I5LHMG~{*Zk*-KUL3b>hN|^R
z3^&5c*@V@!_vmAI*nlyW3|fyrLspv8?4Y%T>2#J94nwpOpYd8+1Q4&wbR@i$_TE~g
z0pXP%k{9U&lUaj=0!BYZcMka2xm&q@$Rr{d|L6cAO`T<bduN>G=ls7+C<21#jzI2M
z0s@_Aqr(uO9xZwnhs5Pi&C~I_&Ooq118cJKR=PT(<GC@7>z(8N%d59PEu`P1LgbLj
z3#tu)O~$lKyO;6_tC`k|by4mnj~bd6@9yS271IfZOycU6_4KG|z$T(JdJT6Pq4sIk
z;8z;uY1N>hJEWx%>>WDaWT;5vnF`;&<mD^ALy8CZwqK#e5c(Nh=F$M&8x`jN&oB#z
z{T*gCjJA`G`Ugs7Ar235|91=rX5*^^1sT$d)ZA59xDvA+a%qpU64u;KmR9WVJM7UL
zRtTK1G<`8NwzsW#Y&%|rz0oB`c6ZM}Z_M4*Q0^#t_-6jRl7JZ&%T86SsmAXSiPjvh
ztgNcv4R~}|-r$Yj(vyFOj3OYoKMaZqP>yfC;^1Vt_iv*5)%xybv6zxJ$DU~K7N7Ut
zPc9C!C68&$JWDMzpAuQ0XKB0M0-lb}P7$tNja-?R50fEX!-z1^w<Cse^NyqScgpN<
z%HbR!Zpp)ik2Shpo$C{etanp&T?J%<Q3@L!CM4wO##K>O|EArh$c4oHFzu5_)FP=A
zykw)X;9chnE6O3EKKiyCSh*F==pkGfRTI+57N@!!x>sucX!vuxR<(lwZ_MzAFUF=N
zOKZU!XTJd1u)EDxU+(#vF>w7KLY<C=qii`dYY`8K96-}2PD^xjzH=qGpE>Y13v(0!
z=}jWkBL}`^C2Q&8I!-iKbXq(1cnc}CcE|UAQ8~TKwf|B1n8)MLK<fysv6vlx^+O=F
zgr1HR!Fm_Fwp4hinZYdabx%ew&PnaLBqy`RqanI%_N~H{H$&Sq>}a2=AJJEet})?H
zmyjA^FHgrs@olpgsO5LW&vm1k4+gJ=!sN{Sw>m`a%FE+CeHg`zFI?}Qu1~ld+#sN1
z74c}_vr{_>H&{wrRi$!OzIE&)!axU__xW(vo(P?90;a6ylS>~tkaHKb#CCDM#20`|
z0AV}vEohThQQphTcW!j6f7AD`Rio9>A`Rd)YxeCTQY)gd+tPU-w8d&NrcX^%^Jd|!
zJZ|Bp{*LdXEXqZj<Oir*oP}+rMbOBw?N{YKiev71DM|hyipHFc<-=%fn*7+9UP9`f
z7;e^(w$lcW#ZciQDfkFm{{y41e@F*EMoSMc>(;0P!WGDqUt~b?R4lMBn)?!oa8S2E
zy-RlOT06{Jn~Qnwvb-CvEPVW9WKNOyQ^wGm&ZNHEHbcLT;IshGhc_Sr`sWuZtF-P?
zck>na<gT?A-<CH7YZ$vYE9K_VTy0IL%v%~sNLpPDe*W;~9m0x+>Foy09*sKFH*GT5
ztlzo9E*;3L1kS=^(g<XH8^{DH7sj95b7oQ&V(?p*%Vz}<3|A5A<D@GeOkJ(mgCfDl
zf_1Q`NJ-yH_T42F7(-c3Ej3mnkx&!a8LQf8R-Q_~cISFwQ>^}12E?mE$`jJaux~@f
zas5iSu4~+FyPPI@Ig^)*R3rV?C!)Ga)S3x?HE;4FlK62H+8nadprO$A8U?%2Gr(p=
z<A){WM;f4p$J>JU9t){e2W#54Do}Oj6zV1XQlZ)@iLiY)(KJwLebnN^Sd~3VPxkF<
zB`wZ2jcvj<zPI+a$Rv)UTj8B$<byT3<&M6@Mv>$(8~DX9v$%+qEnoF5v8lK+v{trW
zsk5$@PMJcVI$zl)A1lzeTCplG;7`T7Xb-R8*I{8V#YY6B48v<XbBCa7bV<w5n1%qK
zKE+VMkvobosv!viT|q{7m0A98+9=ds6<r$eI8^%hMctXj1?EgTrTp7;G{ki)7zvIn
z*1eMtNo2*3_Hxnz!X|i)b?lHNVeu&48Tl%}f?&OUpv8BCNz}rKKxA4%$|?MQ$f1pt
z&1`XIt!cTOKEZHpr8<nC`0ib$LUHyP*!^?8RH3MpSYV7_iVMGgq!-5q3~&TxnFpaJ
z@p3=I{fKieTGG&y&399?6STL2N8uX0QP2?I&;idc6Tz5<)Tugao1Mq^8a~f$-u93i
zb-&<jWOQmCg&9gza6fj2lT1tt5A$j=MWo3m;4b#hj@o-uY<O7!Kj41D&2>WIWE^TI
z;C$nhtPZ{z@9-S3B1cS1UrT%dPh?6FE{L?<+s=fKd|UOA9TMrfYF9oM&vj;3mElzC
zeO^lo7|Vcp{oxvXE$YN!M?#usqG(8i`=geEnNfm+X~>IT13!Peql8Sxjz+#UL{U4`
z@T7C`<wibEZ&QcoZscy*mv&<n07G8J%}Z&V9Q{qiH{jK$o<9b$s;RkIQCo%AVlO(z
zkf#_8P6KljZ5w&e<C%a5e@$Q!<2J+<U&`x(CmyE+9;r0bS@L$hxQN@S+_Px;iyJm!
zMR!P3pJm`%Z9{j&A0@P+K<+>B_XtGLa$`la6k6c7Z`a2w<Biri?;ph3b5dn~ev({(
z{pOsq{-kt(z^-j8E1m4(0u7QpL~Yq?z?0C!<#n(^Y!brLZF5kN-ACE^&U63M2~Yf7
z6KFEX@yfW%n3?#-NxJI5#uy*MqvGadBktYZu@3~pd3&r00U9)!$z8yVN*+ow{h-}r
zVB<<tq!`n#Zc?ne4`{YP=j$$C#lq!S;$F5dB)Cg`vkP#t78iM-A?9ak*m{hH9x;a2
zn=Y^PBcADaUG(1>-eA05(I70m!ynIgyX4QNXP5Uf;D{JjppwSEZZ!Is1%p=`d5d-3
z3PO~<T{hFtUh-ry@y8jPK9yksCSqxL@g_!z0(X0{%dUvzt?84lRg$|(fE9P#W|+|A
z_ii;XPV5CiCF8{aJ9IvvKpzs8MWQNzzI)f`cZg3hjIHl3U0Ol$pLkXEP+yVI3vCk`
zwF3&}S3JwuG(QDspSYC}rRoWX2y4MS)hPU<&1CLdO<*b#c<mlI(th{(gVQ(pgQE#F
zJo=$~{$}nuznJ^We1|A`Si#_O>`?Kl@Dr)t`Sl9_H}4--0OjGD)>rH;P+#DJYjZ`1
zQ;aM-<TUq)oW|6f)7M7EiS>^Z-`NWnqi7yY-42`_lA3d2iD1wTr=l>SIHu4VlMiR$
zDb9~vz}Vz`^}tq@Y|YChUvLYddAj%{MiM{+Kc-i<Z`sQ)9X@#OSbvD>0!-Wl1+LRb
zqvT8U6wL2%Ps+_wVSOP>!;nwvwV}550L00+@SVmfEUK1w=4>r`Afdobb^lDU&bO!g
zoIrCU0AIa>Ljq=RN5sU4pybcU7^vDm@s>1hX(#hlH_aIG;>>20?`bwdh-DU}dsLgG
zcH^3v)cHK!<71-d(d&(yE+(tW$;uP=LGLF5Ojcj2_7GL`uQ*@-h+u!)p+9%|uiu0(
zRI0){n|L>#oZncl<b9IQ14z)|-R~b>J!U)CP+^F90cJ|KNeI&*?XcN!`GUswuHm|W
zaHWK49Cz7lc-?;z*w_Cv$7MlkqG=Js*(vEJob3nr`Bp+jzpvj#$sG=t_%u}2{~7hp
zbDwH=hnr3rdZbp1+~M@Iv~o2!`UQ>7FkGl_0DMCo!yrfE`@`oX(P}<+BN>i!)TKFq
zE~d3#sZZ2NQ^Fl<1v@^YU0$V-om%y+m7}V=6O;Qct>E1}BRVJbo)Xx^GiIDWkyrD(
zv4}H`&Kq~mb^&CkR)?V2{t28eJ4ZyLCE)+C_tsHywac4mhlCIyKp;r)1VTt~cT3RV
z1lPvhU4n!pxCAG-1ZW_*I|TRO?(Xi=^OEy@=bkg?H*4KBf6TgfX7&mz-MyP;_ulnZ
zJ@wR6wHG4Br1>tB`Ke2OMb}+<5W{tv#3Qd0)yV?_MxKAZ51G${HD;nb+_@)E_j+@R
z#mk}Pra-Orh8|L8b{uB#(puXsixalkGOQD3mQI}c;`hcdOblvb;h-Y=3d#E>nkH0`
zGV^6YRW6yYltF&*<Yu9xijZ`zn~D(dDXTV98=~c|kl`olvBJ;CjrwXiaJgUE+VGfi
zVg?6B9u^MKxk*scralu5Z57gAd^bEj{H3D^;Y92?heMCX`uK_-8Pfo39CeIzmA(rE
zh9S@1b3ssiG|xPcG~GQ~kz^VrlbIyL!{V{kAST%vG4LU-`sDY=0OXGwg#6=H=wrq^
z6LXW~6;KfCd6(*jW$54MX7fYXBIerROQD9hCP!t6DO9r;87I_XtUZhKXIfA3U9#3b
zZ4eH8b`sM1X{%1YY8v-`4#Jkmcr`*G@w)ARdWNj1NV?l0l`-roMY&;L<0IpYm%1%&
z+sa?J-;lKH&5$v9=nvX{;&M>^QS;>#ELm}32vQFU?}(X)$8Pz?fU18dyOG+K>2;jD
zxfaLt?C1!HNd{v&9v-*y-BfLqS9>EP9NkrRRjfTfXP>^%yU_=GhBnys4!%%D#BTFf
zT(t@OAmE)rAiK|Q`B<UmsQ~XGY2-Y|RUNQIVJTvECJuW|JWSeg<G=vxo-vSF;Af*z
znKj$6mTBwRx?~r>K(ypxi>WVv756GzF{(+YJH9Z(cAU>4Xhe38ezq=sRyh^fQl5I`
zl@$2ZR6<u%#ePFY*f&&iq86KqI6M3xC51#PJVB^bTr*QGEH=n|i)A7l+&cI<<h{^a
z1=n(n>NnhiX2wr2HOG3b)PAwDNZD&b)k|15xZMRc$oLOK>!_$W<jP?t6}OzbW}gr0
z;ZPVQv|)b(&MWKjwe7`b69+Ow?=wiDu=6#dG4^<2A7tEF|G@bOcs_ymm2hC13Qx`&
zAT_zs$Vo0bxMXCuFsnAKfUpp)rxMma{PY@SAwPDvlWl37U{+Tn3mP<Hli_Cv-j=y7
z+0A5Rl$l(4_(U7wc7l+fCCJ_W9pDBS4#($_6ragw9OP&w@au!r`=3)LnsGQXao98s
zD`<_k3#;>%9|vm)7u6pa)3LtJgJQp_1|jY<1$KHT`1KS0V<aZW*I_;N9^*&;8m`6b
znRGX~7bDVUmp-W{c}BMv7x5wm;ry3Et^ro40>SF%=Ww^RJ<mn*;_|*p1X<Ps_4<#h
zPZ^7)>zq5o1H(K_gz7zWax7?LbGcML-r=3@(XDr<cZg?$i_xNUUX0YgGb%qku0%tN
z?#;}I<EP0Y@Xgu_j@6hs1`U>dfYn8MLC~?uu><TM3vrz$b2DJ$VTQFVA8UxtJccrd
z_C{D=&42&o=dPji8XLs{yoYPea2B*XG`w{^VlHxC6<r(nkt+L|#hP!B(<z_71Y@b_
zd!N|N<H84pCeAWdMF!RnZebJGCQTE2INu9&?Vs~|kVO=xV7+FMN(jePc;(#{%1{BQ
zgWwpE(k~X-4k&M|^`<nUwKl5h(hS`4DV2jV6pnp1_q!Io59*zM`nY-mkaTu*i{|(v
zODVXES(Q>c41bn3=f=`Q-oWTwYVOHe?Bf~i&aiB@aaafnZ$oV+{wHFy-1ik+sK>e%
zy<!e>+%U&9_5d#3v5kGl>9ICZLZ}Y#nSkyAVf{jt=fc!T+4X!xl$4_$ZY0B7)5LpE
zHZHq>6N(UUOSW>(Cu_6XL$l|p6d>M*5YT#2hzp~)A<6QljcwKY7=L^BdzScRs%)=*
zJ{=FJgez`U!Vyz^f`w_D<7a-+2%(=KX8KA+bVxj-<|#Kd>85ZO6cU;pK_CvNzVY@O
zU^Jt68!9d&Es5yEExFQr0)gQ_)r_7q%Q{9|SwCo8M;!7ue%Hcy`YZj-7JiHSMz`8?
zQ!Z48Ko!@Fsm+7+wgc_wF)MQ7$@hl8O!0yu?t^;!UoYGRT@G84TBgp53+gL9)syUl
z`u7j@hB9UwW72jJ(`zWkM$>)CK*AT_SwPxS03YfBQ+`~xLgB0NhoR9<n}xKH5&Ij8
zZe{!|p~ox3flVcN!Tj$BQO5#9>tOoo1#xDRiu$6E;pzkO=%KIbv0uf;{2Q13HmYT{
zv~u`YheN#fKAjtucfV7v?ILNwdbYt5wTES+`4)uI0CKPU6n>8%m$#QRk6HGj32o|{
z&rVmRZjbYrkP`Kgel*4nuumYJO(gumZtv>E%k$9u_SD?)i@{Zjk&|H9r-fNJQyxS2
zU^2wQPUm(V7S=u41=Ni>`l;2~O4=3Xn<n_jC=1$}mxPYrl8aM;nMqgo76H!(2(AJ7
zT&jv+ZnTjBlI(FiZhcBorV4rW7IrI0-hO`Tmy@PU{JFnu41)2a5VMs?tlTDRVsXtC
zY;fW-Sy|XrFQcS{<-WfPsTom`VzhWb-k6r{;@ciXPfx|&Tr2gHlcsI8i^3+BnG;3W
z$`}ppVuPeh&30v=c@D;b;b7er{qeOuU%NZ)=zWhR>g&<zTG5BToWdDxS@>@7IG9WY
z(8q{4APEdiXU4_z8%%Pti2xIohoZV(cc<~?ta#$ejd75UB1E+=MKh+tLT|UTjVk&k
z1l;32j`!IS*-`O{8o8EkmGc(wu%96@?Z@%lbp-(~$@>X`mYZ=^d5w<!<;${jK}Se>
zt9bYFWFPtAK6#J#`hH5+PfD_W=KvLfw3P(Eki>*)j@gW?ArnmXOcBPP{x9O3ctId9
zwd_BH$dcY(gl)9$-e?KWU4|1e4RU>e(|oXph?<n|O~*Py%!5GiOjc)x>$@&_UR_SW
z57!RoZWl<mYJfSFncq@9{n(qkeugj@adlXxF;brI8JA1F-;%v}g^-e|T7c@{hEs?2
zt<N?e)!#2DeiGGKlY~OkjG^I?@z-|(l>%rWHHLxISis#?*NNk5sqX4%WZNG)Dyrua
z%GdKq_@R$m$=3JSge3KXdd0r1Gv+0X=9=W!3T(pzBS{xU(x)%HdNVAJdH06B0NMda
zeaJhw37$c*FWOsvHYGCGoounlfAh9i`G8KZFV)y|c3`+xPraM&1S+T(k!eB&)@{#s
z0=DuuXKiGJ+JN#cduu)=RL4AZHhmQ)d%MbnbuTW?Ngv^rLg|<MwYAXL3k+O2#9A|E
zCMKrVgtXTniN`f(SySWKRceQP1&4rQ=&3?z=eN%48xC7)a(oF=_ipd6w(K<5IOSOx
z<}RB(`+#9v;%~QW6;T>5cA6`X)&w!P)jAd~j|j5EvV?j$g<4cq2H@29o7bQXtXHB;
zUQ!`f&@%rkD8zJ}^(fG53yplmy78E7=YYl0DEb<0XGRyVpM!A(f12cxezh%3_P`qv
zqA(+T6!=btgx8Q0Gm1%Kas6PtC_>(yGV79uQcT6r(^=HeS;WV$L+&UL%g7f3E0*bD
z7K^`~tl7U!%#<H#Rt~V*DUf>Kv(cY5jVe9EUN?TCe?WT=hT(b5-d=xk>UnXb8J|ZB
zA+&6{CX=1ly*7)chtB02ssqzv57<HWBEVyAcx|r<o3gFu_#o~UP0hzRrb`k6-9UR7
zj5>@+FCArjf&zi0ysIU-M2a-|lfb#rhjj!6L|AjwX65CUq|fj3m+AA=XanHn&^DIx
zmfdbBVnVvq>B2Ky8J3>wUc6t@x4LF%ITy!TkX@9Y7bg2=CW+h0^>scK>RJ9+Nqf>b
zG46XmvBldgvdmYjM$);ypYbiYcfLt?=d8r+7iCHCzvIcFJ}G=t&fKPCN?ObK<a;vr
znIX-PJqOY3kfUJ=Ig8V-7ndsy_keSS`#!*8ky7mZ{7zFYQb$h2vO~jqR*FpTZc8mr
zZkH@%_zyf2pN3g;BtSot^KW2;a1vYYU8RPz1|6eJWr4T^ID|-fPm{xnJXa$Fn}x9h
z(j2GqvTdfzXpm?+X$`TE&%Q&HT<^qLk|!q?9CH|O$cu$ww{m|Ll*!wyw)~>Ok(r>+
z^-axTy7pp#Fu<h*d%ZBzz?;FI7cV1qmeGO;6L8XYlP94bXX*u}n(Nj1tSswhX6S*&
zYQWspY#eGwY#r~n*r#<PcKL~&ezP>4mAgvC2_?DB^lpWi*Jc-7R39>*lxZL1WuNRm
zZoKDDwYZ;4ydC%v9QirC+3j-yEyTLzhW?I92)u%NrB=#5-Ws~h$HB8Y@Ln15CyOrS
zKG@ZnGbq5JL->amf}s{A`_WA<(o$u1A08_fz(U^13+gvtKGr7Ym0a~ezg3j^Gzd4W
z;khQe(p`8OMwZU#nl$OMwA#Gwvb1niVF}mvG}FqSDp>CF>*%WpFnb{2{>bxa3*O$u
z5AWumF}Jt}3qd@JDOi2?9=G@2MURRADJG|4MtP4a$tOch7l-9U8PXrLss31x_-3!`
z<dFh5$h;=C>*y3-V>Y`6Qz(6`{aW?O>lK-15hJewf5rpv$tb$e?zaHjQ<MQESYgwn
zd4cyqlv4iN6FV$jCLX)f@e`Q#oOj&Avj7>NDI`rKA5PluVlBRxOAE6LM>P^#d((N_
zOASBV8;+j8!wc)1wbnq>vM2h)!Ed+Q+gXc)b5wnrg#X&|rOHM+hMgb^W!9P+YH%H@
zLqfxVT;&HnNp)#UU)X|w2nCkl!TYmYZYvIcQ9nD%ZZ30iehuc;z7q_5-`JN}$R<Z)
zRzvk{4&02`$rhwM{PXrO*8-E<ymu++u;~XCLP2RcRHi5a>N+?p^F=A*yw*RMILr};
zBd^es^a@^GuZdd0>a)MY%{p-gSWfx?_jod#L%aMA(D1qMqj*yhUiH|@Axv1K;4sRD
z-BxVRoN1p5y0e-{+;oin9@qCUnaV>X7bCLcWeRFR!|~7w>|tmaW_c4ysZ00MM2)VY
znbBXc-)(FGP4m!E6$dAZUOpS@dmV(MXqQJ34!a$eu(I$t89>@nc-Gjq&)Gcjy4GQp
zGo6a#xtuVON7-lh1sIHn``4b=3CID`Yo_5K41z=es`Xl<W&2Bqb*Dw1n$HI#&uTVY
z2ZH%&=X0S{mM35@oz4RRob%+u=Kv!M(fyPWwXLS4i|r?F(u?}ZvvZPiAHuph*nq28
z>k#wtT6jU1Bn`>i3fixwBI$Vx7)t(`r?<HkNeXkce-rk-o&vnMA<-kxy(W(+J&#WI
zu9FF|eT%|?<)tYHQhUfo1hNi9zv(7_-c6ra7wh;sQkwVt=h}Ab{E}#cvV!a*FH>qJ
zfu%~;s4VB|i^w$6??g5{89(F)IXST&B~^bgZxl_+U&ImtE}tp@;CL&K)#Gn#&V0TH
z#!CL@vT)w4RFfo$SFf)6%PR2ZpkWtRXT|)uoVT(shbM89BsoF_?4R*QQ(`7>$IP%3
zwTE*$jIc!Jo%?cp=jI;vEO7sQ8k@i9&SmG?VG;Tge9BJ!fXkd+avkSTTmu#&d%_mx
zBCNJK^SGIHNo%{*-Qbr+9-pfRz2l^_UyNXSmeX2=3Gs%px_&YI_$J)rt(WN`Gd=IZ
zX?!tyqFfaYqC`##+Ejc_|7rrNu<*-gX4UsV&Y1_(0&JTsJ?YJEMVlxU&#!Yax!kHu
zCXNK)kT#K|h7@$77Q-4a6?$9q!!k`QvvD`wX~)-}xsuUgzRdBhek0XPJ|ifPj}O=n
zwQ_q)ZUDbT#C<XmI1bCh1&RE{=NRZIr16R33@;`$1jd3b($Mo-Yn(vYDL3qiulEIO
zl&h@W?fYhphRafsa{tuPOHlO#+jb>93YgQbCo(Xl1|{01KpE^Z^Dw)PcA5fo{|=So
z99~y)c{{GCB$AMa-txYITOX1JM57ma4MrboYpzEh3D~IOMWL6(t|t{5As~9U)5lC9
z&1@(jr6gUCV;2*ie_q5oCE>+(DPG5QOLA4^>oTzRd$~u!#}|1%UbL^b5JQhM_rPoN
zRMhjr$+I5R7i6W5qOD#b1*);t$(G%DEt7WC_ph>%@Sp^om+qiIa;Z~iY+DO6kf@~7
z?;CpUEdD-AB&4OQXa1RlDxEA|g*W(GE1;@*+uPso>FUPCW=6n_v%AwnsKV|<$DTkv
z&F7qX<S*;^c>4N!`Z%7ktd^k85I=^U|H#ntxym??tL0zZY0Hx3=K`J;5*6Ip8=)dp
zU&lQ;nGviiCr>R~8Sgmm7(yYu^5B#B<m4J&;%>eLDw6n(pz~)E%CqZ&m|?Q*-)ASb
zPi5}hAQ={VMMc?;l?AN_jjpck>leD3br%=kH`wjWqVeAuO=(v7Dp0)2TI2fJOy5h`
zS3Z9|U24)<OWq>q;w!b`Yi3EEP5T?%J&_#iL$`l{RHABOT}7@u?VBSzW-EgGd`E1F
z2I~a|$@)qPSQ$6okP)M#nEa%#SGl2lPqyrA;J$m4!^IwzyiQ{ppO@?mg4uhMT`SBi
zyNxhBfx8jr#UC&7xn{R8D>W2P+4u0<jbqLIB{ug&3SdVn^LJwoyj8`;l_|w99&gIy
zgrp!40Fb9%DUy$@$-YR%lxG&FpH6Kq6}uw(4R#zX8SBS6`zu+JUGgoVTA`dmwC_N~
za1Q69&{uctDOn#@o90^`$L?H<!}SHLE`9~7j{b$ZS>itaPmzsq&a4OO%P(gC*yS+t
zT}+jzgs&7vvF+Uy_+pSMmDc}q+?}p!qD}K*efZ;dp+0;0#hTv0Mx>L}hJDV`=C~w2
zKGA@Q9qOkxL5i^lDL%&c!cdUZB<+K_lEeNK!M<L8BcW&D*Z-w-CRD^p4%zI3BCTjm
zoDBohtXhl6PHJhUuNB|nIhQQWn$83VLi7dUusc}%>9Mg{eFxu<&Upx18B+l&4q>>O
z;luW@_uYu5c|U0|6mZT~ArOZgrOeG#HsDe)BroY_`JCGpCZtKo*WaafG+u}qo%&0Z
z+ql>bwTTs;HZ6SeGntV6q4#{Sx~4HAZaS6(Qf}mQ(BFKA`<W3KZQ$}@=7@F`#d>uK
zW2Jh}K{q%oPe=84UQ8yE+{9SJqtuY2X5ATCe<7YQKbag}LSSJIJ66Hx1@-Q~nfB?h
zH=l8kpE1$C#6q7z6Moo-VK_F!W<F&9MXb_&zt%t!`WhRb!&47lwVig5YIlW-_LgR=
z@XeBSnJFpH9A+o&+JxSfcP7qbVVatLw#5+#DT$k#^ZENwObBvCdt_A3A`Su!UFvBV
zC^0h%OT%P<6rkAZuAI}kLNOJ8r$T5@NG{)=sc=3E5ij^{aqx6ABQyKe@$e?#JgCyi
z*m_QKe_*7afXYy&iK*%UOH_7PAg0Z7_^3{cHw*!!00!0|?*;r_9u1xIusM|6{C)g+
z@-L}cQ@$cZ*|<qgThlDc*T&S_Kl2Jh2t)HWpAvRu1`vJ$QBsRdqVD*q`u($p1nWCk
z7Jqz;C(HXO?Wmz$NL-@*kyz?+a1bz<oE{>&j9|QcbafGSq^nul3<3yY&iT^;E+5s(
z3@g751KKC76xKqN`dwvsPc;cy_0SsI#<i9o#~wG#;+oKA0}wWSCO!8g%$L8~p~lG5
z1k9WB%A+pK9jC|-B>E|$zQ@rLhgg3-GKO9sMti-vl|1o_dk|yRMN#V`x#wF*S>E~9
zyRYx$cG`&O1|N7g?JgCWf7OhUFQ*>c7*>f%@LQo*5gS{<+(O*^Kn03cuO$S_xV>C)
zb{&Ywje@$&NbTJ{6k$t}Rw_cn<ROQ6QKxU)2b`)_Og|0GSX<pP^aOR>p7zL`?e@Ek
z8WOlavm~vR+gul<7`C}>+t+p*x6z}(I_OMIy}gfwp1{Y>yf#CPS;l;*Zft?7F23x<
zUa{+cl^i$A9y-Od#8b||#9+R?exApfMJ9a`Wj!8gQ>n(mi`<Wwl}EU_ZaZ!Nb(Ce;
z#-osTPZOfT4j6s#4=5MRzY~y6#6G50WU8ty1;!9olllRa0)7Yzo4`^~q=rg;9_FQ{
zrkccG(nq8$lx_otqN?`zv**=#!4kiQs)jSe)dJA<<Wxys`k){Uu-vIm=wF~V*ZbWu
z7TL#WC!TA%6TfuP_x|_|h``Rm9W0Wh_(PSZC}rf4hkj;GyxDBnX%f(ySTIJW)gboJ
zx6s>^MJ|JklO-g3#7S%Vg=no;uOCmy_1cz;*I-BZ=gWDz2tu@vI){(g`Rmsh=7x6z
z`}z6o?M!dDf!<mcjl6v9Q$Z51>zkPR7i36t&O7TFCD3xdG0vOrv9Sg#$YuyHg#5`i
zFKeO%u#lyqUHPH9^1$JZ;g!XyYM$_7t(5vf#g7!=pbb!P=@4SSH1D`go;o>JKK|B?
zwm!MX`keX?nt9oB?MFW3#W&WZks6YJ`hMv`ZUhIv(nc>uHg&qWxmHD#G>Lw{h1&C7
z*r%hgdtrYSbeGWk2_#!6eHTl1J{#5ga93}4{NClAz_FVe>D#g-5FT4$TkbDTU#eJr
zNVco~<7abbQ1eJmBvQ9a=~m^&^-mLh&zhFN^Bd>W38OmHkD#TF49k(ETP;n9Ol|_D
z=;Aum)Cs~jVapDdJXowlESf_0?iQ>Y<=UD0g*JDL!k~t_t&0gL3kd;Ln0VQ5`CL=X
zCHDb`fl|p;VN=B@xg9nZwuwomt|>c4p8FgisQgemDPx_9i%T_$1Dw=|Y22yv)DXzD
zKymlkJs{$-Q9nX2!zD{A$jfe3AbrXIxmktW9N*zsMyF_PnM<wgq;Zb(^(R-F>q}zJ
zYg{WO2iIQ<UO)3s_!fpa&>L(OP7|8ztxGT1BtSn|^3App&JOxK>RChaA4?Ien}s8k
zN8Sl7kdFxWNH(8_;r%?o-^kU|rCseUuN2eHjBUNwNRwFoB$0#Dq386pNM-&Viv=wk
z!J~(vP`;d50D*qXoBU#jL>g{hv<PnD>#sFgm#)1(Iogq^@Niu^`jOq(EJ!+2f2y$c
z@bn3g2n!~-9t?JzSkZrs77rAg2RBR8>>|FNRlPW!2&_6UKISgy3bd+>d{`*#kQf^h
zOxmeYh%9+lt;xdYk)a$S8M0h&m(gCm{Drr^Ih|Pj(&s@YC=d_o00xo?h>=RH_X=;`
ziUE)?RMhG0=&TX5W?TlaG82ERFZwwFQdA{8TdboPwx&(zNq(6F3N%`LH^--W>Lq6h
zZ8sCs$$W`3wPeNZb}u!M--U&Qh>5m)`StYoySrW~^>967#3bE}X(F`zOmfFS{T5ij
zGOU`>(5{*eaU?u8$}RZv50k32ke#%>9gW2F8^=y-DCGA{gK-j;xcAe})(sUqC5z%K
zrU1T*zl=Bb=Jx?PaEmw4uk<i%<dI9T>Nkxld;o-H4tQhwmY?|O)Z^uk545975pkI-
z2+xZZAYX9JH6-%=I!+Uzp!%Gb_e$}h5Re&6#TlX)p*sKe4vy<kvlX`It*X^W-H=r%
z0EDNS5MW77OlLtY`vY<GwomV}nQ>f_+RSLN#EP+>OGzML3DVe^P&cbbf2W}2Cc#F5
z>?dd$Rc|Kv70>AFp{G($qoz{Db6Q8(I(N^O55ohN8FV?oDlmQV$C+u45fa2Je9eIa
zs`NUAO>Gn$Tj!5&4+uwanVq!i3d$wBXWciSTt<gxxmuzFem?xh;6Ujx`oN{H6OO}{
zJ6yp0$C1Gv?0PC9oHa}Xk;3pwjvd@=tuHYu_c1K9_86_X#`lW$1%b&8Vwls-NL+G#
zQC&$vkQTx?N{Gxv-&4Hy^i>zE4OC4$nYIwMZX#~{XLZ_%B+1elbV8!U1dkrHM7HW+
z96=!Q%9d(_>?EiSA2;3~t>y7vsk;f$xgVc8==pv4Q?t%%1qZ|SRUWBftziC)1ml>e
zLsR<FUTAsP)d5AIZj;NS71~7IBw*BK|CQ_2;7Y;Q+T-99vq7Xyq(6x0E`57)SA|h?
z$`w_hSQNZD<4KYO;?cDBhlai}_<^n%T}-?#rp?QvT~$^`M|KriR=u+|G2aVRBfmIQ
zs(f5O-mQI`f`f%3{Ej$Gtlg{Q4`vsfunE0`Et)>hqc4qff1wO-+lj_F{E10<b4pzY
zRhcIY@9<QNEYkOYvd|_j4a$JsEIFdLEad0M`4VLG;%txK75hJ?VtH%e2yn)J-y@ZV
z?-#UgXE$88&5{tILOG~Zz?+(;@O!ByIaA*SP>PoApA8UFfid~bbIvR|@x=^}+i**8
zhDwU?RtvH0bn;T)N?u@7{UDEkqzwwP^umO0egjfq*?qj2GiU_Y?o#B1s|UW>?bjlM
zVfc~qsMz|q;T)SDKa+@R3krwl8`b%ziJSsLa$!UOK~)G66E-g?$TJ|{_1*z_CWvU2
z@G?43c_G^M)aj-+nzq)ik+}3Z++7f8x3~Oww8<JK_?6m0a&G_o^t}|mJek<mfhBoR
zJvNOq3?slR;?H3j#bb%FdHN`6_!t#ukKD@@^ZfJZBJR2VcG8BT)}@GazdSwse054>
zfgVf#P``Bj7C?@3+d}k7t`Q=RE7e`M>33hN@25CU6pSc*h{27;mPhby_?j@tu8Xsa
z{|^aDW2kKVMbj0M&mtL_|M~BGp`s#r6(1g|d{vwJ6=7<w4Qc`cKPvOrNa`ZlfA<_Z
zmll7Sayx6GTp|mP{b-lzkm&QA;160ROk_6I$QI*a8^TvzW6KS&acNY#a~E-tz4NeS
zf*@>+_6)F&!J;>55+TQaO98HHlG=!TE|3~r1YM|L+{u%^V_zhjPCP*Y6irAG&=J{1
z`e1&NBa(WWr4rDi^g#~$)oSk*pm1o_2bDe#Oi3n7eJ6-}lYCe6Kq<7=mM$(7QRB$_
z68)a4(<ch3oDDvj+V^*;79!HYKDVeKlJ~YSX8!<hwyXEwEX<s6GAPGNM>c102(ac(
ze&*X^K4&+=g`pr0$U{IR-t;}Y5zeuG0U_M#PjWp~CUu{0*specIG>E8fawd}(S~y9
zmOw_DZ|J{XK~Dm%ppCATT+?#abd_Q$C7x7!Qv7;;%I*rg5(3nbfdhEo7IfR#Lxr;}
zhUP(mKIcw6(C-i1)4x5w&reBK_kl7EA*9uNNlyNOXpESaZ)oe~t6FA`KS|Iaua0Y0
zqox*waB+0Xmyce<w4a&2A?j=ul*{OV*o-vB=4NZ~=Euf}bOk8T<rz8hc^>JU(j4#O
zgk7Tp+F<$UsWc-rl^#ksKH5Zyfah8Db`JkuWhKwx##*)PUN*Ih)##wZ3rx;@M3EnU
zt3VPsNiPd}-a&062!ei&jsI-7E(Sk;^`zr9#dM8UXc5&ywp28d01hvL&F$&W?)!~*
zlwRu*-Xq`UVx#_~nX9u3zJrz>_L8X{O_09y2)ulGwb$<3h40PbM8_;Lyf=#BU|%L<
z=o2*ySI~coA#luRKDuX(D2roSeFZK9SvYB$+_1iL`^3qrHe%<i=gXA#*mDP6vecHV
z--=V$tHm7wJORu~uoU3_Fy*qJQjs^;#tAzJ4zaZ7-LZV|D40b&Dm&xouTm~5*Jh&N
zFxoWrUvF5Xg7!nqJw!h#K4Q!H`9!gE$w<flZ!JLA;;*j7ca}C%@^!AYJ0>=BORI_u
zY3x1tM4RqYt2aZZ4VtxXfcJjeFCoY1%@d^cK1`Ig%LsM-pfH1pBlf|g;Y`5bl<!>!
z*!4t=fH0E1NXjtIb#8>H*s`F^J&qSLP9t#gC&O_Fx`GOz89x@Q9)8g+-1lM(L6;=|
zW7<5zXnn8?t*u3AGao$h7%jzeyboPib&er0z5RYVo-H!bQ?m9$RN}xBe~pCoL3xua
zqCo9!W5Z8q2}B6mn%54pyHT!+W><ZSUGdG_2+#mrDv}vmtGRAaaFx*awCTME@*Q&Q
zZ{f{v5Kb}kS%;%*w#bmE74f_rKGckT-EI>OsquJ5xYbi5Aa(f=^c3(RaLlWNjAhT<
z^3meP%*E&OpfwsC{-nGhVRyf6tC&gx8kOZoEmbCZ5FNJ|7x%S3{B(@wh12r<zER3|
zFWef|P|#x@S70RdKVAs}_0<^W2aFZSMppe!=W5jf7;V*nn<3kEXW@0~l~DEVy(F>2
zSuH7{K>Llzmd#{TTO%c)xItLXPTc(FTC%o4Ox%wgCszmERknzrt)Gl+A`(q=OtI@T
z2>i2*DoV8RIt#ti#uvj6<>J?Y#*rErNZe*s5&)WPV?)~Mu{EjR^PgO*AZBE{w1ccR
zEF>&&wqvxUJZ_l90@U-bVVEO2f7IDs9SpQtfhNs4WzMxb8&LP-PkQVC#f10bcjjr9
zKRh{l<>$O#V6evrKWZkH*QOee@J5X70u2Q~pHB>gZf6Ztepk>|kl3(&auVjUE#MKP
zV?s2Q@wgOu&S6(+-H7BxV58c?(lE`r@{R_PEyMxK&uuP*<S0Q_XXj|(@Cs2o{p9+S
z2At=}?h<MhSn7(~u(R}Lul^i)?=>Y6$Y;#oSos|<YvDZ3s@!#gnU+o2ie5gHMbafp
ztWr*Tf%hF@<b&khADh{!BLhc8n08mA<VAK!*g!LazI-*8@Vx?WypHln{(4j+%;8Gi
zemA<Ax#FXrJppehM&i*-#l`p70j(>Kqrz4V(MhEKWYTrN<xsxk_G^Wrjz;a8%TALm
z-0-O{{u5KJbhU3K-D~05qMSo^W2f7v<AnyqYz9R6g#Az>_mHUvvgMhp#IuMfeLcxL
zZkFYHRbZ8vA03TwZw?J!8s`5Fmq}d{WBc8N8(oxVFb2v{?5ul*c`2?&SEO)%(VEqA
z7OAPxMVvHE>-&?X+cRUuPN4WHFQ_*JN{=9X0GiessCX}jhHIM{30u|rSnDdlrJ)^%
zsmZ;0mS&<R2n^)+@6tSwvx?KIs%kK_7sGgCHMD@F9|R+*p&$#(bgU1<Y!$6WrNQ-C
zYF5VPUE6Q6#1GGgN97pCbu}aC19Cpv>01kVh8;Thm7irnOq-H<Pqx92xk)ktPCGv}
zEPA4T?QZFxU07fB6+~WZJ{De3nD^^NP{0JV2G_5f0y1lOuEI3@mthK`Eq@lOwBv8u
zB|i+Dk#Jhf^l=)D7y-0Fb}y-vgr{-`I=C}<?KMCcE=n)NPSdrNR@F~<Wvv8>DlU0*
zX1AG)8g;uSnu%AFq~UuDu5P4u%tiN@ZR%k6Dm`>Eo4~SLpHH*$vDf5>K8Mu&E~oZ7
zM!X+ZAmXdx-ieH%G&1)jB&GWu9;=XKM!BUa+(m60ca)Los{5N4stdE*b}hCT8;8y!
z%Gpa80LN+WM0_`L<m9iX>A#+?@>n{6F3d{X#n|4j`?-XbhZE~Ve~G=Zjn+>2^uwTh
zqiy`}wO>{3@h(4w4Zb^;v}5WLc&elr{cRKdadAn?NZNJS==KZ;NzMg)y`dnpM@J;*
zK`DMI`$NeAQ*M9|4Q*4Nagg0!lj1E*WG0R-mK=y&>%lthB`0Xcy$I`9Huc&21=xx^
zP2$?e(eKa(lqy-Y6(p^fF)>V1v9pZgpcAKbTa6Hz0O(d_Ml#DYtX>0*aVO0{9FAr$
z$aXoneiqZ42XYx4E%`!y0P;hoa4GbPNoDuMCIqAwl5xx#j^@72t_kA!cR<4lKg)k2
zXG4{xn1*+GwA~a8{^M%CL*ts}mshwDBpYfAv{%sYO5ayJ5?l0AeorZXZ-asAHBYpf
z5~{RHzr@_52Ul4Efr$4C(L)kxcCW1!^oO0pcCq|ju#*ZC{A$IyHVEpp)jZxU$2z%e
z2?Zj2ov~LxVyBByLiOsJgddirb@5LSiShK@GVy;DUZxHGRY4NxvE$?#MYB9081@d_
zq+nGSbcpV6M0Thy313c;%^khIu3qVmQSLl3-<mR_U%eeiCaOD*bIKj-9g9cMl3BQ4
zXjv^Ig-549NWo-MD@KZ0wJ3L9ykftgIM^B188&_RNq%T%wza*NK7n3#Edz*Bv+{r1
z(EyAqfJu-ub8nciTfXs-Bok+Iy^H6UI99~X+A2}9AS*wJ@>sH<4AW+v-*J1#SPn^t
zHpC@!@+xee3o`O+NxWp7@SMzWn&9h8*zGUF-Q`_k+ufdZUo)j?#zW=V>uM==5A;Nd
z-QYCx@4Ovi)-p4mNFZdp{aqXXG~%YOUzgdc%A1#Zt?<p7^^Ef>tHWk)b~?l~G465%
zfKI*PudYz3OR6ela!LRWMdgn|fcl1jAf6zu91u{LKyy<au7=a!@z}1thaNN~tipwa
zh5%EgkZ$>g1hC+!Y`8&qHj*_A4K&XNQ!ju2L*|g9?4mi)a&S{?3ByrPA1rzST+G%b
zKn-_oY~h<mWuxDB(zn8Ax<L|i@qH`{g@5+6^dNbvsYLkBuXu<O>2U--B3zY3@Ty!h
zje<UdY|1q~B{q5o86T{K8#Lb(EdJ#pUi~(5{gv{gz!fK0&4(SYQglI}gawC|VaH*Z
z$(EB}>GLs~_7R6AC%2?C&JNX{kUlR)28xv5Gbf9kMQ8itcc%?D{TBpJN3Vmi<ar<p
z>h={ItvBjg!qAK9W3qFr0E>+Rss5Leo>~k8#+&sUt*42t<sjUbN8<Hj@JrY2DjrWo
z7YO!_m{RLsV4X>SVuQkau1~i7<q-q!!_X*c8rGAVGDN9dG>v}>Y;?cr#!0-+75n31
zCPg7g0fq!+D+~_?SRP7){4uM)x2Km-FJpCW&9LwL!_zXS@4hGBMIOC+H4@(DZO6PO
zeHL?NpWe|}X!2P1BMjbXVg?slRm8(M##kY+bKRCVJ|nSH?sa&^8p-kOXTl4==!I1O
zPQZbhBKAhXZWH;m36|M83U!t+^6tM~5=5dQv;xZ2>5cnUv};-a^g6>6=igRfp!CD+
zge^dU&AQXY7}FGd(bPQ)>QGP2#J_SPlPSL&1DaDHG!WJIiNCfFlI-~+5$ziwEIYG5
zjsP-E5NH7Ev_uSw`rzI-cV2EDBKgo)oP^QWa@}>6A1^0@gqIP<2g-wY@`VZns)xd;
zY^G>vrcW_3<X$<n<ScbvCo@=JK@PF6r<&JDDIu~A0bBz%N)Nnf0qG4Q1k*n+zcIce
z&!x9|m09}eLN8?i81iVbdIb}mSTD^etF6ic>dwYfX%H^B40*XL*Hkr1DRi-k4Q(a{
z!+%iL4)Rr-XVs-=Dz7F}<U#H8ez^VOQeT3Zy@#;&#RW1tcIHh}>YYgzkW9^}nG5pP
z#Dj-?d&H+UN2wbxhmLzG11*58$(**FCs28&E4Vg-sSahEOC_&-wh&vYfn{-Eo||{{
zc__Pv1Okk<T;uWBqrmal6Y<^yUBUoWt&GqXjV_>-a!@1erZ&5Cdu|}%fk4VARVp88
z<KaIu4>c0yZuRZkY0x2fg}ceY?>zTRF#1lrPDyNrd<G_A#RxxqIJJk?`ZQm{$Ib0b
zQaYNij|yZVB$4dQQhqP74!)?v6+Z9O89hw!NfhRCYd+}tT`u$y)*zd_JChva*S}zv
z8p2#xVmhs%d)WJZw00phsXK`jcPNEBlsJWB+vf!LE~OcRS}$Gfs&e{Dar$ogl-Q5<
z%ORzhgCODDzg%I!d%m{T)SV%9`2jTtzUI_Vq@2p<4$^|o%xBJLR1N$W_(UXMTT4f0
zb4P)D0NMlMInaOx(t{U(ypnh<w-@WCd!;bu5Y?2rRc#z6y6Y2+()~KNXz=%VX|%?H
zye;5dlckzJ2hhtQv#vJ3-#I%$XMeLf7QylaGBUz;nbf~u0Q6Xoe8F&`#_0CDl5f=K
zD<}mJ5p2%Y_}(Y3Dh&j!W&?O!SF>d})rJ_guixw}wkT*<k}FG9>$%#yd8m`oYY><7
zaTWF}HRAR!s33cB*<PqTkCUho%Mxj?xeA#|c6J5gc3#X@@kb>bIfx>y_9WuMl&>;F
zI3KuxcC^AIKNd635Ts@A?DnOZ@G+gRU4PV3I_}7Ovh;T%7xEa#uv!m0@g)D#aTIr^
zlaFL3u60Krko}*@+>Yq^hk_xs;-0sT{gm|FqjnqV6?@;INbs&%QeY0b>{_VD3BB%7
z7b_D^TD(_jyfEG6#3ZC>D<;q;QCF^YWjnuZRHa6uaedk`u*Z44b}3;*JGbCn5QRm6
zkV2(9^QrIskYaRaQU2`H)5gG9-u<-5(!gD9Z6*U#lV!0UMpnJ+jvpt*vK74B&1%n|
z^y~pingFda3?QK79vPc(=bqEpoDOjF$=TCF_ZJb?HO=y6!)a@pcxWJB*&Ejscj~>P
z=-GXccxz&Imge_e+6R2eOc`RGC>Bm<o2HwycRkf4DZ8CfV!v?pVPLpg!8MQ)otj-y
z9V+(>F)@)V@J33C(c_uW$Aa<j_qiOK>*{talkdI$07R4_L#sziIOeJth9P@aJ)dI~
z<3vD|0_>!ocn3&LlGH4K)DIZJRF`WB?I1wB0H4go#lwL{^O6%TKN4*!E$qQ@`Vpip
z09>!?KUw1Me5L!0IzvhR=BLuAO{j5PIRMjl2pN7Z1417t+YFR?Q;!eV0xj0%+$#~%
zX>b!uCF_3NGSPDLmd{)X!*W@-@_XizTZm3%7}7|x<k#2T5_ef3<Xh9<)|YW-eOhyl
zN6Zf?Zj4<yTpehxY-wB_#^dT=d`jWPY%XYiD<L25!NNOj<FNg8dkZf=*ijOLzqq{o
zrm!sBEIoRl7sQG9CY!B^7l{OQzgZc>+&KqWZC;r2W8&Pz%41@3;+7=&yi5!T331Y7
z(UP&32@(~hA6M^2-SBg9P^qR5apF|?@FC{jaKQS#?jZzHt%eK7{=BJ)m<s5{`i`K#
z@9NxD`>%m;0@J!<ita6X&=pzDdD=x;RZa&i^tu-`b<wn*@{Fi`umg;~gv&wh=-ypt
z<Uz(*fe};-e!7L1=l=}L`MNcy++qJ}l->qN8Qioy*G9s&x5BpOJZ^MD-~131dbz){
z7d2@D8L|H5p=oTp+_7@7-F-kFBg(adrz#ye{QH2qy!)<)vRhA6#974rx?=_OQ@ZIX
zn~$`#(cr=UI7L=!;1uUuuCW<>+ZDcYJ?wL`@9ULpHw5i{6YX;E?=*a%;bP&0=BI3X
z<HcxS6=<wz3tzBybIh&^IOh9P64v&xTs_TGl|WL&m^!U+bz~%Gvi7L7+dZ{6XS397
z<+Zpf+2R<ZJE-&%yQA}3;N(C4Kzh?F+#76$us=0zou4VOV+=J<)yPv#(X;LN-VG<w
zWtja1+3A+;$G<l3Xc}>mr~aPiENHL0_d4Xl|DfFKKDka4t$TXF(>)7}>a&jd@zLV$
z_1g9>aYqLrWxWC_^jN~fb-^YN*PaJT#Il-LTWtywoOCrnouEY3a0Kd5sBDv?60}Jg
zT51yc*xPxYB!oQtbs5jgif0i6z>P>tZlYbZ-*J0Q%I!n7fF(ju?s!#2T;K~qf(>)p
zBS9C*qD>C~{k84Kg?1YP)|=m!kn~2y5ttJ%IEg{?geZ3=?fqDaL|0&v3xgE`zmp1d
z;=d$2LfIOMcYZiP$~F&+w8vmKLfFK0^BP=Ff`}ciwSTF-*Uq7*&srMq1Y-9olh>e#
zMC`Dm{@nT6_O{05#MwdR<?l5GWUuPMqhWRr`=cd#$-R#U7wx{59@`pR=fM&(Tprdn
z?H@Y(%=(vC?6*BH=0(dt1~1pOwEJSwS@`qddR)*J)ZZVE^B{$X^2c3dZRc9;dR}pd
zDwlABW<<gQ_X`R$4PmYu)&7aVzKr|O+uJwo82Red$QoQ*--hJAmuqno@VE^e#>jEq
zbOQtJ<5DKasyv>2ThR|0&Qx*Q@Az?3ExWet2Tsbqc+3*g5;!SCbuOI5Gg$}`<fB{N
z8<9cw)w6BLl1@|fbi_rPDM>^d_Uc>(<<Cv-#^I2gROc(#fIEr=#Bw_cps8B`ap}Bg
zZEidU-?@NLlKbgIv02yqmsktd3u{C73{ADI6+;rPA&>sF6$&6lWzFL@F0dWk-iUQr
zWM7L5EvwtL@*)kcSB1styp)242K7yuI%=qX-Yn$wBEJlb)ay*+dgF3Iut?v~TyAdb
z@}2_9;iCa_@B|FZb?DtF2%sqskpelmD?gSR$l_1SkZ_h$Sy`@4uI47$D)rOZgA#Wu
zjk4;ayIbX_ly+LY?-2>{C?xdTg(UY7IQA+vgKA@w$Zev0Uqda7HNjfeo64go2hroU
zTIrmSn{<sRGNs2Zq?+bUR6P?i1g4Ht4_Cw)*we^!#4Ef#v}H`-Y%_iBfxsaov^;4n
z+9jio_2_i%tNZzOmYw?p>*3t&JrX__)9T5(#!JG`kV`I1&K6Vs^GEYfFvIBF9{RfJ
zHAT?rkY4mgDIsq*eGTjLJliY4qw~$v-n@v%aKB#UZvFf1%9i!+{m^6tsOtsN)eVkt
z;^iLvWd6#xL48Ln1()W!!jpshfcwPmua5gLql01ujsSDi_{oZVk2QYMrK#i9L+O51
zDAa&Rk5#C}2qOA*vdgM3H>n`W{QwP0A;*6^)PhQX3QzfSv|9<s4%OCvBOUvgjSI3p
zxwKiekCShXp)X8lxtsfyYfu5l@EUk3NXjJ8AfL&+65?sc?VD@J@_4l8q5Js~(cIRw
zFUq1k2Sr}?0HuuQVD>b{&F;&WrJv00Q8-(}8g^nr7hMxVA>GvuT0EbHd$P$cl5e?g
z1smK4S=7@FObOjX{d@_Gu4qf#OAl`DL9DJL?m;}TcG-U371y4C5=z8I{6;B-@iOuQ
zBlH?HOS-nydVDuEUp=#n4MMUps^5HGY7QTB3S@r>gI>ICl?&R|OKC?Ljn^y_WKXua
zJtlG+i&UD>@$!NC*8g@MU)|eB^Z9u&DTo6e5>~uNhZ*WtP<X9C5${6GoNMv*=pE=f
zT+Ar>N=AzStv{Y@iZj`oKkM(xEr4Ddd@_3<waM(2!@)SC{+fXNNy4t9gu%()c@$_E
z;`y+DSZb%fYy|iguDi)B!%@!WnaqQjU2tuWRVh{>IsJZ1-~M}%RG#HIn8&ieTHILm
z=8T3Y`WMb3G8c-p6K%I@i-^U?lRzRWljj;EXF>yr%qEfKk;^?~ObpiC<vAvBQp|bm
zY7^Pt-Oc&R0l09SNLtb%K!Y$ibMUmN{Zpf3-rmeLN099lLExrzFk)G@Al-XX*G?I{
z2eNn}t(QV{AdV&zhCun$ilXUAx!*7ifJRyZs!ZAfbfnUyqArJL0=>p{$C9}pnMNX<
zjRO}RfjJQ4#5>s<L?ZpSF`(=?Lr`<zGUynNEH<>ekDSI1&{a0N2mEJ82RA*h^*P<q
zh$el#P0s41>rz=d*<BxO@zPiUrIy=4u=p{KfFg0jAT<^t_u@y*b)lcRy0op;GX6oN
zRX?+c3)bzEt$Qct5m*5QO&*uOHL?pPjeMIXZxniYCkAB3>Iz{>_#@+s4>ICIKSNJ;
zB1ly!{ikO_8IFn)d--d@8j-q1>6_R~p&t)J69!++5=PWt^Eg+HW_n1u)5gr;f&xzf
z@0;k!?U1SKy8=(8pIXxyd1_QEBJDUpO3it;IE|@*GJ0*%?`p12M#84jn`9?@b<T}M
zF`w{x(E8U~WR3;DwXvefYZN7Nj<u8r21$0q2Jf?vquv9onFDyD96ygW)3O?XG~zFJ
z7b3DUb7xQ?>SKQ`f8`oHNO?Z|7K%~td0sQJi|CcNrLNp?lpI<ImO5m-`u|y`kV1z#
z=If8(qaXtQ-d_X3Wv2s8Bz&CM%Jf(VskfC9O{Y!FC)}4$VDuTi-HT1#iaN8~v-j}X
z0B%&q2UW_9>&NI|63ndfNPc_L%3Xs7get>euO85`L~u*`&G@@9rc}U|?)Ln>!5BeD
zmVmq8<LPISLxK})A32c$JjVB9xPAh2nALg0<Lq+d+*3wfIsB)5;*D$GL;p=f>#l%O
z)-YZIK1QII>DCy}%rWV5E0KTruBhbw-~EZZDIWqeevg5CoRi-P2tYv>1%n8gPxAt=
zg=S)2*b_aBPh7t~$FRpvX>jwS=BNY%pnNy{Mp=jGPO)?q$O}9aWSGet^mRf^(SS1c
zUp858Fp|it8dRjY9+DA{>^}55+@D`D?VKx@hf-gk6~b>?ll~et<&|awzx$+rZe^eD
z=GS;_J`YK)Kc9nctNwG(zdYEK1~RXIdH(MY6`ua*jJChuF+TWbTGHQ7|Ly<AXmEe#
ze|zO$o|Lm7aQw^De|uP%@n^8jzrFVN6EE7}{~Y-D_fB(@yJ;)`^`<QWTIBxgU0^5(
z!T<e=|ATY$f3M>IazjN`s=2dN7{5GpYS~l7@)GCY=dJ#Ar3}$V?qcVEy~#@POb~)y
zFysN~kDUK*`s!c5QV#P0lDhxV2Fuj`;}rJcB~I-;VH#J>DHzfGzk2t@ERl&+`UbpS
zJ%0da-5M$WsbBXC-%rY`F+$tqsSFdlwwl_`_HzqsWD&I%YSnbJ@u0@I@(x-gAH`6X
z-EEIAQR9$x%aokzoa*Y_>Ma)pR1EvE>>&r~<eWLqhGzY9=<OW5V4F;T$g+YkSY#sO
z-w*IrqmjWshcN$hSJ%sX|KV!NUj3i;+8NbPAdC6OT!8<V-~H5O;;+yFfy7F~&)^6A
zQJN1@c4u#sw+!&^!OGxko^aA`Is{oWc(W%Q`j1Cb@RPZ&Kd-qS@yBQa^Ni)e`lss^
z_<Fs@58t@G4)wfFZwZH#WsoCuW#IkW55#o0r1;+=)t_r=5dI%d+Dk3s|A()|SJl_k
z)jWnM3^c%R_KE@60o(Af=gpWWIKMzHdGkMgiY{xn?{LmPHeNQI2jW@()%mqty}Yd6
z20**z@~i<){O-H}epyhAiZJZWhxCln9i4%Vi%2`+nDg&Xf;4`){P{1R{MYLO-(SxE
z^DTcpNE!YA(rbqK1qI!7{}J0GSO2#$u&X~;eA(J&|9ZvcQTNr}8EzLHDgz=H9%cKU
z2e(y)6}(|U#pCqma`KmXb_RlJ6YSx$@6v0FXG=~E?&ePI92K7vvQSGzI0Y6gxrxsd
zA&ngN(vIb!lP{bznV2g2f4(y|uQkcXAAD<IPiR!Via!+1YLjB^V<c-(e~Bdf_M6%q
z+WSW@+;d&_tHU1SaV7R`H@~i2y-~5){J}=qb?i2N)Mm^=u+Sa!)SHRV0L4WuT3r~T
z9qz8=2;eeP4Tsnf4f6BgzSgF`qn#MhIz`<VUmUNl*H=!D5<dq@yHMQUYy^iK9FN)8
z`y|UM$d0>XVwRJGvEyFAFem5uL`Yr1;0iE4>@>T>{4;^7o2k9sasX?wGHR)-jDsBw
zd!($dqqtwDv`y|zA-~Pe;&tnK0AcWZKm@FMBVk(whJwAasC~8K!o6Fua|9a3@Rn#W
zWT{118;!w<2p8da@#XYl;50-IUG`gFXWyyZp!*1TN_01{Sa9q5cB)P?SG1kvPVrA6
zUZ(3-dcjVIo`mrz<o>Yu)wMy~er=W{1ajY9>(BRHPuMmmbx}gcUS3S8IIb%yg6mr7
zuzh*Ll^D3B?DVPXC<ZAITtjiH7Yublre2sE3ALLibipe(cw49?B^3>uq|2qjwqEa!
z&$!OG5^wpk+)Lf}g<|eQ7~bNZ&Nj?Z*F0%-pNYck?(SY&TYK^4XtZ_KTOEv4((WM$
z7O9y*XUFFl9W{)&J*9(-oUuZde~&v41?^aIC-%+NA-_#^Ki*t6xrQRie1|{=&;Km?
zAH!7~h0-{$WucegIL>;fwh3Q+nV9%mw;F%5G6m=LG%7JnQ32;`U4zJ}Cg$!fb(W{W
z*4!N~bHvNfRv?J$CtSh7?BICg33G$ncHl`B=HJ1hWcae%<<}N*(mJ)qZCM7c85572
zqnrHT?XlA#barJ9n5}t>0RpUG0;wcP@M@i7<9pcMxqKrAhFeN{^|f|3^*ORLQVZPM
zO>=uevnOia!c$Y|bjIM?m~QO#7@WR=&5xJPSMtk`D8D>${haXocE<^y!%N3rdUj@3
znUzgTXeWia=G2|NEEV}K$+<XD65(;<x@wJ>GWR{q1+LV<d9d!y`w+POkrfRevCDUe
zsimV%T@Am;`T3jMus=s*a(6Te8f*G%7pL9nH3m80B;Cvoz3{5NJhY+_p_fLh^BqtM
zxF^hzb5A6nkUW%h;mZ?T+!}n2hZbTFzOC=cINgsok|7_o`tszRQ~jOL>Ozx<6dcNz
ziL8Fhy|QIx5QfF~h0iiy+Gx$p++w~TFoOgPy6Hx;yy%dB1@ThgqX&sIq_f(L6D-#4
z1{L+#orGHI5`jjA^(kULzIyhn<?bhwFQ#99*78%8dH5|{4voRK#H{zs>@$2jkTQ7c
zEqbHuAM=s8jnE1Sp@at23qT-vNd3?x-zUEgZ?EdUKKa|D%mu$E=<ctFLJ9W^&sK#=
zr<@8<f<>0Pzl=Hj@f%J^I~qi_u1;<(BIyS!C3tRT?gd-+bRM3vKlY&8v^}SFQhG~L
zcC@Rqc$5tl9$WP474|ZChU8ii74NDG0;zjGvy01>h31@OF8AXmqhAW7RHJOz4H9qA
z6<k}~HiL6QAiCpNsY+OcJ7l!XH`&zJJtaAaoY?4Go$f2zJIvo+CkelA&^a04yTire
z`C8%f)pT7&m02C5RqVMDXTzCq-P$JpM|qZWmXbO4lREPty(C*7N6Fe*sZN(IU8@(>
zd2>6Y1I~xSkCH#snkD70DqR_0UC03+mu}7y-QZb7;u5t+^`XF`#evw}^GWpC+}_u>
zyuEdJ4*gqlNF#8D&3rTw_PN+0Ld`^>q+c9+NJDioa%4m4>>HdOqesrkOlT>*2oRh*
z49b8^&`s#gBoH=8i?Br}ZQ*&z{lNW+EF_&E1%bvFscmI(e&uEY*rwiah1a!ojr**J
z%k9*m2eToj%G&b$$iQd3ufHuTrgmc1M1kGKJE*@$n%T|&@hEl1($~m+&%<V)Dou>O
zqMwk^Do!RshO*lDOuWK|f2@6}*r4EBN4QWz+YUD|E7=^?Ax;H_RWfNtwhF91El%Y)
zzJg>{y*J;z;p6ck4|W5auJ&oLZFTy3h`u)bC;U^_YwgIW?5*F|sj~PDe)4%;dvx8N
zwKmeXBziEnd|iVEt7;e3db|F)4aGI-2;|7LJf|b8AIMlML`PYt=el(}qyc8(sBTx?
zYwZ1BCq)aSllUr=Hxp{N_?XpZ=hPp)prP|Ux_9O<Gu0HU*+ybOwy|Yaj8XV-th10*
z{!RbZ!z8b}b6UZwbAI?<QaXW=&p`YKIIW6js&}we?6h_9HvX)acT3?D*-6rWYwxZ9
zqUxeP(DC6BPzePQX$b-86l4%k8VNx{x<NvQ9J&;xK{^LSkPhh>LP_avh5?3-K^g`I
z?iqaE_w)V-_lJA8d=Q2cd#%0lyVgEOpzFB>3FDmAN8lUFNBoCT>cp?_J&}J2YFjr*
zr`>j~-P4<AXv};BlFE{3_7+NODe}E^USnAb-39y-bh$OVnbJYNx#H^U$yEU)w7cUM
z)F&pY#yyH0`ho6k)!mop7w^s&n9A*`I9Zg|EZ8oEz9Bg_+TmqN>|)9v0B%NfaWcOo
zyc6DAOCT!7Q+UG-a1m_QNcKE1_0UoK@gRmFKg>){m#TrJpCK$B74C7oQQ6tAwpLP?
z_#nyOJcQmn8DQ#49ppCSRs9<Ju+Sy1JBWH}LU0-R8@#HvRjV_Inr|Lrt^ZNm;&ZY^
z;BE_x(?N(+)ozGjcDE$3biVO7&P4nqJ-+xd*ZW|d$GV5ExrlB_QD=HnD$*K;h_SX;
zOTizazY(zcH>hAqi+uHw^o#M#dt{5p2w@hT!$*UO@Eg+ng*OTpmrmK8<(Q>z&_gQ6
z-_D=<`Kq|%fg<z)2ZJd84ZLFAl}FX6sgb6zDmsdH0=--4=`<v=SX<tBW#(&oyqe;e
z=i)Sn2rt8VuVvrE&*`-Uh$(+3mp;`srn#mrMPrAkOZQ!1X<EQykl^Dp&oOTau9>D-
z1TE?$O$2x?+<`mW)}5Gw!3rIP_{$C^Up6lYyp_Pa@%a7j%-ea7@=Mnm{Zm|+<8W^&
zoFwjRGV232XUq2fnUJ&$W*f-ap%^8>tow5<*W<|8hV1ZW3d?IBYwQY+1<pDhhaW(}
zGKbgevLvVb1b)3PyWw@UkzRA6-)Ty0?Tdlk8UE4yW6yQUo##Y8eYu)*arYXpAGL7q
zmpI$EGBZ3qEg+|VFBd0qC6`fZ1uZ%A;|_9ka!ezUSmj|2a&#~G+_`@l-ZX1h?c#i~
ziDrUa9eZ2vT-v#~%ir0ei9ATFeCF@@!{Y@0`;aDUx0xpwt~}5(EFrd;F_mq!yGly$
z_OxwGv01-To`CM)pn*PHCfCSchkWuRYU9c1qZwwvxJYo~90M>Sx;kD!7&{tnL8OKo
zx=FkxqTcJg00j^vo}_u(AyybF@7VQ=&_u$G{VrS~)_wE+8t=GvTOjD867f`+;gntT
zF4}~8z?jv(v9=MpP45CdM&}uEwEt+l%DqBZQ%)RXdMUENPsD_tqMMG7AKH2M&mR{$
z(d5o2a=U$!Ny@nrnm<t$3Yz}ZT)B>9D~ITBVOQs8?EM$5cPX-fknLE`s)88NZg!~6
zB5oiyjDQK50&65cy<vrXJ0GoEtF7cdcEz{VuW~X6-rnABHL5&WI&k-e({K5IWeTIU
z{Jw=V^{>&qOY>5Xz0%2=4qX79!5Jc)&Pb<wqgBhq+*6e+9Nt7@G*n>z%>FuY-S0;?
zn|EjC!_n249~W&zCRUSPyWS`8x<eY%*X}9l{wOqYoL3V8tiLU3wIGgBGFKyG^3&Z2
zvJ@*AASBBsdH#73E4$1LJNJ*`3w*1v;t2$4?ykLAfEkLD`UmI9filVZEVih8ua*AZ
zPX!_B$d#29Ih2Ev^7#cWz@0l_yaC$x`jNpH<XB03uoHbUNqST!?sc!paJAJ4@=R)*
zWjR?hxT)579_d}hY)@Ee*Yrtck5{!WW~)Cs$`}z}@G|xl*_ja*l4u$^zO8qFjhyne
zHDw|~{^g}n%|z`}rYnvr$2eb<11Saw&eXA*gpb|in~to`%bR2%f%;1`6<xfeS3e)#
z7w+MxNE^RktTt^9Ay10$0*njNoXBlZtJvp6MfNE1o|#zF=A4bdG?{>7&R`3F1HXT!
zlLTu+(gzemk&O{6r-0^pvy6N=w?!e?%&}?CvEpf{Cl{)z|8gLFX9tMx*LEBV!#g|E
z)>#Fk@{~Jp#)jj|-xS_t6R?R5`mYpPZbbZUoj<p#tgo$|w10|h@vm%|L-axPNk~8@
zjb_@cK6@;-NK`Y0#;`u#=euZbf{?V~tChHXofB^Ru)e{2FR%VxlW&=R#Xl*EX~b9-
zeo^qPzF)b`T%Nzi>rSG|BZmJ;xnMEvZA4DOJ*=VpC01(J8u?C+cc|`}-YtZBDmFu!
zuu?4((JW@`lz&N}q&ddUX)IE7*7oJ|Ig{#J4GH!uxm9}eI`7h#D{u%$PHo@BhSsWT
zoFZjrAZk}lq?T$pTFL7+)d8u{YL^W5iKw4<HSCKYGmBf&$1PT^GM)C*!u;rh^fkBW
zhmQsNfdBxq)Qph7ihly$xd+PsG1$w4lN=C;dAv5BNr(*k{RaC8Ao2(OakH0szOFvy
zi9cfRfWE$tWHZjWLJj0aZx|hS#K{rv+;vF2Cq8c?gjs2-Rnv2=_)$kYFlg>i_kDCD
zsO9RiX&kd-j>Hch7SK{S93Gh%d0uIU<Yj1Td{zczIK2nicVE5}O3Zw&@6+(0T%J=o
z7ssUKYbKr7BEfbknWsKt^UuU~j~R?vOLLSsR`VOOjRsvV_dN@FDzF>7cdev{uB3zk
zIqDg6dw2Od9*`sHb(;M2TAt$vn9o#AmUGxUg~99ma?rs9!{+l`Q?K47aMRwrjlfxh
z%GNLLZo^k<1Cx9{I&lN14CHRJR`AIvrcod;elQ_Iy(YgSkFX58pouvNb%)IE9DQam
zra0nHIA}yi0ii(g@H?NgmCBAIAe-HrbPiNYq2AM#pr@^bRA?{&j7=uQ!B8m51EH0P
z{Y&p6et^Sa`+Vo0!8|4r7}mk`7O<}I7Z4#Rzv{+p#Q<Q#p)dUtlaM=SVkY%ckZP%P
zoI62HKo}&C%jVgCeeJi~)t{O*zC7K(FbbUAIlyJD>_J0-H<a+S5>-y)D&G%Bp^a;z
zYcd%gH$gIrgLlsDUJn|h#TBdl%<K{BwRkFJkY5Rx1cesA)wQ3dO5QiG<6XL&7HGZ@
zw%_+$&h*4Uwg-xg-`9$1TKNGp(n1TRylo<)#Zpm@tRf#a7g2rB!sok=p?JM(4I(3t
z>Kc-LF~v?}e#wF6j+c>D&s8$gAD<*#@KZdE6;iVJ_I<$fp7oeoO2sgsdAI&tM@pcF
zaNZ0d2fuef?76Y4fxmGBR9NWZG?e6Pj*8r-`8mbS#e=*QTEHg)1&G4J<37UHW@mY$
z2L@-!Yguhg%g(I)>xV-wEU$C+>!h7O={D;N_a-f|wUTt=g7QWuZ{&o8m)nRUX7tup
z)S={S{ZvWKwpL813i#r*0NlEs-bhpTEj?l%>P7yvD!Za#^bZEDps|Kbwrv9G5@zjf
z+gu3j<B}l`{%<4)I%X$vSzYWg)HIWCa>uIVL0|X(CO{zb$zJb_Wc#Lr4b+a>88jDo
zO2K#&omSEbay!zR^HIQgC%o(t=;vX*=#(E7|MQ07cR{=@7yI2yOQXEar&z7s%d!tv
zl*g%eyhXsAE$2HId#$C%XT2~sX@@1kH*Bb$kl?i5ywSJw{^hlNG!G;Z+x>F@Vf#X*
z{E3rvE67!?E4=FOXVOlePnrC1>);T43~6x(@@;8C7}6j9KI+IQ(YTHlHh+K4M&q74
zchhbZi`r%6g`tX;A?;ypdAojz(PN3iHb$>r^=<K|#0+ujOv1v6dCIyr(IBa6MIh6R
zK;o>t-t;HXW(gcrjMKWMjDaX7qMIPT2e3;Umr=fZKwq25A(`OQlQsJRCIo7I{$QN|
zGhrumowxa8I<~jl-0MR;_Y{cozyDha_<nETi!lh0?_H!m_MmW9zKhb%dt!kftI^ft
z&B+X;IBddEMz<%fqb{n98^S=`P`ld*1e@ZIZps5jvxh|JCYCh*Jxg}6kdXiHvJl>-
zU@FmuGZAW<t@21-muUXw=acW2;SlPYT4f+5*acF8V|G0G#0#2^$TNA9WL8f}tQ;X8
zK0T?RV@_{NjPqLgA|SVs7A!XL3jGm_Pu}<e5zTA$%mhO~Zv*zyeD_y&)EFe)mM-%t
z)6e!5JC?jbAVo1YFk+h~a>aLABp{|v)WdIUfoXBjOP^ZlJt0U_bJx@9d=z=hf0Aa(
z!pmyT`}{!rqV9=9xzyF%x|Iv${o-C~{+_~{SFkvk5Oe$8m&q{0w{K_d04FqU+<ab}
zg9*fJ*kSRsOZp{zC+e`}K;%+xTE>obx4Gl(-d~ObPxfZxhv~j=NmVu%ToQCzw`Ug@
ze;>YJOnccJvWw3lk{xe!8ELJ?@bs9`D`F8{zaQ4O7PH%I!_s8woTrllR>_s#O<Znt
z8@79Ivk!@|>XS@oO0~E~Ua*nyO;4(;x*Yw`lTAmd$E**2)E-QsctnacBy#s6St7{_
z0?WlMIX)Z%q=o|c7O-E7n+wWzkG@MB)+8#?Ks=BWM(__vv%TZjF@i6*d@r{?31Vy(
z@S@qnS>%k>X5`iGKw-cwKM*K;*w=`504d8Q)-o=VDLE^MYOal@6FHJoAtjkHWKMMK
z8|a1~)FCOX|JaDDHYGAKcWt*~5HIh4%_}n`MLUJ*(G1%~-FqV1vk`@H5GTj1H50{v
z-Tl8`1=>E`W9QuO2ck9QOWq|3ER|^NK7?r8oXBNR_;p{y?!4H9_x_v+u?zYcxal^+
zMu5(%mIYYnG_RiG?N6@49n@Z+QGdzWmf9afaPwCRzlKZNo5g5Fg~E7EjH$wYL$rIk
zH2)}!0QP}%<;ff8K)^tlkl`>I(3pmC(CW&sj_j^>*rm+jKCo+uJlmzhMI_7y2XX3d
zlSK4w&zywTYAs&E5-w;!QlE}8C~Ko@vaPYlD`I^YeI7qi>)zS9Zy6+I)!fmCN%lW8
zzP8o-Q%CzFlg2RT)$(BVF5X+AZ~gGAlbpD{u4Y1}+7F)el{P6T{Bp8RN2ntst<T?q
zF$0w?<U0t@O(1xc6YmN83Yd(2|9eR9BnSK`6e{O}M*8UCr#TlyQsb2PB^H-nFb?>x
ztCn5h<l`hUXL4**?o-%?ZJ6*JDA#4buvWhl^Rmu8FX79FY+h6KnG2e7cZcX+z8q(X
zTr>GtxzI5kMYWVz5?o|T{sI@7e$Bv;v1GQ)&2QWx0s=2+jrWGg@yRmK%+4x<)J(lr
zAa!pKeYX*j#Znii5r6CPDUYl-|A08`O(9nfO8%n$XFh8S-$uR7<^Xf*Ij?V+fJ)gz
z0ndEdov*(V$HdHAEFTXyQGOoUz-8FtY93C*Xjta~EbcOIlg(FaOe6+=5_>${A5adF
zJL;#hm*<xJnA%$>urxmt66A{<ZB)0v9aYhgeMzjgixxizuY1>KoFtY5Jfp_(Ozhh#
zX-79!c5c^U9Pqr2u+!kS!-r?P^-G8Nk^qcZxVw1r@q@TT`1Xm4X^6|<<c8L!YTU+9
zQ$9M?>+`?BTd;*0tNhQOTitVqO1sLRXc)Uno1bVnwwFBJwOSArQ%dt^KAiEZ1)a?J
zI%UH?NaRKkaJ%BHkzv9~{F4ko$&`cshH6sJtR_>>GtuApaQLx$x%q|&DB#w}tT1QL
zLDC?>V1)j>Kja4<ktuYi<1%jt{eB!5Y~@SZ{l4`tNE6vQpB5SIKI(;{e;5p)v2klp
zImS`z>&_(*p#g#iMQf-$)wd7FhwOl}J0S=qF|T?$evJ2PX)JFj10OyqD$95^;hu;c
zbJUDTnHa=yn|s;ewQotyHIjScP+E*$HzG(KixJ=cdL;RpBAd=qcIbjp6~MGS4L+`q
z0`TuPH`U;otj}QUlC$V8cZ}*o<nk{6KGLNlab<pCZ8(K@cQ-gGMv>#cH21*GE1_Z3
zLTnDQt_!H{^tjJn&b?NZJ9S&jVT6He_OZSjBJ^BTnTF?nXX~r9yGWa(jh3=IHVy<p
z%@&S`DyvxP)XB<+1Qp+=Q>1>Zq4r(EGE6cE$IyVm9$f4zWdneb+q@!6ezR}qdHH}n
zp?qerIQOU7|8%~8dRD^`)Vbf8mJ8+t6uYlf7-`QEZGVhh9^`zw;B?s8>3j$Q)*}>E
zau*1#&k@8gW-~7bbbo3BA9m;M)qMv32i_&F_Os2Sj<>5PWPW0wk*Eo5gPiZ%$<}FV
zO6lE-0&dTCGAL4;5iRx(ORCusBCpF&Fm{pk$g?G6x2el)7i#_U)GCCYdH1=_OZBCB
z$WNtR+7lIb-d-1Uy;zeiV74eQG+s3qG>k-}<9!mo$A2sV8%B@Sg?gr1AfZ&zNfEr;
zN2!`uFzqY;b2ZlforOxu@+OsGRMFkAGr8ZzLoqpk>(@MflN)6I4)Fcsf9nSzB_a5a
zNuL%!C+7=v?SaETXL9Ce(<fVY#}0tz_8>E7K3)-ibrh-9f`ua)=5&OcUCnQTlvAgx
zn^tG~@leq_1ZokPJbfh3DpGK1y{^BIz24<z2c~s<yW_>S(U^MK9UPd;o5?*-T*rfk
zrqT`P2A!xequv}%SFfJuiQsiMdvTXqnKm5qPAfGDFwM<6VJ;BenAW>?8D=L|JZ0c8
zbS44P3(!*zIdD&y_*a~sp3BaS(&Xn{OO`WEvb=$~5sSlF(6hqSy~vg+{bXWB7OR}!
z&uE_P)~{_I7P{qi-(QC_cc=R^2W53$Kr(0#t4HqwP}a+P#z$0X%N@UZ;X49Ad~uK1
z%G5X8$<L)_QbMfxt#@Ow4-P#W?vK%!CmG<`7DjSq;Bchfqc_YEgTz;nucuzPkZeVe
z6{Or2w+zd}VX-1EzP^S~phEt4e60cSIb<;|(u0^w0`Z1f{n|f;yjp7G*lZwuT~&C{
z0n{(<{ug#$pTh?OVvku7FfnU!#yU`SIFLqlQ<`%C{RlsW90zYIhlWp_^^^97@Ba&U
zqWG=<TYwxzUjXUzJ|zIziS%~7Jlzj1c?BG_k2~!L;Aqn>Xd)xs&VVLFkkbEYT!?7j
z9<`2~Wtf$p1rt!hnkh_SxRMCMp+E60aHz#f#+_szSolY{%nKmmczhcuXRyz%0UZ)L
zmkYqO{`La=r56Os9RoaHuRJg0DL*mJimmbc*8yN%KN(Pn4^W<Sxe>Ryv~bpYZm4Z8
zEf3(UuW$T~#QvV2tarzI_auoX)PDqES!J!sC*Tk%D|CH8_hCoexna^F4R9dE-v8++
z2(YQa3G#gXDfScrVVK=HMo5ZKNAhO+wp_ZJPV55@9}bwq-FLk^DbjJ&@>)v1umRgM
zzcT%QrL3!*T~t)rYwUlvGAMa@%3P&aa^7NKU_i8(25COSz;~Nr#wlr>c_CkxH@CUa
zx0{;+CJkF&3+&wgyDGV9$&WNMp#UdApf^_kQ}b7jEUH|U^q+rjPkh-c5gwi<iQSS{
zpExf~eRTlmr>H;0?6(cXYfes2^lh3Rmi37C)9U<LTHVU@o73%99}0tkMK(tcEAsVV
z5!JANEQ|CgHOcRNjPD=as^wgX&mCbZ8w%6ZZhh;*QqpC4U3$N-uucEsTUVhro7o9&
zE_@|_ammre&EooMA;FeZ;d42{tZ^{|SZ1c~jIB}RH;XT={WsIsZadJSO9uw~xZcC+
z{Y<N>-sJX;>$((me#}hG^s9!wf1Ya9c#4s?xzA*Hd3MTav$u6drlirFz+lWPahlgQ
zUQ(&ibKMZ3%)|d8zgevIjk+ps2DZ2uC?Lxa_MUr1smV!g%hBD>N--q19B%u-@kT9)
z&KjHNXOLz(w8ynxY*tS%)fn;oB6~YF{@ZGvvFE)Fk=7*)fUAYO9!;Wp>zb>DYY#?O
z+x+7`T+{noO8IVKS?7*sJUMW$01OV{G7;73|9&>k=mI&cg05M!ayR#|z-@W@`$*Ho
zbBo;S#CSuA?yG<;Q|Gqvg<HI)IBd6t)NTDR<cu6xxv2}Ayj-fs`1P2h)4l48ex0L|
zl7wA)%Y`q0Z>S)TyJ5}$+^``Acf*I)`Uh~?t_P$#YuqDdv;#GXVJ&|8@Pk)8ChQT=
zp2Kx?h~}WKuCOFl(P{kR%n1|OF)3LZ|2i?$@oWus$LjztrZd?ST}H`(MJ*1-@q0eN
zagIG2kSSx8*#dSp+^^NR<|-)%fVL3sig1%m3pupCosFB!;@-sKX3+NV4T?ONz+>BH
zW{rL||5yWI|Bz(i`nt;<gIk|c2(4RWmO0P7nuB6ByXI=Qq6HnRp#Fbg*R4P8WaL4c
zMdo0N6<mYc*G`rN5a{>TcIGFD{Sfm_f^64u+UdRkni7Pk4j+*`*wQu~TR8z{8oxMX
zQ#_pVJIJFu6EQQu9(AGoFIJ0ECnx6y)Y77_s;ke{jO%eTm%5LW2(Oy|Bw|w;ClNlS
zsgjcOduPoZ{gY*Cyu7?+=Y~c`8O`$sK8^hQ(!r8C&i;70Kc^>rQkVGw)rsa==<OXH
zolN`4i+mJH1We@S!=mJJ?5aDbId4x+9d5lS6K>xzU4e%X(K?`UHx6Jo`|#R#<_YfS
zv1nYHRxbmI@wIk3INDe_3a$l%IbB>%L;^F?()6c2&llT{RyG3e7Wu|fg;th~->)}U
z<t>x+@i<@qCnHT?7k7l$pCg)L!E|L8FheQIXR6>V<Jde0_D`*AuQ{1uO_D0pua(T`
zBKyC}Av{baIyVQn7Z8Xv^j56Qmwf7~H@M&OnQsAYxBGRNt>{x@hYUXtc%=aQ)brZ?
z<2IYInc|E#W3G{JCj7JT6E24gqujcHHSKx7I$FmY`qqSJ@A5z>ipkxcmY7GIe(#*H
z`s9Angq)JiQK_<aOI*0i52*n%ss;x90w}->*UI%MxX3s<`}j8rW6+ONOxGCak@h^R
zyw25WFEw|VD}X{fd!Ll{kCH+NSflF&fobTGQL(2&Z}VAMUqx)YS4M1ZvF-b<r<E-u
zfqD?%$&QyMJ_)%sj<NupwoCjqgPTgYh5cZu@BX0j`F_bnW|DLgK|`!S7XT_68#nhi
z47k1~qpHMv#JqFgq;J44bgsc3nEy1(63(Q|+8%(ZUO9AFEc{8hmdiPuT#l;Z?(((f
zS|}N9f9a~a82iTknU%b3=sBs;(Xs9?DfteJl=DPj)YPQ*R~Du3Ui^`=2LAG!89`q>
z)waUBX81`)Q+}}n-XLZtQ5*}$qwq{GHmh-OsXoORgFO>dSlMEp*g~bo2s~VzC{7k5
zkC;Nae||*ewK#sW%fIMVb&nsGsU8Jfgq3~Kv~5P@)=ao9yu%o44v&?-*^tdws>eO1
z!r;|XRlsQ{CiZ1)bKHX<@!Q$K+kJ=uOqjv1pVGGi^^gY<+M`!-dgSv<KE<S*O4QFm
zRbJ%UXcAd|Gy6aqf*|#)p2E5Y&9i=kshVj;`{Pe0w_5WD^9pl;HHmO9PW(DwdIR)(
zTi^+GgJ<eaKbeTgZ;Vw^QeK*3%nZQ;5kQeD=@!vad1aNf><5dng&uoup9S)G*O!)7
zZJK+2Lw^A3)`Vj5!KA1PMlQb>sxY#}oA-2baydA%Sz3eb&%M#!<jX`5^2ScxIlJsi
z%SCk3H%w7;6mM)q?HA~QBcYz24fV-21G`6_((z9VZ&$cCJd4;-QuUP+>FS=}jjYxt
zmsjehH(cA>9X}s!ho)*y4ZA)<J;Tx(zHXw1C^-qI=R(8JQyCial!RmJ+#{w2ImX>6
z+&1}nl^Z_4@hw|%_*W5G@X0?gkN4yu39y44cnq51xY-A@6YS%=a7uWKdNtI`_$WJF
znxY|q$^b>0i2$GGw^c}Y4UykTy7cLq!Mz6&)Ojt8wD>Z#!O{P$a%g5O*aC%a2dfAV
z;(Lz7Hf-eg?`6To`_+f$_~*6{C`!w^QUjQs9DaTw03(RPQ}Xa~L{EB^CtWk*9v;aG
zc#x^(-dP)t5j29f6+c!z!#Ff;MmZN4fPGnJg(r<CQ<O5zi2Z8@C5(|0&48h$Me3TD
zjF+=6JoM3$8$6*lB+%&im-fiEk*fUx+c?*+UoLKy=?>@&j;_0|zo~*^@+jU|R;&EH
z=j~HTd@b<D9F*ku;Vs;OQ`jcK`*1aOhl29+?ht1Qk2U_m1rA1^3Gk_Cs=*EZ!}iu3
zSPI(}Zi6jflLXDu-fiIJzWu&ODkEipKT4`ewbns#s1hy~)(f>Xcu(~l%uFnzFZwhD
z4V}scYOGe>UFh<>>&DNj<Y`(}A}x5|Z8^I6*>G+JyN?^B8ux2p%?m$C==ft}{>(y}
zQ)yQFb>@xyPya0L%)$~9>a*ta2gj!;MtpMZ8miOERlB@n+7lnOJ*xp6d+h-Fa@U@7
z%QC6@^Zqd1iq0O?YrM~^|HuOKjv#5--tNh)wSUWA4)}Ln>GJ%qp6#ek%_3&UaRX`(
zEeF5a{dLRH$v~ETT~z<_<|Jx;Nyb2U8YzylOhkR_ne)8wx@ZKLjpT6SLDKV@H&!d~
zb=}kwlrdZO7EWeBayd)<zq;Dd=22WwBzWlX(|qBKDqBV2tK?r=FHg5ur^X3;bHlPS
zSy@sPY5K#$sOI$yu4?jm8G&;HG;k=^7x_8o!cu@+n-_h39{Y|KM!gRi4PCG7?fi}$
z>+leC?{i*3%TlGs6*{ME=d-J!Ye{~pVN;Y(puRSQS;?}w9T+!G9R;zr1mXGcoS$U0
zr98rAFGXilyLF@`BcM#5(C;25IOL_|jmy5<85q3JceTT1eC0Ulp+RqMpyQnX(*1tt
zo1$#T@e5N&#B&Gn(W$`pLVz3WHAwD9DKgEyT)jcFrk}q`N0~CMrl-oBBopB}<v>#L
zN?rdlhakGh;JfrTn3oWW_jz$|$KwP)+D+l>ru?v<dFVB=+~qo4Txo2^E1x9|oMW<t
zL%hHA{v@ImaX4`71L8p8G;5vg5LjNfUgp5h^~-5(MyENN;G*o?t0YwLrUbDWpBmxz
zVZD-<5)Jpsp3>znM6Hw{5OYa0i-S+hmdu+KzR_NNf&W{AU`R}K7U@)1MGeyt7c}ft
z)kQy@o4w5u+>&}$b~!YIwN5pU-cu@-aTm(aqYIVvQ{{JC-}|?2V_WLmc=4XqgG_((
zDJoO{OP6HGrl!{Oxx20>!Hg9@WhZ98=r^h4ot05CUm4<);7HG}-t7E08?d9IrBm4W
z&+{D9s%$?B*=+GA9TzKmf-D8whl2Zm-b<*w0^o4q_@B$Tw&q8f@l>MdDiU{hoi&Jy
zlRZUJxOx86B6FzI!Arb)pkRF3%WhU*uz#8t$2w!+Q4*o>a!62lHmv(sxNDbBj0`o;
zNza!u`r&VCh~Xa`v#R0PcX0g|E8aV3M_I)^bsTYW?m~%rbTOK+#kLEYH#31UxP|2L
z+pRE#&rK{kl5gXStY`5qn|6|^yQmh`J~pT51625s*J1%jRseO8y|UxK?P!8#z3s~w
z+o1WvscXXC<Lg|m8hrm@pZE~tNF};tUV?*KL7-7wO*|l7dp^TT>DTbMn0RA<M|qWy
z1xHk1c^lvY@{*D!4FR`NCXy{;*34~SFn3XS=F*x)Prru2?61C7?4p+Ca)15j_{9re
zUN!PwQTBi|1I~N$w5K=Xzd5I^&sEmdoh=B`y#3UzGm*?2t4W_}k=wg`*-Oyk#Zu|v
zD$lHK1f~;zwc?abHKufDwr@k768vL>2$SFC@fg^f?BzPeWLvFCTJg_V0bT*X3TEhT
zMJs%68}(}$|G5WYs`*9y%rtK_snO1`h39HZRKXDtPN2AN;fUj-D#TW4ZA)zRcA)c~
z98SDZ0>ge5md+8C#(Wy&pi7J#i_9WT03Ne6!SUBD(&2n2xApBPdHVKKjRqgs!-!e&
zNbj!YyH`tFA(M~kI4`9>-lp=*|5X<5ue(G*c3#nrQD8^?1aiTUZU)2xvvp#bl^mz+
zfo=r64u8x9tyc0UoTWpa&ONrEH1_T!hyY6Jsg!;tvo%Gx^11^mmY>!6bYyF*_XV$t
zShkRHnNu+R;mj$}wWYH^)l#WcG~TY0S6?v-4YpBwdQ4(bl_%N$$a01n!n5a=cKlj8
za-BzwpiyisM8xJI-Zu8n*DZm~;y9zYxot91;ZtH~<0uqHFOpkY7kO>#q&5m40^UMR
zLb3e3jyVhw-T$xLvm)V-c0r4k&aD-nI))BJ3QtGwe8jnyH;a`QgTUUrx1L~4siLC)
zO$9T)s~P(r4+g>hd8)>XEN!=p%hdNgXOtkCn8B`PKgTQedmaUzShcwbd`aj6mZXUP
zZt)`rK?Fp0ep8`Otj&C7w7N?2cUs=iJFQ9c>!|WKpB1LkklJVXg~01pp=3OZIVv4p
zzA;4i{bLJe239xR4tv+2mkF++*gc(cu;qG`Tv<9)%!CFunN%Jr6jYy38~R4&vyJGj
z$;;Qq{iw#+qDi{QLhY`VMW~sGH@@2mx?Vk#B*!6AB<Zv_>!)uoncl&i(*8SEKb?YQ
z#&;v9Ps)pwo>a^iiY0R)FDCgZj}(rDuJ`b;8fh`qh9Sa%e<$%ns%&M2|1oRxv!%2G
zg>>dQ^m)xkB_eA>?2-JO`eaMVstyjYl48ZxwTZqDDc>$>LyVt&FR?0I_tMyQl=2V`
zx~uIM!F#;OON!@#%MO5y2GAT}7(*N!9C(278Hj?f>d1~BVC6A95>gD{ai-`XkAtNx
z|Bem^NJ(R4q>V&r6iw*xY@j?XU)ML2J=4L>8FAk<kN1xa-tM_*b^^?coeLmSb6V()
z`>=eo_BleGpJPgu+}fJhLGWYZD8pqs`W#(;11r5Wp2-*dzF5b3ydgVgeAB}|!wASS
z1SU&r>gy9=wYcdqxZW9^4HSP}+nF}FlIfJ7IaSpi=ms+w=69JP3;z1CQ>|_7S!iMH
zKW`G1)9TBbL(f!>#IB932^Ysha$1FLqiu<Q%g;4b4779r1IylKoj~0Cd5r%sol2BU
zVyQ4U78lz7UOm9=FyY%<g-RIsRUfr&n))bkhvQ-p2LFspJE=J^|GP`X*FUy(x4!Pw
zv@V$uB7F{2PBB}%%@_{gvD3kPB9u*eU2}8&{@U6ajlP+Hy_aUsb&h4(Ir%gCGh|ie
z_ZeS>2C4ru@AK32IEg?Z{xLorNPxB-c|cVC^ps=3$u#nuiOOQ=1wr~!K+>S^QNO63
z(GZ<VSr)wBG0uM7L&9o6aqD!l2&M^fLH_LBT#fC)C`KPCmFl?*Rp`-$Pj!EYF;t)v
z`_dph|BWJ1g-OLum6lXa+jYMS>=?W9LwYEmDzBzo=`PacVsBFGq0sO(zWMc^cJ#-o
zrYYZ>${JcacKDNo*F3Tc&0>>%r|)m{`R>`%J#V_4O<RCbpu*j2q>_(#6_@Qbiz8@l
z!5b1KZR2Mp?S{`{q529`-Kn!@j=tE%HsG}w^Uc2JP-|;TU^wuwS{V5%!r5~4<)~UV
zq{kAu6f>n}v_kMi6bs7l>o{FREji(&N8r!b?mW>yD+k7u<BuPY(y0`+PwU!Pb<F*)
zH&*(<xMtV>9wj)zS#OIN{%z|<tD%xom;{iVzBKFeB2D0XlTcje$9roA`*(#ZPAVd`
zK5XM@<7OUUdk;<uK-$|GT(*KAiz-Z~?fDPuVVP!Xmw%GhH^TSwc>Wq<Zd@A2-ZR%T
zOFB6Tc%Sg@HmACVv53g5$L$<N_q!W_{LF#7<NF4$wvR1#Ec_qyJaHNBQ)N*jj8=$%
zV>=ae_iSCb0oAO4PvuZ=YY9!81Zeoj${NA$Rilu4$=Kz!wrA9>wG}0&RfOV;8gOjA
zUbDv7&(5FNEyPbfNK9YWkJf&^09LF`XMNW1TYO6W^|O{A374P$YDZV5{1}>L;VdfT
z6#8hF1H=rUEd||`Zm}`I0+}zzILF3u^b*PAysKG%$#x^w(Kju`yiS=G2y>P+99cR@
zg43~l#XzQ5xjHd1GSb)AcL8)9E_y80-BRx#6DDg}J=k07`tU6uHM7d%`OGeZ*hJ2r
z1zZpfNNCP>joy8Px;n>%YMhm|<%I>gb+|?5)zjEZ)4u;vWzmSjb@GY8!vaH8Ng+`J
zjE~+&b5K!wh)6iyoshaZMQa*)tVL-&NkRJ;u*%jcPhK!FJ0X|ZZe}N+R6W=HT37_o
z7Qa{tSi4I1IVwTN_RNOA(a_=P^&S`H0-TZH1Vg0j8pO_zCXJ(_N%g1AQ%ey389R+7
zUhg(niLS&inFpUdcx+1l!VW<6-0R#SSWl|frUi=^@LWK}R0zj2bY6$K#;gPSDmXr0
z#~b_Nb9#YhrCZ+<t#9B|GQ}o?i6Y*DqH|PP(GLxHI8NwTI2|h%s$3sYL&Ww3_ii`j
z3ur<Uy?6k^wCob5CrEM354d+1&VcRB8><4+WT}<H1|Gq>LYl<>duW3*2SHj1)pmxL
zguyN8O@{T7?)=AND?p(Y0NsCIZ&);$=?7K~Shhcl7+~xl9&Pn^Z~^^e{tj~X>91e!
zUuFUMI^?Nxa!RH#0`Z(nBq~jlS@w5@zw9(ku#D1NKI5RnE%JON%5I@yItpvE8Jnyo
zgH`fgzHZ*|a-;5XK=L#J$veyw<fc>Lcr)!$R2a0kuue)<pOT>4ftv{^MOj(dF{qG|
zD~-hLq$4E&s)r~n)xhJm!`G=doi^C<UFn*1qQoxpTAy4kZYQm4(*2F;-pJylj(>Ay
zsZLxaB{bXZ!mh&#-v9tsx%>P2%Fx<uUf}x9>ONO=5roW7e{zG3yd*Z%sMDdRAzP_>
z)Qqh|xj)Q{!#bqdu}AC+Po#T~rupr)lKlSSzHB_)#D3>$@V^Lq1Z8L)1CpgV)x6?D
zHXXONiZe`3J!JBY)k4T~7pLoFp%9lIy$-u}B@%PXiwc*$O5E^kkhN@rFG#;YjWOvS
zWYDQbxUy>8K4aWdU})Kxh1`l7&b$+EwjPC_+DY=S9FMI4Oz|&IbVF}NrJGDAtC+8s
zeA{uQE)o+Y_K_K}S^#i8QOrN_{<sMy^WfZXyp4n9+D`Cje|t+o9|@baK7gfHlPsxL
zEbly{x7YqMJ<S`O<yHrWi%AL!R#sJM?M{nF{ECOlYuV;?TLvmV{Z99rE;30yG+p_m
zU&rO~eH=~(?2^)Y!b&KeHZr77F}o43ebk=X{KLM7h&IW;rv(=_jeWr3htr0*0mBX9
z+8SWsP^UyIE32CQn{RdIS#uYDCpY-ZoQVMte^GGCQXPO?rj@Ayuney;+xDGbb;UkO
zt_j2<;Dzm^w&p9lYLfm{#-Oj@9tYc5n>nJ0Tk0t?Z^aWm_pa`{KVs%zjI33b-D#>m
zJ-VnPrDEH(<%_xYz>nYsk+(EV%}#W+N?@u^TG1o<ZPp!61h~?`fqsQVv_QIk?~Jnb
zGcfO;RcC*7I=QkWWxOpJi3J?g+ivo9J)E|5HrWLMzWtVI@g1hY%M>fhPaXaJM`bPN
z$Fpbk_06^qySSoqsHer7{PaD(KrXjB>hNkV8;KZqEcrS=mE4jgVWhouK88NF7Di;!
z42{~-)o7DO<08U>*41_M76sdv?wa|xt4F_i)xQuZ=4L1Wbfh7BU)hboxyT{1)*-fq
zO(Q&wD(TUtce@sD9e||!li7!ryIeHH<>xQ8ieRdM=d{7sd63d({pMjJZ1&8?MZ`;Q
z8S&WvXv=B&S?~4YD8sI+**}9)uEhMz119#$p0y<^?xqg2SDY9N`>QIER@9wg4)<|7
z-!_ZGeNUF^DEPibir;=_5;Y!O-@kJaUnki$*za>Bw7&WFszz^8I;Y&2MK}xGWB+o-
zX5vBI*!ar4k7Yz>19ueqU3_jMo?;dddUnt0Alol;zp@$+dHgYb-jxI%Wvt8#2<KS-
zlijE(OK;xXu$Ou$cO7;vKV{_q<xH3u`FYCM3XsOpH{;AU3*SEtk}2E2g|5KsdI5J?
zdDpB+aRz7%0-0d0K?oI%DIS0PsABcToFKZ5j~>?gf`WptfT52Y{gt>;R8%C-8{KOJ
zSU%IeqZI(z=DhW*xtuHYRenrNak^>)i2$u*<DJ{6wj9PFAme5nJPf!oYp-cTG6>k^
zh9NqNB7q%O^z6|$J>`)+ZJ5nN^QKrg`IY2L5|*IXZJqjBR1SB@N=v$q(#QOBBrN6z
zY;S$L{UlI^x}2fxFL7?#Daj^^|MtpBZ#Tb9u*r+l9yf<9JpMG8E64?YgeR{othI3m
z2U2}~gwtY+_=hI%cP<aOy1ZHC_P#OMguZC3`qh?}U87erv4|V!P29KuMql4Z6T8z$
z;m^sdnF1L??gBZviolnQ<dn|EezZM6-30ZD?xo#HaEblou9l09@ZNn63NyNMSq<`r
z><?PGFxC~S2E;6j<&D_YFwMN^hg%87U=Dd=u`hRDVyPiUd(!nE^h`T5m}gM;DMJ+(
zH0TAo2!%Eury)x$eAC2Zp%>Q>ZL(%U=&eNQANE4+=sVg8P%#^==g0U)8mIP3Z55$Z
zq66Ph&;-l}oULHTZFvOgK883s)}`_{QnzXytmUv9EzghK{Fjv6EcgV^T(M|g!`|n4
zuB^)~;n>ob4QU=+?|~tl=APyuKHrI2^-v_I)L1p8a=(ncy9)JFYVeW<fH|8jF3YNO
z0Lsg(6O<=RlhooEkwk~*@VLs~EK&xmU~0e@&yE}<6!lLvZK*p&%PbJ;Pi%(+gYIrl
zdblXr9&|ixUJL1KOFBHurT&=l{;}Fl1eqI_vK*luhG-!ZaZ3a38On$GUGL&gHv7K>
zM?%FsuL)K<wg2hFCc;IN>9P7}Et44q9lZ<2Ev$zO&Oi0V3OWY0jej5Gue#eht?e8j
zYz5h9UocH|w8U7pwo%P~s}G`9Bv;vazcef+E-pTYXqs&|_w!@j$5ZL~8N(4J_Gqss
zglOt3a}ZG1y7+nP9~{OXD2Ug(w8hX>!SCe-M1PxrS*a?l8eP`qkj!aVC--g$Q)#<i
z*HlsR(ZHbBEuCqjK+0$MJ}l?%MvhJYGY+vmZHw1DKZ^_w2(nAB#&lko6~+b5rmJX<
zhLgdc7~Z=D>~;i#i%{)^WRquq#rRzuQuS5AX1T~8vQ}%nXfWWNPj5Mu{6&tWe6O_=
z5kN$nRSlM`9HVXQ`UO->ZLFm3^0Hoaetk6W<NSw@-zVuYAM~{ME9%05kJuNe1c4M~
zUTE^Me=taHkGzR*A=^zppxpP9pD#*`eXr+t-;R%{O0yq?mGD0Z?&KsI2*SWXSlkcz
zczO~JERZm2+1`QOZ9V;~x+NStLVL43{jM>~8Kq|D_K{~w6Lbi6uF+p#ls2X0pcps1
z#<TXdPY;vtGgbd8dA89LW;f7X%XCY>1gOgPucb;pOpsyfh_bCUdjks!jN@Y4fP?ie
zN}V@2@8WoptgZFGf*-EE2h|&AUz?&iJ#Djo5(O?doo>J76g}2BigTm6NjSlOMl`qh
zC};u)n~sc3w5{6`>eSwt_?~h$ZdxOvz$5eW51+x~&%5+DJIQKGR71jzVBe>HW+{XU
zkIcZLIc=PM+k*w;@S?g245*oWz*PygczFcChE1Sy?C$30=a*~k=O^4z6>yhB>=RqW
z6;r_-g!*3Luxgy-^n}BRk%W%5e(00j!rH-!hh@uw1Zfp$cAF#Y!-Vm`c;~;o@Dq}?
zl2f$p?_cKRY<@;sM^Qg*TjDtZC#Y{|fFV*K7gW`=h^5H`8{>?r<JLvnz7EluUG#**
zm5#Lo9xI3Z8qY0877N{Lqi*}18Zhc(Pu@ON1)@1lI84<d{@@C-i!LX#nU?@14SFX1
zgHNwLR0>jwP`=WoF%y^$vWhbyoQ|W8A><68@F2?d=54%z66at*5Myr~y4@sSf2Sp4
z_$dRqV_41mrvX*zgUYbw($B#QzPyGk>qh{{(u4u_t0je7^d(_<$EkONm(wto{Kx8V
zQFk4DSA0o#*~aial@koc^tk>?K6QY;w9Gv>n5J_MR=`6)DbV9uv&kEq!j$Cx(}R%l
zJW3qO`?R9~yxUbODrnwli1*Y$d5Gb$C;dVb@FZQe{3}MAJG1IHqs(AdO|#scBFIVQ
z=9dY~h{v|xu?IvGu{-a4T(B4F&Q%SqYsD?r$k}@ikH!uCW*_Vt)#yYeoG1Qq7rw&&
z&Wtp-?m!K!q+^SIZzg?FTY;Swf<=`4TT0Z*6rLE$`L3Skp1WCN;pG{cM>}$9XhXuR
z*Lb+PK4_tlPL_3w$W&L@;)LRz2aJBPWI*+qo$<m^r}bIa^ZvI_{Rv30xUQXm%I`p#
z{HLFD&4Q?mIGv>?zC}jGkGByJiIb~!C<N+$IVEakRebzByN~+27!@1!{pdG>s`A|u
zYu{2YjvZmY8VfN2?~?nXtAp5rFbuIZKj^d;*e7hF+~H*+&j;lw)uN$E`5TItf^YR}
zvSGmjpE^(~CkGVGn%e4<L--?OE49EoAk4SyDm~Pc=xzBg&z07uMJtkAh+D!g5*o_}
ztKD@MUi{`6K5wYZI8HmE#F1%x59c)k2BVnB?d9xRD^k1BoI~qk&xhkHk8c`T`)E7P
zjfPuRM&)209pHS?KoUNnX6KtktGfM^BzKA9ijQd+<9&NtYg)(#Xo@JbFEd3FF@?*X
z_bT)qw;zxK%|g;|#|e~CvGS6~w=?cidnnAbLNz~`zs<93nWZ}~ELc-CvgwCboxje-
z2vr3NLC@anH+rul-`BrT$t+GpeBS**-TF?n@9JG2>sBzN!~GNer;A{V_nPk-<R4?P
zHD~0XsFQ(-fk11!4+Vb7Rk3mvrENFgCN2fuoAqqVc%iz(+76Xr$^{ozz$#ziih2tX
zHb4&H7~m|pbtQ9?dO9g_X(wV=VDRzO;|vGlCtY#V^GIk~g;7EyZ|e<|Y@3BE9GR4u
zJ<5>q_R-yrQB>K2Zo&DrU_OpAg&oT)DjqTAy<XQDjk}Ns*JQu+_c6ASHDsnyh)Y6N
zbm-63w3c`DZe`JlE_om$!S0HP4C@l$suFd5SYT`Nmyptij!yBMIsS|XDQG9fc@YL9
z?H_^HaPts;*OyyQztNfMVl7S*-Z{WBY<qOi|0!dWm|Y+;pQfb~J-kjafv1uKt3t3r
zwA$;O<|wnrmqKZ0fn=Z)5AHj4Ksri<<#p#vPEq%0k)`dX<oQ*kLaR}9uPkOZqE~w@
zsO1S@%Jgb+D?VeAAGW#+(7`#v_VVNJ{9w=?4;!_-p|V7_GigqMuwa2EtVGqpuYnYy
zAAv$%xRsK9zw^_{gDoDePAtjr;JIDNsa;BQbMeuask}somIt(KKbm#~(jr9s46a!Z
zGO1h@1Qv;j%kIN=Gc|F^Xx#N8jdxk=cRq)|eG12u-`#(^itA>2EkO*lDv(T26<3UZ
zY*D{Yn%C`9r4~Cu)fcznSoW+~uAl#%e-#%rL0w$sa=Bwxx<1V8K+8J*IpjCehLXDM
zx1ggzjLlmu!oDPLacOn&bk^R{Lgd-&6X_A^n(RWMW>r$pTJSH~H(hZTHpP|5C`@*&
z>cukyavFuU>xJv9nXHOeWv{;QsS~r~_9i$x2$vh46!0|OYALtoMi`Apv0=Brw;-SX
zFL6nog1=a+7`g0%wwpO3i~l2zqRSeLFH!ck=E~e%`d$jql?QU15Ly6&HQ7ALRNj+r
zZD~SA8Gwq|6(1qdM+#eq1ZVqgYN*apmL~$X13~+{?w7kX?qxN;!>1ed-{R2VkKejf
zi&fx5j%R4qjlK7wV%T5rnhkj3EX3$KooHpZPI$t;u}7g6T~k?lR{}}jAK3aS=(RyL
zvp!F5bioZTWh*%t#rm%)N19l-&>hJ?+bVppt9ay6;EQK-u|Hew5k92>sq^sUUZFER
zIxrny@hz(T%#uQaI^BZ5yf5n{&UF~xnRHBD+3m3-xo!OF#7V@q3~Dcf!BosQvgFHN
zJ1FNE#G0e6sjeVkIt%?_84f+VSO1jpNc7U4`j|}*>|mq9zj)&8&)SYzy}{Gx1fYO}
zxZKLg$?o&sS$#s&^3%&&9rgR4hPIAM(P#%aR`uo$J+ZVVRaRaI1SxkVxQ|6d&$e}{
zSP!<~KFVzIy)T>PeswSeF$`5J`;DR7I708D9nM!Cr)+GWx$W=opP^;cef<61y=N$R
ztxM;Fb7bx1t?hfmy)*1`_t}813F{`h?;G1klcPx|zwyh8kBOHRcv`f=MMk$5m#mt4
zWkNoutTlWeoJ4d*K7a5sB(HZkDYY_El3uRqv!BGA^Dv`*vM0sUk|2DD{NtKAlpdNF
zpbzxZ2E6;D_(jF}fm6?uwQ`<6N=udG(Uy0<XW{<3CX08STYvq2>=5tR@Yot`m65$i
z=(&EKVFJ=3&~Ck@Dgtc&0f9v4u7scCcm@ivI`k*X-t*d-@(`doivD5G<=c#w_{@Hr
z@zAU3ZfPl+hgW(X3TF(8YC}b}+gEIaRJ}Vz5i(r&B!L3nuu2r+mg_svmA41$!}Yz0
zx=;i85is75*EdDeQ?iW)G6M}X;T=f-wn;!+qSBP+Yqfiic*Y+kn!LJ-JlA)}2{Vvb
z{j*dV+wWJBY4YdAUlV(Q9X-630Z{4))kTDekH^Jf`h-N%$0qDRsw?KAssJnlH1w<&
z>hJsL*|=Oqp00gOSV_6Q8c2mKlS0J1ZrV%!uX&r1`j0;y7|GAW7<o$GvaZR%Y?tZT
zv$GCV*>sdzUS;srd&v@xhL&-rC6lyQLfMO?-$VXh?K^p90vBT1Px^XhUny{!2aK6w
z!BeU-|3h}L<KgbpzQfAB{*I$IE|nJL=PIMq2W>}gCsvKsem@d!2Ccj@Hhwy{bh`Gl
z=^C*C<p-SFt8Hx0s^}tO(WS>UsRKBWXveYKfF~{E8qkvZII^~d#w=>@xjviGkghqq
zBha$#SQ#8{s0HyZavdot-y0A8+->leFaX9589c7&dJjGT<Gk#j#mwC6&O6Xh@xa2`
z=gRWVI9==;U^~Bxpz7XC@MxhDf0}*1Muh(+?j2{qculo8ZPcKnUd}6e{`V3u{^Qxv
zQ|;Br%#2a|)e8MH;h^C^#?P5Zk|ZQcPLLsa12?W`3(Ai_Eyn+kimbm+4wG8HdRR(#
zP~~wUY-h#;KayAUlnol@ODRg^{IxxaS~~Bd{TG*e0r$q+(=~<l8eLw|5|gW}0TT*1
z9s&YIU_1&MSU?^<phb1xgDw!lF5jgPp{l1Jkax6MOn=z_rPs%b+GTJSqQ)Hyl2U26
zmUSz0o-c+9vEax$jmEh(U@+P{{NnfXQl&{zP8={1GG-kWoym^k3tO_ONRN*1Nnr8y
z!_50jpI^7^051s&9qg%Tc!I|H&>;cJ-j#}tXV=&H(INzUTo%`TJ`0$a@Q$VPwQTm=
zS(EUV26*Gn$@uTTymXvsx7Hn*ZY#?>f|JXtuOpC1#QZV%eIxLI9T4b06}i6N-)}<>
zS3-YP^8b7YdKdfO<p0lyRjL2~XaBG9VCD)hJ+-CzyWh<hxFu1PRh21yY4YL!0Y%fG
A3;+NC

diff --git a/_output.yml b/_output.yml
index 6141b2b..ec512fa 100644
--- a/_output.yml
+++ b/_output.yml
@@ -1,4 +1,6 @@
-bookdown::gitbook: 
+bookdown::gitbook:
+  includes:
+    in_header: header.html
   toc_depth: 2
   number_sections: false
   css: style.css
diff --git a/coding-best-practices.html b/coding-best-practices.html
index 24c357b..5b41a6c 100644
--- a/coding-best-practices.html
+++ b/coding-best-practices.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
diff --git a/data-import-from-pdf-sources.html b/data-import-from-pdf-sources.html
index 4023a81..b40c2c5 100644
--- a/data-import-from-pdf-sources.html
+++ b/data-import-from-pdf-sources.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
@@ -439,8 +448,8 @@ <h3>Getting Familiar with the Example Dataset<a href="data-import-from-pdf-sourc
 <p>Here is what a full view of one of the PDFs looks like, with values we want to extract highlighted in yellow:
 <img src="Module4_3_Input/Module4_3_Image1.png" width="850px" style="display: block; margin: auto;" /></p>
 <p>Our goal is to extract these values and end up with a dataframe that looks like this, with each sample in a row and each variable in a column:</p>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-5eb69ec5ed795734863a" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-5eb69ec5ed795734863a">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["20230214_0002_Expt1_A_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0014_Expt1_C_size_488","20230214_0023_Expt1_D_size_488","20230214_0024_Expt1_B_size_488"],["20230214","20230214","20230214","20230214","20230214"],["Expt1","Expt1","Expt1","Expt1","Expt1"],["A","Ctrl","C","D","B"],["488","488","488","488","488"],[200,50,1000,1000,2000],[43000000,32000000,56000000,13000000,17000000],[102.7,109.3,98.40000000000001,75.3,73.7],[174.9,191.4,169.6,145.2,137.6],[292.9,312.2,292.8,260.2,233.5],[191.8,206.8,187.8,169.5,161.7],[80,83.09999999999999,77.8,71.40000000000001,62.3]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Sample Identifier<\/th>\n      <th>Date<\/th>\n      <th>Experiment Number<\/th>\n      <th>Sample_ID<\/th>\n      <th>Wavelength<\/th>\n      <th>Dilution Factor<\/th>\n      <th>Concentration (Particles/mL)<\/th>\n      <th>X10<\/th>\n      <th>X50<\/th>\n      <th>X90<\/th>\n      <th>Mean<\/th>\n      <th>StdDev<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[6,7,8,9,10,11,12]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Sample Identifier","targets":1},{"name":"Date","targets":2},{"name":"Experiment Number","targets":3},{"name":"Sample_ID","targets":4},{"name":"Wavelength","targets":5},{"name":"Dilution Factor","targets":6},{"name":"Concentration (Particles/mL)","targets":7},{"name":"X10","targets":8},{"name":"X50","targets":9},{"name":"X90","targets":10},{"name":"Mean","targets":11},{"name":"StdDev","targets":12}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-cea2f55de9ee4f6566ed" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-cea2f55de9ee4f6566ed">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["20230214_0002_Expt1_A_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0014_Expt1_C_size_488","20230214_0023_Expt1_D_size_488","20230214_0024_Expt1_B_size_488"],["20230214","20230214","20230214","20230214","20230214"],["Expt1","Expt1","Expt1","Expt1","Expt1"],["A","Ctrl","C","D","B"],["488","488","488","488","488"],[200,50,1000,1000,2000],[43000000,32000000,56000000,13000000,17000000],[102.7,109.3,98.40000000000001,75.3,73.7],[174.9,191.4,169.6,145.2,137.6],[292.9,312.2,292.8,260.2,233.5],[191.8,206.8,187.8,169.5,161.7],[80,83.09999999999999,77.8,71.40000000000001,62.3]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Sample Identifier<\/th>\n      <th>Date<\/th>\n      <th>Experiment Number<\/th>\n      <th>Sample_ID<\/th>\n      <th>Wavelength<\/th>\n      <th>Dilution Factor<\/th>\n      <th>Concentration (Particles/mL)<\/th>\n      <th>X10<\/th>\n      <th>X50<\/th>\n      <th>X90<\/th>\n      <th>Mean<\/th>\n      <th>StdDev<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[6,7,8,9,10,11,12]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Sample Identifier","targets":1},{"name":"Date","targets":2},{"name":"Experiment Number","targets":3},{"name":"Sample_ID","targets":4},{"name":"Wavelength","targets":5},{"name":"Dilution Factor","targets":6},{"name":"Concentration (Particles/mL)","targets":7},{"name":"X10","targets":8},{"name":"X50","targets":9},{"name":"X90","targets":10},{"name":"Mean","targets":11},{"name":"StdDev","targets":12}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>If your files are not already named in a way that reflects unique sample information, such as the date of the experiment or sample ID, update your file names to contain this information before proceeding with the script. Here are the names for the example PDF files:
 <img src="Module4_3_Input/Module4_3_Image2.png" width="400px" style="display: block; margin: auto;" /></p>
 <p><br></p>
@@ -499,8 +508,8 @@ <h4>Initial data import from PDF files<a href="data-import-from-pdf-sources.html
 <span id="cb364-18"><a href="data-import-from-pdf-sources.html#cb364-18" tabindex="-1"></a>  <span class="fu">distinct</span>()</span></code></pre></div>
 <p>The new dataframe contains the data from all of the PDFs, with the <code>PDF Identifier</code> column containing the name of the input PDF file that corresponds to the text in the column next to it.</p>
 <div class="sourceCode" id="cb365"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb365-1"><a href="data-import-from-pdf-sources.html#cb365-1" tabindex="-1"></a><span class="fu">datatable</span>(pdf_raw)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-14adf5a0dda1df0719f8" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-14adf5a0dda1df0719f8">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80","81","82","83","84","85","86","87","88","89","90","91","92","93","94","95","96","97","98","99","100","101","102","103","104","105","106","107","108","109","110","111","112","113","114","115","116","117","118","119","120","121","122","123","124","125","126","127","128","129","130","131","132","133","134","135","136","137","138","139","140","141","142","143","144","145","146","147","148","149","150","151","152","153","154","155","156","157","158","159","160","161","162","163","164","165","166","167","168","169","170","171","172","173","174","175","176","177","178","179","180","181","182","183","184","185","186","187","188","189","190","191","192","193","194","195","196","197","198","199","200","201","202","203","204","205","206","207","208","209","210","211","212","213","214","215","216","217","218","219","220","221","222","223","224","225","226","227","228","229","230","231","232","233","234","235","236","237","238","239","240","241","242","243","244","245","246","247","248","249","250","251","252","253","254","255","256","257","258","259","260","261","262","263","264","265","266","267","268","269","270","271","272","273","274","275","276"],["20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488"],["                                                                                                    Electrophoresis &amp; Brownian Motion","                                                                                                              Video Analysis","                                                                                                       Laser Scattering Microscopy","                                           www.particle-metrix.de","","                                                                                      Operator (Report): sop_kabanovlab.inst","                                                                                      Video Operator: sop_kabanovlab.inst"," Sample Parameters                                                                   Result (sizes in nm)"," Sample Name: G104_F1                                                                                   Number      Concentration    Volume"," Comment: Sens=80, Sample Remarks0:                                                  Median (X50)        174.9       174.9       283.6"," Sample Remarks1:                                                                    Span                 80.0       79.9      116.6"," Sample Remarks2:"," Electrolyte:                                                                        Concentration:                 4.3E+7 Particles / mL"," Temperature: 24.64 °C sensed                                                        Dilution Factor:                     200"," pH 7.0 entered                                                                      Original Concentration:        8.6E+9 Particles / mL"," Conductivity: 9241.01 µS/cm sensed"," Instrument Parameters"," Laser Wavelength: 488 nm                                                            Quality"," Filter Wavelength: Scatter                                                          Average Counted Particles per Frame: 94"," Measurement Parameters                                                              Number of Traced Particles: 745"," Cell S/N: ZNTA"," Measurement Mode: Size Distribution 2 Cycles                                        Analysis Parameters"," 11 Positions, 3 Removed for Analysis                                                Max Area: 1000, Min Area: 10, Min Brightness: 20"," 1.3E+6                                                                              1.5E+8"," 1.2E+6"," 1.1E+6                                                                             1.25E+8"," 1.0E+6"," 9.0E+5","                                                                                      1E+8"," 8.0E+5"," 7.0E+5","                                                                                     7.5E+7"," 6.0E+5"," 5.0E+5","                                                                                      5E+7"," 4.0E+5"," 3.0E+5"," 2.0E+5                                                                              2.5E+7"," 1.0E+5"," 0.0E+0                                                                                   0","          0   100   200     300   400 500 600      700     800   900   1000                   0   100   200   300   400  500     600   700    800   900   1000","                                   Diameter / nm                                                                     Diameter / nm","Peak Analysis (Concentration)","Diameter / nm             Particles/mL       FWHM / nm             Percentage","       150.3                   1.8E+6            145.5                   85.6","       284.8                   5.8E+5             41.2                   14.4","X Values (all sizes are given in nm)","               Number             Concentration          Volume","   X10          102.7                    102.7            168.0","   X50          174.9                    174.9            283.6","   X90          292.9                    293.0            463.2","  Span            1.1                       1.1              1.0","                                                                                          Comment"," Mean           191.8                    191.8            305.3","StdDev           80.0                     79.9            116.6","                                                                                                                                                             (Signature)","Analyzed Video: C:\\Zetaview Data\\Nancy\\2023_02_14\\20230214_0002_G104_F1_size_488.avi","ZetaVIEW S/N 19-490, Software ZetaView (version 8.05.12 SP2), Camera 0.713 mum/px                              Experiment: 2023-02-14 14:47, Report: 2023-02-14 14:50","                                                                                                   Electrophoresis &amp; Brownian Motion","                                                                                                             Video Analysis","                                                                                                      Laser Scattering Microscopy","                                           www.particle-metrix.de","","                                                                                    Operator (Report): sop_kabanovlab.inst","                                                                                    Video Operator: sop_kabanovlab.inst"," Sample Parameters                                                                  Result (sizes in nm)"," Sample Name: G104_F4                                                                                  Number      Concentration    Volume"," Comment: Sens=80, Sample Remarks0:                                                 Median (X50)        191.4       191.4       292.6"," Sample Remarks1:                                                                   Span                 83.1       83.0      121.8"," Sample Remarks2:"," Electrolyte:                                                                       Concentration:                 3.2E+7 Particles / mL"," Temperature: 25.05 °C sensed                                                       Dilution Factor:                     50"," pH 7.0 entered                                                                     Original Concentration:        1.6E+9 Particles / mL"," Conductivity: 15000.00 µS/cm sensed"," Instrument Parameters"," Laser Wavelength: 488 nm                                                           Quality"," Filter Wavelength: Scatter                                                         Average Counted Particles per Frame: 70"," Measurement Parameters                                                             Number of Traced Particles: 752"," Cell S/N: ZNTA"," Measurement Mode: Size Distribution 2 Cycles                                       Analysis Parameters"," 11 Positions, 1 Removed for Analysis                                               Max Area: 1000, Min Area: 10, Min Brightness: 20"," 1.2E+6                                                                              2E+8"," 1.1E+6                                                                             1.8E+8"," 1.0E+6","                                                                                    1.6E+8"," 9.0E+5","                                                                                    1.4E+8"," 8.0E+5"," 7.0E+5                                                                             1.2E+8"," 6.0E+5                                                                              1E+8"," 5.0E+5                                                                              8E+7"," 4.0E+5","                                                                                     6E+7"," 3.0E+5","                                                                                     4E+7"," 2.0E+5"," 1.0E+5                                                                              2E+7"," 0.0E+0                                                                                  0","          0   100   200     300   400 500 600      700     800   900   1000                  0   100   200   300   400  500     600   700    800   900   1000","                                   Diameter / nm                                                                    Diameter / nm","Peak Analysis (Concentration)","Diameter / nm             Particles/mL       FWHM / nm             Percentage","       167.3                   1.3E+6            134.2                   95.4","       396.9                   1.6E+5             33.7                     4.6","X Values (all sizes are given in nm)","               Number             Concentration          Volume","   X10          109.3                    109.3            181.5","   X50          191.4                    191.4            292.6","   X90          312.2                    312.2            494.7","  Span            1.1                       1.1              1.1","                                                                                         Comment"," Mean           206.8                    206.8            318.4","StdDev           83.1                     83.0            121.8","                                                                                                                                                            (Signature)","Analyzed Video: C:\\Zetaview Data\\Nancy\\2023_02_14\\20230214_0006_G104_F4_size_488.avi","ZetaVIEW S/N 19-490, Software ZetaView (version 8.05.12 SP2), Camera 0.713 mum/px                             Experiment: 2023-02-14 15:06, Report: 2023-02-14 15:08","                                                                                                   Electrophoresis &amp; Brownian Motion","                                                                                                             Video Analysis","                                                                                                      Laser Scattering Microscopy","                                           www.particle-metrix.de","","                                                                                    Operator (Report): sop_kabanovlab.inst","                                                                                    Video Operator: sop_kabanovlab.inst"," Sample Parameters                                                                  Result (sizes in nm)"," Sample Name: G104_F1-4                                                                                Number      Concentration    Volume"," Comment: Sens=80, Sample Remarks0:                                                 Median (X50)        169.6       169.6       273.3"," Sample Remarks1:                                                                   Span                 77.8       77.8       95.9"," Sample Remarks2:"," Electrolyte:                                                                       Concentration:                 5.6E+7 Particles / mL"," Temperature: 25.68 °C sensed                                                       Dilution Factor:                     1000"," pH 7.0 entered                                                                     Original Concentration:        5.6E+10 Particles / mL"," Conductivity: 12216.54 µS/cm sensed"," Instrument Parameters"," Laser Wavelength: 488 nm                                                           Quality"," Filter Wavelength: Scatter                                                         Average Counted Particles per Frame: 123"," Measurement Parameters                                                             Number of Traced Particles: 994"," Cell S/N: ZNTA"," Measurement Mode: Size Distribution 2 Cycles                                       Analysis Parameters"," 11 Positions, 2 Removed for Analysis                                               Max Area: 1000, Min Area: 10, Min Brightness: 20"," 2.4E+6                                                                              2E+8"," 2.2E+6                                                                             1.8E+8"," 2.0E+6","                                                                                    1.6E+8"," 1.8E+6","                                                                                    1.4E+8"," 1.6E+6"," 1.4E+6                                                                             1.2E+8"," 1.2E+6                                                                              1E+8"," 1.0E+6                                                                              8E+7"," 8.0E+5","                                                                                     6E+7"," 6.0E+5","                                                                                     4E+7"," 4.0E+5"," 2.0E+5                                                                              2E+7"," 0.0E+0                                                                                  0","          0   100   200     300   400 500 600      700     800   900   1000                  0   100   200   300   400  500     600   700    800   900   1000","                                   Diameter / nm                                                                    Diameter / nm","Peak Analysis (Concentration)","Diameter / nm             Particles/mL       FWHM / nm             Percentage","       141.1                   2.7E+6            142.8                   95.0","       353.6                   3.9E+5             25.9                     5.0","X Values (all sizes are given in nm)","               Number             Concentration          Volume","   X10           98.4                     98.4            164.5","   X50          169.6                    169.6            273.3","   X90          292.8                    292.8            411.9","  Span            1.1                       1.1              0.9","                                                                                         Comment"," Mean           187.8                    187.8            287.2","StdDev           77.8                     77.8              95.9","                                                                                                                                                            (Signature)","Analyzed Video: C:\\Zetaview Data\\Nancy\\2023_02_14\\20230214_0014_G104_F1-4_size_488.avi","ZetaVIEW S/N 19-490, Software ZetaView (version 8.05.12 SP2), Camera 0.713 mum/px                             Experiment: 2023-02-14 15:37, Report: 2023-02-14 15:40","                                                                                                    Electrophoresis &amp; Brownian Motion","                                                                                                              Video Analysis","                                                                                                       Laser Scattering Microscopy","                                           www.particle-metrix.de","","                                                                                      Operator (Report): sop_kabanovlab.inst","                                                                                      Video Operator: sop_kabanovlab.inst"," Sample Parameters                                                                    Result (sizes in nm)"," Sample Name: G104_F3_scatter80                                                                         Number      Concentration    Volume"," Comment: Sample Remarks0:                                                            Median (X50)       145.2       145.2       237.4"," Sample Remarks1:                                                                     Span                71.4       71.1       76.4"," Sample Remarks2:"," Electrolyte:                                                                         Concentration:                 1.3E+7 Particles / mL"," Temperature: 27.49 °C sensed                                                         Dilution Factor:                     1000"," pH 7.0 entered                                                                       Original Concentration:        1.3E+10 Particles / mL"," Conductivity: 12726.56 µS/cm sensed"," Instrument Parameters"," Laser Wavelength: 488 nm                                                             Quality"," Filter Wavelength: Scatter                                                           Average Counted Particles per Frame: 28"," Measurement Parameters                                                               Number of Traced Particles: 133"," Cell S/N: ZNTA"," Measurement Mode: Size Distribution 1 Cycle                                          Analysis Parameters"," 11 Positions, 2 Removed for Analysis                                                 Max Area: 1000, Min Area: 10, Min Brightness: 20"," 7.5E+6                                                                             7.5E+7"," 5.0E+6                                                                               5E+7"," 2.5E+6                                                                             2.5E+7"," 0.0E+0                                                                                   0","          0   100   200     300   400 500 600      700     800   900   1000                   0   100   200   300   400  500     600    700    800   900   1000","                                   Diameter / nm                                                                     Diameter / nm","Peak Analysis (Concentration)","Diameter / nm             Particles/mL       FWHM / nm             Percentage","       146.3                   2.2E+6            158.0                  100.0","X Values (all sizes are given in nm)","               Number             Concentration          Volume","   X10           75.3                     75.3            140.6","   X50          145.2                    145.2            237.4","   X90          260.2                    260.2            337.7","  Span            1.3                       1.3              0.8"," Mean           169.5                    169.5            256.0","                                                                                          Comment","StdDev           71.4                     71.1              76.4","                                                                                                                                                              (Signature)","Analyzed Video: D:\\Julia Rager\\0413223\\20230412_0023_G104_F3_scatter80_size_488.avi","ZetaVIEW S/N 19-490, Software ZetaView (version 8.05.12 SP2), Camera 0.713 mum/px                               Experiment: 2023-04-12 14:28, Report: 2023-04-12 14:29","                                                                                                    Electrophoresis &amp; Brownian Motion","                                                                                                              Video Analysis","                                                                                                       Laser Scattering Microscopy","                                           www.particle-metrix.de","","                                                                                      Operator (Report): sop_kabanovlab.inst","                                                                                      Video Operator: sop_kabanovlab.inst"," Sample Parameters                                                                    Result (sizes in nm)"," Sample Name: G104_F2_scatter80                                                                         Number      Concentration    Volume"," Comment: Sample Remarks0:                                                            Median (X50)       137.6       137.6       211.3"," Sample Remarks1:                                                                     Span                62.3       62.1       69.2"," Sample Remarks2:"," Electrolyte:                                                                         Concentration:                 1.7E+7 Particles / mL"," Temperature: 27.56 °C sensed                                                         Dilution Factor:                     2000"," pH 7.0 entered                                                                       Original Concentration:        3.4E+10 Particles / mL"," Conductivity: 14213.85 µS/cm sensed"," Instrument Parameters"," Laser Wavelength: 488 nm                                                             Quality"," Filter Wavelength: Scatter                                                           Average Counted Particles per Frame: 38"," Measurement Parameters                                                               Number of Traced Particles: 183"," Cell S/N: ZNTA"," Measurement Mode: Size Distribution 1 Cycle                                          Analysis Parameters"," 11 Positions, 2 Removed for Analysis                                                 Max Area: 1000, Min Area: 10, Min Brightness: 20"," 1.0E+7                                                                             1.1E+8"," 9.0E+6                                                                               1E+8","                                                                                      9E+7"," 8.0E+6","                                                                                      8E+7"," 7.0E+6","                                                                                      7E+7"," 6.0E+6","                                                                                      6E+7"," 5.0E+6","                                                                                      5E+7"," 4.0E+6","                                                                                      4E+7"," 3.0E+6","                                                                                      3E+7"," 2.0E+6                                                                               2E+7"," 1.0E+6                                                                               1E+7"," 0.0E+0                                                                                   0","          0   100   200     300   400 500 600      700     800   900   1000                   0   100   200   300   400  500     600    700    800   900   1000","                                   Diameter / nm                                                                     Diameter / nm","Peak Analysis (Concentration)","Diameter / nm             Particles/mL       FWHM / nm             Percentage","       145.3                   3.2E+6            148.0                  100.0","X Values (all sizes are given in nm)","               Number             Concentration          Volume","   X10           73.7                     73.7            127.2","   X50          137.6                    137.6            211.3","   X90          233.5                    233.5            322.4","  Span            1.2                       1.2              0.9"," Mean           161.7                    161.7            231.0","                                                                                          Comment","StdDev           62.3                     62.1              69.2","                                                                                                                                                              (Signature)","Analyzed Video: D:\\Julia Rager\\0413223\\20230412_0024_G104_F2_scatter80_size_488.avi","ZetaVIEW S/N 19-490, Software ZetaView (version 8.05.12 SP2), Camera 0.713 mum/px                               Experiment: 2023-04-12 14:35, Report: 2023-04-12 14:37"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>PDF Identifier<\/th>\n      <th>Text<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"PDF Identifier","targets":1},{"name":"Text","targets":2}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-7e588be9b9738f21fa2b" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-7e588be9b9738f21fa2b">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80","81","82","83","84","85","86","87","88","89","90","91","92","93","94","95","96","97","98","99","100","101","102","103","104","105","106","107","108","109","110","111","112","113","114","115","116","117","118","119","120","121","122","123","124","125","126","127","128","129","130","131","132","133","134","135","136","137","138","139","140","141","142","143","144","145","146","147","148","149","150","151","152","153","154","155","156","157","158","159","160","161","162","163","164","165","166","167","168","169","170","171","172","173","174","175","176","177","178","179","180","181","182","183","184","185","186","187","188","189","190","191","192","193","194","195","196","197","198","199","200","201","202","203","204","205","206","207","208","209","210","211","212","213","214","215","216","217","218","219","220","221","222","223","224","225","226","227","228","229","230","231","232","233","234","235","236","237","238","239","240","241","242","243","244","245","246","247","248","249","250","251","252","253","254","255","256","257","258","259","260","261","262","263","264","265","266","267","268","269","270","271","272","273","274","275","276"],["20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0002_Expt1_A_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0014_Expt1_C_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0023_Expt1_D_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488","20230214_0024_Expt1_B_size_488"],["                                                                                                    Electrophoresis &amp; Brownian Motion","                                                                                                              Video Analysis","                                                                                                       Laser Scattering Microscopy","                                           www.particle-metrix.de","","                                                                                      Operator (Report): sop_kabanovlab.inst","                                                                                      Video Operator: sop_kabanovlab.inst"," Sample Parameters                                                                   Result (sizes in nm)"," Sample Name: G104_F1                                                                                   Number      Concentration    Volume"," Comment: Sens=80, Sample Remarks0:                                                  Median (X50)        174.9       174.9       283.6"," Sample Remarks1:                                                                    Span                 80.0       79.9      116.6"," Sample Remarks2:"," Electrolyte:                                                                        Concentration:                 4.3E+7 Particles / mL"," Temperature: 24.64 °C sensed                                                        Dilution Factor:                     200"," pH 7.0 entered                                                                      Original Concentration:        8.6E+9 Particles / mL"," Conductivity: 9241.01 µS/cm sensed"," Instrument Parameters"," Laser Wavelength: 488 nm                                                            Quality"," Filter Wavelength: Scatter                                                          Average Counted Particles per Frame: 94"," Measurement Parameters                                                              Number of Traced Particles: 745"," Cell S/N: ZNTA"," Measurement Mode: Size Distribution 2 Cycles                                        Analysis Parameters"," 11 Positions, 3 Removed for Analysis                                                Max Area: 1000, Min Area: 10, Min Brightness: 20"," 1.3E+6                                                                              1.5E+8"," 1.2E+6"," 1.1E+6                                                                             1.25E+8"," 1.0E+6"," 9.0E+5","                                                                                      1E+8"," 8.0E+5"," 7.0E+5","                                                                                     7.5E+7"," 6.0E+5"," 5.0E+5","                                                                                      5E+7"," 4.0E+5"," 3.0E+5"," 2.0E+5                                                                              2.5E+7"," 1.0E+5"," 0.0E+0                                                                                   0","          0   100   200     300   400 500 600      700     800   900   1000                   0   100   200   300   400  500     600   700    800   900   1000","                                   Diameter / nm                                                                     Diameter / nm","Peak Analysis (Concentration)","Diameter / nm             Particles/mL       FWHM / nm             Percentage","       150.3                   1.8E+6            145.5                   85.6","       284.8                   5.8E+5             41.2                   14.4","X Values (all sizes are given in nm)","               Number             Concentration          Volume","   X10          102.7                    102.7            168.0","   X50          174.9                    174.9            283.6","   X90          292.9                    293.0            463.2","  Span            1.1                       1.1              1.0","                                                                                          Comment"," Mean           191.8                    191.8            305.3","StdDev           80.0                     79.9            116.6","                                                                                                                                                             (Signature)","Analyzed Video: C:\\Zetaview Data\\Nancy\\2023_02_14\\20230214_0002_G104_F1_size_488.avi","ZetaVIEW S/N 19-490, Software ZetaView (version 8.05.12 SP2), Camera 0.713 mum/px                              Experiment: 2023-02-14 14:47, Report: 2023-02-14 14:50","                                                                                                   Electrophoresis &amp; Brownian Motion","                                                                                                             Video Analysis","                                                                                                      Laser Scattering Microscopy","                                           www.particle-metrix.de","","                                                                                    Operator (Report): sop_kabanovlab.inst","                                                                                    Video Operator: sop_kabanovlab.inst"," Sample Parameters                                                                  Result (sizes in nm)"," Sample Name: G104_F4                                                                                  Number      Concentration    Volume"," Comment: Sens=80, Sample Remarks0:                                                 Median (X50)        191.4       191.4       292.6"," Sample Remarks1:                                                                   Span                 83.1       83.0      121.8"," Sample Remarks2:"," Electrolyte:                                                                       Concentration:                 3.2E+7 Particles / mL"," Temperature: 25.05 °C sensed                                                       Dilution Factor:                     50"," pH 7.0 entered                                                                     Original Concentration:        1.6E+9 Particles / mL"," Conductivity: 15000.00 µS/cm sensed"," Instrument Parameters"," Laser Wavelength: 488 nm                                                           Quality"," Filter Wavelength: Scatter                                                         Average Counted Particles per Frame: 70"," Measurement Parameters                                                             Number of Traced Particles: 752"," Cell S/N: ZNTA"," Measurement Mode: Size Distribution 2 Cycles                                       Analysis Parameters"," 11 Positions, 1 Removed for Analysis                                               Max Area: 1000, Min Area: 10, Min Brightness: 20"," 1.2E+6                                                                              2E+8"," 1.1E+6                                                                             1.8E+8"," 1.0E+6","                                                                                    1.6E+8"," 9.0E+5","                                                                                    1.4E+8"," 8.0E+5"," 7.0E+5                                                                             1.2E+8"," 6.0E+5                                                                              1E+8"," 5.0E+5                                                                              8E+7"," 4.0E+5","                                                                                     6E+7"," 3.0E+5","                                                                                     4E+7"," 2.0E+5"," 1.0E+5                                                                              2E+7"," 0.0E+0                                                                                  0","          0   100   200     300   400 500 600      700     800   900   1000                  0   100   200   300   400  500     600   700    800   900   1000","                                   Diameter / nm                                                                    Diameter / nm","Peak Analysis (Concentration)","Diameter / nm             Particles/mL       FWHM / nm             Percentage","       167.3                   1.3E+6            134.2                   95.4","       396.9                   1.6E+5             33.7                     4.6","X Values (all sizes are given in nm)","               Number             Concentration          Volume","   X10          109.3                    109.3            181.5","   X50          191.4                    191.4            292.6","   X90          312.2                    312.2            494.7","  Span            1.1                       1.1              1.1","                                                                                         Comment"," Mean           206.8                    206.8            318.4","StdDev           83.1                     83.0            121.8","                                                                                                                                                            (Signature)","Analyzed Video: C:\\Zetaview Data\\Nancy\\2023_02_14\\20230214_0006_G104_F4_size_488.avi","ZetaVIEW S/N 19-490, Software ZetaView (version 8.05.12 SP2), Camera 0.713 mum/px                             Experiment: 2023-02-14 15:06, Report: 2023-02-14 15:08","                                                                                                   Electrophoresis &amp; Brownian Motion","                                                                                                             Video Analysis","                                                                                                      Laser Scattering Microscopy","                                           www.particle-metrix.de","","                                                                                    Operator (Report): sop_kabanovlab.inst","                                                                                    Video Operator: sop_kabanovlab.inst"," Sample Parameters                                                                  Result (sizes in nm)"," Sample Name: G104_F1-4                                                                                Number      Concentration    Volume"," Comment: Sens=80, Sample Remarks0:                                                 Median (X50)        169.6       169.6       273.3"," Sample Remarks1:                                                                   Span                 77.8       77.8       95.9"," Sample Remarks2:"," Electrolyte:                                                                       Concentration:                 5.6E+7 Particles / mL"," Temperature: 25.68 °C sensed                                                       Dilution Factor:                     1000"," pH 7.0 entered                                                                     Original Concentration:        5.6E+10 Particles / mL"," Conductivity: 12216.54 µS/cm sensed"," Instrument Parameters"," Laser Wavelength: 488 nm                                                           Quality"," Filter Wavelength: Scatter                                                         Average Counted Particles per Frame: 123"," Measurement Parameters                                                             Number of Traced Particles: 994"," Cell S/N: ZNTA"," Measurement Mode: Size Distribution 2 Cycles                                       Analysis Parameters"," 11 Positions, 2 Removed for Analysis                                               Max Area: 1000, Min Area: 10, Min Brightness: 20"," 2.4E+6                                                                              2E+8"," 2.2E+6                                                                             1.8E+8"," 2.0E+6","                                                                                    1.6E+8"," 1.8E+6","                                                                                    1.4E+8"," 1.6E+6"," 1.4E+6                                                                             1.2E+8"," 1.2E+6                                                                              1E+8"," 1.0E+6                                                                              8E+7"," 8.0E+5","                                                                                     6E+7"," 6.0E+5","                                                                                     4E+7"," 4.0E+5"," 2.0E+5                                                                              2E+7"," 0.0E+0                                                                                  0","          0   100   200     300   400 500 600      700     800   900   1000                  0   100   200   300   400  500     600   700    800   900   1000","                                   Diameter / nm                                                                    Diameter / nm","Peak Analysis (Concentration)","Diameter / nm             Particles/mL       FWHM / nm             Percentage","       141.1                   2.7E+6            142.8                   95.0","       353.6                   3.9E+5             25.9                     5.0","X Values (all sizes are given in nm)","               Number             Concentration          Volume","   X10           98.4                     98.4            164.5","   X50          169.6                    169.6            273.3","   X90          292.8                    292.8            411.9","  Span            1.1                       1.1              0.9","                                                                                         Comment"," Mean           187.8                    187.8            287.2","StdDev           77.8                     77.8              95.9","                                                                                                                                                            (Signature)","Analyzed Video: C:\\Zetaview Data\\Nancy\\2023_02_14\\20230214_0014_G104_F1-4_size_488.avi","ZetaVIEW S/N 19-490, Software ZetaView (version 8.05.12 SP2), Camera 0.713 mum/px                             Experiment: 2023-02-14 15:37, Report: 2023-02-14 15:40","                                                                                                    Electrophoresis &amp; Brownian Motion","                                                                                                              Video Analysis","                                                                                                       Laser Scattering Microscopy","                                           www.particle-metrix.de","","                                                                                      Operator (Report): sop_kabanovlab.inst","                                                                                      Video Operator: sop_kabanovlab.inst"," Sample Parameters                                                                    Result (sizes in nm)"," Sample Name: G104_F3_scatter80                                                                         Number      Concentration    Volume"," Comment: Sample Remarks0:                                                            Median (X50)       145.2       145.2       237.4"," Sample Remarks1:                                                                     Span                71.4       71.1       76.4"," Sample Remarks2:"," Electrolyte:                                                                         Concentration:                 1.3E+7 Particles / mL"," Temperature: 27.49 °C sensed                                                         Dilution Factor:                     1000"," pH 7.0 entered                                                                       Original Concentration:        1.3E+10 Particles / mL"," Conductivity: 12726.56 µS/cm sensed"," Instrument Parameters"," Laser Wavelength: 488 nm                                                             Quality"," Filter Wavelength: Scatter                                                           Average Counted Particles per Frame: 28"," Measurement Parameters                                                               Number of Traced Particles: 133"," Cell S/N: ZNTA"," Measurement Mode: Size Distribution 1 Cycle                                          Analysis Parameters"," 11 Positions, 2 Removed for Analysis                                                 Max Area: 1000, Min Area: 10, Min Brightness: 20"," 7.5E+6                                                                             7.5E+7"," 5.0E+6                                                                               5E+7"," 2.5E+6                                                                             2.5E+7"," 0.0E+0                                                                                   0","          0   100   200     300   400 500 600      700     800   900   1000                   0   100   200   300   400  500     600    700    800   900   1000","                                   Diameter / nm                                                                     Diameter / nm","Peak Analysis (Concentration)","Diameter / nm             Particles/mL       FWHM / nm             Percentage","       146.3                   2.2E+6            158.0                  100.0","X Values (all sizes are given in nm)","               Number             Concentration          Volume","   X10           75.3                     75.3            140.6","   X50          145.2                    145.2            237.4","   X90          260.2                    260.2            337.7","  Span            1.3                       1.3              0.8"," Mean           169.5                    169.5            256.0","                                                                                          Comment","StdDev           71.4                     71.1              76.4","                                                                                                                                                              (Signature)","Analyzed Video: D:\\Julia Rager\\0413223\\20230412_0023_G104_F3_scatter80_size_488.avi","ZetaVIEW S/N 19-490, Software ZetaView (version 8.05.12 SP2), Camera 0.713 mum/px                               Experiment: 2023-04-12 14:28, Report: 2023-04-12 14:29","                                                                                                    Electrophoresis &amp; Brownian Motion","                                                                                                              Video Analysis","                                                                                                       Laser Scattering Microscopy","                                           www.particle-metrix.de","","                                                                                      Operator (Report): sop_kabanovlab.inst","                                                                                      Video Operator: sop_kabanovlab.inst"," Sample Parameters                                                                    Result (sizes in nm)"," Sample Name: G104_F2_scatter80                                                                         Number      Concentration    Volume"," Comment: Sample Remarks0:                                                            Median (X50)       137.6       137.6       211.3"," Sample Remarks1:                                                                     Span                62.3       62.1       69.2"," Sample Remarks2:"," Electrolyte:                                                                         Concentration:                 1.7E+7 Particles / mL"," Temperature: 27.56 °C sensed                                                         Dilution Factor:                     2000"," pH 7.0 entered                                                                       Original Concentration:        3.4E+10 Particles / mL"," Conductivity: 14213.85 µS/cm sensed"," Instrument Parameters"," Laser Wavelength: 488 nm                                                             Quality"," Filter Wavelength: Scatter                                                           Average Counted Particles per Frame: 38"," Measurement Parameters                                                               Number of Traced Particles: 183"," Cell S/N: ZNTA"," Measurement Mode: Size Distribution 1 Cycle                                          Analysis Parameters"," 11 Positions, 2 Removed for Analysis                                                 Max Area: 1000, Min Area: 10, Min Brightness: 20"," 1.0E+7                                                                             1.1E+8"," 9.0E+6                                                                               1E+8","                                                                                      9E+7"," 8.0E+6","                                                                                      8E+7"," 7.0E+6","                                                                                      7E+7"," 6.0E+6","                                                                                      6E+7"," 5.0E+6","                                                                                      5E+7"," 4.0E+6","                                                                                      4E+7"," 3.0E+6","                                                                                      3E+7"," 2.0E+6                                                                               2E+7"," 1.0E+6                                                                               1E+7"," 0.0E+0                                                                                   0","          0   100   200     300   400 500 600      700     800   900   1000                   0   100   200   300   400  500     600    700    800   900   1000","                                   Diameter / nm                                                                     Diameter / nm","Peak Analysis (Concentration)","Diameter / nm             Particles/mL       FWHM / nm             Percentage","       145.3                   3.2E+6            148.0                  100.0","X Values (all sizes are given in nm)","               Number             Concentration          Volume","   X10           73.7                     73.7            127.2","   X50          137.6                    137.6            211.3","   X90          233.5                    233.5            322.4","  Span            1.2                       1.2              0.9"," Mean           161.7                    161.7            231.0","                                                                                          Comment","StdDev           62.3                     62.1              69.2","                                                                                                                                                              (Signature)","Analyzed Video: D:\\Julia Rager\\0413223\\20230412_0024_G104_F2_scatter80_size_488.avi","ZetaVIEW S/N 19-490, Software ZetaView (version 8.05.12 SP2), Camera 0.713 mum/px                               Experiment: 2023-04-12 14:35, Report: 2023-04-12 14:37"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>PDF Identifier<\/th>\n      <th>Text<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"PDF Identifier","targets":1},{"name":"Text","targets":2}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 </div>
 </div>
 <div id="extracting-variables-of-interest" class="section level3 hasAnchor">
@@ -512,8 +521,8 @@ <h3>Extracting Variables of Interest<a href="data-import-from-pdf-sources.html#e
 <span id="cb366-2"><a href="data-import-from-pdf-sources.html#cb366-2" tabindex="-1"></a>  <span class="fu">filter</span>(<span class="fu">grepl</span>(<span class="st">&quot;Dilution Factor&quot;</span>, Text))</span>
 <span id="cb366-3"><a href="data-import-from-pdf-sources.html#cb366-3" tabindex="-1"></a></span>
 <span id="cb366-4"><a href="data-import-from-pdf-sources.html#cb366-4" tabindex="-1"></a><span class="fu">datatable</span>(dilution_factor_df)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-62169e3dc11820ee75e1" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-62169e3dc11820ee75e1">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["20230214_0002_Expt1_A_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0014_Expt1_C_size_488","20230214_0023_Expt1_D_size_488","20230214_0024_Expt1_B_size_488"],[" Temperature: 24.64 °C sensed                                                        Dilution Factor:                     200"," Temperature: 25.05 °C sensed                                                       Dilution Factor:                     50"," Temperature: 25.68 °C sensed                                                       Dilution Factor:                     1000"," Temperature: 27.49 °C sensed                                                         Dilution Factor:                     1000"," Temperature: 27.56 °C sensed                                                         Dilution Factor:                     2000"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>PDF Identifier<\/th>\n      <th>Text<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"PDF Identifier","targets":1},{"name":"Text","targets":2}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-197b03ec192b0132ae63" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-197b03ec192b0132ae63">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["20230214_0002_Expt1_A_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0014_Expt1_C_size_488","20230214_0023_Expt1_D_size_488","20230214_0024_Expt1_B_size_488"],[" Temperature: 24.64 °C sensed                                                        Dilution Factor:                     200"," Temperature: 25.05 °C sensed                                                       Dilution Factor:                     50"," Temperature: 25.68 °C sensed                                                       Dilution Factor:                     1000"," Temperature: 27.49 °C sensed                                                         Dilution Factor:                     1000"," Temperature: 27.56 °C sensed                                                         Dilution Factor:                     2000"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>PDF Identifier<\/th>\n      <th>Text<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"PDF Identifier","targets":1},{"name":"Text","targets":2}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>The value we are trying to extract is at the end of a long character string. We will want to use the tidyverse function <code>separate()</code> to isolate those values, but we need to know what part of the character string will separate the dilution factor values from the rest of the text. To determine this, we can call just one of the data cells and copy the semicolon and following spaces for use in the <code>separate()</code> function.</p>
 <div class="sourceCode" id="cb367"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb367-1"><a href="data-import-from-pdf-sources.html#cb367-1" tabindex="-1"></a><span class="co"># Return the value in the first row and second column.</span></span>
 <span id="cb367-2"><a href="data-import-from-pdf-sources.html#cb367-2" tabindex="-1"></a>dilution_factor_df[<span class="dv">1</span>,<span class="dv">2</span>]</span></code></pre></div>
@@ -524,16 +533,16 @@ <h3>Extracting Variables of Interest<a href="data-import-from-pdf-sources.html#e
 <span id="cb369-3"><a href="data-import-from-pdf-sources.html#cb369-3" tabindex="-1"></a>  <span class="fu">separate</span>(Text, <span class="at">into =</span> <span class="fu">c</span>(<span class="cn">NA</span>, <span class="st">&quot;Dilution Factor&quot;</span>), <span class="at">sep =</span> <span class="st">&quot;:                     &quot;</span>)</span>
 <span id="cb369-4"><a href="data-import-from-pdf-sources.html#cb369-4" tabindex="-1"></a></span>
 <span id="cb369-5"><a href="data-import-from-pdf-sources.html#cb369-5" tabindex="-1"></a><span class="fu">datatable</span>(dilution_factor_df)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-94d8f711f0fdc61ff53c" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-94d8f711f0fdc61ff53c">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["20230214_0002_Expt1_A_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0014_Expt1_C_size_488","20230214_0023_Expt1_D_size_488","20230214_0024_Expt1_B_size_488"],["200","50","1000","1000","2000"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>PDF Identifier<\/th>\n      <th>Dilution Factor<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"PDF Identifier","targets":1},{"name":"Dilution Factor","targets":2}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-52ec23a34f861a573793" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-52ec23a34f861a573793">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["20230214_0002_Expt1_A_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0014_Expt1_C_size_488","20230214_0023_Expt1_D_size_488","20230214_0024_Expt1_B_size_488"],["200","50","1000","1000","2000"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>PDF Identifier<\/th>\n      <th>Dilution Factor<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"PDF Identifier","targets":1},{"name":"Dilution Factor","targets":2}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>For the “Original Concentration” variable, we filter rows by the string “pH” because the word concentration is found in multiple locations in the document.</p>
 <div class="sourceCode" id="cb370"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb370-1"><a href="data-import-from-pdf-sources.html#cb370-1" tabindex="-1"></a>concentration_df <span class="ot">=</span> pdf_raw <span class="sc">%&gt;%</span></span>
 <span id="cb370-2"><a href="data-import-from-pdf-sources.html#cb370-2" tabindex="-1"></a>    <span class="fu">filter</span>(<span class="fu">grepl</span>(<span class="st">&quot;pH&quot;</span>, Text)) <span class="sc">%&gt;%</span> </span>
 <span id="cb370-3"><a href="data-import-from-pdf-sources.html#cb370-3" tabindex="-1"></a>    <span class="fu">separate</span>(Text, <span class="fu">c</span>(<span class="cn">NA</span>, <span class="st">&quot;Concentration&quot;</span>), <span class="at">sep =</span> <span class="st">&quot;:        &quot;</span>) </span>
 <span id="cb370-4"><a href="data-import-from-pdf-sources.html#cb370-4" tabindex="-1"></a></span>
 <span id="cb370-5"><a href="data-import-from-pdf-sources.html#cb370-5" tabindex="-1"></a><span class="fu">datatable</span>(concentration_df)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-f89f0342e9464ec03ed7" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-f89f0342e9464ec03ed7">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["20230214_0002_Expt1_A_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0014_Expt1_C_size_488","20230214_0023_Expt1_D_size_488","20230214_0024_Expt1_B_size_488"],["8.6E+9 Particles / mL","1.6E+9 Particles / mL","5.6E+10 Particles / mL","1.3E+10 Particles / mL","3.4E+10 Particles / mL"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>PDF Identifier<\/th>\n      <th>Concentration<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"PDF Identifier","targets":1},{"name":"Concentration","targets":2}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-bd931f24b406eacf3ea6" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-bd931f24b406eacf3ea6">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["20230214_0002_Expt1_A_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0014_Expt1_C_size_488","20230214_0023_Expt1_D_size_488","20230214_0024_Expt1_B_size_488"],["8.6E+9 Particles / mL","1.6E+9 Particles / mL","5.6E+10 Particles / mL","1.3E+10 Particles / mL","3.4E+10 Particles / mL"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>PDF Identifier<\/th>\n      <th>Concentration<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"PDF Identifier","targets":1},{"name":"Concentration","targets":2}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>With the dilution factor variable, there were no additional characters after the value of interest, but here, “Particles / mL” remains and needs to be removed so that the data can be used in downstream analyses. We can add an additional cleaning step to remove “Particles / mL” from the data and add the units to the column title. <code>sep = " P"</code> refers to the space before and first letter of the string to be removed.</p>
 <div class="sourceCode" id="cb371"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb371-1"><a href="data-import-from-pdf-sources.html#cb371-1" tabindex="-1"></a>concentration_df <span class="ot">=</span> pdf_raw <span class="sc">%&gt;%</span></span>
 <span id="cb371-2"><a href="data-import-from-pdf-sources.html#cb371-2" tabindex="-1"></a>    <span class="fu">filter</span>(<span class="fu">grepl</span>(<span class="st">&quot;pH&quot;</span>, Text)) <span class="sc">%&gt;%</span> </span>
@@ -541,8 +550,8 @@ <h3>Extracting Variables of Interest<a href="data-import-from-pdf-sources.html#e
 <span id="cb371-4"><a href="data-import-from-pdf-sources.html#cb371-4" tabindex="-1"></a>    <span class="fu">separate</span>(Concentration, <span class="fu">c</span>(<span class="st">&quot;Concentration (Particles/ mL)&quot;</span>, <span class="cn">NA</span>), <span class="at">sep =</span> <span class="st">&quot; P&quot;</span>)</span>
 <span id="cb371-5"><a href="data-import-from-pdf-sources.html#cb371-5" tabindex="-1"></a></span>
 <span id="cb371-6"><a href="data-import-from-pdf-sources.html#cb371-6" tabindex="-1"></a><span class="fu">datatable</span>(concentration_df)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-ab12e3d3357e19ff667c" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-ab12e3d3357e19ff667c">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["20230214_0002_Expt1_A_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0014_Expt1_C_size_488","20230214_0023_Expt1_D_size_488","20230214_0024_Expt1_B_size_488"],["8.6E+9","1.6E+9","5.6E+10","1.3E+10","3.4E+10"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>PDF Identifier<\/th>\n      <th>Concentration (Particles/ mL)<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"PDF Identifier","targets":1},{"name":"Concentration (Particles/ mL)","targets":2}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-b73587e4875467ffcd03" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-b73587e4875467ffcd03">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["20230214_0002_Expt1_A_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0014_Expt1_C_size_488","20230214_0023_Expt1_D_size_488","20230214_0024_Expt1_B_size_488"],["8.6E+9","1.6E+9","5.6E+10","1.3E+10","3.4E+10"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>PDF Identifier<\/th>\n      <th>Concentration (Particles/ mL)<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"PDF Identifier","targets":1},{"name":"Concentration (Particles/ mL)","targets":2}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>Next, we want to extract size distribution data from the lower table. Note that the space in the first <code>separate()</code> function comes from the space between the “Number” and “Concentration” column in the string, and the space in the second <code>separate()</code> function comes from the space between the variable name and the number of interest. We can also convert values to numeric since they are currently stored as characters.</p>
 <div class="sourceCode" id="cb372"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb372-1"><a href="data-import-from-pdf-sources.html#cb372-1" tabindex="-1"></a>size_distribution_df <span class="ot">=</span> pdf_raw <span class="sc">%&gt;%</span></span>
 <span id="cb372-2"><a href="data-import-from-pdf-sources.html#cb372-2" tabindex="-1"></a>    <span class="fu">filter</span>(<span class="fu">grepl</span>(<span class="st">&quot;X10&quot;</span>, Text)<span class="sc">|</span> <span class="fu">grepl</span>(<span class="st">&quot;X50 &quot;</span>, Text)<span class="sc">|</span> <span class="fu">grepl</span>(<span class="st">&quot;X90&quot;</span>, Text) <span class="sc">|</span> <span class="fu">grepl</span>(<span class="st">&quot;Mean&quot;</span>, Text)<span class="sc">|</span> <span class="fu">grepl</span>(<span class="st">&quot;StdDev&quot;</span>, Text)) <span class="sc">%&gt;%</span></span>
@@ -552,8 +561,8 @@ <h3>Extracting Variables of Interest<a href="data-import-from-pdf-sources.html#e
 <span id="cb372-6"><a href="data-import-from-pdf-sources.html#cb372-6" tabindex="-1"></a>    <span class="fu">pivot_wider</span>(<span class="at">names_from =</span> Text, <span class="at">values_from =</span> Size)</span>
 <span id="cb372-7"><a href="data-import-from-pdf-sources.html#cb372-7" tabindex="-1"></a></span>
 <span id="cb372-8"><a href="data-import-from-pdf-sources.html#cb372-8" tabindex="-1"></a><span class="fu">datatable</span>(size_distribution_df)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-e61e29a6b7c208c064b5" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-e61e29a6b7c208c064b5">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["20230214_0002_Expt1_A_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0014_Expt1_C_size_488","20230214_0023_Expt1_D_size_488","20230214_0024_Expt1_B_size_488"],[102.7,109.3,98.40000000000001,75.3,73.7],[174.9,191.4,169.6,145.2,137.6],[292.9,312.2,292.8,260.2,233.5],[191.8,206.8,187.8,169.5,161.7],[80,83.09999999999999,77.8,71.40000000000001,62.3]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>PDF Identifier<\/th>\n      <th>   X10<\/th>\n      <th>   X50<\/th>\n      <th>   X90<\/th>\n      <th> Mean<\/th>\n      <th>StdDev<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[2,3,4,5,6]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"PDF Identifier","targets":1},{"name":"   X10","targets":2},{"name":"   X50","targets":3},{"name":"   X90","targets":4},{"name":" Mean","targets":5},{"name":"StdDev","targets":6}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-763a8cf39521c3a974e3" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-763a8cf39521c3a974e3">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["20230214_0002_Expt1_A_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0014_Expt1_C_size_488","20230214_0023_Expt1_D_size_488","20230214_0024_Expt1_B_size_488"],[102.7,109.3,98.40000000000001,75.3,73.7],[174.9,191.4,169.6,145.2,137.6],[292.9,312.2,292.8,260.2,233.5],[191.8,206.8,187.8,169.5,161.7],[80,83.09999999999999,77.8,71.40000000000001,62.3]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>PDF Identifier<\/th>\n      <th>   X10<\/th>\n      <th>   X50<\/th>\n      <th>   X90<\/th>\n      <th> Mean<\/th>\n      <th>StdDev<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[2,3,4,5,6]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"PDF Identifier","targets":1},{"name":"   X10","targets":2},{"name":"   X50","targets":3},{"name":"   X90","targets":4},{"name":" Mean","targets":5},{"name":"StdDev","targets":6}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 </div>
 <div id="creating-the-final-dataframe" class="section level3 hasAnchor">
 <h3>Creating the final dataframe<a href="data-import-from-pdf-sources.html#creating-the-final-dataframe" class="anchor-section" aria-label="Anchor link to header"></a></h3>
@@ -568,8 +577,8 @@ <h3>Creating the final dataframe<a href="data-import-from-pdf-sources.html#creat
 <span id="cb373-8"><a href="data-import-from-pdf-sources.html#cb373-8" tabindex="-1"></a></span>
 <span id="cb373-9"><a href="data-import-from-pdf-sources.html#cb373-9" tabindex="-1"></a><span class="co"># View new dataframe</span></span>
 <span id="cb373-10"><a href="data-import-from-pdf-sources.html#cb373-10" tabindex="-1"></a><span class="fu">datatable</span>(full_df)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-7fc6b6ce59c0233ddc6e" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-7fc6b6ce59c0233ddc6e">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["20230214_0002_Expt1_A_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0014_Expt1_C_size_488","20230214_0023_Expt1_D_size_488","20230214_0024_Expt1_B_size_488"],["200","50","1000","1000","2000"],["8.6E+9","1.6E+9","5.6E+10","1.3E+10","3.4E+10"],[102.7,109.3,98.40000000000001,75.3,73.7],[174.9,191.4,169.6,145.2,137.6],[292.9,312.2,292.8,260.2,233.5],[191.8,206.8,187.8,169.5,161.7],[80,83.09999999999999,77.8,71.40000000000001,62.3]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>PDF Identifier<\/th>\n      <th>Dilution Factor<\/th>\n      <th>Concentration (Particles/ mL)<\/th>\n      <th>   X10<\/th>\n      <th>   X50<\/th>\n      <th>   X90<\/th>\n      <th> Mean<\/th>\n      <th>StdDev<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"PDF Identifier","targets":1},{"name":"Dilution Factor","targets":2},{"name":"Concentration (Particles/ mL)","targets":3},{"name":"   X10","targets":4},{"name":"   X50","targets":5},{"name":"   X90","targets":6},{"name":" Mean","targets":7},{"name":"StdDev","targets":8}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-d57c3414a7471ac813b0" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-d57c3414a7471ac813b0">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["20230214_0002_Expt1_A_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0014_Expt1_C_size_488","20230214_0023_Expt1_D_size_488","20230214_0024_Expt1_B_size_488"],["200","50","1000","1000","2000"],["8.6E+9","1.6E+9","5.6E+10","1.3E+10","3.4E+10"],[102.7,109.3,98.40000000000001,75.3,73.7],[174.9,191.4,169.6,145.2,137.6],[292.9,312.2,292.8,260.2,233.5],[191.8,206.8,187.8,169.5,161.7],[80,83.09999999999999,77.8,71.40000000000001,62.3]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>PDF Identifier<\/th>\n      <th>Dilution Factor<\/th>\n      <th>Concentration (Particles/ mL)<\/th>\n      <th>   X10<\/th>\n      <th>   X50<\/th>\n      <th>   X90<\/th>\n      <th> Mean<\/th>\n      <th>StdDev<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"PDF Identifier","targets":1},{"name":"Dilution Factor","targets":2},{"name":"Concentration (Particles/ mL)","targets":3},{"name":"   X10","targets":4},{"name":"   X50","targets":5},{"name":"   X90","targets":6},{"name":" Mean","targets":7},{"name":"StdDev","targets":8}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>For easier downstream analysis, the last step is to separate the <code>PDF Identifier</code> column into an informative sample ID that matches up with other experimental data.</p>
 <div class="sourceCode" id="cb374"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb374-1"><a href="data-import-from-pdf-sources.html#cb374-1" tabindex="-1"></a>final_df <span class="ot">&lt;-</span> full_df <span class="sc">%&gt;%</span></span>
 <span id="cb374-2"><a href="data-import-from-pdf-sources.html#cb374-2" tabindex="-1"></a>  <span class="fu">separate</span>(<span class="st">&#39;PDF Identifier&#39;</span>, </span>
@@ -579,8 +588,8 @@ <h3>Creating the final dataframe<a href="data-import-from-pdf-sources.html#creat
 <span id="cb374-6"><a href="data-import-from-pdf-sources.html#cb374-6" tabindex="-1"></a>  <span class="fu">mutate</span>(<span class="fu">across</span>(<span class="st">&#39;Dilution Factor&#39;</span><span class="sc">:</span><span class="st">&#39;StdDev&#39;</span>, as.numeric)) <span class="co"># Change variables to numeric where appropriate</span></span>
 <span id="cb374-7"><a href="data-import-from-pdf-sources.html#cb374-7" tabindex="-1"></a></span>
 <span id="cb374-8"><a href="data-import-from-pdf-sources.html#cb374-8" tabindex="-1"></a><span class="fu">datatable</span>(final_df)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-ec6ce9dec5c5fbc47d4a" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-ec6ce9dec5c5fbc47d4a">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["20230214_0002_Expt1_A_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0014_Expt1_C_size_488","20230214_0023_Expt1_D_size_488","20230214_0024_Expt1_B_size_488"],["20230214","20230214","20230214","20230214","20230214"],["Expt1","Expt1","Expt1","Expt1","Expt1"],["A","Ctrl","C","D","B"],["488","488","488","488","488"],[200,50,1000,1000,2000],[8600000000,1600000000,56000000000,13000000000,34000000000],[102.7,109.3,98.40000000000001,75.3,73.7],[174.9,191.4,169.6,145.2,137.6],[292.9,312.2,292.8,260.2,233.5],[191.8,206.8,187.8,169.5,161.7],[80,83.09999999999999,77.8,71.40000000000001,62.3]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>PDF Identifier<\/th>\n      <th>Date<\/th>\n      <th>Experiment Number<\/th>\n      <th>Sample_ID<\/th>\n      <th>Wavelength<\/th>\n      <th>Dilution Factor<\/th>\n      <th>Concentration (Particles/ mL)<\/th>\n      <th>   X10<\/th>\n      <th>   X50<\/th>\n      <th>   X90<\/th>\n      <th> Mean<\/th>\n      <th>StdDev<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[6,7,8,9,10,11,12]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"PDF Identifier","targets":1},{"name":"Date","targets":2},{"name":"Experiment Number","targets":3},{"name":"Sample_ID","targets":4},{"name":"Wavelength","targets":5},{"name":"Dilution Factor","targets":6},{"name":"Concentration (Particles/ mL)","targets":7},{"name":"   X10","targets":8},{"name":"   X50","targets":9},{"name":"   X90","targets":10},{"name":" Mean","targets":11},{"name":"StdDev","targets":12}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-ce7123173bea600dd244" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-ce7123173bea600dd244">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["20230214_0002_Expt1_A_size_488","20230214_0006_Expt1_Ctrl_size_488","20230214_0014_Expt1_C_size_488","20230214_0023_Expt1_D_size_488","20230214_0024_Expt1_B_size_488"],["20230214","20230214","20230214","20230214","20230214"],["Expt1","Expt1","Expt1","Expt1","Expt1"],["A","Ctrl","C","D","B"],["488","488","488","488","488"],[200,50,1000,1000,2000],[8600000000,1600000000,56000000000,13000000000,34000000000],[102.7,109.3,98.40000000000001,75.3,73.7],[174.9,191.4,169.6,145.2,137.6],[292.9,312.2,292.8,260.2,233.5],[191.8,206.8,187.8,169.5,161.7],[80,83.09999999999999,77.8,71.40000000000001,62.3]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>PDF Identifier<\/th>\n      <th>Date<\/th>\n      <th>Experiment Number<\/th>\n      <th>Sample_ID<\/th>\n      <th>Wavelength<\/th>\n      <th>Dilution Factor<\/th>\n      <th>Concentration (Particles/ mL)<\/th>\n      <th>   X10<\/th>\n      <th>   X50<\/th>\n      <th>   X90<\/th>\n      <th> Mean<\/th>\n      <th>StdDev<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[6,7,8,9,10,11,12]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"PDF Identifier","targets":1},{"name":"Date","targets":2},{"name":"Experiment Number","targets":3},{"name":"Sample_ID","targets":4},{"name":"Wavelength","targets":5},{"name":"Dilution Factor","targets":6},{"name":"Concentration (Particles/ mL)","targets":7},{"name":"   X10","targets":8},{"name":"   X50","targets":9},{"name":"   X90","targets":10},{"name":" Mean","targets":11},{"name":"StdDev","targets":12}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>Let’s make a graph to help us answer Environmental Health Question 1.</p>
 <div class="sourceCode" id="cb375"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb375-1"><a href="data-import-from-pdf-sources.html#cb375-1" tabindex="-1"></a><span class="fu">theme_set</span>(<span class="fu">theme_bw</span>())</span>
 <span id="cb375-2"><a href="data-import-from-pdf-sources.html#cb375-2" tabindex="-1"></a></span>
@@ -621,8 +630,8 @@ <h3>Getting Familiar with the Example Dataset<a href="data-import-from-pdf-sourc
 <li><code>Label</code>: A description of the variable and values associated with the variable.</li>
 </ul>
 <p>After extracting the data, we want to end up with a dataframe that contains all of the variables, their corresponding columns, and a column that indicates which dataset the variable is associated with:</p>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-03d19c6b0fa22850790a" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-03d19c6b0fa22850790a">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80","81","82","83","84","85","86","87","88","89","90","91","92","93","94","95","96","97","98","99","100","101","102","103","104","105","106","107","108","109","110","111","112","113","114","115","116","117","118","119","120","121","122","123","124","125","126","127","128","129","130","131","132","133","134","135","136","137","138","139","140","141","142","143","144","145","146","147","148","149","150","151","152","153","154","155","156","157","158","159","160","161","162","163","164","165","166","167","168","169","170","171","172","173","174","175","176","177","178","179","180","181","182","183","184","185","186","187","188","189","190","191","192","193","194","195","196","197","198","199","200","201","202","203","204","205","206","207","208","209","210","211","212","213","214","215","216","217","218","219","220","221","222","223","224","225","226","227","228","229","230","231","232","233","234","235","236","237","238","239","240","241","242","243","244","245","246","247","248","249","250","251","252","253","254","255","256","257","258","259","260","261","262","263","264","265","266","267","268","269","270","271","272","273","274","275","276","277","278","279","280","281","282","283","284","285","286","287","288","289","290","291","292","293","294","295","296","297","298","299","300","301","302","303","304","305","306","307","308","309","310","311","312","313","314","315","316","317","318","319","320","321","322","323","324","325","326","327","328","329","330","331","332","333","334","335","336","337","338","339","340","341","342","343","344","345","346","347","348","349","350","351","352","353","354","355","356","357","358","359","360","361","362","363","364","365","366","367","368","369","370","371","372","373","374","375","376","377","378","379","380","381","382","383","384","385","386","387","388","389","390","391","392","393","394","395","396","397","398","399","400","401","402","403","404","405","406","407","408","409","410","411","412","413","414","415","416","417","418","419","420","421","422","423","424","425","426","427","428","429","430","431","432","433","434","435","436","437","438","439","440","441","442","443","444","445","446","447","448","449","450","451","452","453","454","455","456","457","458","459","460","461","462","463","464","465","466","467","468","469","470","471","472","473","474","475","476","477","478","479","480","481","482","483","484","485","486","487","488","489","490","491","492","493","494","495","496","497","498","499","500","501","502","503","504","505","506","507","508","509","510","511","512","513","514","515","516","517","518","519","520","521","522","523","524","525","526","527","528","529","530","531","532","533","534","535","536","537","538","539","540","541","542","543","544","545","546","547","548","549","550","551","552","553","554","555","556","557","558","559","560","561","562","563","564","565","566","567","568","569","570","571","572","573","574","575","576","577","578","579","580","581","582","583","584","585","586","587","588","589","590","591","592","593","594","595","596","597","598","599","600","601","602","603","604","605","606","607","608","609","610","611","612","613","614","615","616","617","618","619","620","621","622","623","624","625","626","627","628","629","630","631","632","633","634","635","636","637","638","639","640","641","642","643","644","645","646","647","648","649","650","651","652","653","654","655","656","657","658","659","660","661","662","663","664","665","666","667","668","669","670","671","672","673","674","675","676","677","678","679","680","681","682","683","684","685","686","687","688","689","690","691","692","693","694","695","696","697","698","699","700","701","702","703","704","705","706","707","708","709","710","711","712","713","714","715","716","717","718","719","720","721","722","723","724","725","726","727","728","729","730","731","732","733","734","735","736","737","738","739","740","741","742","743","744","745","746","747","748","749","750","751","752","753","754","755","756","757","758","759","760","761","762","763","764","765","766","767","768","769","770","771","772","773","774","775","776","777","778","779","780","781","782","783","784","785","786","787","788","789","790","791","792","793","794","795","796","797","798","799","800","801","802","803","804","805","806","807","808","809","810","811","812","813","814","815","816","817","818","819","820","821","822","823","824","825","826","827","828","829","830","831","832","833","834","835","836","837","838","839","840","841","842","843","844","845","846","847","848","849","850","851","852","853","854","855","856","857","858","859","860","861","862","863","864","865","866","867","868","869","870","871","872","873","874","875","876","877","878","879","880","881","882","883","884","885","886","887","888","889","890","891","892","893","894","895","896","897","898","899","900","901","902","903","904","905","906","907","908","909","910","911","912","913","914","915","916","917","918","919","920","921","922","923","924","925","926","927","928","929","930","931","932","933","934","935","936","937","938","939","940","941","942","943","944","945","946","947","948","949","950","951","952","953","954","955","956","957","958","959","960","961","962","963","964","965","966","967","968","969","970","971","972","973","974","975","976","977","978","979","980","981","982","983","984","985","986","987","988","989","990","991","992","993","994","995","996","997","998","999","1000","1001","1002","1003","1004","1005","1006","1007","1008","1009","1010","1011","1012","1013","1014","1015","1016","1017","1018","1019","1020","1021","1022","1023","1024","1025","1026","1027","1028","1029","1030","1031","1032","1033","1034","1035","1036","1037","1038","1039","1040","1041","1042","1043","1044","1045","1046","1047","1048","1049","1050","1051","1052","1053","1054","1055","1056","1057","1058","1059","1060","1061","1062","1063","1064","1065","1066","1067","1068","1069","1070","1071","1072","1073","1074","1075","1076","1077","1078","1079","1080","1081","1082","1083","1084","1085","1086","1087","1088","1089","1090","1091","1092","1093","1094","1095","1096","1097","1098","1099","1100","1101","1102","1103","1104","1105","1106","1107","1108","1109","1110","1111","1112","1113","1114","1115","1116","1117","1118","1119","1120","1121","1122","1123","1124","1125","1126","1127","1128","1129","1130","1131","1132","1133","1134","1135","1136","1137","1138","1139","1140","1141","1142","1143","1144","1145","1146","1147","1148","1149","1150","1151","1152","1153","1154","1155","1156","1157","1158","1159","1160","1161","1162","1163","1164","1165","1166","1167","1168","1169","1170","1171","1172","1173","1174","1175","1176","1177","1178","1179","1180","1181","1182","1183","1184","1185","1186","1187","1188","1189","1190"],["aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","bmi.sas7bdat","bmi.sas7bdat","bmi.sas7bdat","bmi.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat"],[1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,1,2,3,4,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,1,2,3,4,5,6,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28],["AAQ_1000","AAQ_1010","AAQ_1020","AAQ_1030","AAQ_1040","AAQ_1050","AAQ_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ABP_1000","ABP_1010","ABP_1020","ABP_1030","ABP_1040","ABP_1050","ABP_1060","ABP_1070","ABP_1080","ABP_1090","ABP_1100","ABP_1110","ABP_1120","ABP_1130","ABP_1140","ABP_1150","ABP_1160","ABP_1170","ABP_1180","ABP_1190","ABP_1200","ABP_1210","ABP_1220","ABP_1230","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ACT_1","ACT_2","ACT_3","ACT_4","ACT_5","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ICD9_CAT","AEC_1000","AEC_1010","AEC_1040","AEC_1050","AEC_1060","AEC_1070","AEC_1080","AEC_1090","AEC_1100","AEC_1110","AEC_1120","AEC_1020","AEC_1030","VNUM_C","VNUM","ICD9LONG","ICD9SHRT","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","AHA_1000","AHA_1020","AHA_1030","AHA_1040","AHA_1050","AHA_1060","AHA_1070","AHA_1090","AHA_1100","AHA_1110","AHA_1120","AHA_1130","AHA_1140","AHA_1150","AHA_1160","AHA_1170","AHA_1180","AHA_1190","AHA_1200","AHA_1210","AHA_1220","AHA_1250","AHA_1260","AHA_1270","AHA_1280","AHA_1290","AHA_1300","AHA_1310","AHA_1320","AHA_1330","AHA_1340","AHA_1350","AHA_1360","AHA_1370","AHA_1380","AHA_1390","AHA_1400","AHA_1410","AHA_1420","AHA_1430","AHA_1440","AHA_1450","AHA_1460","AHA_1470","AHA_1500","AHA_1570","AHA_1580","AHA_1590","AHA_1600","AHA_1730","AHA_1740","AHA_1750","AHA_1760","AHA_1770","AHA_1780","AHA_1790","AHA_1800","AHA_1810","AHA_1820","AHA_1830","AHA_1840","AHA_1850","AHA_1890","AHA_1900","VNUM_C","VNUM","VDATE","DATE_VAPE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ASU_1000","ASU_1010","ASU_1020","ASU_1030","ASU_1040","ASU_1050","ASU_1060","ASU_1070","ASU_1080","ASU_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","BAS_1000","BAS_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","BMI","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAGE","CHG_1000","CHG_1010","CHG_1020","CHG_1030","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","CME_1040","CME_1050","CME_1055","CME_1000","CME_1010","CME_1020","CME_1030","CME_1080","CME_1090","CME_1060","CME_1070","VNUM_C","VNUM","GEN_NAME","CLASS","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","COM_1000","COM_1010","COM_1020","COM_1030","COM_1040","COM_1050","COM_1060","COM_1070","COM_1080","COM_1090","COM_1100","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","CTX_1000","CTX_1010","CTX_1030","CTX_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E0A_1000","E0A_1020","E0A_1030","E0A_1040","E0A_1043","E0A_1045","E0A_1050","E0A_1060","E0A_1070","E0A_1073","E0A_1075","E0A_1080","E0A_1090","E0A_1100","E0A_1110","E0A_1120","E0A_1130","E0A_1140","E0A_1145","E0A_1150","E0A_1160","E0A_1170","E0A_1180","E0A_1190","E0A_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E0B_1000","E0B_1010","E0B_1020","E0B_1030","E0B_1040","E0B_1050","E0B_1060","E0B_1065","E0B_1070","E0B_1080","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E1_1000","E1_1020","E1_1030","E1_1040","E1_1050","E1_1060","E1_1065","E1_1070","E1_1080","E1_1090","E1_1100","E1_1110","E1_1120","E1_1130","E1_1140","E1_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E2_1000","E2_1010","E2_1020","E2_1030","E2_1040","E2_1050","E2_1060","E2_1070","E2_1080","E2_1090","E2_1100","E2_1110","E2_1120","E2_1130","E2_1140","E2_1145","E2_1150","E2_1160","E2_1170","E2_1173","E2_1175","E2_1180","E2_1190","E2_1200","E2_1210","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E3_1000","E3_1010","E3_1020","E3_1030","E3_1070","E3_1080","E3_1090","E3_1100","E3_1110","E3_1120","E3_1130","E3_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E4_1000","E4_1010","E4_1020","E4_1030","E4_1050","E4_1060","E4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E5_1000","E5_1010","E5_1020","E5_1030","E5_1035","E5_1040","E5_1050","E5_1060","E5_1065","E5_1070","E5_1080","E5_1085","E5_1090","E5_1100","E5_1110","E5_1120","E5_1130","E5_1140","E5_1150","E5_1160","E5_1170","E5_1180","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ENO_1000","ENO_1010","ENO_1020","ENO_1040","ENO_1050","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","HEQ_1000","HEQ_1010","HEQ_1020","HEQ_1030","HEQ_1040","HEQ_1050","HEQ_1060","HEQ_1070","HEQ_1080","HEQ_1090","HEQ_1100","HEQ_1110","HEQ_1120","HEQ_1130","HEQ_1140","HEQ_1150","HEQ_1160","HEQ_1170","HEQ_1180","HEQ_1190","HEQ_1200","HEQ_1210","HEQ_1220","HEQ_1230","HEQ_1240","HEQ_1250","HEQ_1260","HEQ_1270","HEQ_1280","HEQ_1290","HEQ_1300","HEQ_1310","HEQ_1320","HEQ_1330","HEQ_1340","HEQ_1350","HEQ_1360","HEQ_1370","HEQ_1380","HEQ_1390","HEQ_1400","HEQ_1410","HEQ_1420","HEQ_1430","HEQ_1440","HEQ_1450","HEQ_1460","HEQ_1470","HEQ_1480","HEQ_1490","HEQ_1500","HEQ_1510","HEQ_1520","HEQ_1530","HEQ_1540","HEQ_1550","HEQ_1560","HEQ_1570","HEQ_1580","HEQ_1590","HEQ_1600","HEQ_1610","HEQ_1620","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SEI_1000","SEI_1010","SEI_1020","SEI_1030","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","IGE","RE1","RE5","RE72","RE74","RMX1","RI6","RGX2","RTX4","RTX6","RWX1","RW3","RD2","RD1","N_ALLERGENPOS","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","LAB_1000","LAB_1010","LAB_1020","LAB_1030","LAB_1040","LAB_1050","LAB_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PLE_1000","PLE_1010","PLE_1020","PLE_1030","PLE_1040","PLE_1050","PLE_1060","PLE_1070","PLE_1080","PLE_1090","PLE_1100","PLE_1110","PLE_1130","PLE_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MTH_1000","MTH_1010","MTH_1020","MTH_1030","MTH_1040","MTH_1050","MTH_1060","MTH_1070","MTH_1080","MTH_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MAD_1000","MAD_1010","MAD_1020","MAD_1030","MAD_1040","MAD_1050","MAD_1060","MAD_1070","MAD_1090","MAD_1100","MAD_1110","MAD_1120","MAD_1130","MAD_1140","MAD_1150","MAD_1160","MAD_1170","MAD_1180","MAD_1190","MAD_1200","MAD_1220","MAD_1230","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MCA_1000","MCA_1010","MCA_1050","MCA_1060","MCA_1070","MCA_1080","MCA_1100","MCA_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MCP_1000","MCP_1010","MCP_1030","MCP_1040","MCP_1050","MCP_1060","MCP_1070","MCP_1080","MCP_1090","MCP_1100","MCP_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PA4_1000","PA4_1010","PA4_1020","PA4_1030","PA4_1040","PA4_1050","PA4_1060","PA4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PTX_1000","PTX_1010","PTX_1020","PTX_1030","PTX_1040","PTX_1050","PTX_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","VNUM","PERIOSTIN","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PI4_1000","PI4_1010","PI4_1020","PI4_1030","PI4_1040","PI4_1050","PI4_1060","PI4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PRG_1000","PRG_1010","PRG_1020","PRG_1030","PRG_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAD_1000","PAD_1010","PAD_1020","PAD_1030","PAD_1040","PAD_1050","PAD_1060","PAD_1070","PAD_1080","PAD_1090","PAD_1100","PAD_1110","PAD_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAL_1000","PAL_1010","PAL_1020","PAL_1030","PAL_1040","PAL_1050","PAL_1060","PAL_1070","PAL_1080","PAL_1090","PAL_1100","PAL_1110","PAL_1120","PAL_1130","PAL_1150","PAL_1160","PAL_1170","PAL_1180","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PTR_1000","PTR_1010","PTR_1050","PTR_1060","PTR_1100","PTR_1140","PTR_1180","PTR_1220","PTR_1260","PTR_1300","PTR_1340","PTR_1380","PTR_1390","PTR_1430","PTR_1470","PTR_1480","PTR_1490","PTR_1500","PTR_1535","PTR_1540","PTR_1550","PTR_1560","PTR_1600","PTR_1610","PTR_1620","PTR_1630","PTR_1670","PTR_1710","PTR_1750","PTR_1790","PTR_1830","PTR_1870","PTR_1910","VNUM_C","VNUM","VDATE","DATE_Q02","DATE_Q03","DATE_Q04","DATE_Q05","DATE_Q07","DATE_Q08","DATE_Q09","DATE_Q10","DATE_Q11","DATE_Q12","DATE_Q13","DATE_Q14","DATE_Q15","DATE_Q16","DATE_Q17","DATE_Q18","DATE_Q19","DATE_Q20","DATE_Q21","DATE_Q22","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","P10_1000","P10_1010","P10_1020","P10_1030","P10_1040","P10_1050","P10_1060","P10_1070","P10_1080","P10_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PCH_1000","PCH_1010","PCH_1020","PCH_1030","PCH_1040","PCH_1050","PCH_1060","PCH_1070","PCH_1080","PCH_1090","PCH_1100","PCH_1110","PCH_1120","PCH_1130","PCH_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","IAQ_1000","IAQ_1010","IAQ_1020","IAQ_1030","IAQ_1040","IAQ_1050","IAQ_1060","IAQ_1070","IAQ_1080","IAQ_1090","IAQ_1100","IAQ_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","REGIMEN_PERIOD_1","REGIMEN_PERIOD_2","REGIMEN_PERIOD_3","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","REG_1080","REG_1150","AGE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SER_1010","SER_1020","SER_1030","SER_1040","SER_1050","SER_1060","SER_1070","SER_1100","SER_1110","SER_1120","SER_1130","SER_1140","SER_1150","SER_1160","SER_1170","SER_1180","SER_1190","SER_1200","SER_1210","SER_1220","SER_1240","SER_1250","SER_1000","SER_1080","SER_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PSE_1060","PSE_1070","PSE_1080","PSE_1090","PSE_1100","PSE_1110","PSE_1130","PSE_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SIG_1000","SIG_1010","SIG_1020","SIG_1030","SIG_1040","SIG_1050","SIG_1060","SIG_1080","SIG_1090","SIG_1100","SIG_1110","SIG_1120","SIG_1130","SIG_1140","SIG_1150","SIG_1160","SIG_1170","SIG_1180","SIG_1190","SIG_1200","SIG_1210","SIG_1220","SIG_1230","SIG_1240","SIG_1250","SIG_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SFU_1000","SFU_1010","SFU_1020","SFU_1030","SFU_1040","SFU_1050","SFU_1060","SFU_1070","SFU_1080","SFU_1090","SFU_1100","SFU_1110","SFU_1120","SFU_1130","SFU_1140","SFU_1150","SFU_1160","SFU_1170","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SNQ_1000","SNQ_1010","SNQ_1020","SNQ_1030","SNQ_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SPI_1010","SPI_1020","SPI_1040","SPI_1050","SPI_1060","SPI_1070","VNUM_C","VNUM","VDATE","SPI_1030","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","AM_TRIAL_TYPE","AM_FVC","AM_FEV1","AM_PEF","AM_FEF2575","AM_FET","DRY_1","DRY_2","DRY_3","DRY_4","DRY_5","DRY_6","DRY_7","DRY_8","DRY_9","VNUM_C","VNUM","VDATE","DDATE","PM_TRIAL_TYPE","PM_FVC","PM_FEV1","PM_PEF","PM_FEF2575","PM_FET","DRY_10","DRY_11","DRY_12","DRY_13","DRY_14","DRY_15","DRY_16","DRY_17","DRY_18","DRY_19","DRY_20","DRY_21","DRY_22","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SLA_1010","SLA_1020","SLA_1030","SLA_1000","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SRE_1010","SRE_1020","SRE_1030","SRE_1040","SRE_1050","SRE_1060","SRE_1070","SRE_1080","SRE_1090","SRE_1100","SRE_1000","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SPU_1000","SPU_1010","SPU_1020","SPU_1030","SPU_1040","SPU_1050","SPU_1060","SPU_1070","SPU_1080","SPU_1090","SPU_1100","SPU_1110","SPU_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SAD_1000","SAD_1010","SAD_1020","SAD_1030","SAD_1040","SAD_1050","SAD_1060","SAD_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SCH_1000","SCH_1010","SCH_1020","SCH_1030","SCH_1040","SCH_1050","SCH_1055","SCH_1060","SCH_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TER_1000","TER_1010","TER_1020","TER_1040","TER_1050","TER_1070","TER_1080","TER_1090","TER_1100","TER_1110","TER_1120","TER_1130","TER_1140","TER_1150","TER_1160","TER_1170","TER_1180","TER_1185","TER_1200","TER_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TXF_1000","TXF_1010","TXF_1020","TXF_1030","TXF_1040","TXF_1050","TXF_1060","TXF_1070","TXF_1080","TXF_1090","TXF_1100","TXF_1110","TXF_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TCH_1000","TCH_1010","TCH_1020","TCH_1030","TCH_1040","TCH_1050","TCH_1060","TCH_1080","TCH_1090","TCH_1100","TCH_1110","TCH_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","WSH_1000","WSH_1010","WSH_1020","WSH_1040","WSH_1050","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","WPA_1000","WPA_1010","WPA_1020","WPA_1030","WPA_1040","WPA_1050","WPA_1060","WPA_1070","WPA_1080","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","HOW_SICK_TODAY","RUNNY_NOSE","PLUGGED_NOSE","SNEEZING","SORE_THROAT","SCRATCHY_THROAT","COUGH","HOARSENESS","HEAD_CONGESTION","CHEST_CONGESTION","FEELING_TIRED","THINK_CLEARLY","SLEEP_WELL","BREATH_EASILY","WALK","DAILY_ACTIVITIES","WORK_OUTSIDE_HOME","WORK_INSIDE_HOME","INTERACT","LIVE_LIFE","MY_COLD_IS","VNUM_C","VNUM","VDATE","WDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER"],["Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Char","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Char","Char","Num","Num","Num","Num","Char","Char","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Char","Num","Char","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Char","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num"],["In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, school, or at home? 0=None of the time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All of the time","During the past 3 days, how often have you had asthma symptoms? 0=Not at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how often have you used your rescue inhaler or nebulizer medication ? 0=Not at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how many total times did your asthma symptoms wake you up from sleep? 0=Not at all, 1=1 time in the last 3 days, 2=2-3 times in the last 3 days, 3=4-5 times in the last 3 days, 4=&gt;=6 times in the last 3 days","How would you rate the amount of impairment you have experienced due to your asthma in the past 3 days? 0=No impairment, 1=Mild impairment, 2=Moderate impairment, 3=Severe impairment, 4=Very severe impairment","How stressed or frightened were you by your asthma symptoms in the past 3 days? 0=Not at all, 1=Mildly, 2=Moderately, 3=Severely, 4=Very severely","Why do you think your asthma was worse in the past 3 days compared to what is normal for you? 0=I have not been worse over the past 3 days. My asthma symptoms have been usual., 1=Common cold, 2=Allergies, 3=Pollution or chemical irritant, 4=Too little asth","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Are you currently retired? 1=Yes,0=No","Are you retired because of asthma? 1=Yes,0=No","Are you currently unemployed? 1=Yes,0=No","Are you unemployed because of asthma? 1=Yes,0=No","Do you get paid to do work? 1=Yes,0=No","How much does your asthma bother you at your paid work? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Overall, how much does your asthma bother you when you do jobs around the house? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Overall, how much does your asthma bother your social life? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Overall, how much does your asthma bother your personal life? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Are you involved in leisure activities, such as: walking for pleasure, sports, exercise, travelling, taking vacations? 1=Yes,0=No","When involved in leisure activities, how much does your asthma bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Would you say that you can't do some of these sorts of things because of asthma? 1=Yes,0=No","How much does your asthma bother you when you sleep? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much does the cost of your asthma medicines bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Do you get free prescriptions? 1=Yes,0=No","How much does the inconvenience or embarrassment of taking your asthma medicines bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much do coughs and colds bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Feeling upset is also a bother.  Does your asthma make you feel anxious, depressed, tired, or helpless? 1=Yes,0=No","Feeling upset is also a bother.  Does your asthma make you feel anxious, depressed, tired, or helpless? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will have an asthma attack when visiting a new place? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will catch a cold? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will let others down? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that your health may get worse in the future? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you won't be able to cope with an asthma attack? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past 4 weeks, how much of the time did your asthma keep you from getting as much done at work, school or at home? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often have your had shortness of breath? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often did your asthma symptoms wake you up at night or earlier than usual in the morning? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often have you used your rescue inhaler or nebulizer medication? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","How would you rate your asthma control during the past 4 weeks? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","ICD-9 category","Description of Adverse Event","ICD9 Code","Ongoing at current visit","Type (1=Intermitent; 2=Continuous)","Severity (1=Mild; 2=Moderate; 3=Severe)","Serious (0=No, 1=Yes)","Likelyhood of Relationship to Study Drug(s) (1=None; 2=Unlikely (remote); 3=Possible; 4=Probable)","Change in Study Drugs (1=Unchanged; 2=Altered)","Outcome (1=Completely recovered; 2=Recovered, but with lasting effects; 3=Death)","Treatment Required (1=None; 2=Medication; 3=Hospitalization; 4=Other)","Ongoing at final visit","Date Started","Date Stopped","Visit Number (character)","Visit Number (numeric)","Long ICD-9 description","Short ICD-9 description","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","ASTHMA HISTORY1.Approximately how old were you when chest symptoms suggesting asthma first appeared?","Did these symptoms appear immediately after or as a result of:1a. a respiratory infection such as a cold or pneumonia? 1=Yes, 0=No, 8=Don't Know","an occupational or job change? 1=Yes, 0=No, 8=Don't Know","a household move? 1=Yes, 0=No, 8=Don't Know","a pregnancy? 1=Yes, 0=No, 8=Don't Know","a hormonal change ? 1=Yes, 0=No, 8=Don't Know","How old were you when a doctor first diagnosed you with asthma?","Have any of your immediate blood relatives been told by a physician that they have asthma? 1=Yes, 0=No, 8=Don't Know","Father  (1=Yes, 0=No, 8=Don't Know)","Brother(s) or Sister(s)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Child(ren)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","ASTHMA SYMPTOMS1.How do you categorize your asthma symptoms throughout the course of the year? 1=Relatively the same all year, 2=Vary by season","Winter? 1=Yes,0=No","Spring? 1=Yes,0=No","Summer? 1=Yes,0=No","Fall? 1=Yes,0=No","In the last 12 months, how many...  5a.Asthma episodes have you had that required emergency care or an unscheduled office visit?","Overnight hospitalizations have you had due to asthma?","Courses of systemic corticosteroid therapy  for asthma have you taken?","Days of work, school, or housework have you missed due to asthma?","In the past 3 months, how many days of work, school, or housework have you missed due to asthma?","Have you ever been admitted to an intensive care unit for asthma? 1=Yes,0=No","How many times have you been admitted to an intensive care unit for asthma?","Have you ever had invasive mechanical ventilation? 1=Yes, 0=No, 8=Don't Know","Have you ever had non-invasive mechanical ventilation? 1=Yes, 0=No, 8=Don't Know","Do any of the following currently provoke your asthma? 1=Yes, 0=No, 8=Don't Know","Menstrual cycle(If participant is male or a postmenopausal female, leave blank.) (1=Yes, 0=No, 8=Don't Know)","Aspirin or non-steroidal anti-inflammatory drugs (e.g., Aleve, Motrin) (1=Yes, 0=No, 8=Don't Know)","Respiratory infections (e.g., colds)  (1=Yes, 0=No, 8=Don't Know)","Irritants (e.g., pollution, odors, perfumes, chemicals, household cleaners) (1=Yes, 0=No, 8=Don't Know)","Weather conditions (e.g., change in weather, humidity)  (1=Yes, 0=No, 8=Don't Know)","Exposure to cold air  (1=Yes, 0=No, 8=Don't Know)","Emotional factors (e.g., stress, laughing)  (1=Yes, 0=No, 8=Don't Know)","Tobacco smoke  (1=Yes, 0=No, 8=Don't Know)","Food additives/preservatives (e.g., MSG, sulfites)  (1=Yes, 0=No, 8=Don't Know)","Allergies (e.g., dust, animals, pollens)(1=Yes, 0=No, 8=Don't Know)","Other (1=Yes,0=No)","To which of the following did a doctor or other health practitioner say you were allergic? 1=Yes, 0=No, 8=Don't Know","Foods  (1=Yes, 0=No, 8=Don't Know)","Things you breathe in or are exposed to (e.g., dust, pollens, molds, animal fur, feathers, dander)  (1=Yes, 0=No, 8=Don't Know)","Stinging insects such as bees or wasps (1=Yes, 0=No, 8=Don't Know)","Latex  (1=Yes, 0=No, 8=Don't Know)","Other  (1=Yes,0=No)","Have you ever had eczema / atopic dermatitis ? 1=Yes, 0=No, 8=Don't Know","If YES, was your eczema diagnosed by a doctor? 1=Yes,0=No","Have any of your immediate blood relatives been told by a physician that they have allergies/eczema/hay fever? 1=Yes, 0=No, 8=Don't Know","Father  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Brother(s) or Sister(s)(1=Yes, 0=No, 8=Don't Know, 9=N/A)","Child(ren)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Did you grow up in a household where you were exposed to tobacco smoke? 1=Yes,0=No","Do you currently smoke? 1=Yes,0=No","Record smoking history in pack-years*. __ __ . __ pack-years","Were you ever a smoker? 1=Yes,0=No","Record smoking history in pack-years*. __ __ . __ pack-years","Do you currently live in a household where you are exposed to tobacco smoke? 1=Yes,0=No","Do you currently vape  or use a hookah ? 1=Yes,0=No","How frequently do you vape or use a hookah? 1=Infrequently , 2=Occasionally , 3=Weekly , 4=Daily","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many days a week","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many times a day","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many years have y","Have you ever vaped or used a hookah in the past? 1=Yes,0=No","Have you ever vaped or used a hookah in the past?  If NO, skip to Q17. Approximately how many years did you vape or use a hookah?  (1-99 years)","Do you currently live in a household where you are exposed to others vaping or using a hookah? 1=Yes,0=No","Do you spend time in social settings  where you are exposed to others vaping or using a hookah? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","When was the last time that you vaped or used a hookah? (number of days from Visit 1)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","How many days were you bothered by coughing during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe was your coughing during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by wheezing during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe was your wheezing during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by shortness of breath during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","Over average, how severe was your shortness of breath during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you awakened at night during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how much of a problem was being awakened at night during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by side effects of your asthma medication during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe were the side effects of your asthma medication during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Participants baseline peak flow (PEF) value (0-999 L/M) First visit (Visit 0A, or Visit 1): PEF (FEF Max) from prebronchodilator (baseline) spirometry at first visit (convert to L/M)","Participants baseline rescue use value (0-99 puffs/day) First visit (Visit 0A, or Visit 1): Self-reported average daily use of albuterol during the 14 days prior to first visit.","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","body mass index computed from height and weight measured at Visit 1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Page Number","Reason for change in scheduled medications status (1=Adverse Event, 2=Other (specify))","Reason for change in scheduled medications status Related adverse event number (0-99)","Current status of participants scheduled medications (1=Discontinued, 2=Resumed)","Date change took effect","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Code to identify a drug unit of measure; (Units)","Code to identify how frequently a drug is taken or administered. (Frequency)","Code to identify the route used to administer a drug. (Route)","Name of Medication","This number corresponds to the ID number of the AHFS monograph. (Medication Code)","Related Event","Medication Dose","Ongoing at current visit","Ongoing at final visit","Start Date","Stop Date","Visit Number (character)","Visit Number (numeric)","GEN_DRUG_NAME","CLASS_TEXT","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Diary and Peak Flow Compliance: Number of full days since the last visit (0-999 days)","Diary and Peak Flow Compliance: Number of days where AM and PM scheduled sessions are complete (AM and PM PEF and all diary questions for AM and PM answered) (0-999 days)","Diary and Peak Flow Compliance: Percent compliance If the compliance value in Q1c is less than 75Pct, re-emphasize the importance of completing scheduled diary assessments and peak flows. (0-150.0Pct)","Scheduled Respimat Compliance (Visits 2-9 only): Number of scheduled puffs since the last visit (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Number of remaining puffs on scheduled Respimat Reference all returned Respimats (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Number of puffs taken (60 x (# used Respimats)) - Q2b (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Percent compliance=Q2c/Q2a x 100 If the participant took less than 75Pct of the scheduled Respimat puffs, re-emphasize the importance of maintaining the daily dosing schedule. (0-150.0Pct)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of scheduled puffs since the last visit (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of remaining puffs reflected on scheduled Twisthaler counters Total the values reflected on all counters for all returned Twisthalers (i.e., out of their pouches). (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of puffs taken (60 x (# used Twisthalers)) - Q3b (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Percent compliance=Q3c/Q3a x 100 If the participant took less than 75Pct of the scheduled Twisthaler puffs, re-emphasize the importance of maintaining the daily dosing schedule. (0-150.0Pct)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","1=mometasone, 2=placebo","Blinded Scheduled Twisthaler Contents - How sure are you about your answer in Q1? 1=Absolutely sure - I know what the Twisthaler/MDI contains, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","1=tiotropium, 2=placebo","Blinded Scheduled Respimat Contents - How sure are you about your answer in Q4? 1=Absolutely sure - I know what the Respimat capsules contain, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has the participant or parent/legal guardian signed the SIENA Informed Consent document? 1=Yes,0=No","Ages 12-17 Only: Has the participant signed and dated the assent form or, if the participant is less than the local age of assent, has the participant given verbal assent? 1=Yes,0=No","Has the participant used an inhaled corticosteroid in the past 6 weeks? 1=Yes,0=No","Will the participant be 12 years of age, or older, as of Visit 1? 1=Yes,0=No","Has the participant used an inhaled corticosteroid or combination inhaled corticosteroid/LABA in the past 3 weeks? 1=Yes,0=No","If NO, has the participant used a leukotriene modifier in the past 3 weeks? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used inhaled corticosteroid therapy?","Over the past 3 months, on average, how many days per week has the participant used inhaled corticosteroid therapy: &lt; 5? 1=Yes,0=No","Over the past 3 months, has the participant used greater than the equivalent of 80-240 mcg of inhaled beclomethasone daily? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used combination inhaled corticosteroid/LABA therapy?","Over the past 3 months, on average, how many days per week has the participant used combination inhaled corticosteroid/LABA therapy,  &lt; 5 (1=Yes,0=No)","Over the past 3 months, on average, how many days per week has the participant had daytime asthma symptoms?","Over the past 3 months, on average, how many days per week has the participant had daytime asthma symptoms,  &lt;=2? 1=Yes,0=No","Over the past 3 months, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms?","Over the past 3 months, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms, &lt;=2? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used his/her short acting betaagonist  for relief of symptoms?","Over the past 3 months, on average, how many days per week has the participant used his/her short acting betaagonist  for relief of symptoms, &lt; 2? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 80Pct? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 70Pct? 1=Yes,0=No","Does the participant have current evidence of any of the conditions listed on the Exclusionary Medical Conditions for SIENA  reference card, or any chronic diseases that would prevent participation in the trial or put the participant at risk by participati","Is the participant currently taking any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Is the participant able to go off these medications for the required washout period prior to Visit 1 and for the duration of the study? 1=Yes,0=No","Is the participant able to use the spirotel e-diary/PEF meter correctly, as evidenced by achieving a score of 13 on the spirotel Performance Checklist ? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Has the participant or parent/legal guardian signed the SIENA Informed Consent document?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past month, on average, how many days per week has the participant had daytime symptoms?","In the past month, on average, how many days per week has the participant had daytime symptoms, &lt;=2? 1=Yes,0=No","In the past month, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms?","In the past month, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms,  &lt;=2? 1=Yes,0=No","In the past month, on average, how many days per week has the participant used his/her short acting beta-agonist  for relief of symptoms?","In the past month, on average, how many days per week has the participant used his/her short acting beta-agonist  for relief of symptoms, &lt; 2? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 80Pct? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 70Pct? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant or parent/legal guardian sign the SIENA Informed Consent document? 1=Yes,0=No","Ages 12-17 Only: Has the participant signed and dated the assent form or, if the participant is less than the local age of assent, has the participant given verbal assent? 1=Yes,0=No","Is the participant 12 years of age, or older? 1=Yes,0=No","Does the participant plan to move away from the clinical site in the upcoming 11 months such that his/her ability to complete the study will be jeopardized? 1=Yes,0=No","Has the participant used investigative drugs and/or enrolled in an intervention trial in the past 30 days, or have plans to enroll in such a trial during the SIENA study? 1=Yes,0=No","Has the participant had a respiratory infection within the past 6 weeks? 1=Yes,0=No","Has the participant had a respiratory infection within the past 4 weeks? 1=Yes,0=No","During the past 4 weeks, on average, how many days per week has the participant had daytime asthma symptoms?","During the past 4 weeks, on average, how many days per week has the participant had daytime asthma symptoms  &gt; 2? 1=Yes,0=No","During the past 4 weeks, how many nighttime awakenings due to asthma symptoms has the participant had?","During the past 4 weeks, how many nighttime awakenings due to asthma symptoms has the participant had,  &gt; 2? 1=Yes,0=No","During the past 4 weeks, on average, how many days per week has the participant used his/her short-acting beta-agonist  for relief of symptoms?","During the past 4 weeks, on average, how many days per week has the participant used his/her short-acting beta-agonist  for relief of symptoms, &gt; 2? 1=Yes,0=No","Is Q7a, Q8a, or Q9a checked YES? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Did the participant or parent/legal guardian sign the SIENA Informed Consent document?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Does the participant have current evidence of any of the conditions listed on the Exclusionary Medical Conditions for SIENA  reference card, or any chronic diseases  that would prevent participation in the trial or put the participant at risk by participat","Does the participant have a history of bladder-neck obstruction, urinary retention, benign prostatic hyperplasia , or a clinically relevant urologic disorder that precludes study participation? 1=Yes,0=No","Does the participant have a history of narrow angle glaucoma? 1=Yes,0=No","Does the participant have a history of significant cardiovascular disorders or arrhythmias? 1=Yes,0=No","Has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card within the specified time periods? 1=Yes,0=No","Is the participant currently taking prescription or OTC medication other than those listed on the Allowed Medications  reference card? 1=Yes,0=No","Based on input from the participant and the study physician, will the participant need to use intranasal steroids at any time during the study? 1=Yes,0=No","Is the participant willing to use a single intranasal steroid at a stable dose continuously for the duration of the study, starting at Visit 1? 1=Yes,0=No","Is the participant currently receiving allergen immunotherapy  other than an established maintenance regimen implemented continuously for a minimum of 3 months? 1=Yes,0=No","Has the participant taken omalizumab within the past 3 months? 1=Yes,0=No","Has the participant used any smokeless tobacco products  in the past year? 1=Yes,0=No","Has the participant smoked cigarettes, a pipe, cigar, marijuana, electronic cigarettes, or any other substance in the past year? 1=Yes,0=No","Ages 18+ Only: Does the participant have a smoking history of greater than 10 pack-years? 1=Yes,0=No","Ages 12-17 Only: Does the participant have a smoking history of greater than 5 pack-years? 1=Yes,0=No","Has the participant received a physician diagnosis of asthma at least 12 months ago? 1=Yes,0=No","Has the participant received a physician diagnosis of asthma at least 12 months ago or had a history consistent with asthma for the previous 12 months? 1=Yes,0=No","Has the participant experienced a life-threatening asthma exacerbation requiring treatment with intubation or mechanical ventilation in the past 5 years? 1=Yes,0=No","Has the participant had an asthma exacerbation requiring systemic corticosteroid treatment in the past 6 weeks? 1=Yes,0=No","Has the participant used an oral or inhaled corticosteroid, or leukotriene modifier in the past 6 weeks? 1=Yes,0=No","Has the participant used an oral corticosteroid in the past 6 weeks? 1=Yes,0=No","Has the participant used an inhaled corticosteroid or leukotriene modifier in the past 3 weeks? 1=Yes,0=No","Is the participant potentially able to bear children? 1=Yes, 0=No, 9=N/A","Is the participant currently pregnant or lactating? 1=Yes,0=No","Does the participant agree to use one of the approved methods indicated on the Birth Control Methods  reference card for the duration of the study? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was the participants prebronchodilator FEV1 &gt;=70Pct predicted? 1=Yes,0=No","Did the participants FEV1 improve &gt;=12Pct and &gt;=200 ml in response to four puffs of albuterol? 1=Yes,0=No","Does the participant have valid source documentation within the past 6 months for an overread AsthmaNet methacholine challenge  with a PC20 &lt;=16 mg/ml? 1=Yes,0=No","Did the participants FEV1 improve &gt;=12Pct and &gt;=200 ml in response to four puffs of albuterol?","Was the participants methacholine PC20 &lt;=16 mg/ml? 1=Yes,0=No","Is the participant able to use the spirotel e-diary/PEF meter correctly, as evidenced by achieving a score of 13 on the Spirotel Performance Checklist ? 1=Yes, 0=No, 9=N/A","Has the participant used the Twisthaler previously? 1=Yes,0=No","Is the participant able to use a Twisthaler properly, as evidenced by achieving a score of 12 on the Twisthaler Inhalation Technique Checklist ? 1=Yes,0=No","Is the participant able to use a metered dose inhaler properly, as evidenced by achieving a score of 11 on the MDI Inhalation Technique Checklist  ? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Does the participant have valid source documentation within the past 6 months for an overread AsthmaNet methacholine challenge  with a PC20 &lt;=16 mg/ml?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Since Visit 1, has the participant experienced a treatment failure event as defined in the protocol? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Since Visit 1, has the participant experienced one or more asthma exacerbations as defined in the protocol? 1=Yes,0=No","Since Visit 1, has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant provide two acceptable sputum induction samples during the run-in? 1=Yes,0=No","Since Visit 1, has the participant experienced a treatment failure event as defined in the protocol? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Did the participant complete Asmanex220 mcg treatment at least 6 weeks ago? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Since Visit 1, has the participant experienced one or more asthma exacerbations as defined in the protocol? 1=Yes,0=No","Since Visit 1, has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Has the participant experienced a respiratory tract infection in the past 6 weeks? 1=Yes,0=No","Has the participant experienced a respiratory tract infection in the past 4 weeks? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report daytime asthma symptoms daily during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report night-time awakenings due to asthma symptoms more than once per week during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report night-time awakenings due to asthma symptoms more than twice per week during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report daily albuterol use for asthma symptom control during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant record at least 75Pct of AM and PM peak flow measurements and symptoms on his or her spirotel during the run-in? 1=Yes,0=No","Did the participant take at least 75Pct of the required puffs from his or her Respimat during the run-in? 1=Yes,0=No","Was the participants prebronchodilator FEV1 &gt;=70Pct predicted? 1=Yes,0=No","Has the participant used the Respimat previously? 1=Yes,0=No","Is the participant able to use a Respimat inhaler properly, as evidenced by achieving a score of 15 on the Respimat Inhalation Technique Checklist? 1=Yes,0=No","Does the participant wish to withdraw consent from the study? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has QC procedure been performed on the NIOX MINO today? 1=Yes,0=No","Did the participant eat or drink within the past hour? 1=Yes,0=No","Did the participant take part in strenuous activity/exercise within the past hour? 1=Yes,0=No","Time eNO started (based on a 24-hour clock) (0000-2359)","ENO Measurement","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","How long have you lived in the current house? (Estimate if uncertain.) (1010) years","How long have you lived in the current house? (Estimate if uncertain.)(1020) months","Does your house use a wood burning stove as a primary source of heat? 1=Yes, 0=No, 8=Don't Know","Does your house use an air conditioner? 1=Yes, 0=No, 8=Don't Know","Does your house use an evaporative cooler ? 1=Yes, 0=No, 8=Don't Know","Does your house use a humidifier? 1=Yes, 0=No, 8=Don't Know","Does your house use a dehumidifier? 1=Yes, 0=No, 8=Don't Know","Has there been water damage to your house, basement, or its contents during the past 12 months? 1=Yes, 0=No, 8=Don't Know","Has there been any mold or mildew, on any surfaces, inside your house in the past 12 months? 1=Yes, 0=No, 8=Don't Know","Which rooms have or have had mold or mildew?10a.Bathroom(s)  (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew?10b.Basement or attic (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10c.Kitchen (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10d.Your bedroom (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10e.Other bedrooms (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10f.Living or family room (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10g.Other (1 = Yes, 0 = No)","Do you ever see cockroaches in your house? 1=Yes,0=No","In which room(s) have you seen cockroaches?12a.Kitchen (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches?  12b.Basement or attic (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12c.Bathroom(s) (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12d.Living or family room (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12e.Your bedroom  (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12f.Other bedrooms (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12g.Garage (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12h.Other (1 = Yes, 0 = No)","Do you ever see rodents  or rodent droppings in your house? 1=Yes,0=No","14.In which room(s) have you seen rodents or rodent droppings? 14a. Kitchen (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14b. Basement or attic (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14c. Bathroom(s) (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings?14d. Living or family room (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14e. Your bedroom(1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14f. Other bedrooms (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14g. Garage (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14h. Other (1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15a. Barns(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property?15b. Hay(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15c. Woodsheds(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15d. Firewood(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property?15e. Chicken coops (1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15f. Corral (1 = Yes, 0 = No)","CHARACTERISTICS OF THE PARTICIPANT'S BEDROOM16.What is the floor covering in your bedroom? 1=Rug/carpet, 2=Vinyl tile or linoleum, 3=Wood, 4=Ceramic tile, 5=Other , 9=Don't know","What type of mattress is on your bed? 1=None, 2=Inner spring mattress, 3=Foam mattress, 4=Waterbed, 5=Air mattress, 6=Other , 9=Don't know","Is the mattress completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","Does your bed have a box spring? 1=Yes,0=No","Is the box spring completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","What type of pillow do you usually sleep with? 1=None, 2=Feather/down, 3=Foam/Dacron/synthetic, 5=Other , 9=Don't know","Is the pillow completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","Does your household have any pets? 1=Yes,0=No","Enter the number of pets that the household has. (Enter '00' if none. If none to Q24a - Q24d, skip to the next question.)24a.Cat","Cat  (1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has. Dog","Dog(1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has 24c.Rabbit, guinea pig, hamster, gerbil, or mouse","Rabbit, guinea pig, hamster, gerbil, or mouse (1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has 24d.Bird","Bird (1=Indoor, 2=Outdoor, 3=Both)","In general, and on a regular basis, are you exposed to any of the following animals?  25a.Cat(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25b.Dog(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25c.Rabbit, guinea pig, hamster, gerbil, or mouse(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25d.Bird (1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25e.Farm animals (1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals?  25f.Other (1 = Yes, 0 = No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other (specify","Highest education of any household member (1=No High School diploma/GED, 2=High Schol grad, 3=Technical training, 4=Some college, no degree, 5=Assoc degree, 6=Bachelor degree, 7=Masters degree/MD/PhD/JD/PharmD, 9=Decline to answer, 10=Don't know)","Category best describes the combined annual income, before taxes, of all members of your household for the last year. (1=Less than $25,000, 2=$25,000 - $49,999, 3=$50,000 - $99,999, 4=$100,000 or more, 9=Decline to answer, 10=Don't know)","How many people  are supported by this income reported in Q3?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Total IgE","Cat (E1)","Dog (E5)","Mouse (E72)","Rat (E74)","Mold mix (MX1)","Cockroach, German (I6)","Grass mix (GX2)","Tree mix (TX4)","Tree mix (TX6)","Weed mix (WX1)","Weed (W3)","Mite (D2)","Mite (D1)","Number or allergen specific (Cat, Dog, Mouse, etc) that were either POS or greater than 0.35","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","CBC with differential cell count: Eosinophils (absolute count) (0-9999/uL)","CBC with differential cell count: WBC (0-999.9 K/uL)","CBC with differential cell count: Differential Lymphocytes (0-99.9Pct)","CBC with differential cell count: Differential Monocytes (0-99.9Pct)","CBC with differential cell count: Differential Basophils (0-99.9Pct)","CBC with differential cell count: Differential Neutrophils (0-99.9Pct)","CBC with differential cell count: Differential Eosinophils (0-99.9Pct)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","PARENTAL HEIGHT - First study visit only or until both are completed 1.Biological mother's height (complete height or check unknown) __ feet","PARENTAL HEIGHT - First study visit only or until both are completed1.Biological mother's height (complete height or check unknown) __ __ inches","PARENTAL HEIGHT - First study visit only or until both are completed 1.Biological mother's height (complete height or check unknown) (9=Don't Know)","Biological father's height (complete height or check unknown) __ feet","Biological father's height (complete height or check unknown) __ __ inches","Biological father's height (complete height or check unknown) (9=Don't Know)","PARTICIPANT MEASUREMENTS - Complete at all applicable study visits3. What type of height measurement was obtained? 1=Standing height, 2=Length","First measurement __ __ __ . __ cm","Second measurement __ __ __ . __ cm","Third measurement __ __ __ . __ cm","Average height or length measurement __ __ __ . __ cm","In your judgment, was the participant's height or length measurement acceptable? 1=Yes,0=No","Weight (shoes off, light clothing) __ __ __ . __ kg","ORAL CANDIDIASIS5.Does the participant have evidence of oral candidiasis? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Post Diluent FEV1","Did the participant drop20Pct at the diluent stage? 1=Yes,0=No","Last concentration of methacholine administered __ __ . __ __ __ __ mg/ml","FEV1 after last concentration of methacholine administered","Did the participant achieve a PC20? 1=Yes,0=No","PC20 __ __ . __ __ mg/ml","Time methacholine challenge ended (based on 24-hour clock)","FEV1","Time of FEV1 in Q7a (based on 24-hour clock)","Was the FEV1 from Q7athe methacholine reversal reference value  in the gray box above? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was an additional treatment used in the first hour? 1=Yes,0=No","Additional albuterol by MDI (1=Yes,0=No)","Number of additional puffs of albuterol administered (1=2, 2=4, 3=&gt;4)","Nebulized Beta-agonist(1=Yes,0=No)","Subcutaneous epinephrine(1=Yes,0=No)","Implementation of clinic emergency protocol or algorithm (1=Yes,0=No)","Other  (1=Yes,0=No)","Participant's FEV1 after additional treatment within first hour.2a.FEV1","Time of FEV1 in Q2a (based on 24-hour clock)","Was the FEV1 from Q2a &gt;the methacholine reversal reference value  in the gray box on the Methacholine Challenge Testing  form? 1=Yes,0=No","Was additional treatment used after one hour? 1=Yes,0=No","Additional albuterol by MDI (1=Yes,0=No)","Number of additional puffs of albuterol administered (1=2, 2=4, 3=&gt;4)","Nebulized Beta-agonist  (1=Yes,0=No)","Subcutaneous epinephrine(1=Yes,0=No)","Implementation of clinic emergency protocol or algorithm  (1=Yes,0=No)","Treatment in the emergency room  (1=Yes,0=No)","Overnight hospitalization (1=Yes,0=No)","Other (1=Yes,0=No)","Participant's final FEV1 after methacholine challenge4a.FEV1","Time of FEV1 in Q4a (based on 24-hour clock)","Was the FEV1 from Q4a ? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Exclusions and Confounders1.Has the participant had any severe acute illness in the past 4 weeks? 1=Yes,0=No","If YES, has the participant received permission from the supervising physician to proceed with the methacholine challenge testing? 1=Yes,0=No","Has the participant used an oral or injectable steroid  in the last 4 weeks? 1=Yes,0=No","Does the participant have a baseline  FEV1 less than 55Pct of predicted or less than 1.0 L? 1=Yes,0=No","Pregnancy test results:  (Check N/A if the participant is male, or is female and is post-menopausal, had a hysterectomy or tubal ligation.) (1=Positive, 0=Negative, 9=N/A)","Is the participant's systolic blood pressure &gt; 200 mm Hg or diastolic blood pressure &gt; 100 mm Hg? 1=Yes,0=No","Is there any other reason the participant should not proceed with the methacholine challenge testing? 1=Yes,0=No","Is the participant eligible to proceed with the diluent  pulmonary function testing for the methacholine challenge? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Exclusions and Confounders:  1.Has the participant had any severe acute illness in the past 4 weeks? 1=Yes,0=No","If YES, has the participant received permission from the supervising physician to proceed with the methacholine challenge testing? 1=Yes,0=No","During the past 4 weeks, has the participant had any respiratory infections, colds, or bronchitis ? 1=Yes,0=No","If YES, during the past 2 weeks, has the participant had any respiratory infections, colds, or bronchitis ? 1=Yes,0=No","Has the participant used an oral or injectable steroid  in the last 4 weeks? 1=Yes,0=No","Does the participant have a baseline  FEV1 less than 70Pct of predicted? 1=Yes,0=No","Pregnancy test results(Check N/A if the participant is male, or is female and has not started menses.)(1=Positive, 0=Negative, 9=N/A)","If participant's age is ? 1=Yes,0=No","If participant's age is &lt; 12 years: Is the participant's systolic blood pressure &gt; 180 mm Hg or diastolic blood pressure &gt; 90 mm Hg? 1=Yes,0=No","Is there any other reason the participant should not proceed with the methacholine challenge testing? 1=Yes,0=No","Is the participant eligible to proceed with the diluent  pulmonary function testing for the methacholine challenge? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time albuterol administered (based on 24-hour clock)","Time post-albuterol spirometry started (based on 24-hour clock)","Highest FVC","Highest FEV1","Highest FEV1 (Pct predicted)","FEF Max","FEF25-75","In your judgment, was the subject's spirometry technique acceptable? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Please check the box next to the treatment that you believe you received over the past 12 weeks. 1=mometasone, 2=placebo","Blinded Scheduled Twisthaler Contents: How sure are you about your answer to Question 2? 1=Absolutely sure -  I know what the MDI contains, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","Please comment with respect to any observations you made that helped you make your choice in Question 2 . 1=I have no comments, 2=I noticed the following:","Please check the box next to the treatment that you believe you received over the past 12 weeks. 1=tiotropium, 2=placebo","How sure are you about your answer to Question 5? 1=Absolutely sure - I know what the Respimat capsules contain, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","1=I have no comments, 2=I noticed the following:","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Visit Number (numeric)","Level of periostin in peripheral blood (ng/ml)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time ipratropium administered (based on 24-hour clock) (0000-2359)","Time post-ipratropium spirometry started (based on 24-hour clock) (0000-2359)","Highest FVC (0-9.99 L)","Highest FEV1 (0-9.99 L)","Highest FEV1 (Pct predicted) (0-999 Pct predicted)","FEF Max (0-99.99 L/S)","FEF25-75 (0-9.99 L/S)","In your judgment, was the participants spirometry technique acceptable? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Is the participant unable to bear children due to any of the following reasons? 1=Yes,0=No","Post-menopausal (at least one year since last menses) (1=Yes,0=No)","Hysterectomy(1=Yes,0=No)","Tubal ligation(1=Yes,0=No)","Pregnancy test results  (1=Positive, 0=Negative)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Have you had any diseases, illnesses, conditions, or surgeries related Blood, Lymph, or Immune Systems? 1=Yes,0=No","Eyes(1=Yes,0=No)","Breasts (1=Yes,0=No)","Endocrine Systems (1=Yes,0=No)","Heart and Blood Vessels  (1=Yes,0=No)","Liver or Pancreas (1=Yes,0=No)","Kidneys or Urinary Tract System  (1=Yes,0=No)","Reproductive System (1=Yes,0=No)","Muscles or Bones (1=Yes,0=No)","Nervous System (1=Yes,0=No)","Psychiatric(1=Yes,0=No)","Drug Allergies (1=Yes,0=No)","Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Skin (1=Yes,0=No)","Have you ever had allergic rhinitis ? 1=Yes, 0=No, 9=Don't know","Have you ever had nasal polyps? 1=Yes, 0=No, 9=Don't know","Do you have chronic or recurrent sinusitis ? 1=Yes, 0=No, 9=Don't know","Have you ever been diagnosed with vocal cord dysfunction? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the ear, nose, or throat? 1=Yes,0=No","Have you ever had pneumonia? 1=Yes, 0=No, 9=Don't know","If YES, were you diagnosed by chest x-ray? 1=Yes, 0=No, 9=Don't know","If YES, were you treated with antibiotics? 1=Yes, 0=No, 9=Don't know","Have you ever had bronchitis? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the lungs ? 1=Yes, 0=No, 9=Don't know","Do you have gastroesophageal reflux disease ? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the stomach or intestines? 1=Yes,0=No","Have you been diagnosed with sleep disordered breathing ? 1=Yes,0=No","If YES, are you being treated with CPAP or BiPAP? 1=Yes,0=No","Have you ever had other sleep disorders? 1=Yes,0=No","Have you ever had other conditions that have not been mentioned on this form? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Short-acting Inhaled Beta-Agonists by Inhaler (e.g., albuterol, Primatene Mist, Maxair, ProAir, Proventil, Ventolin, Xopenex)  (1=Yes, 0=No, 9=Don't Know)","If YES, indicate average weekly puffs in the past month (Enter '000' if none used) weekly puffs","Rescue treatment via a Nebulizer Machine (e.g., albuterol, ipratropium, Combivent, Xopenex, levalbuterol)  (1=Yes, 0=No, 9=Don't Know)","Long-acting Inhaled Beta-Agonists (e.g., Serevent, Foradil, salmeterol, formoterol) Do not consider combination medications. (1=Yes, 0=No, 9=Don't Know)","Oral Beta-Agonists (e.g., albuterol, Brethine, Bricanyl, metaproterenol, Proventil, Ventolin, Repetabs, Volmax)  (1=Yes, 0=No, 9=Don't Know)","Oral Theophylline (short-acting or sustained release)(e.g., Aminophylline, Slo-Phyllin, Slo-bid, Theo-Dur, Uniphyl)(1=Yes, 0=No, 9=Don't Know)","Inhaled Anticholinergic by Inhaler (e.g., Atrovent, Combivent, Spiriva) (1=Yes, 0=No, 9=Don't Know)","Leukotriene Antagonist / 5LO Inhibitors (e.g., Accolate, Zyflo, Singulair) (1=Yes, 0=No, 9=Don't Know)","IgE Blocker (e.g., Xolair)(1=Yes, 0=No, 9=Don't Know)","Oral Steroids FOR ASTHMA (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron,  dexamethasone)(1=Yes, 0=No, 9=Don't Know)","If YES, in the past 12 months, how many courses of steroids by mouth have you taken FOR ASTHMA? 1=1 course, 2=2 courses, 3=3 courses, 4=4 courses, 5=5 courses, 6=More than 5 courses","Injectable Steroids FOR ASTHMA (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV)(1=Yes, 0=No, 9=Don't Know)","Steroids by Inhaler (e.g., Asmanex Twisthaler, QVAR, Flovent, Pulmicort Flexhaler)(1=Yes, 0=No, 9=Don't Know)","Indicate most recent type of inhaled steroid taken (refer to PRIOR_TRT_CARD reference card) code","Indicate number of daily puffs used daily puffs","Indicate the total number of months that you used the inhaled steroid out of the past 12 months months","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide) (1=Yes, 0=No, 9=Don't Know)","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide). If YES, complete Q13a - Q13c. Indicate most recent type of nebulized steroid taken (refer to PRIOR_TRT_CARD reference card)","Indicate number of daily treatments used daily treatments","Indicate the total number of months that you used the nebulized steroid out of the past 12 months months","Long-Acting Beta-Agonist and Inhaled Steroid Combination Medications (e.g., Advair Diskus, Symbicort MDI, Dulera MDI)(1=Yes, 0=No, 9=Don't Know)","Indicate most recent type of combination medication taken (refer to PRIOR_TRT_CARD reference card) code","Indicate number of daily puffs used daily puffs","Indicate the total number of months that you used the combination medication out of the past 12 months months","Nasal Steroids (e.g., Beconase, Vancenase, Flonase, Nasacort, Nasalide, Nasarel, Omnaris, Rhinocort, Nasonex)(1=Yes, 0=No, 9=Don't Know)","Non-steroidal Anti-allergic Nasal Medications (e.g., Nasalcrom, Astelin, Astepro, ipratropium)(1=Yes, 0=No, 9=Don't Know)","Anti-allergic Oral Medications (e.g., fexofenadine, loratadine, cetirizine, chlorpheniramine) (1=Yes, 0=No, 9=Don't Know)","Topical Steroids - Prescription (e.g., Synalar, Lidex, Dermacin, Fluocinonide) (1=Yes, 0=No, 9=Don't Know)","Topical Steroids - OTC (e.g., Hydrocortisone - multiple strengths and products) (1=Yes, 0=No, 9=Don't Know)","Other Medication FOR ASTHMA OR ALLERGIES (1=Yes, 0=No, 9=Don't Know)","Oral Steroids for Conditions Other Than Asthma (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) (1=Yes, 0=No, 9=Don't Know)","Injectable Steroids for Conditions Other Than Asthma (e.g., Medrol, Solumedrol, Decadron, dexamethasone,  triamcinolone, Kenalog, hydrocortisone IV)(1=Yes, 0=No, 9=Don't Know)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Short-acting Inhaled Beta-Agonists by Inhaler (e.g., albuterol, Primatene Mist, Maxair, ProAir, Proventil, Ventolin, Xopenex) Date","Rescue treatment via a Nebulizer Machine (e.g., albuterol, ipratropium, Combivent, Xopenex, levalbuterol) Date","Long-acting Inhaled Beta-Agonists (e.g., Serevent, Foradil, salmeterol, formoterol) Do not consider combination medications. Date","Oral Beta-Agonists (e.g., albuterol, Brethine, Bricanyl, metaproterenol, Proventil, Ventolin, Repetabs, Volmax) Date","Inhaled Anticholinergic by Inhaler (e.g., Atrovent, Combivent, Spiriva) Date","Leukotriene Antagonist / 5LO Inhibitors (e.g., Accolate, Zyflo, Singulair) Date","IgE Blocker (e.g., Xolair) Date","Oral Steroids FOR ASTHMA (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) Date","Injectable Steroids FOR ASTHMA (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV) Date","Steroids by Inhaler (e.g., Asmanex Twisthaler, QVAR, Flovent, Pulmicort Flexhaler) Date","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide) Date","Long-Acting Beta-Agonist and Inhaled Steroid Combination Medications (e.g., Advair Diskus, Symbicort MDI, Dulera MDI) Date","Nasal Steroids (e.g., Beconase, Vancenase, Flonase, Nasacort, Nasalide, Nasarel, Omnaris, Rhinocort, Nasonex) Date","Non-steroidal Anti-allergic Nasal Medications (e.g., Nasalcrom, Astelin, Astepro, ipratropium) Date","Anti-allergic Oral Medications (e.g., fexofenadine, loratadine, cetirizine, chlorpheniramine) Date","Topical Steroids - Prescription (e.g., Synalar, Lidex, Dermacin, Fluocinonide) Date","Topical Steroids - OTC (e.g., Hydrocortisone - multiple strengths and products) Date","Other Medication FOR ASTHMA OR ALLERGIES Date","Oral Steroids for Conditions Other Than Asthma (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) Date","Injectable Steroids for Conditions Other Than Asthma (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV) Date","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the last month, how often have you been upset because of something that happened unexpectedly? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that you were unable to control the important things in your life? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt nervous and stressed? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt confident about being able to handle your personal problems? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that things were going your way? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you found that you could not cope with all the things that you had to do? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you been able to control irritations in your life? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that you were on top of things? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you been angered because of things that happened that were outside of your control? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that your difficulties were piling up so high that you could not overcome them? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Have you consumed caffeine in the past 4 hours? 1=Yes,0=No","Have you used medications with caffeine in the past 4 hours? 1=Yes,0=No","Have you used any weight loss medications in the past 4 hours? 1=Yes,0=No","Have you consumed any food containing alcohol or beverages containing alcohol in the past 4 hours? 1=Yes,0=No","Have you used a rescue intermediate-acting inhaled betaagonist in the past 6 hours? 1=Yes,0=No","Have you used a short-acting anticholinergic in the past 6 hours? 1=Yes,0=No","Have you used any ophthalmic antihistamines in the past 6 hours? 1=Yes,0=No","Have you used any nasal antihistamines in the past 6 hours? 1=Yes,0=No","Have you used any nasal decongestants in the past 6 hours? 1=Yes,0=No","Have you used any oral antihistamines in the past 48 hours? 1=Yes,0=No","Have you used any oral decongestants or cold remedies in the past 48 hours? 1=Yes,0=No","Have you used any smokeless tobacco products today? 1=Yes,0=No","At this time, is your asthma worse because of recent exposure to triggers? 1=Yes,0=No","Is there any other reason you should not proceed with spirometry testing? 1=Yes,0=No","Is the participant eligible to proceed with the spirometry testing? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past 4 weeks, I worried about the long-term effects of asthma on my health (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I had to worry about asthma triggers (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, my asthma was on my mind (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, it was hard to get a good nights sleep because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt like I couldnt enjoy life because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt that asthma was controlling my life (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt frustrated that I couldnt make plans in advance because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, everyday activities were a struggle (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, asthma placed stress on my relationships with family, friends, significant others, or co-workers (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, I felt frustrated that I have to do things differently than people who dont have asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt like I missed out on doing things with others because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, I had to do a lot of planning to make sure I always had an inhaler ready (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Randomized treatment assignment period 1: (Mometasone, Tiotropium or Placebo)","Randomized treatment assignment period 2: (Mometasone, Tiotropium or Placebo)","Randomized treatment assignment period 3: (Mometasone, Tiotropium or Placebo)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Sex (1=Male, 2=Female)","Primary Racial Identification for Spirometry (1=Amer Indian or Alaskan Native, 2=Asian or Pacific Islander, 3=Black, 4=White, 5=Hispanic or Latino, 6=Other)","Participant's Age at Visit 1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Description of Adverse Event (ICD9 Code)","Is the participant currently taking study drug? 1=Yes,0=No","Time interval between the last administration of the study drug and the Adverse Event","What was the unit of time for the interval in Question #4? 1=Second, 2=Minute, 3=Hour, 4=Day","Fatal event(1=Yes,0=No)","Life-threatening event (1=Yes,0=No)","Inpatient hospitalization required(1=Yes,0=No)","Hospitalization prolonged  (1=Yes,0=No)","Disabling or incapacitating  (1=Yes,0=No)","Overdose (1=Yes,0=No)","Cancer (1=Yes,0=No)","Congenital anomaly(1=Yes,0=No)","Serious laboratory abnormality with clinical symptoms(1=Yes,0=No)","Height failure (per protocol MOP)(1=Yes,0=No)","Pregnancy(1=Yes,0=No)","Other (1=Yes,0=No)","What in your opinion caused the event? 1=Yes,0=No","Withdrawal of study drug(s)(1=Yes,0=No)","Concurrent medication  (1=Yes,0=No)","Other condition or event  (1=Yes,0=No)","Was the event expected or unexpected? 1=Expected, 2=Unexpected","Was the event possibly, probably, or definitely related to study participation? 1=Yes,0=No","Date of Adverse Event","Admission date","Discharge date","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","What type of height measurement was obtained? 1=Standing height, 2=Length","First measurement __ __ __ . __ cm","Second measurement __ __ __ . __ cm","Third measurement __ __ __ . __ cm","Average height or length measurement __ __ __ . __ cm","In your judgment, was the participant's height or length measurement acceptable? 1=Yes,0=No","Weight (shoes off, light clothing) __ __ __ . __ kg","Does the participant have evidence of oral candidiasis? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant fail to respond within 48 hours to the treatment failure rescue algorithm? 1=Yes,0=No","Did the participant use at least 16 puffs as needed albuterol per 24 hours for a period of 48 hours? 1=Yes,0=No","Did the participant experience prebronchodilator FEV1 values &lt; 50Pct of the baseline prebronchodilator value obtained at Visit 1 on two consecutive spirometric determinations made on different days? 1=Yes, 0=No, 9=Not evaluated","Did the participant experience prebronchodilator FEV1 values &lt; 40Pct of predicted on two consecutive spirometric determinations made on different days? 1=Yes, 0=No, 9=Not evaluated","Did the study or treating physician prescribe the participant oral/parenteral corticosteroids for the treatment of his/her asthma? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation in the opinion of the study investigator or personal physician? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation? 1=Yes,0=No","Did the participant seek care for significant asthma exacerbation conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Was the participant hospitalized? 1=Yes,0=No","Duration of hospital stay (0-99.9 days)","Was intubation or ventilation assistance required? 1=Yes,0=No","Was the participant admitted to the intensive care unit? 1=Yes,0=No","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Inhaled corticosteroids (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Nebulized bronchodilator (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Oral corticosteroids (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, IM or IV steroids (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Antibiotics (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Other (1=Yes, 0=No )","Why do you think the participant experienced a significant asthma exacerbation? 1=Common cold, 2=Allergies, 3=Pollution or chemical irritant, 4=Too little asthma maintenance medication, 5=Exercise, 6=Other","Date exacerbation conditions were met","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Which phone contact following Asthma Exacerbation visit is being performed? 1=Phone Contact #1 , 2=Phone Contact #2 , 3=Phone Contact #3","Since the most recent Asthma Exacerbation contact, has the participant sought additional care for significant asthma exacerbation conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Was the participant hospitalized? 1=Yes,0=No","Duration of hospital stay (0-99.9 days)","Was intubation or ventilation assistance required? 1=Yes,0=No","Was the participant admitted to the intensive care unit? 1=Yes,0=No","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Inhaled corticosteroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Nebulized bronchodilator (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Oral corticosteroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, IM or IV steroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Antibiotics (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Runny Nose  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Post nasal drip  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Need to blow your nose  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Facial pain/pressure(0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Nasal obstruction  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time spirometry started (based on 24-hour clock)","The reported FEV1 and FVC are the best volumes of all acceptable maneuvers. 1.Highest FVC","Highest FEV1 (Pct predicted)","The reported flow rates correspond to the maneuver where FEV1 + FVC is maximized. FEF Max","FEF25-75","In your judgment, was the subject's spirometry technique acceptable? 0=No, 1=Yes","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Highest FEV1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","TRIAL_TYPE","FVC","FEV1","PEF","FEF2575","FET","Number of times the participant woke up last night due to asthma symptoms # Times woke up due to asthma (0-9)","Number of puffs the participant will take from the WHITE study inhaler this morning # Puffs from WHITE inhaler in AM (0-9)","only) Number of puffs the participant will take from the BLUE study inhaler this morning # Puffs from BLUE inhaler in AM (0-9)","Has the participant taken any puffs from his/her RED RESCUE albuterol inhaler in the past 4 hours? 1=yes, 0=no","Shortness of Breath Score Shortness of Breath Score overnight (0,1,2,3)","Chest tightness score Chest Tightness Score overnight (0,1,2,3)","Wheezing score Wheezing Score overnight (0,1,2,3)","Coughing score Coughing Score overnight (0,1,2,3)","Phlegm/Mucus score Phlegm/Mucus Score overnight (0,1,2,3)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","date 'fixed' for analysis so that associated AM and PM always have same ddate","TRIAL_TYPE","FVC","FEV1","PEF","FEF2575","FET","Number of puffs the participant will take from the WHITE study inhaler tonight # Puffs from WHITE inhaler in PM (0-9)","Has the participant taken any puffs from his/her RED RESCUE albuterol inhaler during the past 4 hours? 1=yes, 0=no","Shortness of Breath Score Shortness of Breath Score since waking this morning (0,1,2,3)","Chest tightness score Chest Tightness Score since waking this morning (0,1,2,3)","Wheezing score Wheezing Score since waking this morning (0,1,2,3)","Coughing score Coughing Score since waking this morning (0,1,2,3)","Phlegm/Mucus score Phlegm/Mucus Score since waking this morning (0,1,2,3)","Did your regular exercise cause unusually severe shortness of breath during the past 24 hours? 1=Yes, 0=No, 9=N/A","Number of RED RESCUE albuterol puffs taken in the past 24 hours to prevent symptoms# Preventive  RED RESCUE puffs  in past 24 hours?","Number of RED RESCUE albuterol puffs taken for asthma symptoms or low peak flow during past 24 hours  # RED RESCUE  puffs for symptoms  in past 24 hours?","Number of times used RED RESCUE albuterol inhaler for asthma symptoms during past 24 hours  # Times used  RED RESCUE for symptoms in past 24 hours?","Did RED RESCUE albuterol relieve symptoms for less than 4 hours after treatment during past 24 hours? 1=Yes,0=No","Was the participant seen by a healthcare provider  for an unscheduled visit in the past 24 hours due to asthma symptoms? 1=Yes,0=No","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Processing: Time processing started (based on 24-hour clock) __ __ __ __ (0 - 2359)","Processing Sample: Total Cell Count: __ __ __ __. __x 10^4 cells/ml (0 - 9999.9)","Was the participant's sputum sample processed within 4 hours after collection? 1=Yes,0=No","Processing Sample: Processing Date __ __ / __ __ / 20 __ __","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Rate slide's quality:  (1=Very good, 2=Good, 3=Acceptable, 4=Poor but readable, 5=Not readable)","Record the number on the slide(s) that was (were) read (0 - 9)","Record the number on the slide(s) that was (were) read: These are numbers that were assigned to the slides at each site. (0 -9)","Total Cell Count: __ __ __ __ . __ x 10^4 cells/ml (0 - 9999.9)","Differential Cell Counts: Squamous Cells __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Epithelial Cells __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Macrophages __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Neutrophils __ __ . __ Pct (0 -99.9)","Differential Cell Counts: Eosinophils __ __ . __ Pct (0 - 99.9)","Differential Cell Counts: Lymphocytes __ __ . __ Pct (0 -99.9)","Date of Read __ __ / __ __ / 20 __ __","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","For this protocol, what was the duration of sputum induction the first time the participant's sample was processed within 4 hours after collection?","Sputum induction start time (based on 24-hour clock)","Sputum induction stop time (based on 24-hour clock)","Duration of sputum induction collection phase at this visit -- minutes (0 - 99.9)","Was the duration ? 1=Yes,0=No","Volume of sputum sample at this visit -- ml (0.0 - 999.9)","Is the volume adequate for processing? 1=Yes,0=No","Is the sample adequate for laboratory analysis? 1=Yes,0=No","Participant's FEV1 immediately after completion of sputum induction: -- (0.0 -9.99 L)","FEV1 (Pct predicted) (0 - 999)","Time of FEV1 in Q7a (based on 24-hour clock)","Percent difference in FEV1 ((Reference - Q7a) / Reference) X100 (-99.9 - 99.9 Pct)","Did the participant's FEV1 drop &gt; 10Pct from reference FEV1 as indicated in Q7d? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Participant's FEV1 after initial 2 puffs of albuterol: 1a.FEV1 __ . __ __ L (0 -9.99)","Participant's FEV1 after initial 2 puffs of albuterol: 1b. FEV1 (Pct predicted)__ __ __ Pct predicted (0 - 999)","Participant's FEV1 after initial 2 puffs of albuterol: 1c. Time of FEV1 from Q1a (based on 24-hour clock) __ __ __ __ (0 - 2359)","Participant's FEV1 after initial 2 puffs of albuterol: 1d. Was the FEV1 from Q1a &gt;=the sputum induction reversal reference value in the gray box above? 1=Yes,0=No","Participant's FEV1 after 2 additional puffs of albuterol: 2a. FEV1 __ . __ __ L (0 - 9.99)","Participant's FEV1 after 2 additional puffs of albuterol: 2b. FEV1 (Pct predicted) __ __ __ Pct predicted (0 - 999)","Participant's FEV1 after 2 additional puffs of albuterol: 2c. Time of FEV1 from Q2a (based on 24-hour clock)__ __ __ __ (0 - 2359)","Participant's FEV1 after 2 additional puffs of albuterol: 2d.Was the FEV1 from Q2a ? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was the participant's FEV1 after reversal from the methacholine challenge greater yhan 90Pct of the baseline FEV1 (1=Yes,0=No)","If NO, has the participant received permission from the supervising physician to proceed with sputum induction testing? 1=Yes,0=No","Physician's Signature:_________ (1=signature present, NULL=signature missing)","Participant's FEV1 used for assessment of eligibility for sputum induction __ . __ __ L (0 - 9.99)","Participant's FEV1 (Pct predicted) used for assessment of eligibility for sputum induction __ __ __ Pct predicted (O-999)","Was the participant's FEV1  from Q3 ? 1=Yes,0=No","Has the participant used any smokeless tobacco products  today? 1=Yes,0=No","Is there any other reason the participant should not proceed with sputum induction? 1=Yes,0=No","Is the participant eligible for sputum induction? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has the participant completed the study through Visit 9? 1=Yes,0=No","Who initiated termination of the participant? 1=Participant, 2=Clinical Staff","1=no longer interested in participating*, 2=no longer willing to follow protocol*, 3=difficult access to clinic , 4=unable to make visits during clinic hours, 5=moving out of the area, 6=unable to continue due to personal constraints*, 7=unable to continue","Did clinical staff terminate the participant due to pregnancy? 1=Yes, 0=No, 9=N/A","Did clinical staff terminate the participant due to loss to follow-up? 1=Yes,0=No","Did clinical staff terminate the participant due to loss to follow-up, type of contact (1=In-person visit, 2=Phone call)","Did clinical staff terminate the participant due to an asthma-related adverse event? 1=Yes,0=No","Did clinical staff terminate the participant due to a medication-related adverse event? 1=Yes,0=No","Did clinical staff terminate the participant due to an adverse event not related to asthma or medications? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility during the Supervised Washout ? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with medication dosing? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with diary completion? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with visit attendance? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with peak flow monitoring? 1=Yes,0=No","Did clinical staff terminate the participant due to significant asthma exacerbation or two treatment failures during run-in ? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility during the run-in period  for reasons other than compliance or exacerbation/treatment failure or high sputum eosinophils ? 1=Yes,0=No","Did clinical staff terminate the participant due to other reason? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility due to high sputum eosinophils in the run-in ? 1=Yes,0=No","Did clinical staff terminate the participant due to... Indicate the letter corresponding to the primary reason the participant was terminated. (a-n)","Did clinical staff terminate the participant due to loss to follow-up, date of last contact with participant","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant seek care for treatment failure conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact, other than call to notify clinic as instructed in spirotel","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Inhaled corticosteroids (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Nebulized bronchodilator (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started: Oral corticosteroids form.  (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  IM or IV steroids form.  (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started: Antibiotics (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant awaken from asthma three or more times in a two-week period or on two consecutive nights? 1=Yes,0=No","Did the participant use albuterol for relief of symptoms four or more times/day for two or more consecutive days? 1=Yes,0=No","Did albuterol relieve symptoms for less than four hours after treatment? 1=Yes,0=No","Did the participant use albuterol for relief of symptoms daily for seven days, and this use exceeded two times the weekly use of albuterol in the baseline period? 1=Yes,0=No","Did regular exercise cause unusually severe shortness of breath on two or more days during a seven day period? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation? 1=Yes,0=No","Did the participant experience a treatment failure? 1=Yes,0=No","Has the participant experienced two treatment failures in the current treatment period ? 1=Yes,0=No","Does the study physician feel the study treatment contributed directly to the treatment failure, such that it is in the participants best interest to discontinue the current blinded treatment? 1=Yes,0=No","Did the participant begin daily treatment with high-dose ICS for 10 days? 1=Yes,0=No","Did the participant begin daily treatment with high-dose ICS for 10 days? 1=Did not think symptoms werebad enough, 2=Other","Date treatment failure conditions were met","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","During this last treatment period , has the participant experienced a treatment failure or asthma exacerbation? 1=Yes,0=No","Has the participant experienced two or more treatment failures in this last treatment period, or does the study physician feel the study treatment contributed directly to the treatment failure such that it is in the participants best interest to discontinu","As of today, have 2 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","As of today, have 6 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","As of today, have 3 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Are you currently employed ? 1=Yes,0=No","In general, how many hours per week do you usually work?","During the past seven days, how many hours did you miss from work because of problems associated with your asthma?","During the past seven days, how much did asthma affect your productivity while you were working?","Do you currently attend classes in an academic setting ? 1=Yes,0=No","In general, how many hours per week do you usually attend classes?","During the past seven days, how many hours did you miss from class or school because of problems associated with your asthma?","During the past seven days, how much did asthma affect your productivity while in school or attending classes in an academic setting?","During the past seven days, how much did your asthma affect your ability to do your regular daily activities, other than work at a job or attend classes?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","How sick today? 0=Not Sick, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely","Runny Nose:(0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","PLUGGED_NOSE  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SNEEZING  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SORE_THROAT  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SCRATCHY_THROAT  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","COUGH (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","HOARSENESS  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","HEAD_CONGESTION  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","CHEST_CONGESTION  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","FEELING_TIRED (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","THINK_CLEARLY  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SLEEP_WELL  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","BREATH_EASILY  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WALK  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","DAILY_ACTIVITIES  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WORK_OUTSIDE_HOME  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WORK_INSIDE_HOME  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","INTERACT  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","LIVE_LIFE  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","MY_COLD_IS  (1=Very much better, 2=Somewhat better, 3=A little better, 4=The same, 5=A little worse, 6=Somewhat worse, 7=Very much worse)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Date (number of days from Visit 1)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>data_set_name<\/th>\n      <th>variable_number_within_dataset<\/th>\n      <th>variable<\/th>\n      <th>type<\/th>\n      <th>variable_description<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":2},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"data_set_name","targets":1},{"name":"variable_number_within_dataset","targets":2},{"name":"variable","targets":3},{"name":"type","targets":4},{"name":"variable_description","targets":5}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-56f007701c6be38cbe7d" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-56f007701c6be38cbe7d">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80","81","82","83","84","85","86","87","88","89","90","91","92","93","94","95","96","97","98","99","100","101","102","103","104","105","106","107","108","109","110","111","112","113","114","115","116","117","118","119","120","121","122","123","124","125","126","127","128","129","130","131","132","133","134","135","136","137","138","139","140","141","142","143","144","145","146","147","148","149","150","151","152","153","154","155","156","157","158","159","160","161","162","163","164","165","166","167","168","169","170","171","172","173","174","175","176","177","178","179","180","181","182","183","184","185","186","187","188","189","190","191","192","193","194","195","196","197","198","199","200","201","202","203","204","205","206","207","208","209","210","211","212","213","214","215","216","217","218","219","220","221","222","223","224","225","226","227","228","229","230","231","232","233","234","235","236","237","238","239","240","241","242","243","244","245","246","247","248","249","250","251","252","253","254","255","256","257","258","259","260","261","262","263","264","265","266","267","268","269","270","271","272","273","274","275","276","277","278","279","280","281","282","283","284","285","286","287","288","289","290","291","292","293","294","295","296","297","298","299","300","301","302","303","304","305","306","307","308","309","310","311","312","313","314","315","316","317","318","319","320","321","322","323","324","325","326","327","328","329","330","331","332","333","334","335","336","337","338","339","340","341","342","343","344","345","346","347","348","349","350","351","352","353","354","355","356","357","358","359","360","361","362","363","364","365","366","367","368","369","370","371","372","373","374","375","376","377","378","379","380","381","382","383","384","385","386","387","388","389","390","391","392","393","394","395","396","397","398","399","400","401","402","403","404","405","406","407","408","409","410","411","412","413","414","415","416","417","418","419","420","421","422","423","424","425","426","427","428","429","430","431","432","433","434","435","436","437","438","439","440","441","442","443","444","445","446","447","448","449","450","451","452","453","454","455","456","457","458","459","460","461","462","463","464","465","466","467","468","469","470","471","472","473","474","475","476","477","478","479","480","481","482","483","484","485","486","487","488","489","490","491","492","493","494","495","496","497","498","499","500","501","502","503","504","505","506","507","508","509","510","511","512","513","514","515","516","517","518","519","520","521","522","523","524","525","526","527","528","529","530","531","532","533","534","535","536","537","538","539","540","541","542","543","544","545","546","547","548","549","550","551","552","553","554","555","556","557","558","559","560","561","562","563","564","565","566","567","568","569","570","571","572","573","574","575","576","577","578","579","580","581","582","583","584","585","586","587","588","589","590","591","592","593","594","595","596","597","598","599","600","601","602","603","604","605","606","607","608","609","610","611","612","613","614","615","616","617","618","619","620","621","622","623","624","625","626","627","628","629","630","631","632","633","634","635","636","637","638","639","640","641","642","643","644","645","646","647","648","649","650","651","652","653","654","655","656","657","658","659","660","661","662","663","664","665","666","667","668","669","670","671","672","673","674","675","676","677","678","679","680","681","682","683","684","685","686","687","688","689","690","691","692","693","694","695","696","697","698","699","700","701","702","703","704","705","706","707","708","709","710","711","712","713","714","715","716","717","718","719","720","721","722","723","724","725","726","727","728","729","730","731","732","733","734","735","736","737","738","739","740","741","742","743","744","745","746","747","748","749","750","751","752","753","754","755","756","757","758","759","760","761","762","763","764","765","766","767","768","769","770","771","772","773","774","775","776","777","778","779","780","781","782","783","784","785","786","787","788","789","790","791","792","793","794","795","796","797","798","799","800","801","802","803","804","805","806","807","808","809","810","811","812","813","814","815","816","817","818","819","820","821","822","823","824","825","826","827","828","829","830","831","832","833","834","835","836","837","838","839","840","841","842","843","844","845","846","847","848","849","850","851","852","853","854","855","856","857","858","859","860","861","862","863","864","865","866","867","868","869","870","871","872","873","874","875","876","877","878","879","880","881","882","883","884","885","886","887","888","889","890","891","892","893","894","895","896","897","898","899","900","901","902","903","904","905","906","907","908","909","910","911","912","913","914","915","916","917","918","919","920","921","922","923","924","925","926","927","928","929","930","931","932","933","934","935","936","937","938","939","940","941","942","943","944","945","946","947","948","949","950","951","952","953","954","955","956","957","958","959","960","961","962","963","964","965","966","967","968","969","970","971","972","973","974","975","976","977","978","979","980","981","982","983","984","985","986","987","988","989","990","991","992","993","994","995","996","997","998","999","1000","1001","1002","1003","1004","1005","1006","1007","1008","1009","1010","1011","1012","1013","1014","1015","1016","1017","1018","1019","1020","1021","1022","1023","1024","1025","1026","1027","1028","1029","1030","1031","1032","1033","1034","1035","1036","1037","1038","1039","1040","1041","1042","1043","1044","1045","1046","1047","1048","1049","1050","1051","1052","1053","1054","1055","1056","1057","1058","1059","1060","1061","1062","1063","1064","1065","1066","1067","1068","1069","1070","1071","1072","1073","1074","1075","1076","1077","1078","1079","1080","1081","1082","1083","1084","1085","1086","1087","1088","1089","1090","1091","1092","1093","1094","1095","1096","1097","1098","1099","1100","1101","1102","1103","1104","1105","1106","1107","1108","1109","1110","1111","1112","1113","1114","1115","1116","1117","1118","1119","1120","1121","1122","1123","1124","1125","1126","1127","1128","1129","1130","1131","1132","1133","1134","1135","1136","1137","1138","1139","1140","1141","1142","1143","1144","1145","1146","1147","1148","1149","1150","1151","1152","1153","1154","1155","1156","1157","1158","1159","1160","1161","1162","1163","1164","1165","1166","1167","1168","1169","1170","1171","1172","1173","1174","1175","1176","1177","1178","1179","1180","1181","1182","1183","1184","1185","1186","1187","1188","1189","1190"],["aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","bmi.sas7bdat","bmi.sas7bdat","bmi.sas7bdat","bmi.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat"],[1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,1,2,3,4,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,1,2,3,4,5,6,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28],["AAQ_1000","AAQ_1010","AAQ_1020","AAQ_1030","AAQ_1040","AAQ_1050","AAQ_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ABP_1000","ABP_1010","ABP_1020","ABP_1030","ABP_1040","ABP_1050","ABP_1060","ABP_1070","ABP_1080","ABP_1090","ABP_1100","ABP_1110","ABP_1120","ABP_1130","ABP_1140","ABP_1150","ABP_1160","ABP_1170","ABP_1180","ABP_1190","ABP_1200","ABP_1210","ABP_1220","ABP_1230","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ACT_1","ACT_2","ACT_3","ACT_4","ACT_5","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ICD9_CAT","AEC_1000","AEC_1010","AEC_1040","AEC_1050","AEC_1060","AEC_1070","AEC_1080","AEC_1090","AEC_1100","AEC_1110","AEC_1120","AEC_1020","AEC_1030","VNUM_C","VNUM","ICD9LONG","ICD9SHRT","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","AHA_1000","AHA_1020","AHA_1030","AHA_1040","AHA_1050","AHA_1060","AHA_1070","AHA_1090","AHA_1100","AHA_1110","AHA_1120","AHA_1130","AHA_1140","AHA_1150","AHA_1160","AHA_1170","AHA_1180","AHA_1190","AHA_1200","AHA_1210","AHA_1220","AHA_1250","AHA_1260","AHA_1270","AHA_1280","AHA_1290","AHA_1300","AHA_1310","AHA_1320","AHA_1330","AHA_1340","AHA_1350","AHA_1360","AHA_1370","AHA_1380","AHA_1390","AHA_1400","AHA_1410","AHA_1420","AHA_1430","AHA_1440","AHA_1450","AHA_1460","AHA_1470","AHA_1500","AHA_1570","AHA_1580","AHA_1590","AHA_1600","AHA_1730","AHA_1740","AHA_1750","AHA_1760","AHA_1770","AHA_1780","AHA_1790","AHA_1800","AHA_1810","AHA_1820","AHA_1830","AHA_1840","AHA_1850","AHA_1890","AHA_1900","VNUM_C","VNUM","VDATE","DATE_VAPE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ASU_1000","ASU_1010","ASU_1020","ASU_1030","ASU_1040","ASU_1050","ASU_1060","ASU_1070","ASU_1080","ASU_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","BAS_1000","BAS_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","BMI","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAGE","CHG_1000","CHG_1010","CHG_1020","CHG_1030","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","CME_1040","CME_1050","CME_1055","CME_1000","CME_1010","CME_1020","CME_1030","CME_1080","CME_1090","CME_1060","CME_1070","VNUM_C","VNUM","GEN_NAME","CLASS","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","COM_1000","COM_1010","COM_1020","COM_1030","COM_1040","COM_1050","COM_1060","COM_1070","COM_1080","COM_1090","COM_1100","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","CTX_1000","CTX_1010","CTX_1030","CTX_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E0A_1000","E0A_1020","E0A_1030","E0A_1040","E0A_1043","E0A_1045","E0A_1050","E0A_1060","E0A_1070","E0A_1073","E0A_1075","E0A_1080","E0A_1090","E0A_1100","E0A_1110","E0A_1120","E0A_1130","E0A_1140","E0A_1145","E0A_1150","E0A_1160","E0A_1170","E0A_1180","E0A_1190","E0A_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E0B_1000","E0B_1010","E0B_1020","E0B_1030","E0B_1040","E0B_1050","E0B_1060","E0B_1065","E0B_1070","E0B_1080","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E1_1000","E1_1020","E1_1030","E1_1040","E1_1050","E1_1060","E1_1065","E1_1070","E1_1080","E1_1090","E1_1100","E1_1110","E1_1120","E1_1130","E1_1140","E1_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E2_1000","E2_1010","E2_1020","E2_1030","E2_1040","E2_1050","E2_1060","E2_1070","E2_1080","E2_1090","E2_1100","E2_1110","E2_1120","E2_1130","E2_1140","E2_1145","E2_1150","E2_1160","E2_1170","E2_1173","E2_1175","E2_1180","E2_1190","E2_1200","E2_1210","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E3_1000","E3_1010","E3_1020","E3_1030","E3_1070","E3_1080","E3_1090","E3_1100","E3_1110","E3_1120","E3_1130","E3_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E4_1000","E4_1010","E4_1020","E4_1030","E4_1050","E4_1060","E4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E5_1000","E5_1010","E5_1020","E5_1030","E5_1035","E5_1040","E5_1050","E5_1060","E5_1065","E5_1070","E5_1080","E5_1085","E5_1090","E5_1100","E5_1110","E5_1120","E5_1130","E5_1140","E5_1150","E5_1160","E5_1170","E5_1180","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ENO_1000","ENO_1010","ENO_1020","ENO_1040","ENO_1050","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","HEQ_1000","HEQ_1010","HEQ_1020","HEQ_1030","HEQ_1040","HEQ_1050","HEQ_1060","HEQ_1070","HEQ_1080","HEQ_1090","HEQ_1100","HEQ_1110","HEQ_1120","HEQ_1130","HEQ_1140","HEQ_1150","HEQ_1160","HEQ_1170","HEQ_1180","HEQ_1190","HEQ_1200","HEQ_1210","HEQ_1220","HEQ_1230","HEQ_1240","HEQ_1250","HEQ_1260","HEQ_1270","HEQ_1280","HEQ_1290","HEQ_1300","HEQ_1310","HEQ_1320","HEQ_1330","HEQ_1340","HEQ_1350","HEQ_1360","HEQ_1370","HEQ_1380","HEQ_1390","HEQ_1400","HEQ_1410","HEQ_1420","HEQ_1430","HEQ_1440","HEQ_1450","HEQ_1460","HEQ_1470","HEQ_1480","HEQ_1490","HEQ_1500","HEQ_1510","HEQ_1520","HEQ_1530","HEQ_1540","HEQ_1550","HEQ_1560","HEQ_1570","HEQ_1580","HEQ_1590","HEQ_1600","HEQ_1610","HEQ_1620","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SEI_1000","SEI_1010","SEI_1020","SEI_1030","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","IGE","RE1","RE5","RE72","RE74","RMX1","RI6","RGX2","RTX4","RTX6","RWX1","RW3","RD2","RD1","N_ALLERGENPOS","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","LAB_1000","LAB_1010","LAB_1020","LAB_1030","LAB_1040","LAB_1050","LAB_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PLE_1000","PLE_1010","PLE_1020","PLE_1030","PLE_1040","PLE_1050","PLE_1060","PLE_1070","PLE_1080","PLE_1090","PLE_1100","PLE_1110","PLE_1130","PLE_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MTH_1000","MTH_1010","MTH_1020","MTH_1030","MTH_1040","MTH_1050","MTH_1060","MTH_1070","MTH_1080","MTH_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MAD_1000","MAD_1010","MAD_1020","MAD_1030","MAD_1040","MAD_1050","MAD_1060","MAD_1070","MAD_1090","MAD_1100","MAD_1110","MAD_1120","MAD_1130","MAD_1140","MAD_1150","MAD_1160","MAD_1170","MAD_1180","MAD_1190","MAD_1200","MAD_1220","MAD_1230","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MCA_1000","MCA_1010","MCA_1050","MCA_1060","MCA_1070","MCA_1080","MCA_1100","MCA_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MCP_1000","MCP_1010","MCP_1030","MCP_1040","MCP_1050","MCP_1060","MCP_1070","MCP_1080","MCP_1090","MCP_1100","MCP_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PA4_1000","PA4_1010","PA4_1020","PA4_1030","PA4_1040","PA4_1050","PA4_1060","PA4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PTX_1000","PTX_1010","PTX_1020","PTX_1030","PTX_1040","PTX_1050","PTX_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","VNUM","PERIOSTIN","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PI4_1000","PI4_1010","PI4_1020","PI4_1030","PI4_1040","PI4_1050","PI4_1060","PI4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PRG_1000","PRG_1010","PRG_1020","PRG_1030","PRG_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAD_1000","PAD_1010","PAD_1020","PAD_1030","PAD_1040","PAD_1050","PAD_1060","PAD_1070","PAD_1080","PAD_1090","PAD_1100","PAD_1110","PAD_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAL_1000","PAL_1010","PAL_1020","PAL_1030","PAL_1040","PAL_1050","PAL_1060","PAL_1070","PAL_1080","PAL_1090","PAL_1100","PAL_1110","PAL_1120","PAL_1130","PAL_1150","PAL_1160","PAL_1170","PAL_1180","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PTR_1000","PTR_1010","PTR_1050","PTR_1060","PTR_1100","PTR_1140","PTR_1180","PTR_1220","PTR_1260","PTR_1300","PTR_1340","PTR_1380","PTR_1390","PTR_1430","PTR_1470","PTR_1480","PTR_1490","PTR_1500","PTR_1535","PTR_1540","PTR_1550","PTR_1560","PTR_1600","PTR_1610","PTR_1620","PTR_1630","PTR_1670","PTR_1710","PTR_1750","PTR_1790","PTR_1830","PTR_1870","PTR_1910","VNUM_C","VNUM","VDATE","DATE_Q02","DATE_Q03","DATE_Q04","DATE_Q05","DATE_Q07","DATE_Q08","DATE_Q09","DATE_Q10","DATE_Q11","DATE_Q12","DATE_Q13","DATE_Q14","DATE_Q15","DATE_Q16","DATE_Q17","DATE_Q18","DATE_Q19","DATE_Q20","DATE_Q21","DATE_Q22","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","P10_1000","P10_1010","P10_1020","P10_1030","P10_1040","P10_1050","P10_1060","P10_1070","P10_1080","P10_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PCH_1000","PCH_1010","PCH_1020","PCH_1030","PCH_1040","PCH_1050","PCH_1060","PCH_1070","PCH_1080","PCH_1090","PCH_1100","PCH_1110","PCH_1120","PCH_1130","PCH_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","IAQ_1000","IAQ_1010","IAQ_1020","IAQ_1030","IAQ_1040","IAQ_1050","IAQ_1060","IAQ_1070","IAQ_1080","IAQ_1090","IAQ_1100","IAQ_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","REGIMEN_PERIOD_1","REGIMEN_PERIOD_2","REGIMEN_PERIOD_3","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","REG_1080","REG_1150","AGE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SER_1010","SER_1020","SER_1030","SER_1040","SER_1050","SER_1060","SER_1070","SER_1100","SER_1110","SER_1120","SER_1130","SER_1140","SER_1150","SER_1160","SER_1170","SER_1180","SER_1190","SER_1200","SER_1210","SER_1220","SER_1240","SER_1250","SER_1000","SER_1080","SER_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PSE_1060","PSE_1070","PSE_1080","PSE_1090","PSE_1100","PSE_1110","PSE_1130","PSE_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SIG_1000","SIG_1010","SIG_1020","SIG_1030","SIG_1040","SIG_1050","SIG_1060","SIG_1080","SIG_1090","SIG_1100","SIG_1110","SIG_1120","SIG_1130","SIG_1140","SIG_1150","SIG_1160","SIG_1170","SIG_1180","SIG_1190","SIG_1200","SIG_1210","SIG_1220","SIG_1230","SIG_1240","SIG_1250","SIG_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SFU_1000","SFU_1010","SFU_1020","SFU_1030","SFU_1040","SFU_1050","SFU_1060","SFU_1070","SFU_1080","SFU_1090","SFU_1100","SFU_1110","SFU_1120","SFU_1130","SFU_1140","SFU_1150","SFU_1160","SFU_1170","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SNQ_1000","SNQ_1010","SNQ_1020","SNQ_1030","SNQ_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SPI_1010","SPI_1020","SPI_1040","SPI_1050","SPI_1060","SPI_1070","VNUM_C","VNUM","VDATE","SPI_1030","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","AM_TRIAL_TYPE","AM_FVC","AM_FEV1","AM_PEF","AM_FEF2575","AM_FET","DRY_1","DRY_2","DRY_3","DRY_4","DRY_5","DRY_6","DRY_7","DRY_8","DRY_9","VNUM_C","VNUM","VDATE","DDATE","PM_TRIAL_TYPE","PM_FVC","PM_FEV1","PM_PEF","PM_FEF2575","PM_FET","DRY_10","DRY_11","DRY_12","DRY_13","DRY_14","DRY_15","DRY_16","DRY_17","DRY_18","DRY_19","DRY_20","DRY_21","DRY_22","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SLA_1010","SLA_1020","SLA_1030","SLA_1000","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SRE_1010","SRE_1020","SRE_1030","SRE_1040","SRE_1050","SRE_1060","SRE_1070","SRE_1080","SRE_1090","SRE_1100","SRE_1000","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SPU_1000","SPU_1010","SPU_1020","SPU_1030","SPU_1040","SPU_1050","SPU_1060","SPU_1070","SPU_1080","SPU_1090","SPU_1100","SPU_1110","SPU_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SAD_1000","SAD_1010","SAD_1020","SAD_1030","SAD_1040","SAD_1050","SAD_1060","SAD_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SCH_1000","SCH_1010","SCH_1020","SCH_1030","SCH_1040","SCH_1050","SCH_1055","SCH_1060","SCH_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TER_1000","TER_1010","TER_1020","TER_1040","TER_1050","TER_1070","TER_1080","TER_1090","TER_1100","TER_1110","TER_1120","TER_1130","TER_1140","TER_1150","TER_1160","TER_1170","TER_1180","TER_1185","TER_1200","TER_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TXF_1000","TXF_1010","TXF_1020","TXF_1030","TXF_1040","TXF_1050","TXF_1060","TXF_1070","TXF_1080","TXF_1090","TXF_1100","TXF_1110","TXF_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TCH_1000","TCH_1010","TCH_1020","TCH_1030","TCH_1040","TCH_1050","TCH_1060","TCH_1080","TCH_1090","TCH_1100","TCH_1110","TCH_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","WSH_1000","WSH_1010","WSH_1020","WSH_1040","WSH_1050","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","WPA_1000","WPA_1010","WPA_1020","WPA_1030","WPA_1040","WPA_1050","WPA_1060","WPA_1070","WPA_1080","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","HOW_SICK_TODAY","RUNNY_NOSE","PLUGGED_NOSE","SNEEZING","SORE_THROAT","SCRATCHY_THROAT","COUGH","HOARSENESS","HEAD_CONGESTION","CHEST_CONGESTION","FEELING_TIRED","THINK_CLEARLY","SLEEP_WELL","BREATH_EASILY","WALK","DAILY_ACTIVITIES","WORK_OUTSIDE_HOME","WORK_INSIDE_HOME","INTERACT","LIVE_LIFE","MY_COLD_IS","VNUM_C","VNUM","VDATE","WDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER"],["Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Char","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Char","Char","Num","Num","Num","Num","Char","Char","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Char","Num","Char","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Char","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num"],["In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, school, or at home? 0=None of the time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All of the time","During the past 3 days, how often have you had asthma symptoms? 0=Not at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how often have you used your rescue inhaler or nebulizer medication ? 0=Not at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how many total times did your asthma symptoms wake you up from sleep? 0=Not at all, 1=1 time in the last 3 days, 2=2-3 times in the last 3 days, 3=4-5 times in the last 3 days, 4=&gt;=6 times in the last 3 days","How would you rate the amount of impairment you have experienced due to your asthma in the past 3 days? 0=No impairment, 1=Mild impairment, 2=Moderate impairment, 3=Severe impairment, 4=Very severe impairment","How stressed or frightened were you by your asthma symptoms in the past 3 days? 0=Not at all, 1=Mildly, 2=Moderately, 3=Severely, 4=Very severely","Why do you think your asthma was worse in the past 3 days compared to what is normal for you? 0=I have not been worse over the past 3 days. My asthma symptoms have been usual., 1=Common cold, 2=Allergies, 3=Pollution or chemical irritant, 4=Too little asth","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Are you currently retired? 1=Yes,0=No","Are you retired because of asthma? 1=Yes,0=No","Are you currently unemployed? 1=Yes,0=No","Are you unemployed because of asthma? 1=Yes,0=No","Do you get paid to do work? 1=Yes,0=No","How much does your asthma bother you at your paid work? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Overall, how much does your asthma bother you when you do jobs around the house? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Overall, how much does your asthma bother your social life? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Overall, how much does your asthma bother your personal life? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Are you involved in leisure activities, such as: walking for pleasure, sports, exercise, travelling, taking vacations? 1=Yes,0=No","When involved in leisure activities, how much does your asthma bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Would you say that you can't do some of these sorts of things because of asthma? 1=Yes,0=No","How much does your asthma bother you when you sleep? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much does the cost of your asthma medicines bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Do you get free prescriptions? 1=Yes,0=No","How much does the inconvenience or embarrassment of taking your asthma medicines bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much do coughs and colds bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Feeling upset is also a bother.  Does your asthma make you feel anxious, depressed, tired, or helpless? 1=Yes,0=No","Feeling upset is also a bother.  Does your asthma make you feel anxious, depressed, tired, or helpless? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will have an asthma attack when visiting a new place? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will catch a cold? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will let others down? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that your health may get worse in the future? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you won't be able to cope with an asthma attack? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past 4 weeks, how much of the time did your asthma keep you from getting as much done at work, school or at home? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often have your had shortness of breath? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often did your asthma symptoms wake you up at night or earlier than usual in the morning? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often have you used your rescue inhaler or nebulizer medication? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","How would you rate your asthma control during the past 4 weeks? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","ICD-9 category","Description of Adverse Event","ICD9 Code","Ongoing at current visit","Type (1=Intermitent; 2=Continuous)","Severity (1=Mild; 2=Moderate; 3=Severe)","Serious (0=No, 1=Yes)","Likelyhood of Relationship to Study Drug(s) (1=None; 2=Unlikely (remote); 3=Possible; 4=Probable)","Change in Study Drugs (1=Unchanged; 2=Altered)","Outcome (1=Completely recovered; 2=Recovered, but with lasting effects; 3=Death)","Treatment Required (1=None; 2=Medication; 3=Hospitalization; 4=Other)","Ongoing at final visit","Date Started","Date Stopped","Visit Number (character)","Visit Number (numeric)","Long ICD-9 description","Short ICD-9 description","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","ASTHMA HISTORY1.Approximately how old were you when chest symptoms suggesting asthma first appeared?","Did these symptoms appear immediately after or as a result of:1a. a respiratory infection such as a cold or pneumonia? 1=Yes, 0=No, 8=Don't Know","an occupational or job change? 1=Yes, 0=No, 8=Don't Know","a household move? 1=Yes, 0=No, 8=Don't Know","a pregnancy? 1=Yes, 0=No, 8=Don't Know","a hormonal change ? 1=Yes, 0=No, 8=Don't Know","How old were you when a doctor first diagnosed you with asthma?","Have any of your immediate blood relatives been told by a physician that they have asthma? 1=Yes, 0=No, 8=Don't Know","Father  (1=Yes, 0=No, 8=Don't Know)","Brother(s) or Sister(s)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Child(ren)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","ASTHMA SYMPTOMS1.How do you categorize your asthma symptoms throughout the course of the year? 1=Relatively the same all year, 2=Vary by season","Winter? 1=Yes,0=No","Spring? 1=Yes,0=No","Summer? 1=Yes,0=No","Fall? 1=Yes,0=No","In the last 12 months, how many...  5a.Asthma episodes have you had that required emergency care or an unscheduled office visit?","Overnight hospitalizations have you had due to asthma?","Courses of systemic corticosteroid therapy  for asthma have you taken?","Days of work, school, or housework have you missed due to asthma?","In the past 3 months, how many days of work, school, or housework have you missed due to asthma?","Have you ever been admitted to an intensive care unit for asthma? 1=Yes,0=No","How many times have you been admitted to an intensive care unit for asthma?","Have you ever had invasive mechanical ventilation? 1=Yes, 0=No, 8=Don't Know","Have you ever had non-invasive mechanical ventilation? 1=Yes, 0=No, 8=Don't Know","Do any of the following currently provoke your asthma? 1=Yes, 0=No, 8=Don't Know","Menstrual cycle(If participant is male or a postmenopausal female, leave blank.) (1=Yes, 0=No, 8=Don't Know)","Aspirin or non-steroidal anti-inflammatory drugs (e.g., Aleve, Motrin) (1=Yes, 0=No, 8=Don't Know)","Respiratory infections (e.g., colds)  (1=Yes, 0=No, 8=Don't Know)","Irritants (e.g., pollution, odors, perfumes, chemicals, household cleaners) (1=Yes, 0=No, 8=Don't Know)","Weather conditions (e.g., change in weather, humidity)  (1=Yes, 0=No, 8=Don't Know)","Exposure to cold air  (1=Yes, 0=No, 8=Don't Know)","Emotional factors (e.g., stress, laughing)  (1=Yes, 0=No, 8=Don't Know)","Tobacco smoke  (1=Yes, 0=No, 8=Don't Know)","Food additives/preservatives (e.g., MSG, sulfites)  (1=Yes, 0=No, 8=Don't Know)","Allergies (e.g., dust, animals, pollens)(1=Yes, 0=No, 8=Don't Know)","Other (1=Yes,0=No)","To which of the following did a doctor or other health practitioner say you were allergic? 1=Yes, 0=No, 8=Don't Know","Foods  (1=Yes, 0=No, 8=Don't Know)","Things you breathe in or are exposed to (e.g., dust, pollens, molds, animal fur, feathers, dander)  (1=Yes, 0=No, 8=Don't Know)","Stinging insects such as bees or wasps (1=Yes, 0=No, 8=Don't Know)","Latex  (1=Yes, 0=No, 8=Don't Know)","Other  (1=Yes,0=No)","Have you ever had eczema / atopic dermatitis ? 1=Yes, 0=No, 8=Don't Know","If YES, was your eczema diagnosed by a doctor? 1=Yes,0=No","Have any of your immediate blood relatives been told by a physician that they have allergies/eczema/hay fever? 1=Yes, 0=No, 8=Don't Know","Father  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Brother(s) or Sister(s)(1=Yes, 0=No, 8=Don't Know, 9=N/A)","Child(ren)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Did you grow up in a household where you were exposed to tobacco smoke? 1=Yes,0=No","Do you currently smoke? 1=Yes,0=No","Record smoking history in pack-years*. __ __ . __ pack-years","Were you ever a smoker? 1=Yes,0=No","Record smoking history in pack-years*. __ __ . __ pack-years","Do you currently live in a household where you are exposed to tobacco smoke? 1=Yes,0=No","Do you currently vape  or use a hookah ? 1=Yes,0=No","How frequently do you vape or use a hookah? 1=Infrequently , 2=Occasionally , 3=Weekly , 4=Daily","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many days a week","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many times a day","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many years have y","Have you ever vaped or used a hookah in the past? 1=Yes,0=No","Have you ever vaped or used a hookah in the past?  If NO, skip to Q17. Approximately how many years did you vape or use a hookah?  (1-99 years)","Do you currently live in a household where you are exposed to others vaping or using a hookah? 1=Yes,0=No","Do you spend time in social settings  where you are exposed to others vaping or using a hookah? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","When was the last time that you vaped or used a hookah? (number of days from Visit 1)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","How many days were you bothered by coughing during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe was your coughing during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by wheezing during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe was your wheezing during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by shortness of breath during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","Over average, how severe was your shortness of breath during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you awakened at night during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how much of a problem was being awakened at night during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by side effects of your asthma medication during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe were the side effects of your asthma medication during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Participants baseline peak flow (PEF) value (0-999 L/M) First visit (Visit 0A, or Visit 1): PEF (FEF Max) from prebronchodilator (baseline) spirometry at first visit (convert to L/M)","Participants baseline rescue use value (0-99 puffs/day) First visit (Visit 0A, or Visit 1): Self-reported average daily use of albuterol during the 14 days prior to first visit.","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","body mass index computed from height and weight measured at Visit 1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Page Number","Reason for change in scheduled medications status (1=Adverse Event, 2=Other (specify))","Reason for change in scheduled medications status Related adverse event number (0-99)","Current status of participants scheduled medications (1=Discontinued, 2=Resumed)","Date change took effect","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Code to identify a drug unit of measure; (Units)","Code to identify how frequently a drug is taken or administered. (Frequency)","Code to identify the route used to administer a drug. (Route)","Name of Medication","This number corresponds to the ID number of the AHFS monograph. (Medication Code)","Related Event","Medication Dose","Ongoing at current visit","Ongoing at final visit","Start Date","Stop Date","Visit Number (character)","Visit Number (numeric)","GEN_DRUG_NAME","CLASS_TEXT","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Diary and Peak Flow Compliance: Number of full days since the last visit (0-999 days)","Diary and Peak Flow Compliance: Number of days where AM and PM scheduled sessions are complete (AM and PM PEF and all diary questions for AM and PM answered) (0-999 days)","Diary and Peak Flow Compliance: Percent compliance If the compliance value in Q1c is less than 75Pct, re-emphasize the importance of completing scheduled diary assessments and peak flows. (0-150.0Pct)","Scheduled Respimat Compliance (Visits 2-9 only): Number of scheduled puffs since the last visit (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Number of remaining puffs on scheduled Respimat Reference all returned Respimats (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Number of puffs taken (60 x (# used Respimats)) - Q2b (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Percent compliance=Q2c/Q2a x 100 If the participant took less than 75Pct of the scheduled Respimat puffs, re-emphasize the importance of maintaining the daily dosing schedule. (0-150.0Pct)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of scheduled puffs since the last visit (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of remaining puffs reflected on scheduled Twisthaler counters Total the values reflected on all counters for all returned Twisthalers (i.e., out of their pouches). (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of puffs taken (60 x (# used Twisthalers)) - Q3b (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Percent compliance=Q3c/Q3a x 100 If the participant took less than 75Pct of the scheduled Twisthaler puffs, re-emphasize the importance of maintaining the daily dosing schedule. (0-150.0Pct)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","1=mometasone, 2=placebo","Blinded Scheduled Twisthaler Contents - How sure are you about your answer in Q1? 1=Absolutely sure - I know what the Twisthaler/MDI contains, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","1=tiotropium, 2=placebo","Blinded Scheduled Respimat Contents - How sure are you about your answer in Q4? 1=Absolutely sure - I know what the Respimat capsules contain, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has the participant or parent/legal guardian signed the SIENA Informed Consent document? 1=Yes,0=No","Ages 12-17 Only: Has the participant signed and dated the assent form or, if the participant is less than the local age of assent, has the participant given verbal assent? 1=Yes,0=No","Has the participant used an inhaled corticosteroid in the past 6 weeks? 1=Yes,0=No","Will the participant be 12 years of age, or older, as of Visit 1? 1=Yes,0=No","Has the participant used an inhaled corticosteroid or combination inhaled corticosteroid/LABA in the past 3 weeks? 1=Yes,0=No","If NO, has the participant used a leukotriene modifier in the past 3 weeks? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used inhaled corticosteroid therapy?","Over the past 3 months, on average, how many days per week has the participant used inhaled corticosteroid therapy: &lt; 5? 1=Yes,0=No","Over the past 3 months, has the participant used greater than the equivalent of 80-240 mcg of inhaled beclomethasone daily? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used combination inhaled corticosteroid/LABA therapy?","Over the past 3 months, on average, how many days per week has the participant used combination inhaled corticosteroid/LABA therapy,  &lt; 5 (1=Yes,0=No)","Over the past 3 months, on average, how many days per week has the participant had daytime asthma symptoms?","Over the past 3 months, on average, how many days per week has the participant had daytime asthma symptoms,  &lt;=2? 1=Yes,0=No","Over the past 3 months, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms?","Over the past 3 months, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms, &lt;=2? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used his/her short acting betaagonist  for relief of symptoms?","Over the past 3 months, on average, how many days per week has the participant used his/her short acting betaagonist  for relief of symptoms, &lt; 2? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 80Pct? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 70Pct? 1=Yes,0=No","Does the participant have current evidence of any of the conditions listed on the Exclusionary Medical Conditions for SIENA  reference card, or any chronic diseases that would prevent participation in the trial or put the participant at risk by participati","Is the participant currently taking any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Is the participant able to go off these medications for the required washout period prior to Visit 1 and for the duration of the study? 1=Yes,0=No","Is the participant able to use the spirotel e-diary/PEF meter correctly, as evidenced by achieving a score of 13 on the spirotel Performance Checklist ? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Has the participant or parent/legal guardian signed the SIENA Informed Consent document?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past month, on average, how many days per week has the participant had daytime symptoms?","In the past month, on average, how many days per week has the participant had daytime symptoms, &lt;=2? 1=Yes,0=No","In the past month, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms?","In the past month, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms,  &lt;=2? 1=Yes,0=No","In the past month, on average, how many days per week has the participant used his/her short acting beta-agonist  for relief of symptoms?","In the past month, on average, how many days per week has the participant used his/her short acting beta-agonist  for relief of symptoms, &lt; 2? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 80Pct? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 70Pct? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant or parent/legal guardian sign the SIENA Informed Consent document? 1=Yes,0=No","Ages 12-17 Only: Has the participant signed and dated the assent form or, if the participant is less than the local age of assent, has the participant given verbal assent? 1=Yes,0=No","Is the participant 12 years of age, or older? 1=Yes,0=No","Does the participant plan to move away from the clinical site in the upcoming 11 months such that his/her ability to complete the study will be jeopardized? 1=Yes,0=No","Has the participant used investigative drugs and/or enrolled in an intervention trial in the past 30 days, or have plans to enroll in such a trial during the SIENA study? 1=Yes,0=No","Has the participant had a respiratory infection within the past 6 weeks? 1=Yes,0=No","Has the participant had a respiratory infection within the past 4 weeks? 1=Yes,0=No","During the past 4 weeks, on average, how many days per week has the participant had daytime asthma symptoms?","During the past 4 weeks, on average, how many days per week has the participant had daytime asthma symptoms  &gt; 2? 1=Yes,0=No","During the past 4 weeks, how many nighttime awakenings due to asthma symptoms has the participant had?","During the past 4 weeks, how many nighttime awakenings due to asthma symptoms has the participant had,  &gt; 2? 1=Yes,0=No","During the past 4 weeks, on average, how many days per week has the participant used his/her short-acting beta-agonist  for relief of symptoms?","During the past 4 weeks, on average, how many days per week has the participant used his/her short-acting beta-agonist  for relief of symptoms, &gt; 2? 1=Yes,0=No","Is Q7a, Q8a, or Q9a checked YES? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Did the participant or parent/legal guardian sign the SIENA Informed Consent document?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Does the participant have current evidence of any of the conditions listed on the Exclusionary Medical Conditions for SIENA  reference card, or any chronic diseases  that would prevent participation in the trial or put the participant at risk by participat","Does the participant have a history of bladder-neck obstruction, urinary retention, benign prostatic hyperplasia , or a clinically relevant urologic disorder that precludes study participation? 1=Yes,0=No","Does the participant have a history of narrow angle glaucoma? 1=Yes,0=No","Does the participant have a history of significant cardiovascular disorders or arrhythmias? 1=Yes,0=No","Has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card within the specified time periods? 1=Yes,0=No","Is the participant currently taking prescription or OTC medication other than those listed on the Allowed Medications  reference card? 1=Yes,0=No","Based on input from the participant and the study physician, will the participant need to use intranasal steroids at any time during the study? 1=Yes,0=No","Is the participant willing to use a single intranasal steroid at a stable dose continuously for the duration of the study, starting at Visit 1? 1=Yes,0=No","Is the participant currently receiving allergen immunotherapy  other than an established maintenance regimen implemented continuously for a minimum of 3 months? 1=Yes,0=No","Has the participant taken omalizumab within the past 3 months? 1=Yes,0=No","Has the participant used any smokeless tobacco products  in the past year? 1=Yes,0=No","Has the participant smoked cigarettes, a pipe, cigar, marijuana, electronic cigarettes, or any other substance in the past year? 1=Yes,0=No","Ages 18+ Only: Does the participant have a smoking history of greater than 10 pack-years? 1=Yes,0=No","Ages 12-17 Only: Does the participant have a smoking history of greater than 5 pack-years? 1=Yes,0=No","Has the participant received a physician diagnosis of asthma at least 12 months ago? 1=Yes,0=No","Has the participant received a physician diagnosis of asthma at least 12 months ago or had a history consistent with asthma for the previous 12 months? 1=Yes,0=No","Has the participant experienced a life-threatening asthma exacerbation requiring treatment with intubation or mechanical ventilation in the past 5 years? 1=Yes,0=No","Has the participant had an asthma exacerbation requiring systemic corticosteroid treatment in the past 6 weeks? 1=Yes,0=No","Has the participant used an oral or inhaled corticosteroid, or leukotriene modifier in the past 6 weeks? 1=Yes,0=No","Has the participant used an oral corticosteroid in the past 6 weeks? 1=Yes,0=No","Has the participant used an inhaled corticosteroid or leukotriene modifier in the past 3 weeks? 1=Yes,0=No","Is the participant potentially able to bear children? 1=Yes, 0=No, 9=N/A","Is the participant currently pregnant or lactating? 1=Yes,0=No","Does the participant agree to use one of the approved methods indicated on the Birth Control Methods  reference card for the duration of the study? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was the participants prebronchodilator FEV1 &gt;=70Pct predicted? 1=Yes,0=No","Did the participants FEV1 improve &gt;=12Pct and &gt;=200 ml in response to four puffs of albuterol? 1=Yes,0=No","Does the participant have valid source documentation within the past 6 months for an overread AsthmaNet methacholine challenge  with a PC20 &lt;=16 mg/ml? 1=Yes,0=No","Did the participants FEV1 improve &gt;=12Pct and &gt;=200 ml in response to four puffs of albuterol?","Was the participants methacholine PC20 &lt;=16 mg/ml? 1=Yes,0=No","Is the participant able to use the spirotel e-diary/PEF meter correctly, as evidenced by achieving a score of 13 on the Spirotel Performance Checklist ? 1=Yes, 0=No, 9=N/A","Has the participant used the Twisthaler previously? 1=Yes,0=No","Is the participant able to use a Twisthaler properly, as evidenced by achieving a score of 12 on the Twisthaler Inhalation Technique Checklist ? 1=Yes,0=No","Is the participant able to use a metered dose inhaler properly, as evidenced by achieving a score of 11 on the MDI Inhalation Technique Checklist  ? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Does the participant have valid source documentation within the past 6 months for an overread AsthmaNet methacholine challenge  with a PC20 &lt;=16 mg/ml?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Since Visit 1, has the participant experienced a treatment failure event as defined in the protocol? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Since Visit 1, has the participant experienced one or more asthma exacerbations as defined in the protocol? 1=Yes,0=No","Since Visit 1, has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant provide two acceptable sputum induction samples during the run-in? 1=Yes,0=No","Since Visit 1, has the participant experienced a treatment failure event as defined in the protocol? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Did the participant complete Asmanex220 mcg treatment at least 6 weeks ago? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Since Visit 1, has the participant experienced one or more asthma exacerbations as defined in the protocol? 1=Yes,0=No","Since Visit 1, has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Has the participant experienced a respiratory tract infection in the past 6 weeks? 1=Yes,0=No","Has the participant experienced a respiratory tract infection in the past 4 weeks? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report daytime asthma symptoms daily during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report night-time awakenings due to asthma symptoms more than once per week during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report night-time awakenings due to asthma symptoms more than twice per week during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report daily albuterol use for asthma symptom control during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant record at least 75Pct of AM and PM peak flow measurements and symptoms on his or her spirotel during the run-in? 1=Yes,0=No","Did the participant take at least 75Pct of the required puffs from his or her Respimat during the run-in? 1=Yes,0=No","Was the participants prebronchodilator FEV1 &gt;=70Pct predicted? 1=Yes,0=No","Has the participant used the Respimat previously? 1=Yes,0=No","Is the participant able to use a Respimat inhaler properly, as evidenced by achieving a score of 15 on the Respimat Inhalation Technique Checklist? 1=Yes,0=No","Does the participant wish to withdraw consent from the study? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has QC procedure been performed on the NIOX MINO today? 1=Yes,0=No","Did the participant eat or drink within the past hour? 1=Yes,0=No","Did the participant take part in strenuous activity/exercise within the past hour? 1=Yes,0=No","Time eNO started (based on a 24-hour clock) (0000-2359)","ENO Measurement","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","How long have you lived in the current house? (Estimate if uncertain.) (1010) years","How long have you lived in the current house? (Estimate if uncertain.)(1020) months","Does your house use a wood burning stove as a primary source of heat? 1=Yes, 0=No, 8=Don't Know","Does your house use an air conditioner? 1=Yes, 0=No, 8=Don't Know","Does your house use an evaporative cooler ? 1=Yes, 0=No, 8=Don't Know","Does your house use a humidifier? 1=Yes, 0=No, 8=Don't Know","Does your house use a dehumidifier? 1=Yes, 0=No, 8=Don't Know","Has there been water damage to your house, basement, or its contents during the past 12 months? 1=Yes, 0=No, 8=Don't Know","Has there been any mold or mildew, on any surfaces, inside your house in the past 12 months? 1=Yes, 0=No, 8=Don't Know","Which rooms have or have had mold or mildew?10a.Bathroom(s)  (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew?10b.Basement or attic (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10c.Kitchen (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10d.Your bedroom (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10e.Other bedrooms (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10f.Living or family room (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10g.Other (1 = Yes, 0 = No)","Do you ever see cockroaches in your house? 1=Yes,0=No","In which room(s) have you seen cockroaches?12a.Kitchen (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches?  12b.Basement or attic (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12c.Bathroom(s) (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12d.Living or family room (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12e.Your bedroom  (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12f.Other bedrooms (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12g.Garage (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12h.Other (1 = Yes, 0 = No)","Do you ever see rodents  or rodent droppings in your house? 1=Yes,0=No","14.In which room(s) have you seen rodents or rodent droppings? 14a. Kitchen (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14b. Basement or attic (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14c. Bathroom(s) (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings?14d. Living or family room (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14e. Your bedroom(1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14f. Other bedrooms (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14g. Garage (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14h. Other (1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15a. Barns(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property?15b. Hay(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15c. Woodsheds(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15d. Firewood(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property?15e. Chicken coops (1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15f. Corral (1 = Yes, 0 = No)","CHARACTERISTICS OF THE PARTICIPANT'S BEDROOM16.What is the floor covering in your bedroom? 1=Rug/carpet, 2=Vinyl tile or linoleum, 3=Wood, 4=Ceramic tile, 5=Other , 9=Don't know","What type of mattress is on your bed? 1=None, 2=Inner spring mattress, 3=Foam mattress, 4=Waterbed, 5=Air mattress, 6=Other , 9=Don't know","Is the mattress completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","Does your bed have a box spring? 1=Yes,0=No","Is the box spring completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","What type of pillow do you usually sleep with? 1=None, 2=Feather/down, 3=Foam/Dacron/synthetic, 5=Other , 9=Don't know","Is the pillow completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","Does your household have any pets? 1=Yes,0=No","Enter the number of pets that the household has. (Enter '00' if none. If none to Q24a - Q24d, skip to the next question.)24a.Cat","Cat  (1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has. Dog","Dog(1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has 24c.Rabbit, guinea pig, hamster, gerbil, or mouse","Rabbit, guinea pig, hamster, gerbil, or mouse (1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has 24d.Bird","Bird (1=Indoor, 2=Outdoor, 3=Both)","In general, and on a regular basis, are you exposed to any of the following animals?  25a.Cat(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25b.Dog(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25c.Rabbit, guinea pig, hamster, gerbil, or mouse(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25d.Bird (1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25e.Farm animals (1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals?  25f.Other (1 = Yes, 0 = No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other (specify","Highest education of any household member (1=No High School diploma/GED, 2=High Schol grad, 3=Technical training, 4=Some college, no degree, 5=Assoc degree, 6=Bachelor degree, 7=Masters degree/MD/PhD/JD/PharmD, 9=Decline to answer, 10=Don't know)","Category best describes the combined annual income, before taxes, of all members of your household for the last year. (1=Less than $25,000, 2=$25,000 - $49,999, 3=$50,000 - $99,999, 4=$100,000 or more, 9=Decline to answer, 10=Don't know)","How many people  are supported by this income reported in Q3?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Total IgE","Cat (E1)","Dog (E5)","Mouse (E72)","Rat (E74)","Mold mix (MX1)","Cockroach, German (I6)","Grass mix (GX2)","Tree mix (TX4)","Tree mix (TX6)","Weed mix (WX1)","Weed (W3)","Mite (D2)","Mite (D1)","Number or allergen specific (Cat, Dog, Mouse, etc) that were either POS or greater than 0.35","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","CBC with differential cell count: Eosinophils (absolute count) (0-9999/uL)","CBC with differential cell count: WBC (0-999.9 K/uL)","CBC with differential cell count: Differential Lymphocytes (0-99.9Pct)","CBC with differential cell count: Differential Monocytes (0-99.9Pct)","CBC with differential cell count: Differential Basophils (0-99.9Pct)","CBC with differential cell count: Differential Neutrophils (0-99.9Pct)","CBC with differential cell count: Differential Eosinophils (0-99.9Pct)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","PARENTAL HEIGHT - First study visit only or until both are completed 1.Biological mother's height (complete height or check unknown) __ feet","PARENTAL HEIGHT - First study visit only or until both are completed1.Biological mother's height (complete height or check unknown) __ __ inches","PARENTAL HEIGHT - First study visit only or until both are completed 1.Biological mother's height (complete height or check unknown) (9=Don't Know)","Biological father's height (complete height or check unknown) __ feet","Biological father's height (complete height or check unknown) __ __ inches","Biological father's height (complete height or check unknown) (9=Don't Know)","PARTICIPANT MEASUREMENTS - Complete at all applicable study visits3. What type of height measurement was obtained? 1=Standing height, 2=Length","First measurement __ __ __ . __ cm","Second measurement __ __ __ . __ cm","Third measurement __ __ __ . __ cm","Average height or length measurement __ __ __ . __ cm","In your judgment, was the participant's height or length measurement acceptable? 1=Yes,0=No","Weight (shoes off, light clothing) __ __ __ . __ kg","ORAL CANDIDIASIS5.Does the participant have evidence of oral candidiasis? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Post Diluent FEV1","Did the participant drop20Pct at the diluent stage? 1=Yes,0=No","Last concentration of methacholine administered __ __ . __ __ __ __ mg/ml","FEV1 after last concentration of methacholine administered","Did the participant achieve a PC20? 1=Yes,0=No","PC20 __ __ . __ __ mg/ml","Time methacholine challenge ended (based on 24-hour clock)","FEV1","Time of FEV1 in Q7a (based on 24-hour clock)","Was the FEV1 from Q7athe methacholine reversal reference value  in the gray box above? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was an additional treatment used in the first hour? 1=Yes,0=No","Additional albuterol by MDI (1=Yes,0=No)","Number of additional puffs of albuterol administered (1=2, 2=4, 3=&gt;4)","Nebulized Beta-agonist(1=Yes,0=No)","Subcutaneous epinephrine(1=Yes,0=No)","Implementation of clinic emergency protocol or algorithm (1=Yes,0=No)","Other  (1=Yes,0=No)","Participant's FEV1 after additional treatment within first hour.2a.FEV1","Time of FEV1 in Q2a (based on 24-hour clock)","Was the FEV1 from Q2a &gt;the methacholine reversal reference value  in the gray box on the Methacholine Challenge Testing  form? 1=Yes,0=No","Was additional treatment used after one hour? 1=Yes,0=No","Additional albuterol by MDI (1=Yes,0=No)","Number of additional puffs of albuterol administered (1=2, 2=4, 3=&gt;4)","Nebulized Beta-agonist  (1=Yes,0=No)","Subcutaneous epinephrine(1=Yes,0=No)","Implementation of clinic emergency protocol or algorithm  (1=Yes,0=No)","Treatment in the emergency room  (1=Yes,0=No)","Overnight hospitalization (1=Yes,0=No)","Other (1=Yes,0=No)","Participant's final FEV1 after methacholine challenge4a.FEV1","Time of FEV1 in Q4a (based on 24-hour clock)","Was the FEV1 from Q4a ? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Exclusions and Confounders1.Has the participant had any severe acute illness in the past 4 weeks? 1=Yes,0=No","If YES, has the participant received permission from the supervising physician to proceed with the methacholine challenge testing? 1=Yes,0=No","Has the participant used an oral or injectable steroid  in the last 4 weeks? 1=Yes,0=No","Does the participant have a baseline  FEV1 less than 55Pct of predicted or less than 1.0 L? 1=Yes,0=No","Pregnancy test results:  (Check N/A if the participant is male, or is female and is post-menopausal, had a hysterectomy or tubal ligation.) (1=Positive, 0=Negative, 9=N/A)","Is the participant's systolic blood pressure &gt; 200 mm Hg or diastolic blood pressure &gt; 100 mm Hg? 1=Yes,0=No","Is there any other reason the participant should not proceed with the methacholine challenge testing? 1=Yes,0=No","Is the participant eligible to proceed with the diluent  pulmonary function testing for the methacholine challenge? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Exclusions and Confounders:  1.Has the participant had any severe acute illness in the past 4 weeks? 1=Yes,0=No","If YES, has the participant received permission from the supervising physician to proceed with the methacholine challenge testing? 1=Yes,0=No","During the past 4 weeks, has the participant had any respiratory infections, colds, or bronchitis ? 1=Yes,0=No","If YES, during the past 2 weeks, has the participant had any respiratory infections, colds, or bronchitis ? 1=Yes,0=No","Has the participant used an oral or injectable steroid  in the last 4 weeks? 1=Yes,0=No","Does the participant have a baseline  FEV1 less than 70Pct of predicted? 1=Yes,0=No","Pregnancy test results(Check N/A if the participant is male, or is female and has not started menses.)(1=Positive, 0=Negative, 9=N/A)","If participant's age is ? 1=Yes,0=No","If participant's age is &lt; 12 years: Is the participant's systolic blood pressure &gt; 180 mm Hg or diastolic blood pressure &gt; 90 mm Hg? 1=Yes,0=No","Is there any other reason the participant should not proceed with the methacholine challenge testing? 1=Yes,0=No","Is the participant eligible to proceed with the diluent  pulmonary function testing for the methacholine challenge? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time albuterol administered (based on 24-hour clock)","Time post-albuterol spirometry started (based on 24-hour clock)","Highest FVC","Highest FEV1","Highest FEV1 (Pct predicted)","FEF Max","FEF25-75","In your judgment, was the subject's spirometry technique acceptable? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Please check the box next to the treatment that you believe you received over the past 12 weeks. 1=mometasone, 2=placebo","Blinded Scheduled Twisthaler Contents: How sure are you about your answer to Question 2? 1=Absolutely sure -  I know what the MDI contains, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","Please comment with respect to any observations you made that helped you make your choice in Question 2 . 1=I have no comments, 2=I noticed the following:","Please check the box next to the treatment that you believe you received over the past 12 weeks. 1=tiotropium, 2=placebo","How sure are you about your answer to Question 5? 1=Absolutely sure - I know what the Respimat capsules contain, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","1=I have no comments, 2=I noticed the following:","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Visit Number (numeric)","Level of periostin in peripheral blood (ng/ml)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time ipratropium administered (based on 24-hour clock) (0000-2359)","Time post-ipratropium spirometry started (based on 24-hour clock) (0000-2359)","Highest FVC (0-9.99 L)","Highest FEV1 (0-9.99 L)","Highest FEV1 (Pct predicted) (0-999 Pct predicted)","FEF Max (0-99.99 L/S)","FEF25-75 (0-9.99 L/S)","In your judgment, was the participants spirometry technique acceptable? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Is the participant unable to bear children due to any of the following reasons? 1=Yes,0=No","Post-menopausal (at least one year since last menses) (1=Yes,0=No)","Hysterectomy(1=Yes,0=No)","Tubal ligation(1=Yes,0=No)","Pregnancy test results  (1=Positive, 0=Negative)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Have you had any diseases, illnesses, conditions, or surgeries related Blood, Lymph, or Immune Systems? 1=Yes,0=No","Eyes(1=Yes,0=No)","Breasts (1=Yes,0=No)","Endocrine Systems (1=Yes,0=No)","Heart and Blood Vessels  (1=Yes,0=No)","Liver or Pancreas (1=Yes,0=No)","Kidneys or Urinary Tract System  (1=Yes,0=No)","Reproductive System (1=Yes,0=No)","Muscles or Bones (1=Yes,0=No)","Nervous System (1=Yes,0=No)","Psychiatric(1=Yes,0=No)","Drug Allergies (1=Yes,0=No)","Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Skin (1=Yes,0=No)","Have you ever had allergic rhinitis ? 1=Yes, 0=No, 9=Don't know","Have you ever had nasal polyps? 1=Yes, 0=No, 9=Don't know","Do you have chronic or recurrent sinusitis ? 1=Yes, 0=No, 9=Don't know","Have you ever been diagnosed with vocal cord dysfunction? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the ear, nose, or throat? 1=Yes,0=No","Have you ever had pneumonia? 1=Yes, 0=No, 9=Don't know","If YES, were you diagnosed by chest x-ray? 1=Yes, 0=No, 9=Don't know","If YES, were you treated with antibiotics? 1=Yes, 0=No, 9=Don't know","Have you ever had bronchitis? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the lungs ? 1=Yes, 0=No, 9=Don't know","Do you have gastroesophageal reflux disease ? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the stomach or intestines? 1=Yes,0=No","Have you been diagnosed with sleep disordered breathing ? 1=Yes,0=No","If YES, are you being treated with CPAP or BiPAP? 1=Yes,0=No","Have you ever had other sleep disorders? 1=Yes,0=No","Have you ever had other conditions that have not been mentioned on this form? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Short-acting Inhaled Beta-Agonists by Inhaler (e.g., albuterol, Primatene Mist, Maxair, ProAir, Proventil, Ventolin, Xopenex)  (1=Yes, 0=No, 9=Don't Know)","If YES, indicate average weekly puffs in the past month (Enter '000' if none used) weekly puffs","Rescue treatment via a Nebulizer Machine (e.g., albuterol, ipratropium, Combivent, Xopenex, levalbuterol)  (1=Yes, 0=No, 9=Don't Know)","Long-acting Inhaled Beta-Agonists (e.g., Serevent, Foradil, salmeterol, formoterol) Do not consider combination medications. (1=Yes, 0=No, 9=Don't Know)","Oral Beta-Agonists (e.g., albuterol, Brethine, Bricanyl, metaproterenol, Proventil, Ventolin, Repetabs, Volmax)  (1=Yes, 0=No, 9=Don't Know)","Oral Theophylline (short-acting or sustained release)(e.g., Aminophylline, Slo-Phyllin, Slo-bid, Theo-Dur, Uniphyl)(1=Yes, 0=No, 9=Don't Know)","Inhaled Anticholinergic by Inhaler (e.g., Atrovent, Combivent, Spiriva) (1=Yes, 0=No, 9=Don't Know)","Leukotriene Antagonist / 5LO Inhibitors (e.g., Accolate, Zyflo, Singulair) (1=Yes, 0=No, 9=Don't Know)","IgE Blocker (e.g., Xolair)(1=Yes, 0=No, 9=Don't Know)","Oral Steroids FOR ASTHMA (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron,  dexamethasone)(1=Yes, 0=No, 9=Don't Know)","If YES, in the past 12 months, how many courses of steroids by mouth have you taken FOR ASTHMA? 1=1 course, 2=2 courses, 3=3 courses, 4=4 courses, 5=5 courses, 6=More than 5 courses","Injectable Steroids FOR ASTHMA (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV)(1=Yes, 0=No, 9=Don't Know)","Steroids by Inhaler (e.g., Asmanex Twisthaler, QVAR, Flovent, Pulmicort Flexhaler)(1=Yes, 0=No, 9=Don't Know)","Indicate most recent type of inhaled steroid taken (refer to PRIOR_TRT_CARD reference card) code","Indicate number of daily puffs used daily puffs","Indicate the total number of months that you used the inhaled steroid out of the past 12 months months","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide) (1=Yes, 0=No, 9=Don't Know)","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide). If YES, complete Q13a - Q13c. Indicate most recent type of nebulized steroid taken (refer to PRIOR_TRT_CARD reference card)","Indicate number of daily treatments used daily treatments","Indicate the total number of months that you used the nebulized steroid out of the past 12 months months","Long-Acting Beta-Agonist and Inhaled Steroid Combination Medications (e.g., Advair Diskus, Symbicort MDI, Dulera MDI)(1=Yes, 0=No, 9=Don't Know)","Indicate most recent type of combination medication taken (refer to PRIOR_TRT_CARD reference card) code","Indicate number of daily puffs used daily puffs","Indicate the total number of months that you used the combination medication out of the past 12 months months","Nasal Steroids (e.g., Beconase, Vancenase, Flonase, Nasacort, Nasalide, Nasarel, Omnaris, Rhinocort, Nasonex)(1=Yes, 0=No, 9=Don't Know)","Non-steroidal Anti-allergic Nasal Medications (e.g., Nasalcrom, Astelin, Astepro, ipratropium)(1=Yes, 0=No, 9=Don't Know)","Anti-allergic Oral Medications (e.g., fexofenadine, loratadine, cetirizine, chlorpheniramine) (1=Yes, 0=No, 9=Don't Know)","Topical Steroids - Prescription (e.g., Synalar, Lidex, Dermacin, Fluocinonide) (1=Yes, 0=No, 9=Don't Know)","Topical Steroids - OTC (e.g., Hydrocortisone - multiple strengths and products) (1=Yes, 0=No, 9=Don't Know)","Other Medication FOR ASTHMA OR ALLERGIES (1=Yes, 0=No, 9=Don't Know)","Oral Steroids for Conditions Other Than Asthma (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) (1=Yes, 0=No, 9=Don't Know)","Injectable Steroids for Conditions Other Than Asthma (e.g., Medrol, Solumedrol, Decadron, dexamethasone,  triamcinolone, Kenalog, hydrocortisone IV)(1=Yes, 0=No, 9=Don't Know)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Short-acting Inhaled Beta-Agonists by Inhaler (e.g., albuterol, Primatene Mist, Maxair, ProAir, Proventil, Ventolin, Xopenex) Date","Rescue treatment via a Nebulizer Machine (e.g., albuterol, ipratropium, Combivent, Xopenex, levalbuterol) Date","Long-acting Inhaled Beta-Agonists (e.g., Serevent, Foradil, salmeterol, formoterol) Do not consider combination medications. Date","Oral Beta-Agonists (e.g., albuterol, Brethine, Bricanyl, metaproterenol, Proventil, Ventolin, Repetabs, Volmax) Date","Inhaled Anticholinergic by Inhaler (e.g., Atrovent, Combivent, Spiriva) Date","Leukotriene Antagonist / 5LO Inhibitors (e.g., Accolate, Zyflo, Singulair) Date","IgE Blocker (e.g., Xolair) Date","Oral Steroids FOR ASTHMA (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) Date","Injectable Steroids FOR ASTHMA (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV) Date","Steroids by Inhaler (e.g., Asmanex Twisthaler, QVAR, Flovent, Pulmicort Flexhaler) Date","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide) Date","Long-Acting Beta-Agonist and Inhaled Steroid Combination Medications (e.g., Advair Diskus, Symbicort MDI, Dulera MDI) Date","Nasal Steroids (e.g., Beconase, Vancenase, Flonase, Nasacort, Nasalide, Nasarel, Omnaris, Rhinocort, Nasonex) Date","Non-steroidal Anti-allergic Nasal Medications (e.g., Nasalcrom, Astelin, Astepro, ipratropium) Date","Anti-allergic Oral Medications (e.g., fexofenadine, loratadine, cetirizine, chlorpheniramine) Date","Topical Steroids - Prescription (e.g., Synalar, Lidex, Dermacin, Fluocinonide) Date","Topical Steroids - OTC (e.g., Hydrocortisone - multiple strengths and products) Date","Other Medication FOR ASTHMA OR ALLERGIES Date","Oral Steroids for Conditions Other Than Asthma (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) Date","Injectable Steroids for Conditions Other Than Asthma (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV) Date","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the last month, how often have you been upset because of something that happened unexpectedly? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that you were unable to control the important things in your life? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt nervous and stressed? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt confident about being able to handle your personal problems? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that things were going your way? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you found that you could not cope with all the things that you had to do? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you been able to control irritations in your life? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that you were on top of things? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you been angered because of things that happened that were outside of your control? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that your difficulties were piling up so high that you could not overcome them? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Have you consumed caffeine in the past 4 hours? 1=Yes,0=No","Have you used medications with caffeine in the past 4 hours? 1=Yes,0=No","Have you used any weight loss medications in the past 4 hours? 1=Yes,0=No","Have you consumed any food containing alcohol or beverages containing alcohol in the past 4 hours? 1=Yes,0=No","Have you used a rescue intermediate-acting inhaled betaagonist in the past 6 hours? 1=Yes,0=No","Have you used a short-acting anticholinergic in the past 6 hours? 1=Yes,0=No","Have you used any ophthalmic antihistamines in the past 6 hours? 1=Yes,0=No","Have you used any nasal antihistamines in the past 6 hours? 1=Yes,0=No","Have you used any nasal decongestants in the past 6 hours? 1=Yes,0=No","Have you used any oral antihistamines in the past 48 hours? 1=Yes,0=No","Have you used any oral decongestants or cold remedies in the past 48 hours? 1=Yes,0=No","Have you used any smokeless tobacco products today? 1=Yes,0=No","At this time, is your asthma worse because of recent exposure to triggers? 1=Yes,0=No","Is there any other reason you should not proceed with spirometry testing? 1=Yes,0=No","Is the participant eligible to proceed with the spirometry testing? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past 4 weeks, I worried about the long-term effects of asthma on my health (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I had to worry about asthma triggers (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, my asthma was on my mind (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, it was hard to get a good nights sleep because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt like I couldnt enjoy life because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt that asthma was controlling my life (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt frustrated that I couldnt make plans in advance because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, everyday activities were a struggle (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, asthma placed stress on my relationships with family, friends, significant others, or co-workers (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, I felt frustrated that I have to do things differently than people who dont have asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt like I missed out on doing things with others because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, I had to do a lot of planning to make sure I always had an inhaler ready (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Randomized treatment assignment period 1: (Mometasone, Tiotropium or Placebo)","Randomized treatment assignment period 2: (Mometasone, Tiotropium or Placebo)","Randomized treatment assignment period 3: (Mometasone, Tiotropium or Placebo)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Sex (1=Male, 2=Female)","Primary Racial Identification for Spirometry (1=Amer Indian or Alaskan Native, 2=Asian or Pacific Islander, 3=Black, 4=White, 5=Hispanic or Latino, 6=Other)","Participant's Age at Visit 1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Description of Adverse Event (ICD9 Code)","Is the participant currently taking study drug? 1=Yes,0=No","Time interval between the last administration of the study drug and the Adverse Event","What was the unit of time for the interval in Question #4? 1=Second, 2=Minute, 3=Hour, 4=Day","Fatal event(1=Yes,0=No)","Life-threatening event (1=Yes,0=No)","Inpatient hospitalization required(1=Yes,0=No)","Hospitalization prolonged  (1=Yes,0=No)","Disabling or incapacitating  (1=Yes,0=No)","Overdose (1=Yes,0=No)","Cancer (1=Yes,0=No)","Congenital anomaly(1=Yes,0=No)","Serious laboratory abnormality with clinical symptoms(1=Yes,0=No)","Height failure (per protocol MOP)(1=Yes,0=No)","Pregnancy(1=Yes,0=No)","Other (1=Yes,0=No)","What in your opinion caused the event? 1=Yes,0=No","Withdrawal of study drug(s)(1=Yes,0=No)","Concurrent medication  (1=Yes,0=No)","Other condition or event  (1=Yes,0=No)","Was the event expected or unexpected? 1=Expected, 2=Unexpected","Was the event possibly, probably, or definitely related to study participation? 1=Yes,0=No","Date of Adverse Event","Admission date","Discharge date","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","What type of height measurement was obtained? 1=Standing height, 2=Length","First measurement __ __ __ . __ cm","Second measurement __ __ __ . __ cm","Third measurement __ __ __ . __ cm","Average height or length measurement __ __ __ . __ cm","In your judgment, was the participant's height or length measurement acceptable? 1=Yes,0=No","Weight (shoes off, light clothing) __ __ __ . __ kg","Does the participant have evidence of oral candidiasis? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant fail to respond within 48 hours to the treatment failure rescue algorithm? 1=Yes,0=No","Did the participant use at least 16 puffs as needed albuterol per 24 hours for a period of 48 hours? 1=Yes,0=No","Did the participant experience prebronchodilator FEV1 values &lt; 50Pct of the baseline prebronchodilator value obtained at Visit 1 on two consecutive spirometric determinations made on different days? 1=Yes, 0=No, 9=Not evaluated","Did the participant experience prebronchodilator FEV1 values &lt; 40Pct of predicted on two consecutive spirometric determinations made on different days? 1=Yes, 0=No, 9=Not evaluated","Did the study or treating physician prescribe the participant oral/parenteral corticosteroids for the treatment of his/her asthma? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation in the opinion of the study investigator or personal physician? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation? 1=Yes,0=No","Did the participant seek care for significant asthma exacerbation conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Was the participant hospitalized? 1=Yes,0=No","Duration of hospital stay (0-99.9 days)","Was intubation or ventilation assistance required? 1=Yes,0=No","Was the participant admitted to the intensive care unit? 1=Yes,0=No","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Inhaled corticosteroids (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Nebulized bronchodilator (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Oral corticosteroids (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, IM or IV steroids (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Antibiotics (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Other (1=Yes, 0=No )","Why do you think the participant experienced a significant asthma exacerbation? 1=Common cold, 2=Allergies, 3=Pollution or chemical irritant, 4=Too little asthma maintenance medication, 5=Exercise, 6=Other","Date exacerbation conditions were met","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Which phone contact following Asthma Exacerbation visit is being performed? 1=Phone Contact #1 , 2=Phone Contact #2 , 3=Phone Contact #3","Since the most recent Asthma Exacerbation contact, has the participant sought additional care for significant asthma exacerbation conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Was the participant hospitalized? 1=Yes,0=No","Duration of hospital stay (0-99.9 days)","Was intubation or ventilation assistance required? 1=Yes,0=No","Was the participant admitted to the intensive care unit? 1=Yes,0=No","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Inhaled corticosteroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Nebulized bronchodilator (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Oral corticosteroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, IM or IV steroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Antibiotics (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Runny Nose  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Post nasal drip  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Need to blow your nose  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Facial pain/pressure(0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Nasal obstruction  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time spirometry started (based on 24-hour clock)","The reported FEV1 and FVC are the best volumes of all acceptable maneuvers. 1.Highest FVC","Highest FEV1 (Pct predicted)","The reported flow rates correspond to the maneuver where FEV1 + FVC is maximized. FEF Max","FEF25-75","In your judgment, was the subject's spirometry technique acceptable? 0=No, 1=Yes","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Highest FEV1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","TRIAL_TYPE","FVC","FEV1","PEF","FEF2575","FET","Number of times the participant woke up last night due to asthma symptoms # Times woke up due to asthma (0-9)","Number of puffs the participant will take from the WHITE study inhaler this morning # Puffs from WHITE inhaler in AM (0-9)","only) Number of puffs the participant will take from the BLUE study inhaler this morning # Puffs from BLUE inhaler in AM (0-9)","Has the participant taken any puffs from his/her RED RESCUE albuterol inhaler in the past 4 hours? 1=yes, 0=no","Shortness of Breath Score Shortness of Breath Score overnight (0,1,2,3)","Chest tightness score Chest Tightness Score overnight (0,1,2,3)","Wheezing score Wheezing Score overnight (0,1,2,3)","Coughing score Coughing Score overnight (0,1,2,3)","Phlegm/Mucus score Phlegm/Mucus Score overnight (0,1,2,3)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","date 'fixed' for analysis so that associated AM and PM always have same ddate","TRIAL_TYPE","FVC","FEV1","PEF","FEF2575","FET","Number of puffs the participant will take from the WHITE study inhaler tonight # Puffs from WHITE inhaler in PM (0-9)","Has the participant taken any puffs from his/her RED RESCUE albuterol inhaler during the past 4 hours? 1=yes, 0=no","Shortness of Breath Score Shortness of Breath Score since waking this morning (0,1,2,3)","Chest tightness score Chest Tightness Score since waking this morning (0,1,2,3)","Wheezing score Wheezing Score since waking this morning (0,1,2,3)","Coughing score Coughing Score since waking this morning (0,1,2,3)","Phlegm/Mucus score Phlegm/Mucus Score since waking this morning (0,1,2,3)","Did your regular exercise cause unusually severe shortness of breath during the past 24 hours? 1=Yes, 0=No, 9=N/A","Number of RED RESCUE albuterol puffs taken in the past 24 hours to prevent symptoms# Preventive  RED RESCUE puffs  in past 24 hours?","Number of RED RESCUE albuterol puffs taken for asthma symptoms or low peak flow during past 24 hours  # RED RESCUE  puffs for symptoms  in past 24 hours?","Number of times used RED RESCUE albuterol inhaler for asthma symptoms during past 24 hours  # Times used  RED RESCUE for symptoms in past 24 hours?","Did RED RESCUE albuterol relieve symptoms for less than 4 hours after treatment during past 24 hours? 1=Yes,0=No","Was the participant seen by a healthcare provider  for an unscheduled visit in the past 24 hours due to asthma symptoms? 1=Yes,0=No","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Processing: Time processing started (based on 24-hour clock) __ __ __ __ (0 - 2359)","Processing Sample: Total Cell Count: __ __ __ __. __x 10^4 cells/ml (0 - 9999.9)","Was the participant's sputum sample processed within 4 hours after collection? 1=Yes,0=No","Processing Sample: Processing Date __ __ / __ __ / 20 __ __","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Rate slide's quality:  (1=Very good, 2=Good, 3=Acceptable, 4=Poor but readable, 5=Not readable)","Record the number on the slide(s) that was (were) read (0 - 9)","Record the number on the slide(s) that was (were) read: These are numbers that were assigned to the slides at each site. (0 -9)","Total Cell Count: __ __ __ __ . __ x 10^4 cells/ml (0 - 9999.9)","Differential Cell Counts: Squamous Cells __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Epithelial Cells __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Macrophages __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Neutrophils __ __ . __ Pct (0 -99.9)","Differential Cell Counts: Eosinophils __ __ . __ Pct (0 - 99.9)","Differential Cell Counts: Lymphocytes __ __ . __ Pct (0 -99.9)","Date of Read __ __ / __ __ / 20 __ __","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","For this protocol, what was the duration of sputum induction the first time the participant's sample was processed within 4 hours after collection?","Sputum induction start time (based on 24-hour clock)","Sputum induction stop time (based on 24-hour clock)","Duration of sputum induction collection phase at this visit -- minutes (0 - 99.9)","Was the duration ? 1=Yes,0=No","Volume of sputum sample at this visit -- ml (0.0 - 999.9)","Is the volume adequate for processing? 1=Yes,0=No","Is the sample adequate for laboratory analysis? 1=Yes,0=No","Participant's FEV1 immediately after completion of sputum induction: -- (0.0 -9.99 L)","FEV1 (Pct predicted) (0 - 999)","Time of FEV1 in Q7a (based on 24-hour clock)","Percent difference in FEV1 ((Reference - Q7a) / Reference) X100 (-99.9 - 99.9 Pct)","Did the participant's FEV1 drop &gt; 10Pct from reference FEV1 as indicated in Q7d? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Participant's FEV1 after initial 2 puffs of albuterol: 1a.FEV1 __ . __ __ L (0 -9.99)","Participant's FEV1 after initial 2 puffs of albuterol: 1b. FEV1 (Pct predicted)__ __ __ Pct predicted (0 - 999)","Participant's FEV1 after initial 2 puffs of albuterol: 1c. Time of FEV1 from Q1a (based on 24-hour clock) __ __ __ __ (0 - 2359)","Participant's FEV1 after initial 2 puffs of albuterol: 1d. Was the FEV1 from Q1a &gt;=the sputum induction reversal reference value in the gray box above? 1=Yes,0=No","Participant's FEV1 after 2 additional puffs of albuterol: 2a. FEV1 __ . __ __ L (0 - 9.99)","Participant's FEV1 after 2 additional puffs of albuterol: 2b. FEV1 (Pct predicted) __ __ __ Pct predicted (0 - 999)","Participant's FEV1 after 2 additional puffs of albuterol: 2c. Time of FEV1 from Q2a (based on 24-hour clock)__ __ __ __ (0 - 2359)","Participant's FEV1 after 2 additional puffs of albuterol: 2d.Was the FEV1 from Q2a ? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was the participant's FEV1 after reversal from the methacholine challenge greater yhan 90Pct of the baseline FEV1 (1=Yes,0=No)","If NO, has the participant received permission from the supervising physician to proceed with sputum induction testing? 1=Yes,0=No","Physician's Signature:_________ (1=signature present, NULL=signature missing)","Participant's FEV1 used for assessment of eligibility for sputum induction __ . __ __ L (0 - 9.99)","Participant's FEV1 (Pct predicted) used for assessment of eligibility for sputum induction __ __ __ Pct predicted (O-999)","Was the participant's FEV1  from Q3 ? 1=Yes,0=No","Has the participant used any smokeless tobacco products  today? 1=Yes,0=No","Is there any other reason the participant should not proceed with sputum induction? 1=Yes,0=No","Is the participant eligible for sputum induction? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has the participant completed the study through Visit 9? 1=Yes,0=No","Who initiated termination of the participant? 1=Participant, 2=Clinical Staff","1=no longer interested in participating*, 2=no longer willing to follow protocol*, 3=difficult access to clinic , 4=unable to make visits during clinic hours, 5=moving out of the area, 6=unable to continue due to personal constraints*, 7=unable to continue","Did clinical staff terminate the participant due to pregnancy? 1=Yes, 0=No, 9=N/A","Did clinical staff terminate the participant due to loss to follow-up? 1=Yes,0=No","Did clinical staff terminate the participant due to loss to follow-up, type of contact (1=In-person visit, 2=Phone call)","Did clinical staff terminate the participant due to an asthma-related adverse event? 1=Yes,0=No","Did clinical staff terminate the participant due to a medication-related adverse event? 1=Yes,0=No","Did clinical staff terminate the participant due to an adverse event not related to asthma or medications? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility during the Supervised Washout ? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with medication dosing? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with diary completion? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with visit attendance? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with peak flow monitoring? 1=Yes,0=No","Did clinical staff terminate the participant due to significant asthma exacerbation or two treatment failures during run-in ? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility during the run-in period  for reasons other than compliance or exacerbation/treatment failure or high sputum eosinophils ? 1=Yes,0=No","Did clinical staff terminate the participant due to other reason? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility due to high sputum eosinophils in the run-in ? 1=Yes,0=No","Did clinical staff terminate the participant due to... Indicate the letter corresponding to the primary reason the participant was terminated. (a-n)","Did clinical staff terminate the participant due to loss to follow-up, date of last contact with participant","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant seek care for treatment failure conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact, other than call to notify clinic as instructed in spirotel","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Inhaled corticosteroids (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Nebulized bronchodilator (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started: Oral corticosteroids form.  (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  IM or IV steroids form.  (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started: Antibiotics (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant awaken from asthma three or more times in a two-week period or on two consecutive nights? 1=Yes,0=No","Did the participant use albuterol for relief of symptoms four or more times/day for two or more consecutive days? 1=Yes,0=No","Did albuterol relieve symptoms for less than four hours after treatment? 1=Yes,0=No","Did the participant use albuterol for relief of symptoms daily for seven days, and this use exceeded two times the weekly use of albuterol in the baseline period? 1=Yes,0=No","Did regular exercise cause unusually severe shortness of breath on two or more days during a seven day period? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation? 1=Yes,0=No","Did the participant experience a treatment failure? 1=Yes,0=No","Has the participant experienced two treatment failures in the current treatment period ? 1=Yes,0=No","Does the study physician feel the study treatment contributed directly to the treatment failure, such that it is in the participants best interest to discontinue the current blinded treatment? 1=Yes,0=No","Did the participant begin daily treatment with high-dose ICS for 10 days? 1=Yes,0=No","Did the participant begin daily treatment with high-dose ICS for 10 days? 1=Did not think symptoms werebad enough, 2=Other","Date treatment failure conditions were met","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","During this last treatment period , has the participant experienced a treatment failure or asthma exacerbation? 1=Yes,0=No","Has the participant experienced two or more treatment failures in this last treatment period, or does the study physician feel the study treatment contributed directly to the treatment failure such that it is in the participants best interest to discontinu","As of today, have 2 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","As of today, have 6 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","As of today, have 3 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Are you currently employed ? 1=Yes,0=No","In general, how many hours per week do you usually work?","During the past seven days, how many hours did you miss from work because of problems associated with your asthma?","During the past seven days, how much did asthma affect your productivity while you were working?","Do you currently attend classes in an academic setting ? 1=Yes,0=No","In general, how many hours per week do you usually attend classes?","During the past seven days, how many hours did you miss from class or school because of problems associated with your asthma?","During the past seven days, how much did asthma affect your productivity while in school or attending classes in an academic setting?","During the past seven days, how much did your asthma affect your ability to do your regular daily activities, other than work at a job or attend classes?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","How sick today? 0=Not Sick, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely","Runny Nose:(0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","PLUGGED_NOSE  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SNEEZING  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SORE_THROAT  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SCRATCHY_THROAT  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","COUGH (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","HOARSENESS  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","HEAD_CONGESTION  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","CHEST_CONGESTION  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","FEELING_TIRED (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","THINK_CLEARLY  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SLEEP_WELL  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","BREATH_EASILY  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WALK  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","DAILY_ACTIVITIES  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WORK_OUTSIDE_HOME  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WORK_INSIDE_HOME  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","INTERACT  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","LIVE_LIFE  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","MY_COLD_IS  (1=Very much better, 2=Somewhat better, 3=A little better, 4=The same, 5=A little worse, 6=Somewhat worse, 7=Very much worse)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Date (number of days from Visit 1)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>data_set_name<\/th>\n      <th>variable_number_within_dataset<\/th>\n      <th>variable<\/th>\n      <th>type<\/th>\n      <th>variable_description<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":2},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"data_set_name","targets":1},{"name":"variable_number_within_dataset","targets":2},{"name":"variable","targets":3},{"name":"type","targets":4},{"name":"variable_description","targets":5}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 </div>
 <div id="workspace-preparation-and-data-import-3" class="section level3 hasAnchor">
 <h3>Workspace Preparation and Data Import<a href="data-import-from-pdf-sources.html#workspace-preparation-and-data-import-3" class="anchor-section" aria-label="Anchor link to header"></a></h3>
@@ -645,7 +654,7 @@ <h4>Installing and loading required R packages<a href="data-import-from-pdf-sour
 <span id="cb377-7"><a href="data-import-from-pdf-sources.html#cb377-7" tabindex="-1"></a></span>
 <span id="cb377-8"><a href="data-import-from-pdf-sources.html#cb377-8" tabindex="-1"></a>remotes<span class="sc">::</span><span class="fu">install_github</span>(<span class="fu">c</span>(<span class="st">&quot;ropensci/tabulizerjars&quot;</span>, <span class="st">&quot;ropensci/tabulapdf&quot;</span>), <span class="at">force=</span><span class="cn">TRUE</span>, <span class="at">INSTALL_opts =</span> <span class="st">&quot;--no-multiarch&quot;</span>)</span></code></pre></div>
 <pre><code>## 
-## * checking for file &#39;C:\Users\Jessie PC\AppData\Local\Temp\Rtmp6pdi2m\remotes57483a856b72\ropensci-tabulizerjars-d1924e0/DESCRIPTION&#39; ... OK
+## * checking for file &#39;C:\Users\Jessie PC\AppData\Local\Temp\Rtmpm0tPsD\remotes62a857c530a8\ropensci-tabulizerjars-d1924e0/DESCRIPTION&#39; ... OK
 ## * preparing &#39;tabulizerjars&#39;:
 ## * checking DESCRIPTION meta-information ... OK
 ## * checking for LF line-endings in source and make files and shell scripts
@@ -668,8 +677,8 @@ <h4>Installing and loading required R packages<a href="data-import-from-pdf-sour
 <pre><code>## Warning: restored &#39;glue&#39;</code></pre>
 <pre><code>## 
 ## The downloaded binary packages are in
-##  C:\Users\Jessie PC\AppData\Local\Temp\Rtmp6pdi2m\downloaded_packages
-## * checking for file &#39;C:\Users\Jessie PC\AppData\Local\Temp\Rtmp6pdi2m\remotes5748719d13b\ropensci-tabulapdf-7325a8d/DESCRIPTION&#39; ... OK
+##  C:\Users\Jessie PC\AppData\Local\Temp\Rtmpm0tPsD\downloaded_packages
+## * checking for file &#39;C:\Users\Jessie PC\AppData\Local\Temp\Rtmpm0tPsD\remotes62a83035ee4\ropensci-tabulapdf-7325a8d/DESCRIPTION&#39; ... OK
 ## * preparing &#39;tabulapdf&#39;:
 ## * checking DESCRIPTION meta-information ... OK
 ## * checking for LF line-endings in source and make files and shell scripts
@@ -749,8 +758,8 @@ <h4>Initial data import from PDF file<a href="data-import-from-pdf-sources.html#
 ##   ..- attr(*, &quot;problems&quot;)=&lt;externalptr&gt;</code></pre>
 <p>Exploring further, here is how each dataframe is formatted:</p>
 <div class="sourceCode" id="cb387"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb387-1"><a href="data-import-from-pdf-sources.html#cb387-1" tabindex="-1"></a><span class="fu">datatable</span>(tables[[<span class="dv">1</span>]])</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-9092ff85fe5c2611f10e" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-9092ff85fe5c2611f10e">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30"],[1,null,null,null,2,null,null,3,null,null,4,null,null,null,5,null,null,6,null,null,7,null,null,null,8,9,10,11,12,13],["AAQ_1000",null,null,null,"AAQ_1010",null,null,"AAQ_1020",null,null,"AAQ_1030",null,null,null,"AAQ_1040",null,null,"AAQ_1050",null,null,"AAQ_1060",null,null,null,"VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER"],["Num",null,null,null,"Num",null,null,"Num",null,null,"Num",null,null,null,"Num",null,null,"Num",null,null,"Num",null,null,null,"Char","Num","Num","Char","Char","Num"],[8,null,null,null,8,null,null,8,null,null,8,null,null,null,8,null,null,8,null,null,8,null,null,null,3,8,8,6,15,8],["2.",null,null,null,"2.",null,null,"2.",null,null,"2.",null,null,null,"2.",null,null,"2.",null,null,"2.",null,null,null,"$3.",null,null,null,null,null],["2.",null,null,null,"2.",null,null,"2.",null,null,"2.",null,null,null,"2.",null,null,"2.",null,null,"2.",null,null,null,"$3.",null,null,null,null,null],["In the past 3 days, how much of the time did your asthma keep you from","doing your usual activities at work, school, or at home? 0=None of the","time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All","of the time","During the past 3 days, how often have you had asthma symptoms? 0=Not","at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or","more times per day","During the past 3 days, how often have you used your rescue inhaler or","nebulizer medication ? 0=Not at all, 1=Once per day, 2=2-3 times per day,","3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how many total times did your asthma symptoms","wake you up from sleep? 0=Not at all, 1=1 time in the last 3 days, 2=2-3","times in the last 3 days, 3=4-5 times in the last 3 days, 4=&gt;=6 times in the","last 3 days","How would you rate the amount of impairment you have experienced due","to your asthma in the past 3 days? 0=No impairment, 1=Mild impairment,","2=Moderate impairment, 3=Severe impairment, 4=Very severe impairment","How stressed or frightened were you by your asthma symptoms in the past","3 days? 0=Not at all, 1=Mildly, 2=Moderately, 3=Severely, 4=Very","severely","Why do you think your asthma was worse in the past 3 days compared to","what is normal for you? 0=I have not been worse over the past 3 days. My","asthma symptoms have been usual., 1=Common cold, 2=Allergies,","3=Pollution or chemical irritant, 4=Too little asth","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Num<\/th>\n      <th>Variable<\/th>\n      <th>Type<\/th>\n      <th>Len<\/th>\n      <th>Format<\/th>\n      <th>Informat<\/th>\n      <th>Label<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[1,4]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Num","targets":1},{"name":"Variable","targets":2},{"name":"Type","targets":3},{"name":"Len","targets":4},{"name":"Format","targets":5},{"name":"Informat","targets":6},{"name":"Label","targets":7}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-6c8fd506691d0eea2b01" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-6c8fd506691d0eea2b01">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30"],[1,null,null,null,2,null,null,3,null,null,4,null,null,null,5,null,null,6,null,null,7,null,null,null,8,9,10,11,12,13],["AAQ_1000",null,null,null,"AAQ_1010",null,null,"AAQ_1020",null,null,"AAQ_1030",null,null,null,"AAQ_1040",null,null,"AAQ_1050",null,null,"AAQ_1060",null,null,null,"VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER"],["Num",null,null,null,"Num",null,null,"Num",null,null,"Num",null,null,null,"Num",null,null,"Num",null,null,"Num",null,null,null,"Char","Num","Num","Char","Char","Num"],[8,null,null,null,8,null,null,8,null,null,8,null,null,null,8,null,null,8,null,null,8,null,null,null,3,8,8,6,15,8],["2.",null,null,null,"2.",null,null,"2.",null,null,"2.",null,null,null,"2.",null,null,"2.",null,null,"2.",null,null,null,"$3.",null,null,null,null,null],["2.",null,null,null,"2.",null,null,"2.",null,null,"2.",null,null,null,"2.",null,null,"2.",null,null,"2.",null,null,null,"$3.",null,null,null,null,null],["In the past 3 days, how much of the time did your asthma keep you from","doing your usual activities at work, school, or at home? 0=None of the","time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All","of the time","During the past 3 days, how often have you had asthma symptoms? 0=Not","at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or","more times per day","During the past 3 days, how often have you used your rescue inhaler or","nebulizer medication ? 0=Not at all, 1=Once per day, 2=2-3 times per day,","3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how many total times did your asthma symptoms","wake you up from sleep? 0=Not at all, 1=1 time in the last 3 days, 2=2-3","times in the last 3 days, 3=4-5 times in the last 3 days, 4=&gt;=6 times in the","last 3 days","How would you rate the amount of impairment you have experienced due","to your asthma in the past 3 days? 0=No impairment, 1=Mild impairment,","2=Moderate impairment, 3=Severe impairment, 4=Very severe impairment","How stressed or frightened were you by your asthma symptoms in the past","3 days? 0=Not at all, 1=Mildly, 2=Moderately, 3=Severely, 4=Very","severely","Why do you think your asthma was worse in the past 3 days compared to","what is normal for you? 0=I have not been worse over the past 3 days. My","asthma symptoms have been usual., 1=Common cold, 2=Allergies,","3=Pollution or chemical irritant, 4=Too little asth","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Num<\/th>\n      <th>Variable<\/th>\n      <th>Type<\/th>\n      <th>Len<\/th>\n      <th>Format<\/th>\n      <th>Informat<\/th>\n      <th>Label<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[1,4]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Num","targets":1},{"name":"Variable","targets":2},{"name":"Type","targets":3},{"name":"Len","targets":4},{"name":"Format","targets":5},{"name":"Informat","targets":6},{"name":"Label","targets":7}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>Notice that, although the dataframe format mirrors the PDF table format, the label column is stored across multiple rows with NAs in the other columns of that row because the text was across multiple lines. In our final dataframe, we will want the entire block of text in one cell. We can also remove the “Len”, “Format”, and “Informat” columns because they are not informative and they are not found in every table. Next, we will walk through how to clean up this table using a series of steps in tidyverse.</p>
 </div>
 </div>
@@ -769,8 +778,8 @@ <h3>Cleaning dataframes<a href="data-import-from-pdf-sources.html#cleaning-dataf
 <span id="cb388-10"><a href="data-import-from-pdf-sources.html#cb388-10" tabindex="-1"></a>  <span class="fu">fill</span>(Num, <span class="at">.direction =</span> <span class="st">&quot;down&quot;</span>) </span>
 <span id="cb388-11"><a href="data-import-from-pdf-sources.html#cb388-11" tabindex="-1"></a></span>
 <span id="cb388-12"><a href="data-import-from-pdf-sources.html#cb388-12" tabindex="-1"></a><span class="fu">datatable</span>(cleaned_table1)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-35174a45ec9b983f2f9f" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-35174a45ec9b983f2f9f">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30"],[1,1,1,1,2,2,2,3,3,3,4,4,4,4,5,5,5,6,6,6,7,7,7,7,8,9,10,11,12,13],["AAQ_1000",null,null,null,"AAQ_1010",null,null,"AAQ_1020",null,null,"AAQ_1030",null,null,null,"AAQ_1040",null,null,"AAQ_1050",null,null,"AAQ_1060",null,null,null,"VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER"],["Num",null,null,null,"Num",null,null,"Num",null,null,"Num",null,null,null,"Num",null,null,"Num",null,null,"Num",null,null,null,"Char","Num","Num","Char","Char","Num"],["In the past 3 days, how much of the time did your asthma keep you from","doing your usual activities at work, school, or at home? 0=None of the","time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All","of the time","During the past 3 days, how often have you had asthma symptoms? 0=Not","at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or","more times per day","During the past 3 days, how often have you used your rescue inhaler or","nebulizer medication ? 0=Not at all, 1=Once per day, 2=2-3 times per day,","3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how many total times did your asthma symptoms","wake you up from sleep? 0=Not at all, 1=1 time in the last 3 days, 2=2-3","times in the last 3 days, 3=4-5 times in the last 3 days, 4=&gt;=6 times in the","last 3 days","How would you rate the amount of impairment you have experienced due","to your asthma in the past 3 days? 0=No impairment, 1=Mild impairment,","2=Moderate impairment, 3=Severe impairment, 4=Very severe impairment","How stressed or frightened were you by your asthma symptoms in the past","3 days? 0=Not at all, 1=Mildly, 2=Moderately, 3=Severely, 4=Very","severely","Why do you think your asthma was worse in the past 3 days compared to","what is normal for you? 0=I have not been worse over the past 3 days. My","asthma symptoms have been usual., 1=Common cold, 2=Allergies,","3=Pollution or chemical irritant, 4=Too little asth","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Num<\/th>\n      <th>Variable<\/th>\n      <th>Type<\/th>\n      <th>Label<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":1},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Num","targets":1},{"name":"Variable","targets":2},{"name":"Type","targets":3},{"name":"Label","targets":4}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-cab55136f1340ee4b10b" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-cab55136f1340ee4b10b">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30"],[1,1,1,1,2,2,2,3,3,3,4,4,4,4,5,5,5,6,6,6,7,7,7,7,8,9,10,11,12,13],["AAQ_1000",null,null,null,"AAQ_1010",null,null,"AAQ_1020",null,null,"AAQ_1030",null,null,null,"AAQ_1040",null,null,"AAQ_1050",null,null,"AAQ_1060",null,null,null,"VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER"],["Num",null,null,null,"Num",null,null,"Num",null,null,"Num",null,null,null,"Num",null,null,"Num",null,null,"Num",null,null,null,"Char","Num","Num","Char","Char","Num"],["In the past 3 days, how much of the time did your asthma keep you from","doing your usual activities at work, school, or at home? 0=None of the","time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All","of the time","During the past 3 days, how often have you had asthma symptoms? 0=Not","at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or","more times per day","During the past 3 days, how often have you used your rescue inhaler or","nebulizer medication ? 0=Not at all, 1=Once per day, 2=2-3 times per day,","3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how many total times did your asthma symptoms","wake you up from sleep? 0=Not at all, 1=1 time in the last 3 days, 2=2-3","times in the last 3 days, 3=4-5 times in the last 3 days, 4=&gt;=6 times in the","last 3 days","How would you rate the amount of impairment you have experienced due","to your asthma in the past 3 days? 0=No impairment, 1=Mild impairment,","2=Moderate impairment, 3=Severe impairment, 4=Very severe impairment","How stressed or frightened were you by your asthma symptoms in the past","3 days? 0=Not at all, 1=Mildly, 2=Moderately, 3=Severely, 4=Very","severely","Why do you think your asthma was worse in the past 3 days compared to","what is normal for you? 0=I have not been worse over the past 3 days. My","asthma symptoms have been usual., 1=Common cold, 2=Allergies,","3=Pollution or chemical irritant, 4=Too little asth","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Num<\/th>\n      <th>Variable<\/th>\n      <th>Type<\/th>\n      <th>Label<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":1},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Num","targets":1},{"name":"Variable","targets":2},{"name":"Type","targets":3},{"name":"Label","targets":4}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>We still need to move all of the Label text for each variable into one cell in one row instead of across multiple rows. For this, we can use the <code>unlist()</code> function. Here is a demonstration of how the <code>unlist()</code> function works using just the first variable:</p>
 <div class="sourceCode" id="cb389"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb389-1"><a href="data-import-from-pdf-sources.html#cb389-1" tabindex="-1"></a>cleaned_table1_var1 <span class="ot">&lt;-</span> cleaned_table1 <span class="sc">%&gt;%</span></span>
 <span id="cb389-2"><a href="data-import-from-pdf-sources.html#cb389-2" tabindex="-1"></a>  </span>
@@ -781,16 +790,16 @@ <h3>Cleaning dataframes<a href="data-import-from-pdf-sources.html#cleaning-dataf
 <span id="cb389-7"><a href="data-import-from-pdf-sources.html#cb389-7" tabindex="-1"></a>  <span class="fu">mutate</span>(<span class="at">new_label =</span>  <span class="fu">paste</span>(<span class="fu">unlist</span>(Label), <span class="at">collapse =</span> <span class="st">&quot; &quot;</span>)) </span>
 <span id="cb389-8"><a href="data-import-from-pdf-sources.html#cb389-8" tabindex="-1"></a></span>
 <span id="cb389-9"><a href="data-import-from-pdf-sources.html#cb389-9" tabindex="-1"></a><span class="fu">datatable</span>(cleaned_table1_var1)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-ba031c8486a14880fd83" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-ba031c8486a14880fd83">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4"],[1,1,1,1],["AAQ_1000",null,null,null],["Num",null,null,null],["In the past 3 days, how much of the time did your asthma keep you from","doing your usual activities at work, school, or at home? 0=None of the","time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All","of the time"],["In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, school, or at home? 0=None of the time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All of the time","In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, school, or at home? 0=None of the time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All of the time","In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, school, or at home? 0=None of the time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All of the time","In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, school, or at home? 0=None of the time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All of the time"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Num<\/th>\n      <th>Variable<\/th>\n      <th>Type<\/th>\n      <th>Label<\/th>\n      <th>new_label<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":1},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Num","targets":1},{"name":"Variable","targets":2},{"name":"Type","targets":3},{"name":"Label","targets":4},{"name":"new_label","targets":5}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-0ee87af95ed817537591" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-0ee87af95ed817537591">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4"],[1,1,1,1],["AAQ_1000",null,null,null],["Num",null,null,null],["In the past 3 days, how much of the time did your asthma keep you from","doing your usual activities at work, school, or at home? 0=None of the","time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All","of the time"],["In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, school, or at home? 0=None of the time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All of the time","In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, school, or at home? 0=None of the time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All of the time","In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, school, or at home? 0=None of the time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All of the time","In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, school, or at home? 0=None of the time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All of the time"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Num<\/th>\n      <th>Variable<\/th>\n      <th>Type<\/th>\n      <th>Label<\/th>\n      <th>new_label<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":1},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Num","targets":1},{"name":"Variable","targets":2},{"name":"Type","targets":3},{"name":"Label","targets":4},{"name":"new_label","targets":5}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>We now have all of the text we want in one cell, but we have duplicate rows that we don’t need. We can get rid of these rows by assigning blank values “NA” and then omitting rows that contain NAs.</p>
 <div class="sourceCode" id="cb390"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb390-1"><a href="data-import-from-pdf-sources.html#cb390-1" tabindex="-1"></a>cleaned_table1_var1 <span class="ot">&lt;-</span> cleaned_table1_var1 <span class="sc">%&gt;%</span></span>
 <span id="cb390-2"><a href="data-import-from-pdf-sources.html#cb390-2" tabindex="-1"></a>  <span class="fu">mutate</span>(<span class="fu">across</span>(Variable, na_if, <span class="st">&quot;&quot;</span>))  <span class="sc">%&gt;%</span></span>
 <span id="cb390-3"><a href="data-import-from-pdf-sources.html#cb390-3" tabindex="-1"></a>    <span class="fu">na.omit</span>()</span>
 <span id="cb390-4"><a href="data-import-from-pdf-sources.html#cb390-4" tabindex="-1"></a></span>
 <span id="cb390-5"><a href="data-import-from-pdf-sources.html#cb390-5" tabindex="-1"></a><span class="fu">datatable</span>(cleaned_table1_var1)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-8467b93c74317eaa9fe2" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-8467b93c74317eaa9fe2">{"x":{"filter":"none","vertical":false,"data":[["1"],[1],["AAQ_1000"],["Num"],["In the past 3 days, how much of the time did your asthma keep you from"],["In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, school, or at home? 0=None of the time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All of the time"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Num<\/th>\n      <th>Variable<\/th>\n      <th>Type<\/th>\n      <th>Label<\/th>\n      <th>new_label<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":1},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Num","targets":1},{"name":"Variable","targets":2},{"name":"Type","targets":3},{"name":"Label","targets":4},{"name":"new_label","targets":5}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-cf6393a229582c092c9e" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-cf6393a229582c092c9e">{"x":{"filter":"none","vertical":false,"data":[["1"],[1],["AAQ_1000"],["Num"],["In the past 3 days, how much of the time did your asthma keep you from"],["In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, school, or at home? 0=None of the time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All of the time"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Num<\/th>\n      <th>Variable<\/th>\n      <th>Type<\/th>\n      <th>Label<\/th>\n      <th>new_label<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":1},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Num","targets":1},{"name":"Variable","targets":2},{"name":"Type","targets":3},{"name":"Label","targets":4},{"name":"new_label","targets":5}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>We need to apply this code to the whole dataframe and not just one variable, so we can add <code>group_by(Num)</code> to our cleaning workflow, followed by the code we just applied to our filtered dataframe.</p>
 <div class="sourceCode" id="cb391"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb391-1"><a href="data-import-from-pdf-sources.html#cb391-1" tabindex="-1"></a>cleaned_table1 <span class="ot">&lt;-</span> <span class="fu">data.frame</span>(tables[[<span class="dv">1</span>]]) <span class="sc">%&gt;%</span> <span class="co"># Extract the first table in the list</span></span>
 <span id="cb391-2"><a href="data-import-from-pdf-sources.html#cb391-2" tabindex="-1"></a>  </span>
@@ -815,8 +824,8 @@ <h3>Cleaning dataframes<a href="data-import-from-pdf-sources.html#cleaning-dataf
 <span id="cb391-21"><a href="data-import-from-pdf-sources.html#cb391-21" tabindex="-1"></a>  <span class="fu">na.omit</span>() </span>
 <span id="cb391-22"><a href="data-import-from-pdf-sources.html#cb391-22" tabindex="-1"></a></span>
 <span id="cb391-23"><a href="data-import-from-pdf-sources.html#cb391-23" tabindex="-1"></a><span class="fu">datatable</span>(cleaned_table1)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-5e79551be2d464acdea4" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-5e79551be2d464acdea4">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13"],[1,2,3,4,5,6,7,8,9,10,11,12,13],["AAQ_1000","AAQ_1010","AAQ_1020","AAQ_1030","AAQ_1040","AAQ_1050","AAQ_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER"],["Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num"],["In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, school, or at home? 0=None of the time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All of the time","During the past 3 days, how often have you had asthma symptoms? 0=Not at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how often have you used your rescue inhaler or nebulizer medication ? 0=Not at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how many total times did your asthma symptoms wake you up from sleep? 0=Not at all, 1=1 time in the last 3 days, 2=2-3 times in the last 3 days, 3=4-5 times in the last 3 days, 4=&gt;=6 times in the last 3 days","How would you rate the amount of impairment you have experienced due to your asthma in the past 3 days? 0=No impairment, 1=Mild impairment, 2=Moderate impairment, 3=Severe impairment, 4=Very severe impairment","How stressed or frightened were you by your asthma symptoms in the past 3 days? 0=Not at all, 1=Mildly, 2=Moderately, 3=Severely, 4=Very severely","Why do you think your asthma was worse in the past 3 days compared to what is normal for you? 0=I have not been worse over the past 3 days. My asthma symptoms have been usual., 1=Common cold, 2=Allergies, 3=Pollution or chemical irritant, 4=Too little asth","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Num<\/th>\n      <th>Variable<\/th>\n      <th>Type<\/th>\n      <th>Label<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":1},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Num","targets":1},{"name":"Variable","targets":2},{"name":"Type","targets":3},{"name":"Label","targets":4}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-b60229acd66991f064d6" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-b60229acd66991f064d6">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13"],[1,2,3,4,5,6,7,8,9,10,11,12,13],["AAQ_1000","AAQ_1010","AAQ_1020","AAQ_1030","AAQ_1040","AAQ_1050","AAQ_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER"],["Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num"],["In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, school, or at home? 0=None of the time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All of the time","During the past 3 days, how often have you had asthma symptoms? 0=Not at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how often have you used your rescue inhaler or nebulizer medication ? 0=Not at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how many total times did your asthma symptoms wake you up from sleep? 0=Not at all, 1=1 time in the last 3 days, 2=2-3 times in the last 3 days, 3=4-5 times in the last 3 days, 4=&gt;=6 times in the last 3 days","How would you rate the amount of impairment you have experienced due to your asthma in the past 3 days? 0=No impairment, 1=Mild impairment, 2=Moderate impairment, 3=Severe impairment, 4=Very severe impairment","How stressed or frightened were you by your asthma symptoms in the past 3 days? 0=Not at all, 1=Mildly, 2=Moderately, 3=Severely, 4=Very severely","Why do you think your asthma was worse in the past 3 days compared to what is normal for you? 0=I have not been worse over the past 3 days. My asthma symptoms have been usual., 1=Common cold, 2=Allergies, 3=Pollution or chemical irritant, 4=Too little asth","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Num<\/th>\n      <th>Variable<\/th>\n      <th>Type<\/th>\n      <th>Label<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":1},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Num","targets":1},{"name":"Variable","targets":2},{"name":"Type","targets":3},{"name":"Label","targets":4}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>Ultimately, we need to clean up each dataframe in the list the same way, and we need all of the dataframes to be in one dataframe, instead of in a list. There are a couple of different ways to do this. Both rely on the code shown above for cleaning up each dataframe. Option #1 uses a for loop, while Option #2 uses application of a function on the list of dataframes. Both result in the same ending dataframe!</p>
 <p><strong>Option #1</strong></p>
 <div class="sourceCode" id="cb392"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb392-1"><a href="data-import-from-pdf-sources.html#cb392-1" tabindex="-1"></a><span class="co"># Create a dataframe for storing variables</span></span>
@@ -839,8 +848,8 @@ <h3>Cleaning dataframes<a href="data-import-from-pdf-sources.html#cleaning-dataf
 <span id="cb392-18"><a href="data-import-from-pdf-sources.html#cb392-18" tabindex="-1"></a></span>
 <span id="cb392-19"><a href="data-import-from-pdf-sources.html#cb392-19" tabindex="-1"></a><span class="co"># View resulting dataframe</span></span>
 <span id="cb392-20"><a href="data-import-from-pdf-sources.html#cb392-20" tabindex="-1"></a><span class="fu">datatable</span>(variables)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-37e05e54199846aa2717" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-37e05e54199846aa2717">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80","81","82","83","84","85","86","87","88","89","90","91","92","93","94","95","96","97","98","99","100","101","102","103","104","105","106","107","108","109","110","111","112","113","114","115","116","117","118","119","120","121","122","123","124","125","126","127","128","129","130","131","132","133","134","135","136","137","138","139","140","141","142","143","144","145","146","147","148","149","150","151","152","153","154","155","156","157","158","159","160","161","162","163","164","165","166","167","168","169","170","171","172","173","174","175","176","177","178","179","180","181","182","183","184","185","186","187","188","189","190","191","192","193","194","195","196","197","198","199","200","201","202","203","204","205","206","207","208","209","210","211","212","213","214","215","216","217","218","219","220","221","222","223","224","225","226","227","228","229","230","231","232","233","234","235","236","237","238","239","240","241","242","243","244","245","246","247","248","249","250","251","252","253","254","255","256","257","258","259","260","261","262","263","264","265","266","267","268","269","270","271","272","273","274","275","276","277","278","279","280","281","282","283","284","285","286","287","288","289","290","291","292","293","294","295","296","297","298","299","300","301","302","303","304","305","306","307","308","309","310","311","312","313","314","315","316","317","318","319","320","321","322","323","324","325","326","327","328","329","330","331","332","333","334","335","336","337","338","339","340","341","342","343","344","345","346","347","348","349","350","351","352","353","354","355","356","357","358","359","360","361","362","363","364","365","366","367","368","369","370","371","372","373","374","375","376","377","378","379","380","381","382","383","384","385","386","387","388","389","390","391","392","393","394","395","396","397","398","399","400","401","402","403","404","405","406","407","408","409","410","411","412","413","414","415","416","417","418","419","420","421","422","423","424","425","426","427","428","429","430","431","432","433","434","435","436","437","438","439","440","441","442","443","444","445","446","447","448","449","450","451","452","453","454","455","456","457","458","459","460","461","462","463","464","465","466","467","468","469","470","471","472","473","474","475","476","477","478","479","480","481","482","483","484","485","486","487","488","489","490","491","492","493","494","495","496","497","498","499","500","501","502","503","504","505","506","507","508","509","510","511","512","513","514","515","516","517","518","519","520","521","522","523","524","525","526","527","528","529","530","531","532","533","534","535","536","537","538","539","540","541","542","543","544","545","546","547","548","549","550","551","552","553","554","555","556","557","558","559","560","561","562","563","564","565","566","567","568","569","570","571","572","573","574","575","576","577","578","579","580","581","582","583","584","585","586","587","588","589","590","591","592","593","594","595","596","597","598","599","600","601","602","603","604","605","606","607","608","609","610","611","612","613","614","615","616","617","618","619","620","621","622","623","624","625","626","627","628","629","630","631","632","633","634","635","636","637","638","639","640","641","642","643","644","645","646","647","648","649","650","651","652","653","654","655","656","657","658","659","660","661","662","663","664","665","666","667","668","669","670","671","672","673","674","675","676","677","678","679","680","681","682","683","684","685","686","687","688","689","690","691","692","693","694","695","696","697","698","699","700","701","702","703","704","705","706","707","708","709","710","711","712","713","714","715","716","717","718","719","720","721","722","723","724","725","726","727","728","729","730","731","732","733","734","735","736","737","738","739","740","741","742","743","744","745","746","747","748","749","750","751","752","753","754","755","756","757","758","759","760","761","762","763","764","765","766","767","768","769","770","771","772","773","774","775","776","777","778","779","780","781","782","783","784","785","786","787","788","789","790","791","792","793","794","795","796","797","798","799","800","801","802","803","804","805","806","807","808","809","810","811","812","813","814","815","816","817","818","819","820","821","822","823","824","825","826","827","828","829","830","831","832","833","834","835","836","837","838","839","840","841","842","843","844","845","846","847","848","849","850","851","852","853","854","855","856","857","858","859","860","861","862","863","864","865","866","867","868","869","870","871","872","873","874","875","876","877","878","879","880","881","882","883","884","885","886","887","888","889","890","891","892","893","894","895","896","897","898","899","900","901","902","903","904","905","906","907","908","909","910","911","912","913","914","915","916","917","918","919","920","921","922","923","924","925","926","927","928","929","930","931","932","933","934","935","936","937","938","939","940","941","942","943","944","945","946","947","948","949","950","951","952","953","954","955","956","957","958","959","960","961","962","963","964","965","966","967","968","969","970","971","972","973","974","975","976","977","978","979","980","981","982","983","984","985","986","987","988","989","990","991","992","993","994","995","996","997","998","999","1000","1001","1002","1003","1004","1005","1006","1007","1008","1009","1010","1011","1012","1013","1014","1015","1016","1017","1018","1019","1020","1021","1022","1023","1024","1025","1026","1027","1028","1029","1030","1031","1032","1033","1034","1035","1036","1037","1038","1039","1040","1041","1042","1043","1044","1045","1046","1047","1048","1049","1050","1051","1052","1053","1054","1055","1056","1057","1058","1059","1060","1061","1062","1063","1064","1065","1066","1067","1068","1069","1070","1071","1072","1073","1074","1075","1076","1077","1078","1079","1080","1081","1082","1083","1084","1085","1086","1087","1088","1089","1090","1091","1092","1093","1094","1095","1096","1097","1098","1099","1100","1101","1102","1103","1104","1105","1106","1107","1108","1109","1110","1111","1112","1113","1114","1115","1116","1117","1118","1119","1120","1121","1122","1123","1124","1125","1126","1127","1128","1129","1130","1131","1132","1133","1134","1135","1136","1137","1138","1139","1140","1141","1142","1143","1144","1145","1146","1147","1148","1149","1150","1151","1152","1153","1154","1155","1156","1157","1158","1159","1160","1161","1162","1163","1164","1165","1166","1167","1168","1169","1170","1171","1172","1173","1174","1175","1176","1177","1178","1179","1180","1181","1182","1183","1184","1185","1186","1187","1188","1189","1190"],[1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,1,2,3,4,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,1,2,3,4,5,6,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28],["AAQ_1000","AAQ_1010","AAQ_1020","AAQ_1030","AAQ_1040","AAQ_1050","AAQ_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ABP_1000","ABP_1010","ABP_1020","ABP_1030","ABP_1040","ABP_1050","ABP_1060","ABP_1070","ABP_1080","ABP_1090","ABP_1100","ABP_1110","ABP_1120","ABP_1130","ABP_1140","ABP_1150","ABP_1160","ABP_1170","ABP_1180","ABP_1190","ABP_1200","ABP_1210","ABP_1220","ABP_1230","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ACT_1","ACT_2","ACT_3","ACT_4","ACT_5","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ICD9_CAT","AEC_1000","AEC_1010","AEC_1040","AEC_1050","AEC_1060","AEC_1070","AEC_1080","AEC_1090","AEC_1100","AEC_1110","AEC_1120","AEC_1020","AEC_1030","VNUM_C","VNUM","ICD9LONG","ICD9SHRT","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","AHA_1000","AHA_1020","AHA_1030","AHA_1040","AHA_1050","AHA_1060","AHA_1070","AHA_1090","AHA_1100","AHA_1110","AHA_1120","AHA_1130","AHA_1140","AHA_1150","AHA_1160","AHA_1170","AHA_1180","AHA_1190","AHA_1200","AHA_1210","AHA_1220","AHA_1250","AHA_1260","AHA_1270","AHA_1280","AHA_1290","AHA_1300","AHA_1310","AHA_1320","AHA_1330","AHA_1340","AHA_1350","AHA_1360","AHA_1370","AHA_1380","AHA_1390","AHA_1400","AHA_1410","AHA_1420","AHA_1430","AHA_1440","AHA_1450","AHA_1460","AHA_1470","AHA_1500","AHA_1570","AHA_1580","AHA_1590","AHA_1600","AHA_1730","AHA_1740","AHA_1750","AHA_1760","AHA_1770","AHA_1780","AHA_1790","AHA_1800","AHA_1810","AHA_1820","AHA_1830","AHA_1840","AHA_1850","AHA_1890","AHA_1900","VNUM_C","VNUM","VDATE","DATE_VAPE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ASU_1000","ASU_1010","ASU_1020","ASU_1030","ASU_1040","ASU_1050","ASU_1060","ASU_1070","ASU_1080","ASU_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","BAS_1000","BAS_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","BMI","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAGE","CHG_1000","CHG_1010","CHG_1020","CHG_1030","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","CME_1040","CME_1050","CME_1055","CME_1000","CME_1010","CME_1020","CME_1030","CME_1080","CME_1090","CME_1060","CME_1070","VNUM_C","VNUM","GEN_NAME","CLASS","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","COM_1000","COM_1010","COM_1020","COM_1030","COM_1040","COM_1050","COM_1060","COM_1070","COM_1080","COM_1090","COM_1100","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","CTX_1000","CTX_1010","CTX_1030","CTX_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E0A_1000","E0A_1020","E0A_1030","E0A_1040","E0A_1043","E0A_1045","E0A_1050","E0A_1060","E0A_1070","E0A_1073","E0A_1075","E0A_1080","E0A_1090","E0A_1100","E0A_1110","E0A_1120","E0A_1130","E0A_1140","E0A_1145","E0A_1150","E0A_1160","E0A_1170","E0A_1180","E0A_1190","E0A_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E0B_1000","E0B_1010","E0B_1020","E0B_1030","E0B_1040","E0B_1050","E0B_1060","E0B_1065","E0B_1070","E0B_1080","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E1_1000","E1_1020","E1_1030","E1_1040","E1_1050","E1_1060","E1_1065","E1_1070","E1_1080","E1_1090","E1_1100","E1_1110","E1_1120","E1_1130","E1_1140","E1_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E2_1000","E2_1010","E2_1020","E2_1030","E2_1040","E2_1050","E2_1060","E2_1070","E2_1080","E2_1090","E2_1100","E2_1110","E2_1120","E2_1130","E2_1140","E2_1145","E2_1150","E2_1160","E2_1170","E2_1173","E2_1175","E2_1180","E2_1190","E2_1200","E2_1210","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E3_1000","E3_1010","E3_1020","E3_1030","E3_1070","E3_1080","E3_1090","E3_1100","E3_1110","E3_1120","E3_1130","E3_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E4_1000","E4_1010","E4_1020","E4_1030","E4_1050","E4_1060","E4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E5_1000","E5_1010","E5_1020","E5_1030","E5_1035","E5_1040","E5_1050","E5_1060","E5_1065","E5_1070","E5_1080","E5_1085","E5_1090","E5_1100","E5_1110","E5_1120","E5_1130","E5_1140","E5_1150","E5_1160","E5_1170","E5_1180","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ENO_1000","ENO_1010","ENO_1020","ENO_1040","ENO_1050","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","HEQ_1000","HEQ_1010","HEQ_1020","HEQ_1030","HEQ_1040","HEQ_1050","HEQ_1060","HEQ_1070","HEQ_1080","HEQ_1090","HEQ_1100","HEQ_1110","HEQ_1120","HEQ_1130","HEQ_1140","HEQ_1150","HEQ_1160","HEQ_1170","HEQ_1180","HEQ_1190","HEQ_1200","HEQ_1210","HEQ_1220","HEQ_1230","HEQ_1240","HEQ_1250","HEQ_1260","HEQ_1270","HEQ_1280","HEQ_1290","HEQ_1300","HEQ_1310","HEQ_1320","HEQ_1330","HEQ_1340","HEQ_1350","HEQ_1360","HEQ_1370","HEQ_1380","HEQ_1390","HEQ_1400","HEQ_1410","HEQ_1420","HEQ_1430","HEQ_1440","HEQ_1450","HEQ_1460","HEQ_1470","HEQ_1480","HEQ_1490","HEQ_1500","HEQ_1510","HEQ_1520","HEQ_1530","HEQ_1540","HEQ_1550","HEQ_1560","HEQ_1570","HEQ_1580","HEQ_1590","HEQ_1600","HEQ_1610","HEQ_1620","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SEI_1000","SEI_1010","SEI_1020","SEI_1030","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","IGE","RE1","RE5","RE72","RE74","RMX1","RI6","RGX2","RTX4","RTX6","RWX1","RW3","RD2","RD1","N_ALLERGENPOS","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","LAB_1000","LAB_1010","LAB_1020","LAB_1030","LAB_1040","LAB_1050","LAB_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PLE_1000","PLE_1010","PLE_1020","PLE_1030","PLE_1040","PLE_1050","PLE_1060","PLE_1070","PLE_1080","PLE_1090","PLE_1100","PLE_1110","PLE_1130","PLE_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MTH_1000","MTH_1010","MTH_1020","MTH_1030","MTH_1040","MTH_1050","MTH_1060","MTH_1070","MTH_1080","MTH_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MAD_1000","MAD_1010","MAD_1020","MAD_1030","MAD_1040","MAD_1050","MAD_1060","MAD_1070","MAD_1090","MAD_1100","MAD_1110","MAD_1120","MAD_1130","MAD_1140","MAD_1150","MAD_1160","MAD_1170","MAD_1180","MAD_1190","MAD_1200","MAD_1220","MAD_1230","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MCA_1000","MCA_1010","MCA_1050","MCA_1060","MCA_1070","MCA_1080","MCA_1100","MCA_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MCP_1000","MCP_1010","MCP_1030","MCP_1040","MCP_1050","MCP_1060","MCP_1070","MCP_1080","MCP_1090","MCP_1100","MCP_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PA4_1000","PA4_1010","PA4_1020","PA4_1030","PA4_1040","PA4_1050","PA4_1060","PA4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PTX_1000","PTX_1010","PTX_1020","PTX_1030","PTX_1040","PTX_1050","PTX_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","VNUM","PERIOSTIN","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PI4_1000","PI4_1010","PI4_1020","PI4_1030","PI4_1040","PI4_1050","PI4_1060","PI4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PRG_1000","PRG_1010","PRG_1020","PRG_1030","PRG_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAD_1000","PAD_1010","PAD_1020","PAD_1030","PAD_1040","PAD_1050","PAD_1060","PAD_1070","PAD_1080","PAD_1090","PAD_1100","PAD_1110","PAD_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAL_1000","PAL_1010","PAL_1020","PAL_1030","PAL_1040","PAL_1050","PAL_1060","PAL_1070","PAL_1080","PAL_1090","PAL_1100","PAL_1110","PAL_1120","PAL_1130","PAL_1150","PAL_1160","PAL_1170","PAL_1180","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PTR_1000","PTR_1010","PTR_1050","PTR_1060","PTR_1100","PTR_1140","PTR_1180","PTR_1220","PTR_1260","PTR_1300","PTR_1340","PTR_1380","PTR_1390","PTR_1430","PTR_1470","PTR_1480","PTR_1490","PTR_1500","PTR_1535","PTR_1540","PTR_1550","PTR_1560","PTR_1600","PTR_1610","PTR_1620","PTR_1630","PTR_1670","PTR_1710","PTR_1750","PTR_1790","PTR_1830","PTR_1870","PTR_1910","VNUM_C","VNUM","VDATE","DATE_Q02","DATE_Q03","DATE_Q04","DATE_Q05","DATE_Q07","DATE_Q08","DATE_Q09","DATE_Q10","DATE_Q11","DATE_Q12","DATE_Q13","DATE_Q14","DATE_Q15","DATE_Q16","DATE_Q17","DATE_Q18","DATE_Q19","DATE_Q20","DATE_Q21","DATE_Q22","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","P10_1000","P10_1010","P10_1020","P10_1030","P10_1040","P10_1050","P10_1060","P10_1070","P10_1080","P10_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PCH_1000","PCH_1010","PCH_1020","PCH_1030","PCH_1040","PCH_1050","PCH_1060","PCH_1070","PCH_1080","PCH_1090","PCH_1100","PCH_1110","PCH_1120","PCH_1130","PCH_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","IAQ_1000","IAQ_1010","IAQ_1020","IAQ_1030","IAQ_1040","IAQ_1050","IAQ_1060","IAQ_1070","IAQ_1080","IAQ_1090","IAQ_1100","IAQ_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","REGIMEN_PERIOD_1","REGIMEN_PERIOD_2","REGIMEN_PERIOD_3","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","REG_1080","REG_1150","AGE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SER_1010","SER_1020","SER_1030","SER_1040","SER_1050","SER_1060","SER_1070","SER_1100","SER_1110","SER_1120","SER_1130","SER_1140","SER_1150","SER_1160","SER_1170","SER_1180","SER_1190","SER_1200","SER_1210","SER_1220","SER_1240","SER_1250","SER_1000","SER_1080","SER_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PSE_1060","PSE_1070","PSE_1080","PSE_1090","PSE_1100","PSE_1110","PSE_1130","PSE_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SIG_1000","SIG_1010","SIG_1020","SIG_1030","SIG_1040","SIG_1050","SIG_1060","SIG_1080","SIG_1090","SIG_1100","SIG_1110","SIG_1120","SIG_1130","SIG_1140","SIG_1150","SIG_1160","SIG_1170","SIG_1180","SIG_1190","SIG_1200","SIG_1210","SIG_1220","SIG_1230","SIG_1240","SIG_1250","SIG_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SFU_1000","SFU_1010","SFU_1020","SFU_1030","SFU_1040","SFU_1050","SFU_1060","SFU_1070","SFU_1080","SFU_1090","SFU_1100","SFU_1110","SFU_1120","SFU_1130","SFU_1140","SFU_1150","SFU_1160","SFU_1170","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SNQ_1000","SNQ_1010","SNQ_1020","SNQ_1030","SNQ_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SPI_1010","SPI_1020","SPI_1040","SPI_1050","SPI_1060","SPI_1070","VNUM_C","VNUM","VDATE","SPI_1030","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","AM_TRIAL_TYPE","AM_FVC","AM_FEV1","AM_PEF","AM_FEF2575","AM_FET","DRY_1","DRY_2","DRY_3","DRY_4","DRY_5","DRY_6","DRY_7","DRY_8","DRY_9","VNUM_C","VNUM","VDATE","DDATE","PM_TRIAL_TYPE","PM_FVC","PM_FEV1","PM_PEF","PM_FEF2575","PM_FET","DRY_10","DRY_11","DRY_12","DRY_13","DRY_14","DRY_15","DRY_16","DRY_17","DRY_18","DRY_19","DRY_20","DRY_21","DRY_22","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SLA_1010","SLA_1020","SLA_1030","SLA_1000","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SRE_1010","SRE_1020","SRE_1030","SRE_1040","SRE_1050","SRE_1060","SRE_1070","SRE_1080","SRE_1090","SRE_1100","SRE_1000","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SPU_1000","SPU_1010","SPU_1020","SPU_1030","SPU_1040","SPU_1050","SPU_1060","SPU_1070","SPU_1080","SPU_1090","SPU_1100","SPU_1110","SPU_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SAD_1000","SAD_1010","SAD_1020","SAD_1030","SAD_1040","SAD_1050","SAD_1060","SAD_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SCH_1000","SCH_1010","SCH_1020","SCH_1030","SCH_1040","SCH_1050","SCH_1055","SCH_1060","SCH_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TER_1000","TER_1010","TER_1020","TER_1040","TER_1050","TER_1070","TER_1080","TER_1090","TER_1100","TER_1110","TER_1120","TER_1130","TER_1140","TER_1150","TER_1160","TER_1170","TER_1180","TER_1185","TER_1200","TER_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TXF_1000","TXF_1010","TXF_1020","TXF_1030","TXF_1040","TXF_1050","TXF_1060","TXF_1070","TXF_1080","TXF_1090","TXF_1100","TXF_1110","TXF_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TCH_1000","TCH_1010","TCH_1020","TCH_1030","TCH_1040","TCH_1050","TCH_1060","TCH_1080","TCH_1090","TCH_1100","TCH_1110","TCH_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","WSH_1000","WSH_1010","WSH_1020","WSH_1040","WSH_1050","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","WPA_1000","WPA_1010","WPA_1020","WPA_1030","WPA_1040","WPA_1050","WPA_1060","WPA_1070","WPA_1080","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","HOW_SICK_TODAY","RUNNY_NOSE","PLUGGED_NOSE","SNEEZING","SORE_THROAT","SCRATCHY_THROAT","COUGH","HOARSENESS","HEAD_CONGESTION","CHEST_CONGESTION","FEELING_TIRED","THINK_CLEARLY","SLEEP_WELL","BREATH_EASILY","WALK","DAILY_ACTIVITIES","WORK_OUTSIDE_HOME","WORK_INSIDE_HOME","INTERACT","LIVE_LIFE","MY_COLD_IS","VNUM_C","VNUM","VDATE","WDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER"],["Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Char","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Char","Char","Num","Num","Num","Num","Char","Char","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Char","Num","Char","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Char","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num"],["In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, school, or at home? 0=None of the time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All of the time","During the past 3 days, how often have you had asthma symptoms? 0=Not at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how often have you used your rescue inhaler or nebulizer medication ? 0=Not at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how many total times did your asthma symptoms wake you up from sleep? 0=Not at all, 1=1 time in the last 3 days, 2=2-3 times in the last 3 days, 3=4-5 times in the last 3 days, 4=&gt;=6 times in the last 3 days","How would you rate the amount of impairment you have experienced due to your asthma in the past 3 days? 0=No impairment, 1=Mild impairment, 2=Moderate impairment, 3=Severe impairment, 4=Very severe impairment","How stressed or frightened were you by your asthma symptoms in the past 3 days? 0=Not at all, 1=Mildly, 2=Moderately, 3=Severely, 4=Very severely","Why do you think your asthma was worse in the past 3 days compared to what is normal for you? 0=I have not been worse over the past 3 days. My asthma symptoms have been usual., 1=Common cold, 2=Allergies, 3=Pollution or chemical irritant, 4=Too little asth","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Are you currently retired? 1=Yes,0=No","Are you retired because of asthma? 1=Yes,0=No","Are you currently unemployed? 1=Yes,0=No","Are you unemployed because of asthma? 1=Yes,0=No","Do you get paid to do work? 1=Yes,0=No","How much does your asthma bother you at your paid work? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Overall, how much does your asthma bother you when you do jobs around the house? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Overall, how much does your asthma bother your social life? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Overall, how much does your asthma bother your personal life? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Are you involved in leisure activities, such as: walking for pleasure, sports, exercise, travelling, taking vacations? 1=Yes,0=No","When involved in leisure activities, how much does your asthma bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Would you say that you can't do some of these sorts of things because of asthma? 1=Yes,0=No","How much does your asthma bother you when you sleep? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much does the cost of your asthma medicines bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Do you get free prescriptions? 1=Yes,0=No","How much does the inconvenience or embarrassment of taking your asthma medicines bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much do coughs and colds bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Feeling upset is also a bother.  Does your asthma make you feel anxious, depressed, tired, or helpless? 1=Yes,0=No","Feeling upset is also a bother.  Does your asthma make you feel anxious, depressed, tired, or helpless? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will have an asthma attack when visiting a new place? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will catch a cold? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will let others down? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that your health may get worse in the future? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you won't be able to cope with an asthma attack? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past 4 weeks, how much of the time did your asthma keep you from getting as much done at work, school or at home? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often have your had shortness of breath? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often did your asthma symptoms wake you up at night or earlier than usual in the morning? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often have you used your rescue inhaler or nebulizer medication? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","How would you rate your asthma control during the past 4 weeks? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","ICD-9 category","Description of Adverse Event","ICD9 Code","Ongoing at current visit","Type (1=Intermitent; 2=Continuous)","Severity (1=Mild; 2=Moderate; 3=Severe)","Serious (0=No, 1=Yes)","Likelyhood of Relationship to Study Drug(s) (1=None; 2=Unlikely (remote); 3=Possible; 4=Probable)","Change in Study Drugs (1=Unchanged; 2=Altered)","Outcome (1=Completely recovered; 2=Recovered, but with lasting effects; 3=Death)","Treatment Required (1=None; 2=Medication; 3=Hospitalization; 4=Other)","Ongoing at final visit","Date Started","Date Stopped","Visit Number (character)","Visit Number (numeric)","Long ICD-9 description","Short ICD-9 description","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","ASTHMA HISTORY1.Approximately how old were you when chest symptoms suggesting asthma first appeared?","Did these symptoms appear immediately after or as a result of:1a. a respiratory infection such as a cold or pneumonia? 1=Yes, 0=No, 8=Don't Know","an occupational or job change? 1=Yes, 0=No, 8=Don't Know","a household move? 1=Yes, 0=No, 8=Don't Know","a pregnancy? 1=Yes, 0=No, 8=Don't Know","a hormonal change ? 1=Yes, 0=No, 8=Don't Know","How old were you when a doctor first diagnosed you with asthma?","Have any of your immediate blood relatives been told by a physician that they have asthma? 1=Yes, 0=No, 8=Don't Know","Father  (1=Yes, 0=No, 8=Don't Know)","Brother(s) or Sister(s)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Child(ren)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","ASTHMA SYMPTOMS1.How do you categorize your asthma symptoms throughout the course of the year? 1=Relatively the same all year, 2=Vary by season","Winter? 1=Yes,0=No","Spring? 1=Yes,0=No","Summer? 1=Yes,0=No","Fall? 1=Yes,0=No","In the last 12 months, how many...  5a.Asthma episodes have you had that required emergency care or an unscheduled office visit?","Overnight hospitalizations have you had due to asthma?","Courses of systemic corticosteroid therapy  for asthma have you taken?","Days of work, school, or housework have you missed due to asthma?","In the past 3 months, how many days of work, school, or housework have you missed due to asthma?","Have you ever been admitted to an intensive care unit for asthma? 1=Yes,0=No","How many times have you been admitted to an intensive care unit for asthma?","Have you ever had invasive mechanical ventilation? 1=Yes, 0=No, 8=Don't Know","Have you ever had non-invasive mechanical ventilation? 1=Yes, 0=No, 8=Don't Know","Do any of the following currently provoke your asthma? 1=Yes, 0=No, 8=Don't Know","Menstrual cycle(If participant is male or a postmenopausal female, leave blank.) (1=Yes, 0=No, 8=Don't Know)","Aspirin or non-steroidal anti-inflammatory drugs (e.g., Aleve, Motrin) (1=Yes, 0=No, 8=Don't Know)","Respiratory infections (e.g., colds)  (1=Yes, 0=No, 8=Don't Know)","Irritants (e.g., pollution, odors, perfumes, chemicals, household cleaners) (1=Yes, 0=No, 8=Don't Know)","Weather conditions (e.g., change in weather, humidity)  (1=Yes, 0=No, 8=Don't Know)","Exposure to cold air  (1=Yes, 0=No, 8=Don't Know)","Emotional factors (e.g., stress, laughing)  (1=Yes, 0=No, 8=Don't Know)","Tobacco smoke  (1=Yes, 0=No, 8=Don't Know)","Food additives/preservatives (e.g., MSG, sulfites)  (1=Yes, 0=No, 8=Don't Know)","Allergies (e.g., dust, animals, pollens)(1=Yes, 0=No, 8=Don't Know)","Other (1=Yes,0=No)","To which of the following did a doctor or other health practitioner say you were allergic? 1=Yes, 0=No, 8=Don't Know","Foods  (1=Yes, 0=No, 8=Don't Know)","Things you breathe in or are exposed to (e.g., dust, pollens, molds, animal fur, feathers, dander)  (1=Yes, 0=No, 8=Don't Know)","Stinging insects such as bees or wasps (1=Yes, 0=No, 8=Don't Know)","Latex  (1=Yes, 0=No, 8=Don't Know)","Other  (1=Yes,0=No)","Have you ever had eczema / atopic dermatitis ? 1=Yes, 0=No, 8=Don't Know","If YES, was your eczema diagnosed by a doctor? 1=Yes,0=No","Have any of your immediate blood relatives been told by a physician that they have allergies/eczema/hay fever? 1=Yes, 0=No, 8=Don't Know","Father  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Brother(s) or Sister(s)(1=Yes, 0=No, 8=Don't Know, 9=N/A)","Child(ren)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Did you grow up in a household where you were exposed to tobacco smoke? 1=Yes,0=No","Do you currently smoke? 1=Yes,0=No","Record smoking history in pack-years*. __ __ . __ pack-years","Were you ever a smoker? 1=Yes,0=No","Record smoking history in pack-years*. __ __ . __ pack-years","Do you currently live in a household where you are exposed to tobacco smoke? 1=Yes,0=No","Do you currently vape  or use a hookah ? 1=Yes,0=No","How frequently do you vape or use a hookah? 1=Infrequently , 2=Occasionally , 3=Weekly , 4=Daily","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many days a week","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many times a day","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many years have y","Have you ever vaped or used a hookah in the past? 1=Yes,0=No","Have you ever vaped or used a hookah in the past?  If NO, skip to Q17. Approximately how many years did you vape or use a hookah?  (1-99 years)","Do you currently live in a household where you are exposed to others vaping or using a hookah? 1=Yes,0=No","Do you spend time in social settings  where you are exposed to others vaping or using a hookah? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","When was the last time that you vaped or used a hookah? (number of days from Visit 1)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","How many days were you bothered by coughing during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe was your coughing during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by wheezing during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe was your wheezing during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by shortness of breath during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","Over average, how severe was your shortness of breath during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you awakened at night during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how much of a problem was being awakened at night during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by side effects of your asthma medication during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe were the side effects of your asthma medication during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Participants baseline peak flow (PEF) value (0-999 L/M) First visit (Visit 0A, or Visit 1): PEF (FEF Max) from prebronchodilator (baseline) spirometry at first visit (convert to L/M)","Participants baseline rescue use value (0-99 puffs/day) First visit (Visit 0A, or Visit 1): Self-reported average daily use of albuterol during the 14 days prior to first visit.","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","body mass index computed from height and weight measured at Visit 1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Page Number","Reason for change in scheduled medications status (1=Adverse Event, 2=Other (specify))","Reason for change in scheduled medications status Related adverse event number (0-99)","Current status of participants scheduled medications (1=Discontinued, 2=Resumed)","Date change took effect","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Code to identify a drug unit of measure; (Units)","Code to identify how frequently a drug is taken or administered. (Frequency)","Code to identify the route used to administer a drug. (Route)","Name of Medication","This number corresponds to the ID number of the AHFS monograph. (Medication Code)","Related Event","Medication Dose","Ongoing at current visit","Ongoing at final visit","Start Date","Stop Date","Visit Number (character)","Visit Number (numeric)","GEN_DRUG_NAME","CLASS_TEXT","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Diary and Peak Flow Compliance: Number of full days since the last visit (0-999 days)","Diary and Peak Flow Compliance: Number of days where AM and PM scheduled sessions are complete (AM and PM PEF and all diary questions for AM and PM answered) (0-999 days)","Diary and Peak Flow Compliance: Percent compliance If the compliance value in Q1c is less than 75Pct, re-emphasize the importance of completing scheduled diary assessments and peak flows. (0-150.0Pct)","Scheduled Respimat Compliance (Visits 2-9 only): Number of scheduled puffs since the last visit (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Number of remaining puffs on scheduled Respimat Reference all returned Respimats (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Number of puffs taken (60 x (# used Respimats)) - Q2b (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Percent compliance=Q2c/Q2a x 100 If the participant took less than 75Pct of the scheduled Respimat puffs, re-emphasize the importance of maintaining the daily dosing schedule. (0-150.0Pct)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of scheduled puffs since the last visit (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of remaining puffs reflected on scheduled Twisthaler counters Total the values reflected on all counters for all returned Twisthalers (i.e., out of their pouches). (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of puffs taken (60 x (# used Twisthalers)) - Q3b (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Percent compliance=Q3c/Q3a x 100 If the participant took less than 75Pct of the scheduled Twisthaler puffs, re-emphasize the importance of maintaining the daily dosing schedule. (0-150.0Pct)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","1=mometasone, 2=placebo","Blinded Scheduled Twisthaler Contents - How sure are you about your answer in Q1? 1=Absolutely sure - I know what the Twisthaler/MDI contains, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","1=tiotropium, 2=placebo","Blinded Scheduled Respimat Contents - How sure are you about your answer in Q4? 1=Absolutely sure - I know what the Respimat capsules contain, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has the participant or parent/legal guardian signed the SIENA Informed Consent document? 1=Yes,0=No","Ages 12-17 Only: Has the participant signed and dated the assent form or, if the participant is less than the local age of assent, has the participant given verbal assent? 1=Yes,0=No","Has the participant used an inhaled corticosteroid in the past 6 weeks? 1=Yes,0=No","Will the participant be 12 years of age, or older, as of Visit 1? 1=Yes,0=No","Has the participant used an inhaled corticosteroid or combination inhaled corticosteroid/LABA in the past 3 weeks? 1=Yes,0=No","If NO, has the participant used a leukotriene modifier in the past 3 weeks? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used inhaled corticosteroid therapy?","Over the past 3 months, on average, how many days per week has the participant used inhaled corticosteroid therapy: &lt; 5? 1=Yes,0=No","Over the past 3 months, has the participant used greater than the equivalent of 80-240 mcg of inhaled beclomethasone daily? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used combination inhaled corticosteroid/LABA therapy?","Over the past 3 months, on average, how many days per week has the participant used combination inhaled corticosteroid/LABA therapy,  &lt; 5 (1=Yes,0=No)","Over the past 3 months, on average, how many days per week has the participant had daytime asthma symptoms?","Over the past 3 months, on average, how many days per week has the participant had daytime asthma symptoms,  &lt;=2? 1=Yes,0=No","Over the past 3 months, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms?","Over the past 3 months, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms, &lt;=2? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used his/her short acting betaagonist  for relief of symptoms?","Over the past 3 months, on average, how many days per week has the participant used his/her short acting betaagonist  for relief of symptoms, &lt; 2? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 80Pct? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 70Pct? 1=Yes,0=No","Does the participant have current evidence of any of the conditions listed on the Exclusionary Medical Conditions for SIENA  reference card, or any chronic diseases that would prevent participation in the trial or put the participant at risk by participati","Is the participant currently taking any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Is the participant able to go off these medications for the required washout period prior to Visit 1 and for the duration of the study? 1=Yes,0=No","Is the participant able to use the spirotel e-diary/PEF meter correctly, as evidenced by achieving a score of 13 on the spirotel Performance Checklist ? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Has the participant or parent/legal guardian signed the SIENA Informed Consent document?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past month, on average, how many days per week has the participant had daytime symptoms?","In the past month, on average, how many days per week has the participant had daytime symptoms, &lt;=2? 1=Yes,0=No","In the past month, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms?","In the past month, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms,  &lt;=2? 1=Yes,0=No","In the past month, on average, how many days per week has the participant used his/her short acting beta-agonist  for relief of symptoms?","In the past month, on average, how many days per week has the participant used his/her short acting beta-agonist  for relief of symptoms, &lt; 2? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 80Pct? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 70Pct? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant or parent/legal guardian sign the SIENA Informed Consent document? 1=Yes,0=No","Ages 12-17 Only: Has the participant signed and dated the assent form or, if the participant is less than the local age of assent, has the participant given verbal assent? 1=Yes,0=No","Is the participant 12 years of age, or older? 1=Yes,0=No","Does the participant plan to move away from the clinical site in the upcoming 11 months such that his/her ability to complete the study will be jeopardized? 1=Yes,0=No","Has the participant used investigative drugs and/or enrolled in an intervention trial in the past 30 days, or have plans to enroll in such a trial during the SIENA study? 1=Yes,0=No","Has the participant had a respiratory infection within the past 6 weeks? 1=Yes,0=No","Has the participant had a respiratory infection within the past 4 weeks? 1=Yes,0=No","During the past 4 weeks, on average, how many days per week has the participant had daytime asthma symptoms?","During the past 4 weeks, on average, how many days per week has the participant had daytime asthma symptoms  &gt; 2? 1=Yes,0=No","During the past 4 weeks, how many nighttime awakenings due to asthma symptoms has the participant had?","During the past 4 weeks, how many nighttime awakenings due to asthma symptoms has the participant had,  &gt; 2? 1=Yes,0=No","During the past 4 weeks, on average, how many days per week has the participant used his/her short-acting beta-agonist  for relief of symptoms?","During the past 4 weeks, on average, how many days per week has the participant used his/her short-acting beta-agonist  for relief of symptoms, &gt; 2? 1=Yes,0=No","Is Q7a, Q8a, or Q9a checked YES? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Did the participant or parent/legal guardian sign the SIENA Informed Consent document?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Does the participant have current evidence of any of the conditions listed on the Exclusionary Medical Conditions for SIENA  reference card, or any chronic diseases  that would prevent participation in the trial or put the participant at risk by participat","Does the participant have a history of bladder-neck obstruction, urinary retention, benign prostatic hyperplasia , or a clinically relevant urologic disorder that precludes study participation? 1=Yes,0=No","Does the participant have a history of narrow angle glaucoma? 1=Yes,0=No","Does the participant have a history of significant cardiovascular disorders or arrhythmias? 1=Yes,0=No","Has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card within the specified time periods? 1=Yes,0=No","Is the participant currently taking prescription or OTC medication other than those listed on the Allowed Medications  reference card? 1=Yes,0=No","Based on input from the participant and the study physician, will the participant need to use intranasal steroids at any time during the study? 1=Yes,0=No","Is the participant willing to use a single intranasal steroid at a stable dose continuously for the duration of the study, starting at Visit 1? 1=Yes,0=No","Is the participant currently receiving allergen immunotherapy  other than an established maintenance regimen implemented continuously for a minimum of 3 months? 1=Yes,0=No","Has the participant taken omalizumab within the past 3 months? 1=Yes,0=No","Has the participant used any smokeless tobacco products  in the past year? 1=Yes,0=No","Has the participant smoked cigarettes, a pipe, cigar, marijuana, electronic cigarettes, or any other substance in the past year? 1=Yes,0=No","Ages 18+ Only: Does the participant have a smoking history of greater than 10 pack-years? 1=Yes,0=No","Ages 12-17 Only: Does the participant have a smoking history of greater than 5 pack-years? 1=Yes,0=No","Has the participant received a physician diagnosis of asthma at least 12 months ago? 1=Yes,0=No","Has the participant received a physician diagnosis of asthma at least 12 months ago or had a history consistent with asthma for the previous 12 months? 1=Yes,0=No","Has the participant experienced a life-threatening asthma exacerbation requiring treatment with intubation or mechanical ventilation in the past 5 years? 1=Yes,0=No","Has the participant had an asthma exacerbation requiring systemic corticosteroid treatment in the past 6 weeks? 1=Yes,0=No","Has the participant used an oral or inhaled corticosteroid, or leukotriene modifier in the past 6 weeks? 1=Yes,0=No","Has the participant used an oral corticosteroid in the past 6 weeks? 1=Yes,0=No","Has the participant used an inhaled corticosteroid or leukotriene modifier in the past 3 weeks? 1=Yes,0=No","Is the participant potentially able to bear children? 1=Yes, 0=No, 9=N/A","Is the participant currently pregnant or lactating? 1=Yes,0=No","Does the participant agree to use one of the approved methods indicated on the Birth Control Methods  reference card for the duration of the study? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was the participants prebronchodilator FEV1 &gt;=70Pct predicted? 1=Yes,0=No","Did the participants FEV1 improve &gt;=12Pct and &gt;=200 ml in response to four puffs of albuterol? 1=Yes,0=No","Does the participant have valid source documentation within the past 6 months for an overread AsthmaNet methacholine challenge  with a PC20 &lt;=16 mg/ml? 1=Yes,0=No","Did the participants FEV1 improve &gt;=12Pct and &gt;=200 ml in response to four puffs of albuterol?","Was the participants methacholine PC20 &lt;=16 mg/ml? 1=Yes,0=No","Is the participant able to use the spirotel e-diary/PEF meter correctly, as evidenced by achieving a score of 13 on the Spirotel Performance Checklist ? 1=Yes, 0=No, 9=N/A","Has the participant used the Twisthaler previously? 1=Yes,0=No","Is the participant able to use a Twisthaler properly, as evidenced by achieving a score of 12 on the Twisthaler Inhalation Technique Checklist ? 1=Yes,0=No","Is the participant able to use a metered dose inhaler properly, as evidenced by achieving a score of 11 on the MDI Inhalation Technique Checklist  ? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Does the participant have valid source documentation within the past 6 months for an overread AsthmaNet methacholine challenge  with a PC20 &lt;=16 mg/ml?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Since Visit 1, has the participant experienced a treatment failure event as defined in the protocol? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Since Visit 1, has the participant experienced one or more asthma exacerbations as defined in the protocol? 1=Yes,0=No","Since Visit 1, has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant provide two acceptable sputum induction samples during the run-in? 1=Yes,0=No","Since Visit 1, has the participant experienced a treatment failure event as defined in the protocol? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Did the participant complete Asmanex220 mcg treatment at least 6 weeks ago? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Since Visit 1, has the participant experienced one or more asthma exacerbations as defined in the protocol? 1=Yes,0=No","Since Visit 1, has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Has the participant experienced a respiratory tract infection in the past 6 weeks? 1=Yes,0=No","Has the participant experienced a respiratory tract infection in the past 4 weeks? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report daytime asthma symptoms daily during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report night-time awakenings due to asthma symptoms more than once per week during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report night-time awakenings due to asthma symptoms more than twice per week during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report daily albuterol use for asthma symptom control during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant record at least 75Pct of AM and PM peak flow measurements and symptoms on his or her spirotel during the run-in? 1=Yes,0=No","Did the participant take at least 75Pct of the required puffs from his or her Respimat during the run-in? 1=Yes,0=No","Was the participants prebronchodilator FEV1 &gt;=70Pct predicted? 1=Yes,0=No","Has the participant used the Respimat previously? 1=Yes,0=No","Is the participant able to use a Respimat inhaler properly, as evidenced by achieving a score of 15 on the Respimat Inhalation Technique Checklist? 1=Yes,0=No","Does the participant wish to withdraw consent from the study? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has QC procedure been performed on the NIOX MINO today? 1=Yes,0=No","Did the participant eat or drink within the past hour? 1=Yes,0=No","Did the participant take part in strenuous activity/exercise within the past hour? 1=Yes,0=No","Time eNO started (based on a 24-hour clock) (0000-2359)","ENO Measurement","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","How long have you lived in the current house? (Estimate if uncertain.) (1010) years","How long have you lived in the current house? (Estimate if uncertain.)(1020) months","Does your house use a wood burning stove as a primary source of heat? 1=Yes, 0=No, 8=Don't Know","Does your house use an air conditioner? 1=Yes, 0=No, 8=Don't Know","Does your house use an evaporative cooler ? 1=Yes, 0=No, 8=Don't Know","Does your house use a humidifier? 1=Yes, 0=No, 8=Don't Know","Does your house use a dehumidifier? 1=Yes, 0=No, 8=Don't Know","Has there been water damage to your house, basement, or its contents during the past 12 months? 1=Yes, 0=No, 8=Don't Know","Has there been any mold or mildew, on any surfaces, inside your house in the past 12 months? 1=Yes, 0=No, 8=Don't Know","Which rooms have or have had mold or mildew?10a.Bathroom(s)  (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew?10b.Basement or attic (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10c.Kitchen (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10d.Your bedroom (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10e.Other bedrooms (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10f.Living or family room (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10g.Other (1 = Yes, 0 = No)","Do you ever see cockroaches in your house? 1=Yes,0=No","In which room(s) have you seen cockroaches?12a.Kitchen (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches?  12b.Basement or attic (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12c.Bathroom(s) (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12d.Living or family room (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12e.Your bedroom  (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12f.Other bedrooms (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12g.Garage (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12h.Other (1 = Yes, 0 = No)","Do you ever see rodents  or rodent droppings in your house? 1=Yes,0=No","14.In which room(s) have you seen rodents or rodent droppings? 14a. Kitchen (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14b. Basement or attic (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14c. Bathroom(s) (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings?14d. Living or family room (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14e. Your bedroom(1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14f. Other bedrooms (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14g. Garage (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14h. Other (1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15a. Barns(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property?15b. Hay(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15c. Woodsheds(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15d. Firewood(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property?15e. Chicken coops (1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15f. Corral (1 = Yes, 0 = No)","CHARACTERISTICS OF THE PARTICIPANT'S BEDROOM16.What is the floor covering in your bedroom? 1=Rug/carpet, 2=Vinyl tile or linoleum, 3=Wood, 4=Ceramic tile, 5=Other , 9=Don't know","What type of mattress is on your bed? 1=None, 2=Inner spring mattress, 3=Foam mattress, 4=Waterbed, 5=Air mattress, 6=Other , 9=Don't know","Is the mattress completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","Does your bed have a box spring? 1=Yes,0=No","Is the box spring completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","What type of pillow do you usually sleep with? 1=None, 2=Feather/down, 3=Foam/Dacron/synthetic, 5=Other , 9=Don't know","Is the pillow completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","Does your household have any pets? 1=Yes,0=No","Enter the number of pets that the household has. (Enter '00' if none. If none to Q24a - Q24d, skip to the next question.)24a.Cat","Cat  (1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has. Dog","Dog(1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has 24c.Rabbit, guinea pig, hamster, gerbil, or mouse","Rabbit, guinea pig, hamster, gerbil, or mouse (1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has 24d.Bird","Bird (1=Indoor, 2=Outdoor, 3=Both)","In general, and on a regular basis, are you exposed to any of the following animals?  25a.Cat(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25b.Dog(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25c.Rabbit, guinea pig, hamster, gerbil, or mouse(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25d.Bird (1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25e.Farm animals (1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals?  25f.Other (1 = Yes, 0 = No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other (specify","Highest education of any household member (1=No High School diploma/GED, 2=High Schol grad, 3=Technical training, 4=Some college, no degree, 5=Assoc degree, 6=Bachelor degree, 7=Masters degree/MD/PhD/JD/PharmD, 9=Decline to answer, 10=Don't know)","Category best describes the combined annual income, before taxes, of all members of your household for the last year. (1=Less than $25,000, 2=$25,000 - $49,999, 3=$50,000 - $99,999, 4=$100,000 or more, 9=Decline to answer, 10=Don't know)","How many people  are supported by this income reported in Q3?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Total IgE","Cat (E1)","Dog (E5)","Mouse (E72)","Rat (E74)","Mold mix (MX1)","Cockroach, German (I6)","Grass mix (GX2)","Tree mix (TX4)","Tree mix (TX6)","Weed mix (WX1)","Weed (W3)","Mite (D2)","Mite (D1)","Number or allergen specific (Cat, Dog, Mouse, etc) that were either POS or greater than 0.35","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","CBC with differential cell count: Eosinophils (absolute count) (0-9999/uL)","CBC with differential cell count: WBC (0-999.9 K/uL)","CBC with differential cell count: Differential Lymphocytes (0-99.9Pct)","CBC with differential cell count: Differential Monocytes (0-99.9Pct)","CBC with differential cell count: Differential Basophils (0-99.9Pct)","CBC with differential cell count: Differential Neutrophils (0-99.9Pct)","CBC with differential cell count: Differential Eosinophils (0-99.9Pct)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","PARENTAL HEIGHT - First study visit only or until both are completed 1.Biological mother's height (complete height or check unknown) __ feet","PARENTAL HEIGHT - First study visit only or until both are completed1.Biological mother's height (complete height or check unknown) __ __ inches","PARENTAL HEIGHT - First study visit only or until both are completed 1.Biological mother's height (complete height or check unknown) (9=Don't Know)","Biological father's height (complete height or check unknown) __ feet","Biological father's height (complete height or check unknown) __ __ inches","Biological father's height (complete height or check unknown) (9=Don't Know)","PARTICIPANT MEASUREMENTS - Complete at all applicable study visits3. What type of height measurement was obtained? 1=Standing height, 2=Length","First measurement __ __ __ . __ cm","Second measurement __ __ __ . __ cm","Third measurement __ __ __ . __ cm","Average height or length measurement __ __ __ . __ cm","In your judgment, was the participant's height or length measurement acceptable? 1=Yes,0=No","Weight (shoes off, light clothing) __ __ __ . __ kg","ORAL CANDIDIASIS5.Does the participant have evidence of oral candidiasis? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Post Diluent FEV1","Did the participant drop20Pct at the diluent stage? 1=Yes,0=No","Last concentration of methacholine administered __ __ . __ __ __ __ mg/ml","FEV1 after last concentration of methacholine administered","Did the participant achieve a PC20? 1=Yes,0=No","PC20 __ __ . __ __ mg/ml","Time methacholine challenge ended (based on 24-hour clock)","FEV1","Time of FEV1 in Q7a (based on 24-hour clock)","Was the FEV1 from Q7athe methacholine reversal reference value  in the gray box above? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was an additional treatment used in the first hour? 1=Yes,0=No","Additional albuterol by MDI (1=Yes,0=No)","Number of additional puffs of albuterol administered (1=2, 2=4, 3=&gt;4)","Nebulized Beta-agonist(1=Yes,0=No)","Subcutaneous epinephrine(1=Yes,0=No)","Implementation of clinic emergency protocol or algorithm (1=Yes,0=No)","Other  (1=Yes,0=No)","Participant's FEV1 after additional treatment within first hour.2a.FEV1","Time of FEV1 in Q2a (based on 24-hour clock)","Was the FEV1 from Q2a &gt;the methacholine reversal reference value  in the gray box on the Methacholine Challenge Testing  form? 1=Yes,0=No","Was additional treatment used after one hour? 1=Yes,0=No","Additional albuterol by MDI (1=Yes,0=No)","Number of additional puffs of albuterol administered (1=2, 2=4, 3=&gt;4)","Nebulized Beta-agonist  (1=Yes,0=No)","Subcutaneous epinephrine(1=Yes,0=No)","Implementation of clinic emergency protocol or algorithm  (1=Yes,0=No)","Treatment in the emergency room  (1=Yes,0=No)","Overnight hospitalization (1=Yes,0=No)","Other (1=Yes,0=No)","Participant's final FEV1 after methacholine challenge4a.FEV1","Time of FEV1 in Q4a (based on 24-hour clock)","Was the FEV1 from Q4a ? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Exclusions and Confounders1.Has the participant had any severe acute illness in the past 4 weeks? 1=Yes,0=No","If YES, has the participant received permission from the supervising physician to proceed with the methacholine challenge testing? 1=Yes,0=No","Has the participant used an oral or injectable steroid  in the last 4 weeks? 1=Yes,0=No","Does the participant have a baseline  FEV1 less than 55Pct of predicted or less than 1.0 L? 1=Yes,0=No","Pregnancy test results:  (Check N/A if the participant is male, or is female and is post-menopausal, had a hysterectomy or tubal ligation.) (1=Positive, 0=Negative, 9=N/A)","Is the participant's systolic blood pressure &gt; 200 mm Hg or diastolic blood pressure &gt; 100 mm Hg? 1=Yes,0=No","Is there any other reason the participant should not proceed with the methacholine challenge testing? 1=Yes,0=No","Is the participant eligible to proceed with the diluent  pulmonary function testing for the methacholine challenge? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Exclusions and Confounders:  1.Has the participant had any severe acute illness in the past 4 weeks? 1=Yes,0=No","If YES, has the participant received permission from the supervising physician to proceed with the methacholine challenge testing? 1=Yes,0=No","During the past 4 weeks, has the participant had any respiratory infections, colds, or bronchitis ? 1=Yes,0=No","If YES, during the past 2 weeks, has the participant had any respiratory infections, colds, or bronchitis ? 1=Yes,0=No","Has the participant used an oral or injectable steroid  in the last 4 weeks? 1=Yes,0=No","Does the participant have a baseline  FEV1 less than 70Pct of predicted? 1=Yes,0=No","Pregnancy test results(Check N/A if the participant is male, or is female and has not started menses.)(1=Positive, 0=Negative, 9=N/A)","If participant's age is ? 1=Yes,0=No","If participant's age is &lt; 12 years: Is the participant's systolic blood pressure &gt; 180 mm Hg or diastolic blood pressure &gt; 90 mm Hg? 1=Yes,0=No","Is there any other reason the participant should not proceed with the methacholine challenge testing? 1=Yes,0=No","Is the participant eligible to proceed with the diluent  pulmonary function testing for the methacholine challenge? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time albuterol administered (based on 24-hour clock)","Time post-albuterol spirometry started (based on 24-hour clock)","Highest FVC","Highest FEV1","Highest FEV1 (Pct predicted)","FEF Max","FEF25-75","In your judgment, was the subject's spirometry technique acceptable? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Please check the box next to the treatment that you believe you received over the past 12 weeks. 1=mometasone, 2=placebo","Blinded Scheduled Twisthaler Contents: How sure are you about your answer to Question 2? 1=Absolutely sure -  I know what the MDI contains, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","Please comment with respect to any observations you made that helped you make your choice in Question 2 . 1=I have no comments, 2=I noticed the following:","Please check the box next to the treatment that you believe you received over the past 12 weeks. 1=tiotropium, 2=placebo","How sure are you about your answer to Question 5? 1=Absolutely sure - I know what the Respimat capsules contain, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","1=I have no comments, 2=I noticed the following:","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Visit Number (numeric)","Level of periostin in peripheral blood (ng/ml)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time ipratropium administered (based on 24-hour clock) (0000-2359)","Time post-ipratropium spirometry started (based on 24-hour clock) (0000-2359)","Highest FVC (0-9.99 L)","Highest FEV1 (0-9.99 L)","Highest FEV1 (Pct predicted) (0-999 Pct predicted)","FEF Max (0-99.99 L/S)","FEF25-75 (0-9.99 L/S)","In your judgment, was the participants spirometry technique acceptable? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Is the participant unable to bear children due to any of the following reasons? 1=Yes,0=No","Post-menopausal (at least one year since last menses) (1=Yes,0=No)","Hysterectomy(1=Yes,0=No)","Tubal ligation(1=Yes,0=No)","Pregnancy test results  (1=Positive, 0=Negative)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Have you had any diseases, illnesses, conditions, or surgeries related Blood, Lymph, or Immune Systems? 1=Yes,0=No","Eyes(1=Yes,0=No)","Breasts (1=Yes,0=No)","Endocrine Systems (1=Yes,0=No)","Heart and Blood Vessels  (1=Yes,0=No)","Liver or Pancreas (1=Yes,0=No)","Kidneys or Urinary Tract System  (1=Yes,0=No)","Reproductive System (1=Yes,0=No)","Muscles or Bones (1=Yes,0=No)","Nervous System (1=Yes,0=No)","Psychiatric(1=Yes,0=No)","Drug Allergies (1=Yes,0=No)","Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Skin (1=Yes,0=No)","Have you ever had allergic rhinitis ? 1=Yes, 0=No, 9=Don't know","Have you ever had nasal polyps? 1=Yes, 0=No, 9=Don't know","Do you have chronic or recurrent sinusitis ? 1=Yes, 0=No, 9=Don't know","Have you ever been diagnosed with vocal cord dysfunction? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the ear, nose, or throat? 1=Yes,0=No","Have you ever had pneumonia? 1=Yes, 0=No, 9=Don't know","If YES, were you diagnosed by chest x-ray? 1=Yes, 0=No, 9=Don't know","If YES, were you treated with antibiotics? 1=Yes, 0=No, 9=Don't know","Have you ever had bronchitis? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the lungs ? 1=Yes, 0=No, 9=Don't know","Do you have gastroesophageal reflux disease ? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the stomach or intestines? 1=Yes,0=No","Have you been diagnosed with sleep disordered breathing ? 1=Yes,0=No","If YES, are you being treated with CPAP or BiPAP? 1=Yes,0=No","Have you ever had other sleep disorders? 1=Yes,0=No","Have you ever had other conditions that have not been mentioned on this form? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Short-acting Inhaled Beta-Agonists by Inhaler (e.g., albuterol, Primatene Mist, Maxair, ProAir, Proventil, Ventolin, Xopenex)  (1=Yes, 0=No, 9=Don't Know)","If YES, indicate average weekly puffs in the past month (Enter '000' if none used) weekly puffs","Rescue treatment via a Nebulizer Machine (e.g., albuterol, ipratropium, Combivent, Xopenex, levalbuterol)  (1=Yes, 0=No, 9=Don't Know)","Long-acting Inhaled Beta-Agonists (e.g., Serevent, Foradil, salmeterol, formoterol) Do not consider combination medications. (1=Yes, 0=No, 9=Don't Know)","Oral Beta-Agonists (e.g., albuterol, Brethine, Bricanyl, metaproterenol, Proventil, Ventolin, Repetabs, Volmax)  (1=Yes, 0=No, 9=Don't Know)","Oral Theophylline (short-acting or sustained release)(e.g., Aminophylline, Slo-Phyllin, Slo-bid, Theo-Dur, Uniphyl)(1=Yes, 0=No, 9=Don't Know)","Inhaled Anticholinergic by Inhaler (e.g., Atrovent, Combivent, Spiriva) (1=Yes, 0=No, 9=Don't Know)","Leukotriene Antagonist / 5LO Inhibitors (e.g., Accolate, Zyflo, Singulair) (1=Yes, 0=No, 9=Don't Know)","IgE Blocker (e.g., Xolair)(1=Yes, 0=No, 9=Don't Know)","Oral Steroids FOR ASTHMA (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron,  dexamethasone)(1=Yes, 0=No, 9=Don't Know)","If YES, in the past 12 months, how many courses of steroids by mouth have you taken FOR ASTHMA? 1=1 course, 2=2 courses, 3=3 courses, 4=4 courses, 5=5 courses, 6=More than 5 courses","Injectable Steroids FOR ASTHMA (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV)(1=Yes, 0=No, 9=Don't Know)","Steroids by Inhaler (e.g., Asmanex Twisthaler, QVAR, Flovent, Pulmicort Flexhaler)(1=Yes, 0=No, 9=Don't Know)","Indicate most recent type of inhaled steroid taken (refer to PRIOR_TRT_CARD reference card) code","Indicate number of daily puffs used daily puffs","Indicate the total number of months that you used the inhaled steroid out of the past 12 months months","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide) (1=Yes, 0=No, 9=Don't Know)","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide). If YES, complete Q13a - Q13c. Indicate most recent type of nebulized steroid taken (refer to PRIOR_TRT_CARD reference card)","Indicate number of daily treatments used daily treatments","Indicate the total number of months that you used the nebulized steroid out of the past 12 months months","Long-Acting Beta-Agonist and Inhaled Steroid Combination Medications (e.g., Advair Diskus, Symbicort MDI, Dulera MDI)(1=Yes, 0=No, 9=Don't Know)","Indicate most recent type of combination medication taken (refer to PRIOR_TRT_CARD reference card) code","Indicate number of daily puffs used daily puffs","Indicate the total number of months that you used the combination medication out of the past 12 months months","Nasal Steroids (e.g., Beconase, Vancenase, Flonase, Nasacort, Nasalide, Nasarel, Omnaris, Rhinocort, Nasonex)(1=Yes, 0=No, 9=Don't Know)","Non-steroidal Anti-allergic Nasal Medications (e.g., Nasalcrom, Astelin, Astepro, ipratropium)(1=Yes, 0=No, 9=Don't Know)","Anti-allergic Oral Medications (e.g., fexofenadine, loratadine, cetirizine, chlorpheniramine) (1=Yes, 0=No, 9=Don't Know)","Topical Steroids - Prescription (e.g., Synalar, Lidex, Dermacin, Fluocinonide) (1=Yes, 0=No, 9=Don't Know)","Topical Steroids - OTC (e.g., Hydrocortisone - multiple strengths and products) (1=Yes, 0=No, 9=Don't Know)","Other Medication FOR ASTHMA OR ALLERGIES (1=Yes, 0=No, 9=Don't Know)","Oral Steroids for Conditions Other Than Asthma (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) (1=Yes, 0=No, 9=Don't Know)","Injectable Steroids for Conditions Other Than Asthma (e.g., Medrol, Solumedrol, Decadron, dexamethasone,  triamcinolone, Kenalog, hydrocortisone IV)(1=Yes, 0=No, 9=Don't Know)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Short-acting Inhaled Beta-Agonists by Inhaler (e.g., albuterol, Primatene Mist, Maxair, ProAir, Proventil, Ventolin, Xopenex) Date","Rescue treatment via a Nebulizer Machine (e.g., albuterol, ipratropium, Combivent, Xopenex, levalbuterol) Date","Long-acting Inhaled Beta-Agonists (e.g., Serevent, Foradil, salmeterol, formoterol) Do not consider combination medications. Date","Oral Beta-Agonists (e.g., albuterol, Brethine, Bricanyl, metaproterenol, Proventil, Ventolin, Repetabs, Volmax) Date","Inhaled Anticholinergic by Inhaler (e.g., Atrovent, Combivent, Spiriva) Date","Leukotriene Antagonist / 5LO Inhibitors (e.g., Accolate, Zyflo, Singulair) Date","IgE Blocker (e.g., Xolair) Date","Oral Steroids FOR ASTHMA (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) Date","Injectable Steroids FOR ASTHMA (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV) Date","Steroids by Inhaler (e.g., Asmanex Twisthaler, QVAR, Flovent, Pulmicort Flexhaler) Date","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide) Date","Long-Acting Beta-Agonist and Inhaled Steroid Combination Medications (e.g., Advair Diskus, Symbicort MDI, Dulera MDI) Date","Nasal Steroids (e.g., Beconase, Vancenase, Flonase, Nasacort, Nasalide, Nasarel, Omnaris, Rhinocort, Nasonex) Date","Non-steroidal Anti-allergic Nasal Medications (e.g., Nasalcrom, Astelin, Astepro, ipratropium) Date","Anti-allergic Oral Medications (e.g., fexofenadine, loratadine, cetirizine, chlorpheniramine) Date","Topical Steroids - Prescription (e.g., Synalar, Lidex, Dermacin, Fluocinonide) Date","Topical Steroids - OTC (e.g., Hydrocortisone - multiple strengths and products) Date","Other Medication FOR ASTHMA OR ALLERGIES Date","Oral Steroids for Conditions Other Than Asthma (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) Date","Injectable Steroids for Conditions Other Than Asthma (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV) Date","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the last month, how often have you been upset because of something that happened unexpectedly? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that you were unable to control the important things in your life? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt nervous and stressed? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt confident about being able to handle your personal problems? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that things were going your way? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you found that you could not cope with all the things that you had to do? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you been able to control irritations in your life? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that you were on top of things? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you been angered because of things that happened that were outside of your control? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that your difficulties were piling up so high that you could not overcome them? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Have you consumed caffeine in the past 4 hours? 1=Yes,0=No","Have you used medications with caffeine in the past 4 hours? 1=Yes,0=No","Have you used any weight loss medications in the past 4 hours? 1=Yes,0=No","Have you consumed any food containing alcohol or beverages containing alcohol in the past 4 hours? 1=Yes,0=No","Have you used a rescue intermediate-acting inhaled betaagonist in the past 6 hours? 1=Yes,0=No","Have you used a short-acting anticholinergic in the past 6 hours? 1=Yes,0=No","Have you used any ophthalmic antihistamines in the past 6 hours? 1=Yes,0=No","Have you used any nasal antihistamines in the past 6 hours? 1=Yes,0=No","Have you used any nasal decongestants in the past 6 hours? 1=Yes,0=No","Have you used any oral antihistamines in the past 48 hours? 1=Yes,0=No","Have you used any oral decongestants or cold remedies in the past 48 hours? 1=Yes,0=No","Have you used any smokeless tobacco products today? 1=Yes,0=No","At this time, is your asthma worse because of recent exposure to triggers? 1=Yes,0=No","Is there any other reason you should not proceed with spirometry testing? 1=Yes,0=No","Is the participant eligible to proceed with the spirometry testing? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past 4 weeks, I worried about the long-term effects of asthma on my health (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I had to worry about asthma triggers (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, my asthma was on my mind (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, it was hard to get a good nights sleep because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt like I couldnt enjoy life because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt that asthma was controlling my life (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt frustrated that I couldnt make plans in advance because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, everyday activities were a struggle (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, asthma placed stress on my relationships with family, friends, significant others, or co-workers (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, I felt frustrated that I have to do things differently than people who dont have asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt like I missed out on doing things with others because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, I had to do a lot of planning to make sure I always had an inhaler ready (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Randomized treatment assignment period 1: (Mometasone, Tiotropium or Placebo)","Randomized treatment assignment period 2: (Mometasone, Tiotropium or Placebo)","Randomized treatment assignment period 3: (Mometasone, Tiotropium or Placebo)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Sex (1=Male, 2=Female)","Primary Racial Identification for Spirometry (1=Amer Indian or Alaskan Native, 2=Asian or Pacific Islander, 3=Black, 4=White, 5=Hispanic or Latino, 6=Other)","Participant's Age at Visit 1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Description of Adverse Event (ICD9 Code)","Is the participant currently taking study drug? 1=Yes,0=No","Time interval between the last administration of the study drug and the Adverse Event","What was the unit of time for the interval in Question #4? 1=Second, 2=Minute, 3=Hour, 4=Day","Fatal event(1=Yes,0=No)","Life-threatening event (1=Yes,0=No)","Inpatient hospitalization required(1=Yes,0=No)","Hospitalization prolonged  (1=Yes,0=No)","Disabling or incapacitating  (1=Yes,0=No)","Overdose (1=Yes,0=No)","Cancer (1=Yes,0=No)","Congenital anomaly(1=Yes,0=No)","Serious laboratory abnormality with clinical symptoms(1=Yes,0=No)","Height failure (per protocol MOP)(1=Yes,0=No)","Pregnancy(1=Yes,0=No)","Other (1=Yes,0=No)","What in your opinion caused the event? 1=Yes,0=No","Withdrawal of study drug(s)(1=Yes,0=No)","Concurrent medication  (1=Yes,0=No)","Other condition or event  (1=Yes,0=No)","Was the event expected or unexpected? 1=Expected, 2=Unexpected","Was the event possibly, probably, or definitely related to study participation? 1=Yes,0=No","Date of Adverse Event","Admission date","Discharge date","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","What type of height measurement was obtained? 1=Standing height, 2=Length","First measurement __ __ __ . __ cm","Second measurement __ __ __ . __ cm","Third measurement __ __ __ . __ cm","Average height or length measurement __ __ __ . __ cm","In your judgment, was the participant's height or length measurement acceptable? 1=Yes,0=No","Weight (shoes off, light clothing) __ __ __ . __ kg","Does the participant have evidence of oral candidiasis? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant fail to respond within 48 hours to the treatment failure rescue algorithm? 1=Yes,0=No","Did the participant use at least 16 puffs as needed albuterol per 24 hours for a period of 48 hours? 1=Yes,0=No","Did the participant experience prebronchodilator FEV1 values &lt; 50Pct of the baseline prebronchodilator value obtained at Visit 1 on two consecutive spirometric determinations made on different days? 1=Yes, 0=No, 9=Not evaluated","Did the participant experience prebronchodilator FEV1 values &lt; 40Pct of predicted on two consecutive spirometric determinations made on different days? 1=Yes, 0=No, 9=Not evaluated","Did the study or treating physician prescribe the participant oral/parenteral corticosteroids for the treatment of his/her asthma? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation in the opinion of the study investigator or personal physician? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation? 1=Yes,0=No","Did the participant seek care for significant asthma exacerbation conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Was the participant hospitalized? 1=Yes,0=No","Duration of hospital stay (0-99.9 days)","Was intubation or ventilation assistance required? 1=Yes,0=No","Was the participant admitted to the intensive care unit? 1=Yes,0=No","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Inhaled corticosteroids (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Nebulized bronchodilator (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Oral corticosteroids (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, IM or IV steroids (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Antibiotics (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Other (1=Yes, 0=No )","Why do you think the participant experienced a significant asthma exacerbation? 1=Common cold, 2=Allergies, 3=Pollution or chemical irritant, 4=Too little asthma maintenance medication, 5=Exercise, 6=Other","Date exacerbation conditions were met","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Which phone contact following Asthma Exacerbation visit is being performed? 1=Phone Contact #1 , 2=Phone Contact #2 , 3=Phone Contact #3","Since the most recent Asthma Exacerbation contact, has the participant sought additional care for significant asthma exacerbation conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Was the participant hospitalized? 1=Yes,0=No","Duration of hospital stay (0-99.9 days)","Was intubation or ventilation assistance required? 1=Yes,0=No","Was the participant admitted to the intensive care unit? 1=Yes,0=No","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Inhaled corticosteroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Nebulized bronchodilator (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Oral corticosteroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, IM or IV steroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Antibiotics (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Runny Nose  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Post nasal drip  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Need to blow your nose  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Facial pain/pressure(0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Nasal obstruction  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time spirometry started (based on 24-hour clock)","The reported FEV1 and FVC are the best volumes of all acceptable maneuvers. 1.Highest FVC","Highest FEV1 (Pct predicted)","The reported flow rates correspond to the maneuver where FEV1 + FVC is maximized. FEF Max","FEF25-75","In your judgment, was the subject's spirometry technique acceptable? 0=No, 1=Yes","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Highest FEV1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","TRIAL_TYPE","FVC","FEV1","PEF","FEF2575","FET","Number of times the participant woke up last night due to asthma symptoms # Times woke up due to asthma (0-9)","Number of puffs the participant will take from the WHITE study inhaler this morning # Puffs from WHITE inhaler in AM (0-9)","only) Number of puffs the participant will take from the BLUE study inhaler this morning # Puffs from BLUE inhaler in AM (0-9)","Has the participant taken any puffs from his/her RED RESCUE albuterol inhaler in the past 4 hours? 1=yes, 0=no","Shortness of Breath Score Shortness of Breath Score overnight (0,1,2,3)","Chest tightness score Chest Tightness Score overnight (0,1,2,3)","Wheezing score Wheezing Score overnight (0,1,2,3)","Coughing score Coughing Score overnight (0,1,2,3)","Phlegm/Mucus score Phlegm/Mucus Score overnight (0,1,2,3)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","date 'fixed' for analysis so that associated AM and PM always have same ddate","TRIAL_TYPE","FVC","FEV1","PEF","FEF2575","FET","Number of puffs the participant will take from the WHITE study inhaler tonight # Puffs from WHITE inhaler in PM (0-9)","Has the participant taken any puffs from his/her RED RESCUE albuterol inhaler during the past 4 hours? 1=yes, 0=no","Shortness of Breath Score Shortness of Breath Score since waking this morning (0,1,2,3)","Chest tightness score Chest Tightness Score since waking this morning (0,1,2,3)","Wheezing score Wheezing Score since waking this morning (0,1,2,3)","Coughing score Coughing Score since waking this morning (0,1,2,3)","Phlegm/Mucus score Phlegm/Mucus Score since waking this morning (0,1,2,3)","Did your regular exercise cause unusually severe shortness of breath during the past 24 hours? 1=Yes, 0=No, 9=N/A","Number of RED RESCUE albuterol puffs taken in the past 24 hours to prevent symptoms# Preventive  RED RESCUE puffs  in past 24 hours?","Number of RED RESCUE albuterol puffs taken for asthma symptoms or low peak flow during past 24 hours  # RED RESCUE  puffs for symptoms  in past 24 hours?","Number of times used RED RESCUE albuterol inhaler for asthma symptoms during past 24 hours  # Times used  RED RESCUE for symptoms in past 24 hours?","Did RED RESCUE albuterol relieve symptoms for less than 4 hours after treatment during past 24 hours? 1=Yes,0=No","Was the participant seen by a healthcare provider  for an unscheduled visit in the past 24 hours due to asthma symptoms? 1=Yes,0=No","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Processing: Time processing started (based on 24-hour clock) __ __ __ __ (0 - 2359)","Processing Sample: Total Cell Count: __ __ __ __. __x 10^4 cells/ml (0 - 9999.9)","Was the participant's sputum sample processed within 4 hours after collection? 1=Yes,0=No","Processing Sample: Processing Date __ __ / __ __ / 20 __ __","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Rate slide's quality:  (1=Very good, 2=Good, 3=Acceptable, 4=Poor but readable, 5=Not readable)","Record the number on the slide(s) that was (were) read (0 - 9)","Record the number on the slide(s) that was (were) read: These are numbers that were assigned to the slides at each site. (0 -9)","Total Cell Count: __ __ __ __ . __ x 10^4 cells/ml (0 - 9999.9)","Differential Cell Counts: Squamous Cells __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Epithelial Cells __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Macrophages __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Neutrophils __ __ . __ Pct (0 -99.9)","Differential Cell Counts: Eosinophils __ __ . __ Pct (0 - 99.9)","Differential Cell Counts: Lymphocytes __ __ . __ Pct (0 -99.9)","Date of Read __ __ / __ __ / 20 __ __","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","For this protocol, what was the duration of sputum induction the first time the participant's sample was processed within 4 hours after collection?","Sputum induction start time (based on 24-hour clock)","Sputum induction stop time (based on 24-hour clock)","Duration of sputum induction collection phase at this visit -- minutes (0 - 99.9)","Was the duration ? 1=Yes,0=No","Volume of sputum sample at this visit -- ml (0.0 - 999.9)","Is the volume adequate for processing? 1=Yes,0=No","Is the sample adequate for laboratory analysis? 1=Yes,0=No","Participant's FEV1 immediately after completion of sputum induction: -- (0.0 -9.99 L)","FEV1 (Pct predicted) (0 - 999)","Time of FEV1 in Q7a (based on 24-hour clock)","Percent difference in FEV1 ((Reference - Q7a) / Reference) X100 (-99.9 - 99.9 Pct)","Did the participant's FEV1 drop &gt; 10Pct from reference FEV1 as indicated in Q7d? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Participant's FEV1 after initial 2 puffs of albuterol: 1a.FEV1 __ . __ __ L (0 -9.99)","Participant's FEV1 after initial 2 puffs of albuterol: 1b. FEV1 (Pct predicted)__ __ __ Pct predicted (0 - 999)","Participant's FEV1 after initial 2 puffs of albuterol: 1c. Time of FEV1 from Q1a (based on 24-hour clock) __ __ __ __ (0 - 2359)","Participant's FEV1 after initial 2 puffs of albuterol: 1d. Was the FEV1 from Q1a &gt;=the sputum induction reversal reference value in the gray box above? 1=Yes,0=No","Participant's FEV1 after 2 additional puffs of albuterol: 2a. FEV1 __ . __ __ L (0 - 9.99)","Participant's FEV1 after 2 additional puffs of albuterol: 2b. FEV1 (Pct predicted) __ __ __ Pct predicted (0 - 999)","Participant's FEV1 after 2 additional puffs of albuterol: 2c. Time of FEV1 from Q2a (based on 24-hour clock)__ __ __ __ (0 - 2359)","Participant's FEV1 after 2 additional puffs of albuterol: 2d.Was the FEV1 from Q2a ? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was the participant's FEV1 after reversal from the methacholine challenge greater yhan 90Pct of the baseline FEV1 (1=Yes,0=No)","If NO, has the participant received permission from the supervising physician to proceed with sputum induction testing? 1=Yes,0=No","Physician's Signature:_________ (1=signature present, NULL=signature missing)","Participant's FEV1 used for assessment of eligibility for sputum induction __ . __ __ L (0 - 9.99)","Participant's FEV1 (Pct predicted) used for assessment of eligibility for sputum induction __ __ __ Pct predicted (O-999)","Was the participant's FEV1  from Q3 ? 1=Yes,0=No","Has the participant used any smokeless tobacco products  today? 1=Yes,0=No","Is there any other reason the participant should not proceed with sputum induction? 1=Yes,0=No","Is the participant eligible for sputum induction? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has the participant completed the study through Visit 9? 1=Yes,0=No","Who initiated termination of the participant? 1=Participant, 2=Clinical Staff","1=no longer interested in participating*, 2=no longer willing to follow protocol*, 3=difficult access to clinic , 4=unable to make visits during clinic hours, 5=moving out of the area, 6=unable to continue due to personal constraints*, 7=unable to continue","Did clinical staff terminate the participant due to pregnancy? 1=Yes, 0=No, 9=N/A","Did clinical staff terminate the participant due to loss to follow-up? 1=Yes,0=No","Did clinical staff terminate the participant due to loss to follow-up, type of contact (1=In-person visit, 2=Phone call)","Did clinical staff terminate the participant due to an asthma-related adverse event? 1=Yes,0=No","Did clinical staff terminate the participant due to a medication-related adverse event? 1=Yes,0=No","Did clinical staff terminate the participant due to an adverse event not related to asthma or medications? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility during the Supervised Washout ? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with medication dosing? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with diary completion? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with visit attendance? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with peak flow monitoring? 1=Yes,0=No","Did clinical staff terminate the participant due to significant asthma exacerbation or two treatment failures during run-in ? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility during the run-in period  for reasons other than compliance or exacerbation/treatment failure or high sputum eosinophils ? 1=Yes,0=No","Did clinical staff terminate the participant due to other reason? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility due to high sputum eosinophils in the run-in ? 1=Yes,0=No","Did clinical staff terminate the participant due to... Indicate the letter corresponding to the primary reason the participant was terminated. (a-n)","Did clinical staff terminate the participant due to loss to follow-up, date of last contact with participant","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant seek care for treatment failure conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact, other than call to notify clinic as instructed in spirotel","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Inhaled corticosteroids (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Nebulized bronchodilator (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started: Oral corticosteroids form.  (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  IM or IV steroids form.  (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started: Antibiotics (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant awaken from asthma three or more times in a two-week period or on two consecutive nights? 1=Yes,0=No","Did the participant use albuterol for relief of symptoms four or more times/day for two or more consecutive days? 1=Yes,0=No","Did albuterol relieve symptoms for less than four hours after treatment? 1=Yes,0=No","Did the participant use albuterol for relief of symptoms daily for seven days, and this use exceeded two times the weekly use of albuterol in the baseline period? 1=Yes,0=No","Did regular exercise cause unusually severe shortness of breath on two or more days during a seven day period? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation? 1=Yes,0=No","Did the participant experience a treatment failure? 1=Yes,0=No","Has the participant experienced two treatment failures in the current treatment period ? 1=Yes,0=No","Does the study physician feel the study treatment contributed directly to the treatment failure, such that it is in the participants best interest to discontinue the current blinded treatment? 1=Yes,0=No","Did the participant begin daily treatment with high-dose ICS for 10 days? 1=Yes,0=No","Did the participant begin daily treatment with high-dose ICS for 10 days? 1=Did not think symptoms werebad enough, 2=Other","Date treatment failure conditions were met","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","During this last treatment period , has the participant experienced a treatment failure or asthma exacerbation? 1=Yes,0=No","Has the participant experienced two or more treatment failures in this last treatment period, or does the study physician feel the study treatment contributed directly to the treatment failure such that it is in the participants best interest to discontinu","As of today, have 2 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","As of today, have 6 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","As of today, have 3 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Are you currently employed ? 1=Yes,0=No","In general, how many hours per week do you usually work?","During the past seven days, how many hours did you miss from work because of problems associated with your asthma?","During the past seven days, how much did asthma affect your productivity while you were working?","Do you currently attend classes in an academic setting ? 1=Yes,0=No","In general, how many hours per week do you usually attend classes?","During the past seven days, how many hours did you miss from class or school because of problems associated with your asthma?","During the past seven days, how much did asthma affect your productivity while in school or attending classes in an academic setting?","During the past seven days, how much did your asthma affect your ability to do your regular daily activities, other than work at a job or attend classes?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","How sick today? 0=Not Sick, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely","Runny Nose:(0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","PLUGGED_NOSE  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SNEEZING  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SORE_THROAT  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SCRATCHY_THROAT  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","COUGH (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","HOARSENESS  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","HEAD_CONGESTION  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","CHEST_CONGESTION  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","FEELING_TIRED (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","THINK_CLEARLY  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SLEEP_WELL  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","BREATH_EASILY  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WALK  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","DAILY_ACTIVITIES  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WORK_OUTSIDE_HOME  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WORK_INSIDE_HOME  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","INTERACT  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","LIVE_LIFE  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","MY_COLD_IS  (1=Very much better, 2=Somewhat better, 3=A little better, 4=The same, 5=A little worse, 6=Somewhat worse, 7=Very much worse)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Date (number of days from Visit 1)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Num<\/th>\n      <th>Variable<\/th>\n      <th>Type<\/th>\n      <th>Label<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":1},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Num","targets":1},{"name":"Variable","targets":2},{"name":"Type","targets":3},{"name":"Label","targets":4}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-b4f84bb5793ae524ebb5" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-b4f84bb5793ae524ebb5">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80","81","82","83","84","85","86","87","88","89","90","91","92","93","94","95","96","97","98","99","100","101","102","103","104","105","106","107","108","109","110","111","112","113","114","115","116","117","118","119","120","121","122","123","124","125","126","127","128","129","130","131","132","133","134","135","136","137","138","139","140","141","142","143","144","145","146","147","148","149","150","151","152","153","154","155","156","157","158","159","160","161","162","163","164","165","166","167","168","169","170","171","172","173","174","175","176","177","178","179","180","181","182","183","184","185","186","187","188","189","190","191","192","193","194","195","196","197","198","199","200","201","202","203","204","205","206","207","208","209","210","211","212","213","214","215","216","217","218","219","220","221","222","223","224","225","226","227","228","229","230","231","232","233","234","235","236","237","238","239","240","241","242","243","244","245","246","247","248","249","250","251","252","253","254","255","256","257","258","259","260","261","262","263","264","265","266","267","268","269","270","271","272","273","274","275","276","277","278","279","280","281","282","283","284","285","286","287","288","289","290","291","292","293","294","295","296","297","298","299","300","301","302","303","304","305","306","307","308","309","310","311","312","313","314","315","316","317","318","319","320","321","322","323","324","325","326","327","328","329","330","331","332","333","334","335","336","337","338","339","340","341","342","343","344","345","346","347","348","349","350","351","352","353","354","355","356","357","358","359","360","361","362","363","364","365","366","367","368","369","370","371","372","373","374","375","376","377","378","379","380","381","382","383","384","385","386","387","388","389","390","391","392","393","394","395","396","397","398","399","400","401","402","403","404","405","406","407","408","409","410","411","412","413","414","415","416","417","418","419","420","421","422","423","424","425","426","427","428","429","430","431","432","433","434","435","436","437","438","439","440","441","442","443","444","445","446","447","448","449","450","451","452","453","454","455","456","457","458","459","460","461","462","463","464","465","466","467","468","469","470","471","472","473","474","475","476","477","478","479","480","481","482","483","484","485","486","487","488","489","490","491","492","493","494","495","496","497","498","499","500","501","502","503","504","505","506","507","508","509","510","511","512","513","514","515","516","517","518","519","520","521","522","523","524","525","526","527","528","529","530","531","532","533","534","535","536","537","538","539","540","541","542","543","544","545","546","547","548","549","550","551","552","553","554","555","556","557","558","559","560","561","562","563","564","565","566","567","568","569","570","571","572","573","574","575","576","577","578","579","580","581","582","583","584","585","586","587","588","589","590","591","592","593","594","595","596","597","598","599","600","601","602","603","604","605","606","607","608","609","610","611","612","613","614","615","616","617","618","619","620","621","622","623","624","625","626","627","628","629","630","631","632","633","634","635","636","637","638","639","640","641","642","643","644","645","646","647","648","649","650","651","652","653","654","655","656","657","658","659","660","661","662","663","664","665","666","667","668","669","670","671","672","673","674","675","676","677","678","679","680","681","682","683","684","685","686","687","688","689","690","691","692","693","694","695","696","697","698","699","700","701","702","703","704","705","706","707","708","709","710","711","712","713","714","715","716","717","718","719","720","721","722","723","724","725","726","727","728","729","730","731","732","733","734","735","736","737","738","739","740","741","742","743","744","745","746","747","748","749","750","751","752","753","754","755","756","757","758","759","760","761","762","763","764","765","766","767","768","769","770","771","772","773","774","775","776","777","778","779","780","781","782","783","784","785","786","787","788","789","790","791","792","793","794","795","796","797","798","799","800","801","802","803","804","805","806","807","808","809","810","811","812","813","814","815","816","817","818","819","820","821","822","823","824","825","826","827","828","829","830","831","832","833","834","835","836","837","838","839","840","841","842","843","844","845","846","847","848","849","850","851","852","853","854","855","856","857","858","859","860","861","862","863","864","865","866","867","868","869","870","871","872","873","874","875","876","877","878","879","880","881","882","883","884","885","886","887","888","889","890","891","892","893","894","895","896","897","898","899","900","901","902","903","904","905","906","907","908","909","910","911","912","913","914","915","916","917","918","919","920","921","922","923","924","925","926","927","928","929","930","931","932","933","934","935","936","937","938","939","940","941","942","943","944","945","946","947","948","949","950","951","952","953","954","955","956","957","958","959","960","961","962","963","964","965","966","967","968","969","970","971","972","973","974","975","976","977","978","979","980","981","982","983","984","985","986","987","988","989","990","991","992","993","994","995","996","997","998","999","1000","1001","1002","1003","1004","1005","1006","1007","1008","1009","1010","1011","1012","1013","1014","1015","1016","1017","1018","1019","1020","1021","1022","1023","1024","1025","1026","1027","1028","1029","1030","1031","1032","1033","1034","1035","1036","1037","1038","1039","1040","1041","1042","1043","1044","1045","1046","1047","1048","1049","1050","1051","1052","1053","1054","1055","1056","1057","1058","1059","1060","1061","1062","1063","1064","1065","1066","1067","1068","1069","1070","1071","1072","1073","1074","1075","1076","1077","1078","1079","1080","1081","1082","1083","1084","1085","1086","1087","1088","1089","1090","1091","1092","1093","1094","1095","1096","1097","1098","1099","1100","1101","1102","1103","1104","1105","1106","1107","1108","1109","1110","1111","1112","1113","1114","1115","1116","1117","1118","1119","1120","1121","1122","1123","1124","1125","1126","1127","1128","1129","1130","1131","1132","1133","1134","1135","1136","1137","1138","1139","1140","1141","1142","1143","1144","1145","1146","1147","1148","1149","1150","1151","1152","1153","1154","1155","1156","1157","1158","1159","1160","1161","1162","1163","1164","1165","1166","1167","1168","1169","1170","1171","1172","1173","1174","1175","1176","1177","1178","1179","1180","1181","1182","1183","1184","1185","1186","1187","1188","1189","1190"],[1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,1,2,3,4,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,1,2,3,4,5,6,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28],["AAQ_1000","AAQ_1010","AAQ_1020","AAQ_1030","AAQ_1040","AAQ_1050","AAQ_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ABP_1000","ABP_1010","ABP_1020","ABP_1030","ABP_1040","ABP_1050","ABP_1060","ABP_1070","ABP_1080","ABP_1090","ABP_1100","ABP_1110","ABP_1120","ABP_1130","ABP_1140","ABP_1150","ABP_1160","ABP_1170","ABP_1180","ABP_1190","ABP_1200","ABP_1210","ABP_1220","ABP_1230","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ACT_1","ACT_2","ACT_3","ACT_4","ACT_5","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ICD9_CAT","AEC_1000","AEC_1010","AEC_1040","AEC_1050","AEC_1060","AEC_1070","AEC_1080","AEC_1090","AEC_1100","AEC_1110","AEC_1120","AEC_1020","AEC_1030","VNUM_C","VNUM","ICD9LONG","ICD9SHRT","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","AHA_1000","AHA_1020","AHA_1030","AHA_1040","AHA_1050","AHA_1060","AHA_1070","AHA_1090","AHA_1100","AHA_1110","AHA_1120","AHA_1130","AHA_1140","AHA_1150","AHA_1160","AHA_1170","AHA_1180","AHA_1190","AHA_1200","AHA_1210","AHA_1220","AHA_1250","AHA_1260","AHA_1270","AHA_1280","AHA_1290","AHA_1300","AHA_1310","AHA_1320","AHA_1330","AHA_1340","AHA_1350","AHA_1360","AHA_1370","AHA_1380","AHA_1390","AHA_1400","AHA_1410","AHA_1420","AHA_1430","AHA_1440","AHA_1450","AHA_1460","AHA_1470","AHA_1500","AHA_1570","AHA_1580","AHA_1590","AHA_1600","AHA_1730","AHA_1740","AHA_1750","AHA_1760","AHA_1770","AHA_1780","AHA_1790","AHA_1800","AHA_1810","AHA_1820","AHA_1830","AHA_1840","AHA_1850","AHA_1890","AHA_1900","VNUM_C","VNUM","VDATE","DATE_VAPE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ASU_1000","ASU_1010","ASU_1020","ASU_1030","ASU_1040","ASU_1050","ASU_1060","ASU_1070","ASU_1080","ASU_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","BAS_1000","BAS_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","BMI","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAGE","CHG_1000","CHG_1010","CHG_1020","CHG_1030","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","CME_1040","CME_1050","CME_1055","CME_1000","CME_1010","CME_1020","CME_1030","CME_1080","CME_1090","CME_1060","CME_1070","VNUM_C","VNUM","GEN_NAME","CLASS","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","COM_1000","COM_1010","COM_1020","COM_1030","COM_1040","COM_1050","COM_1060","COM_1070","COM_1080","COM_1090","COM_1100","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","CTX_1000","CTX_1010","CTX_1030","CTX_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E0A_1000","E0A_1020","E0A_1030","E0A_1040","E0A_1043","E0A_1045","E0A_1050","E0A_1060","E0A_1070","E0A_1073","E0A_1075","E0A_1080","E0A_1090","E0A_1100","E0A_1110","E0A_1120","E0A_1130","E0A_1140","E0A_1145","E0A_1150","E0A_1160","E0A_1170","E0A_1180","E0A_1190","E0A_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E0B_1000","E0B_1010","E0B_1020","E0B_1030","E0B_1040","E0B_1050","E0B_1060","E0B_1065","E0B_1070","E0B_1080","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E1_1000","E1_1020","E1_1030","E1_1040","E1_1050","E1_1060","E1_1065","E1_1070","E1_1080","E1_1090","E1_1100","E1_1110","E1_1120","E1_1130","E1_1140","E1_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E2_1000","E2_1010","E2_1020","E2_1030","E2_1040","E2_1050","E2_1060","E2_1070","E2_1080","E2_1090","E2_1100","E2_1110","E2_1120","E2_1130","E2_1140","E2_1145","E2_1150","E2_1160","E2_1170","E2_1173","E2_1175","E2_1180","E2_1190","E2_1200","E2_1210","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E3_1000","E3_1010","E3_1020","E3_1030","E3_1070","E3_1080","E3_1090","E3_1100","E3_1110","E3_1120","E3_1130","E3_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E4_1000","E4_1010","E4_1020","E4_1030","E4_1050","E4_1060","E4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E5_1000","E5_1010","E5_1020","E5_1030","E5_1035","E5_1040","E5_1050","E5_1060","E5_1065","E5_1070","E5_1080","E5_1085","E5_1090","E5_1100","E5_1110","E5_1120","E5_1130","E5_1140","E5_1150","E5_1160","E5_1170","E5_1180","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ENO_1000","ENO_1010","ENO_1020","ENO_1040","ENO_1050","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","HEQ_1000","HEQ_1010","HEQ_1020","HEQ_1030","HEQ_1040","HEQ_1050","HEQ_1060","HEQ_1070","HEQ_1080","HEQ_1090","HEQ_1100","HEQ_1110","HEQ_1120","HEQ_1130","HEQ_1140","HEQ_1150","HEQ_1160","HEQ_1170","HEQ_1180","HEQ_1190","HEQ_1200","HEQ_1210","HEQ_1220","HEQ_1230","HEQ_1240","HEQ_1250","HEQ_1260","HEQ_1270","HEQ_1280","HEQ_1290","HEQ_1300","HEQ_1310","HEQ_1320","HEQ_1330","HEQ_1340","HEQ_1350","HEQ_1360","HEQ_1370","HEQ_1380","HEQ_1390","HEQ_1400","HEQ_1410","HEQ_1420","HEQ_1430","HEQ_1440","HEQ_1450","HEQ_1460","HEQ_1470","HEQ_1480","HEQ_1490","HEQ_1500","HEQ_1510","HEQ_1520","HEQ_1530","HEQ_1540","HEQ_1550","HEQ_1560","HEQ_1570","HEQ_1580","HEQ_1590","HEQ_1600","HEQ_1610","HEQ_1620","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SEI_1000","SEI_1010","SEI_1020","SEI_1030","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","IGE","RE1","RE5","RE72","RE74","RMX1","RI6","RGX2","RTX4","RTX6","RWX1","RW3","RD2","RD1","N_ALLERGENPOS","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","LAB_1000","LAB_1010","LAB_1020","LAB_1030","LAB_1040","LAB_1050","LAB_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PLE_1000","PLE_1010","PLE_1020","PLE_1030","PLE_1040","PLE_1050","PLE_1060","PLE_1070","PLE_1080","PLE_1090","PLE_1100","PLE_1110","PLE_1130","PLE_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MTH_1000","MTH_1010","MTH_1020","MTH_1030","MTH_1040","MTH_1050","MTH_1060","MTH_1070","MTH_1080","MTH_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MAD_1000","MAD_1010","MAD_1020","MAD_1030","MAD_1040","MAD_1050","MAD_1060","MAD_1070","MAD_1090","MAD_1100","MAD_1110","MAD_1120","MAD_1130","MAD_1140","MAD_1150","MAD_1160","MAD_1170","MAD_1180","MAD_1190","MAD_1200","MAD_1220","MAD_1230","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MCA_1000","MCA_1010","MCA_1050","MCA_1060","MCA_1070","MCA_1080","MCA_1100","MCA_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MCP_1000","MCP_1010","MCP_1030","MCP_1040","MCP_1050","MCP_1060","MCP_1070","MCP_1080","MCP_1090","MCP_1100","MCP_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PA4_1000","PA4_1010","PA4_1020","PA4_1030","PA4_1040","PA4_1050","PA4_1060","PA4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PTX_1000","PTX_1010","PTX_1020","PTX_1030","PTX_1040","PTX_1050","PTX_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","VNUM","PERIOSTIN","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PI4_1000","PI4_1010","PI4_1020","PI4_1030","PI4_1040","PI4_1050","PI4_1060","PI4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PRG_1000","PRG_1010","PRG_1020","PRG_1030","PRG_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAD_1000","PAD_1010","PAD_1020","PAD_1030","PAD_1040","PAD_1050","PAD_1060","PAD_1070","PAD_1080","PAD_1090","PAD_1100","PAD_1110","PAD_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAL_1000","PAL_1010","PAL_1020","PAL_1030","PAL_1040","PAL_1050","PAL_1060","PAL_1070","PAL_1080","PAL_1090","PAL_1100","PAL_1110","PAL_1120","PAL_1130","PAL_1150","PAL_1160","PAL_1170","PAL_1180","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PTR_1000","PTR_1010","PTR_1050","PTR_1060","PTR_1100","PTR_1140","PTR_1180","PTR_1220","PTR_1260","PTR_1300","PTR_1340","PTR_1380","PTR_1390","PTR_1430","PTR_1470","PTR_1480","PTR_1490","PTR_1500","PTR_1535","PTR_1540","PTR_1550","PTR_1560","PTR_1600","PTR_1610","PTR_1620","PTR_1630","PTR_1670","PTR_1710","PTR_1750","PTR_1790","PTR_1830","PTR_1870","PTR_1910","VNUM_C","VNUM","VDATE","DATE_Q02","DATE_Q03","DATE_Q04","DATE_Q05","DATE_Q07","DATE_Q08","DATE_Q09","DATE_Q10","DATE_Q11","DATE_Q12","DATE_Q13","DATE_Q14","DATE_Q15","DATE_Q16","DATE_Q17","DATE_Q18","DATE_Q19","DATE_Q20","DATE_Q21","DATE_Q22","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","P10_1000","P10_1010","P10_1020","P10_1030","P10_1040","P10_1050","P10_1060","P10_1070","P10_1080","P10_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PCH_1000","PCH_1010","PCH_1020","PCH_1030","PCH_1040","PCH_1050","PCH_1060","PCH_1070","PCH_1080","PCH_1090","PCH_1100","PCH_1110","PCH_1120","PCH_1130","PCH_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","IAQ_1000","IAQ_1010","IAQ_1020","IAQ_1030","IAQ_1040","IAQ_1050","IAQ_1060","IAQ_1070","IAQ_1080","IAQ_1090","IAQ_1100","IAQ_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","REGIMEN_PERIOD_1","REGIMEN_PERIOD_2","REGIMEN_PERIOD_3","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","REG_1080","REG_1150","AGE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SER_1010","SER_1020","SER_1030","SER_1040","SER_1050","SER_1060","SER_1070","SER_1100","SER_1110","SER_1120","SER_1130","SER_1140","SER_1150","SER_1160","SER_1170","SER_1180","SER_1190","SER_1200","SER_1210","SER_1220","SER_1240","SER_1250","SER_1000","SER_1080","SER_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PSE_1060","PSE_1070","PSE_1080","PSE_1090","PSE_1100","PSE_1110","PSE_1130","PSE_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SIG_1000","SIG_1010","SIG_1020","SIG_1030","SIG_1040","SIG_1050","SIG_1060","SIG_1080","SIG_1090","SIG_1100","SIG_1110","SIG_1120","SIG_1130","SIG_1140","SIG_1150","SIG_1160","SIG_1170","SIG_1180","SIG_1190","SIG_1200","SIG_1210","SIG_1220","SIG_1230","SIG_1240","SIG_1250","SIG_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SFU_1000","SFU_1010","SFU_1020","SFU_1030","SFU_1040","SFU_1050","SFU_1060","SFU_1070","SFU_1080","SFU_1090","SFU_1100","SFU_1110","SFU_1120","SFU_1130","SFU_1140","SFU_1150","SFU_1160","SFU_1170","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SNQ_1000","SNQ_1010","SNQ_1020","SNQ_1030","SNQ_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SPI_1010","SPI_1020","SPI_1040","SPI_1050","SPI_1060","SPI_1070","VNUM_C","VNUM","VDATE","SPI_1030","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","AM_TRIAL_TYPE","AM_FVC","AM_FEV1","AM_PEF","AM_FEF2575","AM_FET","DRY_1","DRY_2","DRY_3","DRY_4","DRY_5","DRY_6","DRY_7","DRY_8","DRY_9","VNUM_C","VNUM","VDATE","DDATE","PM_TRIAL_TYPE","PM_FVC","PM_FEV1","PM_PEF","PM_FEF2575","PM_FET","DRY_10","DRY_11","DRY_12","DRY_13","DRY_14","DRY_15","DRY_16","DRY_17","DRY_18","DRY_19","DRY_20","DRY_21","DRY_22","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SLA_1010","SLA_1020","SLA_1030","SLA_1000","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SRE_1010","SRE_1020","SRE_1030","SRE_1040","SRE_1050","SRE_1060","SRE_1070","SRE_1080","SRE_1090","SRE_1100","SRE_1000","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SPU_1000","SPU_1010","SPU_1020","SPU_1030","SPU_1040","SPU_1050","SPU_1060","SPU_1070","SPU_1080","SPU_1090","SPU_1100","SPU_1110","SPU_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SAD_1000","SAD_1010","SAD_1020","SAD_1030","SAD_1040","SAD_1050","SAD_1060","SAD_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SCH_1000","SCH_1010","SCH_1020","SCH_1030","SCH_1040","SCH_1050","SCH_1055","SCH_1060","SCH_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TER_1000","TER_1010","TER_1020","TER_1040","TER_1050","TER_1070","TER_1080","TER_1090","TER_1100","TER_1110","TER_1120","TER_1130","TER_1140","TER_1150","TER_1160","TER_1170","TER_1180","TER_1185","TER_1200","TER_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TXF_1000","TXF_1010","TXF_1020","TXF_1030","TXF_1040","TXF_1050","TXF_1060","TXF_1070","TXF_1080","TXF_1090","TXF_1100","TXF_1110","TXF_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TCH_1000","TCH_1010","TCH_1020","TCH_1030","TCH_1040","TCH_1050","TCH_1060","TCH_1080","TCH_1090","TCH_1100","TCH_1110","TCH_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","WSH_1000","WSH_1010","WSH_1020","WSH_1040","WSH_1050","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","WPA_1000","WPA_1010","WPA_1020","WPA_1030","WPA_1040","WPA_1050","WPA_1060","WPA_1070","WPA_1080","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","HOW_SICK_TODAY","RUNNY_NOSE","PLUGGED_NOSE","SNEEZING","SORE_THROAT","SCRATCHY_THROAT","COUGH","HOARSENESS","HEAD_CONGESTION","CHEST_CONGESTION","FEELING_TIRED","THINK_CLEARLY","SLEEP_WELL","BREATH_EASILY","WALK","DAILY_ACTIVITIES","WORK_OUTSIDE_HOME","WORK_INSIDE_HOME","INTERACT","LIVE_LIFE","MY_COLD_IS","VNUM_C","VNUM","VDATE","WDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER"],["Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Char","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Char","Char","Num","Num","Num","Num","Char","Char","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Char","Num","Char","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Char","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num"],["In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, school, or at home? 0=None of the time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All of the time","During the past 3 days, how often have you had asthma symptoms? 0=Not at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how often have you used your rescue inhaler or nebulizer medication ? 0=Not at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how many total times did your asthma symptoms wake you up from sleep? 0=Not at all, 1=1 time in the last 3 days, 2=2-3 times in the last 3 days, 3=4-5 times in the last 3 days, 4=&gt;=6 times in the last 3 days","How would you rate the amount of impairment you have experienced due to your asthma in the past 3 days? 0=No impairment, 1=Mild impairment, 2=Moderate impairment, 3=Severe impairment, 4=Very severe impairment","How stressed or frightened were you by your asthma symptoms in the past 3 days? 0=Not at all, 1=Mildly, 2=Moderately, 3=Severely, 4=Very severely","Why do you think your asthma was worse in the past 3 days compared to what is normal for you? 0=I have not been worse over the past 3 days. My asthma symptoms have been usual., 1=Common cold, 2=Allergies, 3=Pollution or chemical irritant, 4=Too little asth","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Are you currently retired? 1=Yes,0=No","Are you retired because of asthma? 1=Yes,0=No","Are you currently unemployed? 1=Yes,0=No","Are you unemployed because of asthma? 1=Yes,0=No","Do you get paid to do work? 1=Yes,0=No","How much does your asthma bother you at your paid work? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Overall, how much does your asthma bother you when you do jobs around the house? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Overall, how much does your asthma bother your social life? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Overall, how much does your asthma bother your personal life? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Are you involved in leisure activities, such as: walking for pleasure, sports, exercise, travelling, taking vacations? 1=Yes,0=No","When involved in leisure activities, how much does your asthma bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Would you say that you can't do some of these sorts of things because of asthma? 1=Yes,0=No","How much does your asthma bother you when you sleep? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much does the cost of your asthma medicines bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Do you get free prescriptions? 1=Yes,0=No","How much does the inconvenience or embarrassment of taking your asthma medicines bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much do coughs and colds bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Feeling upset is also a bother.  Does your asthma make you feel anxious, depressed, tired, or helpless? 1=Yes,0=No","Feeling upset is also a bother.  Does your asthma make you feel anxious, depressed, tired, or helpless? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will have an asthma attack when visiting a new place? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will catch a cold? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will let others down? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that your health may get worse in the future? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you won't be able to cope with an asthma attack? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past 4 weeks, how much of the time did your asthma keep you from getting as much done at work, school or at home? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often have your had shortness of breath? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often did your asthma symptoms wake you up at night or earlier than usual in the morning? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often have you used your rescue inhaler or nebulizer medication? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","How would you rate your asthma control during the past 4 weeks? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","ICD-9 category","Description of Adverse Event","ICD9 Code","Ongoing at current visit","Type (1=Intermitent; 2=Continuous)","Severity (1=Mild; 2=Moderate; 3=Severe)","Serious (0=No, 1=Yes)","Likelyhood of Relationship to Study Drug(s) (1=None; 2=Unlikely (remote); 3=Possible; 4=Probable)","Change in Study Drugs (1=Unchanged; 2=Altered)","Outcome (1=Completely recovered; 2=Recovered, but with lasting effects; 3=Death)","Treatment Required (1=None; 2=Medication; 3=Hospitalization; 4=Other)","Ongoing at final visit","Date Started","Date Stopped","Visit Number (character)","Visit Number (numeric)","Long ICD-9 description","Short ICD-9 description","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","ASTHMA HISTORY1.Approximately how old were you when chest symptoms suggesting asthma first appeared?","Did these symptoms appear immediately after or as a result of:1a. a respiratory infection such as a cold or pneumonia? 1=Yes, 0=No, 8=Don't Know","an occupational or job change? 1=Yes, 0=No, 8=Don't Know","a household move? 1=Yes, 0=No, 8=Don't Know","a pregnancy? 1=Yes, 0=No, 8=Don't Know","a hormonal change ? 1=Yes, 0=No, 8=Don't Know","How old were you when a doctor first diagnosed you with asthma?","Have any of your immediate blood relatives been told by a physician that they have asthma? 1=Yes, 0=No, 8=Don't Know","Father  (1=Yes, 0=No, 8=Don't Know)","Brother(s) or Sister(s)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Child(ren)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","ASTHMA SYMPTOMS1.How do you categorize your asthma symptoms throughout the course of the year? 1=Relatively the same all year, 2=Vary by season","Winter? 1=Yes,0=No","Spring? 1=Yes,0=No","Summer? 1=Yes,0=No","Fall? 1=Yes,0=No","In the last 12 months, how many...  5a.Asthma episodes have you had that required emergency care or an unscheduled office visit?","Overnight hospitalizations have you had due to asthma?","Courses of systemic corticosteroid therapy  for asthma have you taken?","Days of work, school, or housework have you missed due to asthma?","In the past 3 months, how many days of work, school, or housework have you missed due to asthma?","Have you ever been admitted to an intensive care unit for asthma? 1=Yes,0=No","How many times have you been admitted to an intensive care unit for asthma?","Have you ever had invasive mechanical ventilation? 1=Yes, 0=No, 8=Don't Know","Have you ever had non-invasive mechanical ventilation? 1=Yes, 0=No, 8=Don't Know","Do any of the following currently provoke your asthma? 1=Yes, 0=No, 8=Don't Know","Menstrual cycle(If participant is male or a postmenopausal female, leave blank.) (1=Yes, 0=No, 8=Don't Know)","Aspirin or non-steroidal anti-inflammatory drugs (e.g., Aleve, Motrin) (1=Yes, 0=No, 8=Don't Know)","Respiratory infections (e.g., colds)  (1=Yes, 0=No, 8=Don't Know)","Irritants (e.g., pollution, odors, perfumes, chemicals, household cleaners) (1=Yes, 0=No, 8=Don't Know)","Weather conditions (e.g., change in weather, humidity)  (1=Yes, 0=No, 8=Don't Know)","Exposure to cold air  (1=Yes, 0=No, 8=Don't Know)","Emotional factors (e.g., stress, laughing)  (1=Yes, 0=No, 8=Don't Know)","Tobacco smoke  (1=Yes, 0=No, 8=Don't Know)","Food additives/preservatives (e.g., MSG, sulfites)  (1=Yes, 0=No, 8=Don't Know)","Allergies (e.g., dust, animals, pollens)(1=Yes, 0=No, 8=Don't Know)","Other (1=Yes,0=No)","To which of the following did a doctor or other health practitioner say you were allergic? 1=Yes, 0=No, 8=Don't Know","Foods  (1=Yes, 0=No, 8=Don't Know)","Things you breathe in or are exposed to (e.g., dust, pollens, molds, animal fur, feathers, dander)  (1=Yes, 0=No, 8=Don't Know)","Stinging insects such as bees or wasps (1=Yes, 0=No, 8=Don't Know)","Latex  (1=Yes, 0=No, 8=Don't Know)","Other  (1=Yes,0=No)","Have you ever had eczema / atopic dermatitis ? 1=Yes, 0=No, 8=Don't Know","If YES, was your eczema diagnosed by a doctor? 1=Yes,0=No","Have any of your immediate blood relatives been told by a physician that they have allergies/eczema/hay fever? 1=Yes, 0=No, 8=Don't Know","Father  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Brother(s) or Sister(s)(1=Yes, 0=No, 8=Don't Know, 9=N/A)","Child(ren)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Did you grow up in a household where you were exposed to tobacco smoke? 1=Yes,0=No","Do you currently smoke? 1=Yes,0=No","Record smoking history in pack-years*. __ __ . __ pack-years","Were you ever a smoker? 1=Yes,0=No","Record smoking history in pack-years*. __ __ . __ pack-years","Do you currently live in a household where you are exposed to tobacco smoke? 1=Yes,0=No","Do you currently vape  or use a hookah ? 1=Yes,0=No","How frequently do you vape or use a hookah? 1=Infrequently , 2=Occasionally , 3=Weekly , 4=Daily","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many days a week","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many times a day","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many years have y","Have you ever vaped or used a hookah in the past? 1=Yes,0=No","Have you ever vaped or used a hookah in the past?  If NO, skip to Q17. Approximately how many years did you vape or use a hookah?  (1-99 years)","Do you currently live in a household where you are exposed to others vaping or using a hookah? 1=Yes,0=No","Do you spend time in social settings  where you are exposed to others vaping or using a hookah? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","When was the last time that you vaped or used a hookah? (number of days from Visit 1)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","How many days were you bothered by coughing during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe was your coughing during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by wheezing during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe was your wheezing during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by shortness of breath during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","Over average, how severe was your shortness of breath during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you awakened at night during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how much of a problem was being awakened at night during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by side effects of your asthma medication during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe were the side effects of your asthma medication during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Participants baseline peak flow (PEF) value (0-999 L/M) First visit (Visit 0A, or Visit 1): PEF (FEF Max) from prebronchodilator (baseline) spirometry at first visit (convert to L/M)","Participants baseline rescue use value (0-99 puffs/day) First visit (Visit 0A, or Visit 1): Self-reported average daily use of albuterol during the 14 days prior to first visit.","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","body mass index computed from height and weight measured at Visit 1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Page Number","Reason for change in scheduled medications status (1=Adverse Event, 2=Other (specify))","Reason for change in scheduled medications status Related adverse event number (0-99)","Current status of participants scheduled medications (1=Discontinued, 2=Resumed)","Date change took effect","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Code to identify a drug unit of measure; (Units)","Code to identify how frequently a drug is taken or administered. (Frequency)","Code to identify the route used to administer a drug. (Route)","Name of Medication","This number corresponds to the ID number of the AHFS monograph. (Medication Code)","Related Event","Medication Dose","Ongoing at current visit","Ongoing at final visit","Start Date","Stop Date","Visit Number (character)","Visit Number (numeric)","GEN_DRUG_NAME","CLASS_TEXT","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Diary and Peak Flow Compliance: Number of full days since the last visit (0-999 days)","Diary and Peak Flow Compliance: Number of days where AM and PM scheduled sessions are complete (AM and PM PEF and all diary questions for AM and PM answered) (0-999 days)","Diary and Peak Flow Compliance: Percent compliance If the compliance value in Q1c is less than 75Pct, re-emphasize the importance of completing scheduled diary assessments and peak flows. (0-150.0Pct)","Scheduled Respimat Compliance (Visits 2-9 only): Number of scheduled puffs since the last visit (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Number of remaining puffs on scheduled Respimat Reference all returned Respimats (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Number of puffs taken (60 x (# used Respimats)) - Q2b (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Percent compliance=Q2c/Q2a x 100 If the participant took less than 75Pct of the scheduled Respimat puffs, re-emphasize the importance of maintaining the daily dosing schedule. (0-150.0Pct)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of scheduled puffs since the last visit (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of remaining puffs reflected on scheduled Twisthaler counters Total the values reflected on all counters for all returned Twisthalers (i.e., out of their pouches). (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of puffs taken (60 x (# used Twisthalers)) - Q3b (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Percent compliance=Q3c/Q3a x 100 If the participant took less than 75Pct of the scheduled Twisthaler puffs, re-emphasize the importance of maintaining the daily dosing schedule. (0-150.0Pct)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","1=mometasone, 2=placebo","Blinded Scheduled Twisthaler Contents - How sure are you about your answer in Q1? 1=Absolutely sure - I know what the Twisthaler/MDI contains, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","1=tiotropium, 2=placebo","Blinded Scheduled Respimat Contents - How sure are you about your answer in Q4? 1=Absolutely sure - I know what the Respimat capsules contain, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has the participant or parent/legal guardian signed the SIENA Informed Consent document? 1=Yes,0=No","Ages 12-17 Only: Has the participant signed and dated the assent form or, if the participant is less than the local age of assent, has the participant given verbal assent? 1=Yes,0=No","Has the participant used an inhaled corticosteroid in the past 6 weeks? 1=Yes,0=No","Will the participant be 12 years of age, or older, as of Visit 1? 1=Yes,0=No","Has the participant used an inhaled corticosteroid or combination inhaled corticosteroid/LABA in the past 3 weeks? 1=Yes,0=No","If NO, has the participant used a leukotriene modifier in the past 3 weeks? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used inhaled corticosteroid therapy?","Over the past 3 months, on average, how many days per week has the participant used inhaled corticosteroid therapy: &lt; 5? 1=Yes,0=No","Over the past 3 months, has the participant used greater than the equivalent of 80-240 mcg of inhaled beclomethasone daily? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used combination inhaled corticosteroid/LABA therapy?","Over the past 3 months, on average, how many days per week has the participant used combination inhaled corticosteroid/LABA therapy,  &lt; 5 (1=Yes,0=No)","Over the past 3 months, on average, how many days per week has the participant had daytime asthma symptoms?","Over the past 3 months, on average, how many days per week has the participant had daytime asthma symptoms,  &lt;=2? 1=Yes,0=No","Over the past 3 months, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms?","Over the past 3 months, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms, &lt;=2? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used his/her short acting betaagonist  for relief of symptoms?","Over the past 3 months, on average, how many days per week has the participant used his/her short acting betaagonist  for relief of symptoms, &lt; 2? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 80Pct? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 70Pct? 1=Yes,0=No","Does the participant have current evidence of any of the conditions listed on the Exclusionary Medical Conditions for SIENA  reference card, or any chronic diseases that would prevent participation in the trial or put the participant at risk by participati","Is the participant currently taking any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Is the participant able to go off these medications for the required washout period prior to Visit 1 and for the duration of the study? 1=Yes,0=No","Is the participant able to use the spirotel e-diary/PEF meter correctly, as evidenced by achieving a score of 13 on the spirotel Performance Checklist ? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Has the participant or parent/legal guardian signed the SIENA Informed Consent document?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past month, on average, how many days per week has the participant had daytime symptoms?","In the past month, on average, how many days per week has the participant had daytime symptoms, &lt;=2? 1=Yes,0=No","In the past month, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms?","In the past month, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms,  &lt;=2? 1=Yes,0=No","In the past month, on average, how many days per week has the participant used his/her short acting beta-agonist  for relief of symptoms?","In the past month, on average, how many days per week has the participant used his/her short acting beta-agonist  for relief of symptoms, &lt; 2? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 80Pct? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 70Pct? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant or parent/legal guardian sign the SIENA Informed Consent document? 1=Yes,0=No","Ages 12-17 Only: Has the participant signed and dated the assent form or, if the participant is less than the local age of assent, has the participant given verbal assent? 1=Yes,0=No","Is the participant 12 years of age, or older? 1=Yes,0=No","Does the participant plan to move away from the clinical site in the upcoming 11 months such that his/her ability to complete the study will be jeopardized? 1=Yes,0=No","Has the participant used investigative drugs and/or enrolled in an intervention trial in the past 30 days, or have plans to enroll in such a trial during the SIENA study? 1=Yes,0=No","Has the participant had a respiratory infection within the past 6 weeks? 1=Yes,0=No","Has the participant had a respiratory infection within the past 4 weeks? 1=Yes,0=No","During the past 4 weeks, on average, how many days per week has the participant had daytime asthma symptoms?","During the past 4 weeks, on average, how many days per week has the participant had daytime asthma symptoms  &gt; 2? 1=Yes,0=No","During the past 4 weeks, how many nighttime awakenings due to asthma symptoms has the participant had?","During the past 4 weeks, how many nighttime awakenings due to asthma symptoms has the participant had,  &gt; 2? 1=Yes,0=No","During the past 4 weeks, on average, how many days per week has the participant used his/her short-acting beta-agonist  for relief of symptoms?","During the past 4 weeks, on average, how many days per week has the participant used his/her short-acting beta-agonist  for relief of symptoms, &gt; 2? 1=Yes,0=No","Is Q7a, Q8a, or Q9a checked YES? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Did the participant or parent/legal guardian sign the SIENA Informed Consent document?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Does the participant have current evidence of any of the conditions listed on the Exclusionary Medical Conditions for SIENA  reference card, or any chronic diseases  that would prevent participation in the trial or put the participant at risk by participat","Does the participant have a history of bladder-neck obstruction, urinary retention, benign prostatic hyperplasia , or a clinically relevant urologic disorder that precludes study participation? 1=Yes,0=No","Does the participant have a history of narrow angle glaucoma? 1=Yes,0=No","Does the participant have a history of significant cardiovascular disorders or arrhythmias? 1=Yes,0=No","Has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card within the specified time periods? 1=Yes,0=No","Is the participant currently taking prescription or OTC medication other than those listed on the Allowed Medications  reference card? 1=Yes,0=No","Based on input from the participant and the study physician, will the participant need to use intranasal steroids at any time during the study? 1=Yes,0=No","Is the participant willing to use a single intranasal steroid at a stable dose continuously for the duration of the study, starting at Visit 1? 1=Yes,0=No","Is the participant currently receiving allergen immunotherapy  other than an established maintenance regimen implemented continuously for a minimum of 3 months? 1=Yes,0=No","Has the participant taken omalizumab within the past 3 months? 1=Yes,0=No","Has the participant used any smokeless tobacco products  in the past year? 1=Yes,0=No","Has the participant smoked cigarettes, a pipe, cigar, marijuana, electronic cigarettes, or any other substance in the past year? 1=Yes,0=No","Ages 18+ Only: Does the participant have a smoking history of greater than 10 pack-years? 1=Yes,0=No","Ages 12-17 Only: Does the participant have a smoking history of greater than 5 pack-years? 1=Yes,0=No","Has the participant received a physician diagnosis of asthma at least 12 months ago? 1=Yes,0=No","Has the participant received a physician diagnosis of asthma at least 12 months ago or had a history consistent with asthma for the previous 12 months? 1=Yes,0=No","Has the participant experienced a life-threatening asthma exacerbation requiring treatment with intubation or mechanical ventilation in the past 5 years? 1=Yes,0=No","Has the participant had an asthma exacerbation requiring systemic corticosteroid treatment in the past 6 weeks? 1=Yes,0=No","Has the participant used an oral or inhaled corticosteroid, or leukotriene modifier in the past 6 weeks? 1=Yes,0=No","Has the participant used an oral corticosteroid in the past 6 weeks? 1=Yes,0=No","Has the participant used an inhaled corticosteroid or leukotriene modifier in the past 3 weeks? 1=Yes,0=No","Is the participant potentially able to bear children? 1=Yes, 0=No, 9=N/A","Is the participant currently pregnant or lactating? 1=Yes,0=No","Does the participant agree to use one of the approved methods indicated on the Birth Control Methods  reference card for the duration of the study? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was the participants prebronchodilator FEV1 &gt;=70Pct predicted? 1=Yes,0=No","Did the participants FEV1 improve &gt;=12Pct and &gt;=200 ml in response to four puffs of albuterol? 1=Yes,0=No","Does the participant have valid source documentation within the past 6 months for an overread AsthmaNet methacholine challenge  with a PC20 &lt;=16 mg/ml? 1=Yes,0=No","Did the participants FEV1 improve &gt;=12Pct and &gt;=200 ml in response to four puffs of albuterol?","Was the participants methacholine PC20 &lt;=16 mg/ml? 1=Yes,0=No","Is the participant able to use the spirotel e-diary/PEF meter correctly, as evidenced by achieving a score of 13 on the Spirotel Performance Checklist ? 1=Yes, 0=No, 9=N/A","Has the participant used the Twisthaler previously? 1=Yes,0=No","Is the participant able to use a Twisthaler properly, as evidenced by achieving a score of 12 on the Twisthaler Inhalation Technique Checklist ? 1=Yes,0=No","Is the participant able to use a metered dose inhaler properly, as evidenced by achieving a score of 11 on the MDI Inhalation Technique Checklist  ? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Does the participant have valid source documentation within the past 6 months for an overread AsthmaNet methacholine challenge  with a PC20 &lt;=16 mg/ml?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Since Visit 1, has the participant experienced a treatment failure event as defined in the protocol? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Since Visit 1, has the participant experienced one or more asthma exacerbations as defined in the protocol? 1=Yes,0=No","Since Visit 1, has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant provide two acceptable sputum induction samples during the run-in? 1=Yes,0=No","Since Visit 1, has the participant experienced a treatment failure event as defined in the protocol? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Did the participant complete Asmanex220 mcg treatment at least 6 weeks ago? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Since Visit 1, has the participant experienced one or more asthma exacerbations as defined in the protocol? 1=Yes,0=No","Since Visit 1, has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Has the participant experienced a respiratory tract infection in the past 6 weeks? 1=Yes,0=No","Has the participant experienced a respiratory tract infection in the past 4 weeks? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report daytime asthma symptoms daily during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report night-time awakenings due to asthma symptoms more than once per week during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report night-time awakenings due to asthma symptoms more than twice per week during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report daily albuterol use for asthma symptom control during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant record at least 75Pct of AM and PM peak flow measurements and symptoms on his or her spirotel during the run-in? 1=Yes,0=No","Did the participant take at least 75Pct of the required puffs from his or her Respimat during the run-in? 1=Yes,0=No","Was the participants prebronchodilator FEV1 &gt;=70Pct predicted? 1=Yes,0=No","Has the participant used the Respimat previously? 1=Yes,0=No","Is the participant able to use a Respimat inhaler properly, as evidenced by achieving a score of 15 on the Respimat Inhalation Technique Checklist? 1=Yes,0=No","Does the participant wish to withdraw consent from the study? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has QC procedure been performed on the NIOX MINO today? 1=Yes,0=No","Did the participant eat or drink within the past hour? 1=Yes,0=No","Did the participant take part in strenuous activity/exercise within the past hour? 1=Yes,0=No","Time eNO started (based on a 24-hour clock) (0000-2359)","ENO Measurement","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","How long have you lived in the current house? (Estimate if uncertain.) (1010) years","How long have you lived in the current house? (Estimate if uncertain.)(1020) months","Does your house use a wood burning stove as a primary source of heat? 1=Yes, 0=No, 8=Don't Know","Does your house use an air conditioner? 1=Yes, 0=No, 8=Don't Know","Does your house use an evaporative cooler ? 1=Yes, 0=No, 8=Don't Know","Does your house use a humidifier? 1=Yes, 0=No, 8=Don't Know","Does your house use a dehumidifier? 1=Yes, 0=No, 8=Don't Know","Has there been water damage to your house, basement, or its contents during the past 12 months? 1=Yes, 0=No, 8=Don't Know","Has there been any mold or mildew, on any surfaces, inside your house in the past 12 months? 1=Yes, 0=No, 8=Don't Know","Which rooms have or have had mold or mildew?10a.Bathroom(s)  (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew?10b.Basement or attic (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10c.Kitchen (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10d.Your bedroom (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10e.Other bedrooms (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10f.Living or family room (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10g.Other (1 = Yes, 0 = No)","Do you ever see cockroaches in your house? 1=Yes,0=No","In which room(s) have you seen cockroaches?12a.Kitchen (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches?  12b.Basement or attic (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12c.Bathroom(s) (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12d.Living or family room (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12e.Your bedroom  (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12f.Other bedrooms (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12g.Garage (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12h.Other (1 = Yes, 0 = No)","Do you ever see rodents  or rodent droppings in your house? 1=Yes,0=No","14.In which room(s) have you seen rodents or rodent droppings? 14a. Kitchen (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14b. Basement or attic (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14c. Bathroom(s) (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings?14d. Living or family room (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14e. Your bedroom(1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14f. Other bedrooms (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14g. Garage (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14h. Other (1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15a. Barns(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property?15b. Hay(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15c. Woodsheds(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15d. Firewood(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property?15e. Chicken coops (1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15f. Corral (1 = Yes, 0 = No)","CHARACTERISTICS OF THE PARTICIPANT'S BEDROOM16.What is the floor covering in your bedroom? 1=Rug/carpet, 2=Vinyl tile or linoleum, 3=Wood, 4=Ceramic tile, 5=Other , 9=Don't know","What type of mattress is on your bed? 1=None, 2=Inner spring mattress, 3=Foam mattress, 4=Waterbed, 5=Air mattress, 6=Other , 9=Don't know","Is the mattress completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","Does your bed have a box spring? 1=Yes,0=No","Is the box spring completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","What type of pillow do you usually sleep with? 1=None, 2=Feather/down, 3=Foam/Dacron/synthetic, 5=Other , 9=Don't know","Is the pillow completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","Does your household have any pets? 1=Yes,0=No","Enter the number of pets that the household has. (Enter '00' if none. If none to Q24a - Q24d, skip to the next question.)24a.Cat","Cat  (1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has. Dog","Dog(1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has 24c.Rabbit, guinea pig, hamster, gerbil, or mouse","Rabbit, guinea pig, hamster, gerbil, or mouse (1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has 24d.Bird","Bird (1=Indoor, 2=Outdoor, 3=Both)","In general, and on a regular basis, are you exposed to any of the following animals?  25a.Cat(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25b.Dog(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25c.Rabbit, guinea pig, hamster, gerbil, or mouse(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25d.Bird (1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25e.Farm animals (1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals?  25f.Other (1 = Yes, 0 = No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other (specify","Highest education of any household member (1=No High School diploma/GED, 2=High Schol grad, 3=Technical training, 4=Some college, no degree, 5=Assoc degree, 6=Bachelor degree, 7=Masters degree/MD/PhD/JD/PharmD, 9=Decline to answer, 10=Don't know)","Category best describes the combined annual income, before taxes, of all members of your household for the last year. (1=Less than $25,000, 2=$25,000 - $49,999, 3=$50,000 - $99,999, 4=$100,000 or more, 9=Decline to answer, 10=Don't know)","How many people  are supported by this income reported in Q3?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Total IgE","Cat (E1)","Dog (E5)","Mouse (E72)","Rat (E74)","Mold mix (MX1)","Cockroach, German (I6)","Grass mix (GX2)","Tree mix (TX4)","Tree mix (TX6)","Weed mix (WX1)","Weed (W3)","Mite (D2)","Mite (D1)","Number or allergen specific (Cat, Dog, Mouse, etc) that were either POS or greater than 0.35","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","CBC with differential cell count: Eosinophils (absolute count) (0-9999/uL)","CBC with differential cell count: WBC (0-999.9 K/uL)","CBC with differential cell count: Differential Lymphocytes (0-99.9Pct)","CBC with differential cell count: Differential Monocytes (0-99.9Pct)","CBC with differential cell count: Differential Basophils (0-99.9Pct)","CBC with differential cell count: Differential Neutrophils (0-99.9Pct)","CBC with differential cell count: Differential Eosinophils (0-99.9Pct)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","PARENTAL HEIGHT - First study visit only or until both are completed 1.Biological mother's height (complete height or check unknown) __ feet","PARENTAL HEIGHT - First study visit only or until both are completed1.Biological mother's height (complete height or check unknown) __ __ inches","PARENTAL HEIGHT - First study visit only or until both are completed 1.Biological mother's height (complete height or check unknown) (9=Don't Know)","Biological father's height (complete height or check unknown) __ feet","Biological father's height (complete height or check unknown) __ __ inches","Biological father's height (complete height or check unknown) (9=Don't Know)","PARTICIPANT MEASUREMENTS - Complete at all applicable study visits3. What type of height measurement was obtained? 1=Standing height, 2=Length","First measurement __ __ __ . __ cm","Second measurement __ __ __ . __ cm","Third measurement __ __ __ . __ cm","Average height or length measurement __ __ __ . __ cm","In your judgment, was the participant's height or length measurement acceptable? 1=Yes,0=No","Weight (shoes off, light clothing) __ __ __ . __ kg","ORAL CANDIDIASIS5.Does the participant have evidence of oral candidiasis? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Post Diluent FEV1","Did the participant drop20Pct at the diluent stage? 1=Yes,0=No","Last concentration of methacholine administered __ __ . __ __ __ __ mg/ml","FEV1 after last concentration of methacholine administered","Did the participant achieve a PC20? 1=Yes,0=No","PC20 __ __ . __ __ mg/ml","Time methacholine challenge ended (based on 24-hour clock)","FEV1","Time of FEV1 in Q7a (based on 24-hour clock)","Was the FEV1 from Q7athe methacholine reversal reference value  in the gray box above? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was an additional treatment used in the first hour? 1=Yes,0=No","Additional albuterol by MDI (1=Yes,0=No)","Number of additional puffs of albuterol administered (1=2, 2=4, 3=&gt;4)","Nebulized Beta-agonist(1=Yes,0=No)","Subcutaneous epinephrine(1=Yes,0=No)","Implementation of clinic emergency protocol or algorithm (1=Yes,0=No)","Other  (1=Yes,0=No)","Participant's FEV1 after additional treatment within first hour.2a.FEV1","Time of FEV1 in Q2a (based on 24-hour clock)","Was the FEV1 from Q2a &gt;the methacholine reversal reference value  in the gray box on the Methacholine Challenge Testing  form? 1=Yes,0=No","Was additional treatment used after one hour? 1=Yes,0=No","Additional albuterol by MDI (1=Yes,0=No)","Number of additional puffs of albuterol administered (1=2, 2=4, 3=&gt;4)","Nebulized Beta-agonist  (1=Yes,0=No)","Subcutaneous epinephrine(1=Yes,0=No)","Implementation of clinic emergency protocol or algorithm  (1=Yes,0=No)","Treatment in the emergency room  (1=Yes,0=No)","Overnight hospitalization (1=Yes,0=No)","Other (1=Yes,0=No)","Participant's final FEV1 after methacholine challenge4a.FEV1","Time of FEV1 in Q4a (based on 24-hour clock)","Was the FEV1 from Q4a ? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Exclusions and Confounders1.Has the participant had any severe acute illness in the past 4 weeks? 1=Yes,0=No","If YES, has the participant received permission from the supervising physician to proceed with the methacholine challenge testing? 1=Yes,0=No","Has the participant used an oral or injectable steroid  in the last 4 weeks? 1=Yes,0=No","Does the participant have a baseline  FEV1 less than 55Pct of predicted or less than 1.0 L? 1=Yes,0=No","Pregnancy test results:  (Check N/A if the participant is male, or is female and is post-menopausal, had a hysterectomy or tubal ligation.) (1=Positive, 0=Negative, 9=N/A)","Is the participant's systolic blood pressure &gt; 200 mm Hg or diastolic blood pressure &gt; 100 mm Hg? 1=Yes,0=No","Is there any other reason the participant should not proceed with the methacholine challenge testing? 1=Yes,0=No","Is the participant eligible to proceed with the diluent  pulmonary function testing for the methacholine challenge? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Exclusions and Confounders:  1.Has the participant had any severe acute illness in the past 4 weeks? 1=Yes,0=No","If YES, has the participant received permission from the supervising physician to proceed with the methacholine challenge testing? 1=Yes,0=No","During the past 4 weeks, has the participant had any respiratory infections, colds, or bronchitis ? 1=Yes,0=No","If YES, during the past 2 weeks, has the participant had any respiratory infections, colds, or bronchitis ? 1=Yes,0=No","Has the participant used an oral or injectable steroid  in the last 4 weeks? 1=Yes,0=No","Does the participant have a baseline  FEV1 less than 70Pct of predicted? 1=Yes,0=No","Pregnancy test results(Check N/A if the participant is male, or is female and has not started menses.)(1=Positive, 0=Negative, 9=N/A)","If participant's age is ? 1=Yes,0=No","If participant's age is &lt; 12 years: Is the participant's systolic blood pressure &gt; 180 mm Hg or diastolic blood pressure &gt; 90 mm Hg? 1=Yes,0=No","Is there any other reason the participant should not proceed with the methacholine challenge testing? 1=Yes,0=No","Is the participant eligible to proceed with the diluent  pulmonary function testing for the methacholine challenge? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time albuterol administered (based on 24-hour clock)","Time post-albuterol spirometry started (based on 24-hour clock)","Highest FVC","Highest FEV1","Highest FEV1 (Pct predicted)","FEF Max","FEF25-75","In your judgment, was the subject's spirometry technique acceptable? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Please check the box next to the treatment that you believe you received over the past 12 weeks. 1=mometasone, 2=placebo","Blinded Scheduled Twisthaler Contents: How sure are you about your answer to Question 2? 1=Absolutely sure -  I know what the MDI contains, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","Please comment with respect to any observations you made that helped you make your choice in Question 2 . 1=I have no comments, 2=I noticed the following:","Please check the box next to the treatment that you believe you received over the past 12 weeks. 1=tiotropium, 2=placebo","How sure are you about your answer to Question 5? 1=Absolutely sure - I know what the Respimat capsules contain, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","1=I have no comments, 2=I noticed the following:","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Visit Number (numeric)","Level of periostin in peripheral blood (ng/ml)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time ipratropium administered (based on 24-hour clock) (0000-2359)","Time post-ipratropium spirometry started (based on 24-hour clock) (0000-2359)","Highest FVC (0-9.99 L)","Highest FEV1 (0-9.99 L)","Highest FEV1 (Pct predicted) (0-999 Pct predicted)","FEF Max (0-99.99 L/S)","FEF25-75 (0-9.99 L/S)","In your judgment, was the participants spirometry technique acceptable? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Is the participant unable to bear children due to any of the following reasons? 1=Yes,0=No","Post-menopausal (at least one year since last menses) (1=Yes,0=No)","Hysterectomy(1=Yes,0=No)","Tubal ligation(1=Yes,0=No)","Pregnancy test results  (1=Positive, 0=Negative)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Have you had any diseases, illnesses, conditions, or surgeries related Blood, Lymph, or Immune Systems? 1=Yes,0=No","Eyes(1=Yes,0=No)","Breasts (1=Yes,0=No)","Endocrine Systems (1=Yes,0=No)","Heart and Blood Vessels  (1=Yes,0=No)","Liver or Pancreas (1=Yes,0=No)","Kidneys or Urinary Tract System  (1=Yes,0=No)","Reproductive System (1=Yes,0=No)","Muscles or Bones (1=Yes,0=No)","Nervous System (1=Yes,0=No)","Psychiatric(1=Yes,0=No)","Drug Allergies (1=Yes,0=No)","Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Skin (1=Yes,0=No)","Have you ever had allergic rhinitis ? 1=Yes, 0=No, 9=Don't know","Have you ever had nasal polyps? 1=Yes, 0=No, 9=Don't know","Do you have chronic or recurrent sinusitis ? 1=Yes, 0=No, 9=Don't know","Have you ever been diagnosed with vocal cord dysfunction? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the ear, nose, or throat? 1=Yes,0=No","Have you ever had pneumonia? 1=Yes, 0=No, 9=Don't know","If YES, were you diagnosed by chest x-ray? 1=Yes, 0=No, 9=Don't know","If YES, were you treated with antibiotics? 1=Yes, 0=No, 9=Don't know","Have you ever had bronchitis? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the lungs ? 1=Yes, 0=No, 9=Don't know","Do you have gastroesophageal reflux disease ? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the stomach or intestines? 1=Yes,0=No","Have you been diagnosed with sleep disordered breathing ? 1=Yes,0=No","If YES, are you being treated with CPAP or BiPAP? 1=Yes,0=No","Have you ever had other sleep disorders? 1=Yes,0=No","Have you ever had other conditions that have not been mentioned on this form? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Short-acting Inhaled Beta-Agonists by Inhaler (e.g., albuterol, Primatene Mist, Maxair, ProAir, Proventil, Ventolin, Xopenex)  (1=Yes, 0=No, 9=Don't Know)","If YES, indicate average weekly puffs in the past month (Enter '000' if none used) weekly puffs","Rescue treatment via a Nebulizer Machine (e.g., albuterol, ipratropium, Combivent, Xopenex, levalbuterol)  (1=Yes, 0=No, 9=Don't Know)","Long-acting Inhaled Beta-Agonists (e.g., Serevent, Foradil, salmeterol, formoterol) Do not consider combination medications. (1=Yes, 0=No, 9=Don't Know)","Oral Beta-Agonists (e.g., albuterol, Brethine, Bricanyl, metaproterenol, Proventil, Ventolin, Repetabs, Volmax)  (1=Yes, 0=No, 9=Don't Know)","Oral Theophylline (short-acting or sustained release)(e.g., Aminophylline, Slo-Phyllin, Slo-bid, Theo-Dur, Uniphyl)(1=Yes, 0=No, 9=Don't Know)","Inhaled Anticholinergic by Inhaler (e.g., Atrovent, Combivent, Spiriva) (1=Yes, 0=No, 9=Don't Know)","Leukotriene Antagonist / 5LO Inhibitors (e.g., Accolate, Zyflo, Singulair) (1=Yes, 0=No, 9=Don't Know)","IgE Blocker (e.g., Xolair)(1=Yes, 0=No, 9=Don't Know)","Oral Steroids FOR ASTHMA (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron,  dexamethasone)(1=Yes, 0=No, 9=Don't Know)","If YES, in the past 12 months, how many courses of steroids by mouth have you taken FOR ASTHMA? 1=1 course, 2=2 courses, 3=3 courses, 4=4 courses, 5=5 courses, 6=More than 5 courses","Injectable Steroids FOR ASTHMA (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV)(1=Yes, 0=No, 9=Don't Know)","Steroids by Inhaler (e.g., Asmanex Twisthaler, QVAR, Flovent, Pulmicort Flexhaler)(1=Yes, 0=No, 9=Don't Know)","Indicate most recent type of inhaled steroid taken (refer to PRIOR_TRT_CARD reference card) code","Indicate number of daily puffs used daily puffs","Indicate the total number of months that you used the inhaled steroid out of the past 12 months months","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide) (1=Yes, 0=No, 9=Don't Know)","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide). If YES, complete Q13a - Q13c. Indicate most recent type of nebulized steroid taken (refer to PRIOR_TRT_CARD reference card)","Indicate number of daily treatments used daily treatments","Indicate the total number of months that you used the nebulized steroid out of the past 12 months months","Long-Acting Beta-Agonist and Inhaled Steroid Combination Medications (e.g., Advair Diskus, Symbicort MDI, Dulera MDI)(1=Yes, 0=No, 9=Don't Know)","Indicate most recent type of combination medication taken (refer to PRIOR_TRT_CARD reference card) code","Indicate number of daily puffs used daily puffs","Indicate the total number of months that you used the combination medication out of the past 12 months months","Nasal Steroids (e.g., Beconase, Vancenase, Flonase, Nasacort, Nasalide, Nasarel, Omnaris, Rhinocort, Nasonex)(1=Yes, 0=No, 9=Don't Know)","Non-steroidal Anti-allergic Nasal Medications (e.g., Nasalcrom, Astelin, Astepro, ipratropium)(1=Yes, 0=No, 9=Don't Know)","Anti-allergic Oral Medications (e.g., fexofenadine, loratadine, cetirizine, chlorpheniramine) (1=Yes, 0=No, 9=Don't Know)","Topical Steroids - Prescription (e.g., Synalar, Lidex, Dermacin, Fluocinonide) (1=Yes, 0=No, 9=Don't Know)","Topical Steroids - OTC (e.g., Hydrocortisone - multiple strengths and products) (1=Yes, 0=No, 9=Don't Know)","Other Medication FOR ASTHMA OR ALLERGIES (1=Yes, 0=No, 9=Don't Know)","Oral Steroids for Conditions Other Than Asthma (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) (1=Yes, 0=No, 9=Don't Know)","Injectable Steroids for Conditions Other Than Asthma (e.g., Medrol, Solumedrol, Decadron, dexamethasone,  triamcinolone, Kenalog, hydrocortisone IV)(1=Yes, 0=No, 9=Don't Know)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Short-acting Inhaled Beta-Agonists by Inhaler (e.g., albuterol, Primatene Mist, Maxair, ProAir, Proventil, Ventolin, Xopenex) Date","Rescue treatment via a Nebulizer Machine (e.g., albuterol, ipratropium, Combivent, Xopenex, levalbuterol) Date","Long-acting Inhaled Beta-Agonists (e.g., Serevent, Foradil, salmeterol, formoterol) Do not consider combination medications. Date","Oral Beta-Agonists (e.g., albuterol, Brethine, Bricanyl, metaproterenol, Proventil, Ventolin, Repetabs, Volmax) Date","Inhaled Anticholinergic by Inhaler (e.g., Atrovent, Combivent, Spiriva) Date","Leukotriene Antagonist / 5LO Inhibitors (e.g., Accolate, Zyflo, Singulair) Date","IgE Blocker (e.g., Xolair) Date","Oral Steroids FOR ASTHMA (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) Date","Injectable Steroids FOR ASTHMA (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV) Date","Steroids by Inhaler (e.g., Asmanex Twisthaler, QVAR, Flovent, Pulmicort Flexhaler) Date","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide) Date","Long-Acting Beta-Agonist and Inhaled Steroid Combination Medications (e.g., Advair Diskus, Symbicort MDI, Dulera MDI) Date","Nasal Steroids (e.g., Beconase, Vancenase, Flonase, Nasacort, Nasalide, Nasarel, Omnaris, Rhinocort, Nasonex) Date","Non-steroidal Anti-allergic Nasal Medications (e.g., Nasalcrom, Astelin, Astepro, ipratropium) Date","Anti-allergic Oral Medications (e.g., fexofenadine, loratadine, cetirizine, chlorpheniramine) Date","Topical Steroids - Prescription (e.g., Synalar, Lidex, Dermacin, Fluocinonide) Date","Topical Steroids - OTC (e.g., Hydrocortisone - multiple strengths and products) Date","Other Medication FOR ASTHMA OR ALLERGIES Date","Oral Steroids for Conditions Other Than Asthma (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) Date","Injectable Steroids for Conditions Other Than Asthma (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV) Date","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the last month, how often have you been upset because of something that happened unexpectedly? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that you were unable to control the important things in your life? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt nervous and stressed? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt confident about being able to handle your personal problems? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that things were going your way? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you found that you could not cope with all the things that you had to do? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you been able to control irritations in your life? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that you were on top of things? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you been angered because of things that happened that were outside of your control? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that your difficulties were piling up so high that you could not overcome them? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Have you consumed caffeine in the past 4 hours? 1=Yes,0=No","Have you used medications with caffeine in the past 4 hours? 1=Yes,0=No","Have you used any weight loss medications in the past 4 hours? 1=Yes,0=No","Have you consumed any food containing alcohol or beverages containing alcohol in the past 4 hours? 1=Yes,0=No","Have you used a rescue intermediate-acting inhaled betaagonist in the past 6 hours? 1=Yes,0=No","Have you used a short-acting anticholinergic in the past 6 hours? 1=Yes,0=No","Have you used any ophthalmic antihistamines in the past 6 hours? 1=Yes,0=No","Have you used any nasal antihistamines in the past 6 hours? 1=Yes,0=No","Have you used any nasal decongestants in the past 6 hours? 1=Yes,0=No","Have you used any oral antihistamines in the past 48 hours? 1=Yes,0=No","Have you used any oral decongestants or cold remedies in the past 48 hours? 1=Yes,0=No","Have you used any smokeless tobacco products today? 1=Yes,0=No","At this time, is your asthma worse because of recent exposure to triggers? 1=Yes,0=No","Is there any other reason you should not proceed with spirometry testing? 1=Yes,0=No","Is the participant eligible to proceed with the spirometry testing? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past 4 weeks, I worried about the long-term effects of asthma on my health (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I had to worry about asthma triggers (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, my asthma was on my mind (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, it was hard to get a good nights sleep because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt like I couldnt enjoy life because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt that asthma was controlling my life (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt frustrated that I couldnt make plans in advance because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, everyday activities were a struggle (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, asthma placed stress on my relationships with family, friends, significant others, or co-workers (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, I felt frustrated that I have to do things differently than people who dont have asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt like I missed out on doing things with others because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, I had to do a lot of planning to make sure I always had an inhaler ready (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Randomized treatment assignment period 1: (Mometasone, Tiotropium or Placebo)","Randomized treatment assignment period 2: (Mometasone, Tiotropium or Placebo)","Randomized treatment assignment period 3: (Mometasone, Tiotropium or Placebo)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Sex (1=Male, 2=Female)","Primary Racial Identification for Spirometry (1=Amer Indian or Alaskan Native, 2=Asian or Pacific Islander, 3=Black, 4=White, 5=Hispanic or Latino, 6=Other)","Participant's Age at Visit 1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Description of Adverse Event (ICD9 Code)","Is the participant currently taking study drug? 1=Yes,0=No","Time interval between the last administration of the study drug and the Adverse Event","What was the unit of time for the interval in Question #4? 1=Second, 2=Minute, 3=Hour, 4=Day","Fatal event(1=Yes,0=No)","Life-threatening event (1=Yes,0=No)","Inpatient hospitalization required(1=Yes,0=No)","Hospitalization prolonged  (1=Yes,0=No)","Disabling or incapacitating  (1=Yes,0=No)","Overdose (1=Yes,0=No)","Cancer (1=Yes,0=No)","Congenital anomaly(1=Yes,0=No)","Serious laboratory abnormality with clinical symptoms(1=Yes,0=No)","Height failure (per protocol MOP)(1=Yes,0=No)","Pregnancy(1=Yes,0=No)","Other (1=Yes,0=No)","What in your opinion caused the event? 1=Yes,0=No","Withdrawal of study drug(s)(1=Yes,0=No)","Concurrent medication  (1=Yes,0=No)","Other condition or event  (1=Yes,0=No)","Was the event expected or unexpected? 1=Expected, 2=Unexpected","Was the event possibly, probably, or definitely related to study participation? 1=Yes,0=No","Date of Adverse Event","Admission date","Discharge date","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","What type of height measurement was obtained? 1=Standing height, 2=Length","First measurement __ __ __ . __ cm","Second measurement __ __ __ . __ cm","Third measurement __ __ __ . __ cm","Average height or length measurement __ __ __ . __ cm","In your judgment, was the participant's height or length measurement acceptable? 1=Yes,0=No","Weight (shoes off, light clothing) __ __ __ . __ kg","Does the participant have evidence of oral candidiasis? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant fail to respond within 48 hours to the treatment failure rescue algorithm? 1=Yes,0=No","Did the participant use at least 16 puffs as needed albuterol per 24 hours for a period of 48 hours? 1=Yes,0=No","Did the participant experience prebronchodilator FEV1 values &lt; 50Pct of the baseline prebronchodilator value obtained at Visit 1 on two consecutive spirometric determinations made on different days? 1=Yes, 0=No, 9=Not evaluated","Did the participant experience prebronchodilator FEV1 values &lt; 40Pct of predicted on two consecutive spirometric determinations made on different days? 1=Yes, 0=No, 9=Not evaluated","Did the study or treating physician prescribe the participant oral/parenteral corticosteroids for the treatment of his/her asthma? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation in the opinion of the study investigator or personal physician? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation? 1=Yes,0=No","Did the participant seek care for significant asthma exacerbation conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Was the participant hospitalized? 1=Yes,0=No","Duration of hospital stay (0-99.9 days)","Was intubation or ventilation assistance required? 1=Yes,0=No","Was the participant admitted to the intensive care unit? 1=Yes,0=No","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Inhaled corticosteroids (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Nebulized bronchodilator (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Oral corticosteroids (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, IM or IV steroids (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Antibiotics (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Other (1=Yes, 0=No )","Why do you think the participant experienced a significant asthma exacerbation? 1=Common cold, 2=Allergies, 3=Pollution or chemical irritant, 4=Too little asthma maintenance medication, 5=Exercise, 6=Other","Date exacerbation conditions were met","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Which phone contact following Asthma Exacerbation visit is being performed? 1=Phone Contact #1 , 2=Phone Contact #2 , 3=Phone Contact #3","Since the most recent Asthma Exacerbation contact, has the participant sought additional care for significant asthma exacerbation conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Was the participant hospitalized? 1=Yes,0=No","Duration of hospital stay (0-99.9 days)","Was intubation or ventilation assistance required? 1=Yes,0=No","Was the participant admitted to the intensive care unit? 1=Yes,0=No","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Inhaled corticosteroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Nebulized bronchodilator (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Oral corticosteroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, IM or IV steroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Antibiotics (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Runny Nose  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Post nasal drip  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Need to blow your nose  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Facial pain/pressure(0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Nasal obstruction  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time spirometry started (based on 24-hour clock)","The reported FEV1 and FVC are the best volumes of all acceptable maneuvers. 1.Highest FVC","Highest FEV1 (Pct predicted)","The reported flow rates correspond to the maneuver where FEV1 + FVC is maximized. FEF Max","FEF25-75","In your judgment, was the subject's spirometry technique acceptable? 0=No, 1=Yes","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Highest FEV1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","TRIAL_TYPE","FVC","FEV1","PEF","FEF2575","FET","Number of times the participant woke up last night due to asthma symptoms # Times woke up due to asthma (0-9)","Number of puffs the participant will take from the WHITE study inhaler this morning # Puffs from WHITE inhaler in AM (0-9)","only) Number of puffs the participant will take from the BLUE study inhaler this morning # Puffs from BLUE inhaler in AM (0-9)","Has the participant taken any puffs from his/her RED RESCUE albuterol inhaler in the past 4 hours? 1=yes, 0=no","Shortness of Breath Score Shortness of Breath Score overnight (0,1,2,3)","Chest tightness score Chest Tightness Score overnight (0,1,2,3)","Wheezing score Wheezing Score overnight (0,1,2,3)","Coughing score Coughing Score overnight (0,1,2,3)","Phlegm/Mucus score Phlegm/Mucus Score overnight (0,1,2,3)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","date 'fixed' for analysis so that associated AM and PM always have same ddate","TRIAL_TYPE","FVC","FEV1","PEF","FEF2575","FET","Number of puffs the participant will take from the WHITE study inhaler tonight # Puffs from WHITE inhaler in PM (0-9)","Has the participant taken any puffs from his/her RED RESCUE albuterol inhaler during the past 4 hours? 1=yes, 0=no","Shortness of Breath Score Shortness of Breath Score since waking this morning (0,1,2,3)","Chest tightness score Chest Tightness Score since waking this morning (0,1,2,3)","Wheezing score Wheezing Score since waking this morning (0,1,2,3)","Coughing score Coughing Score since waking this morning (0,1,2,3)","Phlegm/Mucus score Phlegm/Mucus Score since waking this morning (0,1,2,3)","Did your regular exercise cause unusually severe shortness of breath during the past 24 hours? 1=Yes, 0=No, 9=N/A","Number of RED RESCUE albuterol puffs taken in the past 24 hours to prevent symptoms# Preventive  RED RESCUE puffs  in past 24 hours?","Number of RED RESCUE albuterol puffs taken for asthma symptoms or low peak flow during past 24 hours  # RED RESCUE  puffs for symptoms  in past 24 hours?","Number of times used RED RESCUE albuterol inhaler for asthma symptoms during past 24 hours  # Times used  RED RESCUE for symptoms in past 24 hours?","Did RED RESCUE albuterol relieve symptoms for less than 4 hours after treatment during past 24 hours? 1=Yes,0=No","Was the participant seen by a healthcare provider  for an unscheduled visit in the past 24 hours due to asthma symptoms? 1=Yes,0=No","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Processing: Time processing started (based on 24-hour clock) __ __ __ __ (0 - 2359)","Processing Sample: Total Cell Count: __ __ __ __. __x 10^4 cells/ml (0 - 9999.9)","Was the participant's sputum sample processed within 4 hours after collection? 1=Yes,0=No","Processing Sample: Processing Date __ __ / __ __ / 20 __ __","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Rate slide's quality:  (1=Very good, 2=Good, 3=Acceptable, 4=Poor but readable, 5=Not readable)","Record the number on the slide(s) that was (were) read (0 - 9)","Record the number on the slide(s) that was (were) read: These are numbers that were assigned to the slides at each site. (0 -9)","Total Cell Count: __ __ __ __ . __ x 10^4 cells/ml (0 - 9999.9)","Differential Cell Counts: Squamous Cells __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Epithelial Cells __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Macrophages __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Neutrophils __ __ . __ Pct (0 -99.9)","Differential Cell Counts: Eosinophils __ __ . __ Pct (0 - 99.9)","Differential Cell Counts: Lymphocytes __ __ . __ Pct (0 -99.9)","Date of Read __ __ / __ __ / 20 __ __","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","For this protocol, what was the duration of sputum induction the first time the participant's sample was processed within 4 hours after collection?","Sputum induction start time (based on 24-hour clock)","Sputum induction stop time (based on 24-hour clock)","Duration of sputum induction collection phase at this visit -- minutes (0 - 99.9)","Was the duration ? 1=Yes,0=No","Volume of sputum sample at this visit -- ml (0.0 - 999.9)","Is the volume adequate for processing? 1=Yes,0=No","Is the sample adequate for laboratory analysis? 1=Yes,0=No","Participant's FEV1 immediately after completion of sputum induction: -- (0.0 -9.99 L)","FEV1 (Pct predicted) (0 - 999)","Time of FEV1 in Q7a (based on 24-hour clock)","Percent difference in FEV1 ((Reference - Q7a) / Reference) X100 (-99.9 - 99.9 Pct)","Did the participant's FEV1 drop &gt; 10Pct from reference FEV1 as indicated in Q7d? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Participant's FEV1 after initial 2 puffs of albuterol: 1a.FEV1 __ . __ __ L (0 -9.99)","Participant's FEV1 after initial 2 puffs of albuterol: 1b. FEV1 (Pct predicted)__ __ __ Pct predicted (0 - 999)","Participant's FEV1 after initial 2 puffs of albuterol: 1c. Time of FEV1 from Q1a (based on 24-hour clock) __ __ __ __ (0 - 2359)","Participant's FEV1 after initial 2 puffs of albuterol: 1d. Was the FEV1 from Q1a &gt;=the sputum induction reversal reference value in the gray box above? 1=Yes,0=No","Participant's FEV1 after 2 additional puffs of albuterol: 2a. FEV1 __ . __ __ L (0 - 9.99)","Participant's FEV1 after 2 additional puffs of albuterol: 2b. FEV1 (Pct predicted) __ __ __ Pct predicted (0 - 999)","Participant's FEV1 after 2 additional puffs of albuterol: 2c. Time of FEV1 from Q2a (based on 24-hour clock)__ __ __ __ (0 - 2359)","Participant's FEV1 after 2 additional puffs of albuterol: 2d.Was the FEV1 from Q2a ? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was the participant's FEV1 after reversal from the methacholine challenge greater yhan 90Pct of the baseline FEV1 (1=Yes,0=No)","If NO, has the participant received permission from the supervising physician to proceed with sputum induction testing? 1=Yes,0=No","Physician's Signature:_________ (1=signature present, NULL=signature missing)","Participant's FEV1 used for assessment of eligibility for sputum induction __ . __ __ L (0 - 9.99)","Participant's FEV1 (Pct predicted) used for assessment of eligibility for sputum induction __ __ __ Pct predicted (O-999)","Was the participant's FEV1  from Q3 ? 1=Yes,0=No","Has the participant used any smokeless tobacco products  today? 1=Yes,0=No","Is there any other reason the participant should not proceed with sputum induction? 1=Yes,0=No","Is the participant eligible for sputum induction? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has the participant completed the study through Visit 9? 1=Yes,0=No","Who initiated termination of the participant? 1=Participant, 2=Clinical Staff","1=no longer interested in participating*, 2=no longer willing to follow protocol*, 3=difficult access to clinic , 4=unable to make visits during clinic hours, 5=moving out of the area, 6=unable to continue due to personal constraints*, 7=unable to continue","Did clinical staff terminate the participant due to pregnancy? 1=Yes, 0=No, 9=N/A","Did clinical staff terminate the participant due to loss to follow-up? 1=Yes,0=No","Did clinical staff terminate the participant due to loss to follow-up, type of contact (1=In-person visit, 2=Phone call)","Did clinical staff terminate the participant due to an asthma-related adverse event? 1=Yes,0=No","Did clinical staff terminate the participant due to a medication-related adverse event? 1=Yes,0=No","Did clinical staff terminate the participant due to an adverse event not related to asthma or medications? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility during the Supervised Washout ? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with medication dosing? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with diary completion? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with visit attendance? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with peak flow monitoring? 1=Yes,0=No","Did clinical staff terminate the participant due to significant asthma exacerbation or two treatment failures during run-in ? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility during the run-in period  for reasons other than compliance or exacerbation/treatment failure or high sputum eosinophils ? 1=Yes,0=No","Did clinical staff terminate the participant due to other reason? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility due to high sputum eosinophils in the run-in ? 1=Yes,0=No","Did clinical staff terminate the participant due to... Indicate the letter corresponding to the primary reason the participant was terminated. (a-n)","Did clinical staff terminate the participant due to loss to follow-up, date of last contact with participant","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant seek care for treatment failure conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact, other than call to notify clinic as instructed in spirotel","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Inhaled corticosteroids (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Nebulized bronchodilator (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started: Oral corticosteroids form.  (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  IM or IV steroids form.  (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started: Antibiotics (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant awaken from asthma three or more times in a two-week period or on two consecutive nights? 1=Yes,0=No","Did the participant use albuterol for relief of symptoms four or more times/day for two or more consecutive days? 1=Yes,0=No","Did albuterol relieve symptoms for less than four hours after treatment? 1=Yes,0=No","Did the participant use albuterol for relief of symptoms daily for seven days, and this use exceeded two times the weekly use of albuterol in the baseline period? 1=Yes,0=No","Did regular exercise cause unusually severe shortness of breath on two or more days during a seven day period? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation? 1=Yes,0=No","Did the participant experience a treatment failure? 1=Yes,0=No","Has the participant experienced two treatment failures in the current treatment period ? 1=Yes,0=No","Does the study physician feel the study treatment contributed directly to the treatment failure, such that it is in the participants best interest to discontinue the current blinded treatment? 1=Yes,0=No","Did the participant begin daily treatment with high-dose ICS for 10 days? 1=Yes,0=No","Did the participant begin daily treatment with high-dose ICS for 10 days? 1=Did not think symptoms werebad enough, 2=Other","Date treatment failure conditions were met","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","During this last treatment period , has the participant experienced a treatment failure or asthma exacerbation? 1=Yes,0=No","Has the participant experienced two or more treatment failures in this last treatment period, or does the study physician feel the study treatment contributed directly to the treatment failure such that it is in the participants best interest to discontinu","As of today, have 2 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","As of today, have 6 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","As of today, have 3 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Are you currently employed ? 1=Yes,0=No","In general, how many hours per week do you usually work?","During the past seven days, how many hours did you miss from work because of problems associated with your asthma?","During the past seven days, how much did asthma affect your productivity while you were working?","Do you currently attend classes in an academic setting ? 1=Yes,0=No","In general, how many hours per week do you usually attend classes?","During the past seven days, how many hours did you miss from class or school because of problems associated with your asthma?","During the past seven days, how much did asthma affect your productivity while in school or attending classes in an academic setting?","During the past seven days, how much did your asthma affect your ability to do your regular daily activities, other than work at a job or attend classes?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","How sick today? 0=Not Sick, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely","Runny Nose:(0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","PLUGGED_NOSE  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SNEEZING  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SORE_THROAT  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SCRATCHY_THROAT  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","COUGH (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","HOARSENESS  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","HEAD_CONGESTION  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","CHEST_CONGESTION  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","FEELING_TIRED (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","THINK_CLEARLY  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SLEEP_WELL  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","BREATH_EASILY  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WALK  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","DAILY_ACTIVITIES  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WORK_OUTSIDE_HOME  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WORK_INSIDE_HOME  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","INTERACT  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","LIVE_LIFE  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","MY_COLD_IS  (1=Very much better, 2=Somewhat better, 3=A little better, 4=The same, 5=A little worse, 6=Somewhat worse, 7=Very much worse)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Date (number of days from Visit 1)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Num<\/th>\n      <th>Variable<\/th>\n      <th>Type<\/th>\n      <th>Label<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":1},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Num","targets":1},{"name":"Variable","targets":2},{"name":"Type","targets":3},{"name":"Label","targets":4}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p><strong>Option #2</strong></p>
 <div class="sourceCode" id="cb393"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb393-1"><a href="data-import-from-pdf-sources.html#cb393-1" tabindex="-1"></a><span class="co"># Write a function that applies all of the cleaning steps to an dataframe (output = cleaned dataframe)</span></span>
 <span id="cb393-2"><a href="data-import-from-pdf-sources.html#cb393-2" tabindex="-1"></a>clean_tables <span class="ot">&lt;-</span> <span class="cf">function</span>(data) {</span>
@@ -865,8 +874,8 @@ <h3>Cleaning dataframes<a href="data-import-from-pdf-sources.html#cleaning-dataf
 <span id="cb393-21"><a href="data-import-from-pdf-sources.html#cb393-21" tabindex="-1"></a></span>
 <span id="cb393-22"><a href="data-import-from-pdf-sources.html#cb393-22" tabindex="-1"></a><span class="co"># View resulting dataframe</span></span>
 <span id="cb393-23"><a href="data-import-from-pdf-sources.html#cb393-23" tabindex="-1"></a><span class="fu">datatable</span>(tables_clean_unlisted)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-5273ab80765ae03f607c" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-5273ab80765ae03f607c">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80","81","82","83","84","85","86","87","88","89","90","91","92","93","94","95","96","97","98","99","100","101","102","103","104","105","106","107","108","109","110","111","112","113","114","115","116","117","118","119","120","121","122","123","124","125","126","127","128","129","130","131","132","133","134","135","136","137","138","139","140","141","142","143","144","145","146","147","148","149","150","151","152","153","154","155","156","157","158","159","160","161","162","163","164","165","166","167","168","169","170","171","172","173","174","175","176","177","178","179","180","181","182","183","184","185","186","187","188","189","190","191","192","193","194","195","196","197","198","199","200","201","202","203","204","205","206","207","208","209","210","211","212","213","214","215","216","217","218","219","220","221","222","223","224","225","226","227","228","229","230","231","232","233","234","235","236","237","238","239","240","241","242","243","244","245","246","247","248","249","250","251","252","253","254","255","256","257","258","259","260","261","262","263","264","265","266","267","268","269","270","271","272","273","274","275","276","277","278","279","280","281","282","283","284","285","286","287","288","289","290","291","292","293","294","295","296","297","298","299","300","301","302","303","304","305","306","307","308","309","310","311","312","313","314","315","316","317","318","319","320","321","322","323","324","325","326","327","328","329","330","331","332","333","334","335","336","337","338","339","340","341","342","343","344","345","346","347","348","349","350","351","352","353","354","355","356","357","358","359","360","361","362","363","364","365","366","367","368","369","370","371","372","373","374","375","376","377","378","379","380","381","382","383","384","385","386","387","388","389","390","391","392","393","394","395","396","397","398","399","400","401","402","403","404","405","406","407","408","409","410","411","412","413","414","415","416","417","418","419","420","421","422","423","424","425","426","427","428","429","430","431","432","433","434","435","436","437","438","439","440","441","442","443","444","445","446","447","448","449","450","451","452","453","454","455","456","457","458","459","460","461","462","463","464","465","466","467","468","469","470","471","472","473","474","475","476","477","478","479","480","481","482","483","484","485","486","487","488","489","490","491","492","493","494","495","496","497","498","499","500","501","502","503","504","505","506","507","508","509","510","511","512","513","514","515","516","517","518","519","520","521","522","523","524","525","526","527","528","529","530","531","532","533","534","535","536","537","538","539","540","541","542","543","544","545","546","547","548","549","550","551","552","553","554","555","556","557","558","559","560","561","562","563","564","565","566","567","568","569","570","571","572","573","574","575","576","577","578","579","580","581","582","583","584","585","586","587","588","589","590","591","592","593","594","595","596","597","598","599","600","601","602","603","604","605","606","607","608","609","610","611","612","613","614","615","616","617","618","619","620","621","622","623","624","625","626","627","628","629","630","631","632","633","634","635","636","637","638","639","640","641","642","643","644","645","646","647","648","649","650","651","652","653","654","655","656","657","658","659","660","661","662","663","664","665","666","667","668","669","670","671","672","673","674","675","676","677","678","679","680","681","682","683","684","685","686","687","688","689","690","691","692","693","694","695","696","697","698","699","700","701","702","703","704","705","706","707","708","709","710","711","712","713","714","715","716","717","718","719","720","721","722","723","724","725","726","727","728","729","730","731","732","733","734","735","736","737","738","739","740","741","742","743","744","745","746","747","748","749","750","751","752","753","754","755","756","757","758","759","760","761","762","763","764","765","766","767","768","769","770","771","772","773","774","775","776","777","778","779","780","781","782","783","784","785","786","787","788","789","790","791","792","793","794","795","796","797","798","799","800","801","802","803","804","805","806","807","808","809","810","811","812","813","814","815","816","817","818","819","820","821","822","823","824","825","826","827","828","829","830","831","832","833","834","835","836","837","838","839","840","841","842","843","844","845","846","847","848","849","850","851","852","853","854","855","856","857","858","859","860","861","862","863","864","865","866","867","868","869","870","871","872","873","874","875","876","877","878","879","880","881","882","883","884","885","886","887","888","889","890","891","892","893","894","895","896","897","898","899","900","901","902","903","904","905","906","907","908","909","910","911","912","913","914","915","916","917","918","919","920","921","922","923","924","925","926","927","928","929","930","931","932","933","934","935","936","937","938","939","940","941","942","943","944","945","946","947","948","949","950","951","952","953","954","955","956","957","958","959","960","961","962","963","964","965","966","967","968","969","970","971","972","973","974","975","976","977","978","979","980","981","982","983","984","985","986","987","988","989","990","991","992","993","994","995","996","997","998","999","1000","1001","1002","1003","1004","1005","1006","1007","1008","1009","1010","1011","1012","1013","1014","1015","1016","1017","1018","1019","1020","1021","1022","1023","1024","1025","1026","1027","1028","1029","1030","1031","1032","1033","1034","1035","1036","1037","1038","1039","1040","1041","1042","1043","1044","1045","1046","1047","1048","1049","1050","1051","1052","1053","1054","1055","1056","1057","1058","1059","1060","1061","1062","1063","1064","1065","1066","1067","1068","1069","1070","1071","1072","1073","1074","1075","1076","1077","1078","1079","1080","1081","1082","1083","1084","1085","1086","1087","1088","1089","1090","1091","1092","1093","1094","1095","1096","1097","1098","1099","1100","1101","1102","1103","1104","1105","1106","1107","1108","1109","1110","1111","1112","1113","1114","1115","1116","1117","1118","1119","1120","1121","1122","1123","1124","1125","1126","1127","1128","1129","1130","1131","1132","1133","1134","1135","1136","1137","1138","1139","1140","1141","1142","1143","1144","1145","1146","1147","1148","1149","1150","1151","1152","1153","1154","1155","1156","1157","1158","1159","1160","1161","1162","1163","1164","1165","1166","1167","1168","1169","1170","1171","1172","1173","1174","1175","1176","1177","1178","1179","1180","1181","1182","1183","1184","1185","1186","1187","1188","1189","1190"],[1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,1,2,3,4,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,1,2,3,4,5,6,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28],["AAQ_1000","AAQ_1010","AAQ_1020","AAQ_1030","AAQ_1040","AAQ_1050","AAQ_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ABP_1000","ABP_1010","ABP_1020","ABP_1030","ABP_1040","ABP_1050","ABP_1060","ABP_1070","ABP_1080","ABP_1090","ABP_1100","ABP_1110","ABP_1120","ABP_1130","ABP_1140","ABP_1150","ABP_1160","ABP_1170","ABP_1180","ABP_1190","ABP_1200","ABP_1210","ABP_1220","ABP_1230","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ACT_1","ACT_2","ACT_3","ACT_4","ACT_5","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ICD9_CAT","AEC_1000","AEC_1010","AEC_1040","AEC_1050","AEC_1060","AEC_1070","AEC_1080","AEC_1090","AEC_1100","AEC_1110","AEC_1120","AEC_1020","AEC_1030","VNUM_C","VNUM","ICD9LONG","ICD9SHRT","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","AHA_1000","AHA_1020","AHA_1030","AHA_1040","AHA_1050","AHA_1060","AHA_1070","AHA_1090","AHA_1100","AHA_1110","AHA_1120","AHA_1130","AHA_1140","AHA_1150","AHA_1160","AHA_1170","AHA_1180","AHA_1190","AHA_1200","AHA_1210","AHA_1220","AHA_1250","AHA_1260","AHA_1270","AHA_1280","AHA_1290","AHA_1300","AHA_1310","AHA_1320","AHA_1330","AHA_1340","AHA_1350","AHA_1360","AHA_1370","AHA_1380","AHA_1390","AHA_1400","AHA_1410","AHA_1420","AHA_1430","AHA_1440","AHA_1450","AHA_1460","AHA_1470","AHA_1500","AHA_1570","AHA_1580","AHA_1590","AHA_1600","AHA_1730","AHA_1740","AHA_1750","AHA_1760","AHA_1770","AHA_1780","AHA_1790","AHA_1800","AHA_1810","AHA_1820","AHA_1830","AHA_1840","AHA_1850","AHA_1890","AHA_1900","VNUM_C","VNUM","VDATE","DATE_VAPE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ASU_1000","ASU_1010","ASU_1020","ASU_1030","ASU_1040","ASU_1050","ASU_1060","ASU_1070","ASU_1080","ASU_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","BAS_1000","BAS_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","BMI","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAGE","CHG_1000","CHG_1010","CHG_1020","CHG_1030","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","CME_1040","CME_1050","CME_1055","CME_1000","CME_1010","CME_1020","CME_1030","CME_1080","CME_1090","CME_1060","CME_1070","VNUM_C","VNUM","GEN_NAME","CLASS","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","COM_1000","COM_1010","COM_1020","COM_1030","COM_1040","COM_1050","COM_1060","COM_1070","COM_1080","COM_1090","COM_1100","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","CTX_1000","CTX_1010","CTX_1030","CTX_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E0A_1000","E0A_1020","E0A_1030","E0A_1040","E0A_1043","E0A_1045","E0A_1050","E0A_1060","E0A_1070","E0A_1073","E0A_1075","E0A_1080","E0A_1090","E0A_1100","E0A_1110","E0A_1120","E0A_1130","E0A_1140","E0A_1145","E0A_1150","E0A_1160","E0A_1170","E0A_1180","E0A_1190","E0A_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E0B_1000","E0B_1010","E0B_1020","E0B_1030","E0B_1040","E0B_1050","E0B_1060","E0B_1065","E0B_1070","E0B_1080","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E1_1000","E1_1020","E1_1030","E1_1040","E1_1050","E1_1060","E1_1065","E1_1070","E1_1080","E1_1090","E1_1100","E1_1110","E1_1120","E1_1130","E1_1140","E1_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E2_1000","E2_1010","E2_1020","E2_1030","E2_1040","E2_1050","E2_1060","E2_1070","E2_1080","E2_1090","E2_1100","E2_1110","E2_1120","E2_1130","E2_1140","E2_1145","E2_1150","E2_1160","E2_1170","E2_1173","E2_1175","E2_1180","E2_1190","E2_1200","E2_1210","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E3_1000","E3_1010","E3_1020","E3_1030","E3_1070","E3_1080","E3_1090","E3_1100","E3_1110","E3_1120","E3_1130","E3_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E4_1000","E4_1010","E4_1020","E4_1030","E4_1050","E4_1060","E4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E5_1000","E5_1010","E5_1020","E5_1030","E5_1035","E5_1040","E5_1050","E5_1060","E5_1065","E5_1070","E5_1080","E5_1085","E5_1090","E5_1100","E5_1110","E5_1120","E5_1130","E5_1140","E5_1150","E5_1160","E5_1170","E5_1180","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ENO_1000","ENO_1010","ENO_1020","ENO_1040","ENO_1050","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","HEQ_1000","HEQ_1010","HEQ_1020","HEQ_1030","HEQ_1040","HEQ_1050","HEQ_1060","HEQ_1070","HEQ_1080","HEQ_1090","HEQ_1100","HEQ_1110","HEQ_1120","HEQ_1130","HEQ_1140","HEQ_1150","HEQ_1160","HEQ_1170","HEQ_1180","HEQ_1190","HEQ_1200","HEQ_1210","HEQ_1220","HEQ_1230","HEQ_1240","HEQ_1250","HEQ_1260","HEQ_1270","HEQ_1280","HEQ_1290","HEQ_1300","HEQ_1310","HEQ_1320","HEQ_1330","HEQ_1340","HEQ_1350","HEQ_1360","HEQ_1370","HEQ_1380","HEQ_1390","HEQ_1400","HEQ_1410","HEQ_1420","HEQ_1430","HEQ_1440","HEQ_1450","HEQ_1460","HEQ_1470","HEQ_1480","HEQ_1490","HEQ_1500","HEQ_1510","HEQ_1520","HEQ_1530","HEQ_1540","HEQ_1550","HEQ_1560","HEQ_1570","HEQ_1580","HEQ_1590","HEQ_1600","HEQ_1610","HEQ_1620","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SEI_1000","SEI_1010","SEI_1020","SEI_1030","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","IGE","RE1","RE5","RE72","RE74","RMX1","RI6","RGX2","RTX4","RTX6","RWX1","RW3","RD2","RD1","N_ALLERGENPOS","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","LAB_1000","LAB_1010","LAB_1020","LAB_1030","LAB_1040","LAB_1050","LAB_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PLE_1000","PLE_1010","PLE_1020","PLE_1030","PLE_1040","PLE_1050","PLE_1060","PLE_1070","PLE_1080","PLE_1090","PLE_1100","PLE_1110","PLE_1130","PLE_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MTH_1000","MTH_1010","MTH_1020","MTH_1030","MTH_1040","MTH_1050","MTH_1060","MTH_1070","MTH_1080","MTH_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MAD_1000","MAD_1010","MAD_1020","MAD_1030","MAD_1040","MAD_1050","MAD_1060","MAD_1070","MAD_1090","MAD_1100","MAD_1110","MAD_1120","MAD_1130","MAD_1140","MAD_1150","MAD_1160","MAD_1170","MAD_1180","MAD_1190","MAD_1200","MAD_1220","MAD_1230","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MCA_1000","MCA_1010","MCA_1050","MCA_1060","MCA_1070","MCA_1080","MCA_1100","MCA_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MCP_1000","MCP_1010","MCP_1030","MCP_1040","MCP_1050","MCP_1060","MCP_1070","MCP_1080","MCP_1090","MCP_1100","MCP_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PA4_1000","PA4_1010","PA4_1020","PA4_1030","PA4_1040","PA4_1050","PA4_1060","PA4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PTX_1000","PTX_1010","PTX_1020","PTX_1030","PTX_1040","PTX_1050","PTX_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","VNUM","PERIOSTIN","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PI4_1000","PI4_1010","PI4_1020","PI4_1030","PI4_1040","PI4_1050","PI4_1060","PI4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PRG_1000","PRG_1010","PRG_1020","PRG_1030","PRG_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAD_1000","PAD_1010","PAD_1020","PAD_1030","PAD_1040","PAD_1050","PAD_1060","PAD_1070","PAD_1080","PAD_1090","PAD_1100","PAD_1110","PAD_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAL_1000","PAL_1010","PAL_1020","PAL_1030","PAL_1040","PAL_1050","PAL_1060","PAL_1070","PAL_1080","PAL_1090","PAL_1100","PAL_1110","PAL_1120","PAL_1130","PAL_1150","PAL_1160","PAL_1170","PAL_1180","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PTR_1000","PTR_1010","PTR_1050","PTR_1060","PTR_1100","PTR_1140","PTR_1180","PTR_1220","PTR_1260","PTR_1300","PTR_1340","PTR_1380","PTR_1390","PTR_1430","PTR_1470","PTR_1480","PTR_1490","PTR_1500","PTR_1535","PTR_1540","PTR_1550","PTR_1560","PTR_1600","PTR_1610","PTR_1620","PTR_1630","PTR_1670","PTR_1710","PTR_1750","PTR_1790","PTR_1830","PTR_1870","PTR_1910","VNUM_C","VNUM","VDATE","DATE_Q02","DATE_Q03","DATE_Q04","DATE_Q05","DATE_Q07","DATE_Q08","DATE_Q09","DATE_Q10","DATE_Q11","DATE_Q12","DATE_Q13","DATE_Q14","DATE_Q15","DATE_Q16","DATE_Q17","DATE_Q18","DATE_Q19","DATE_Q20","DATE_Q21","DATE_Q22","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","P10_1000","P10_1010","P10_1020","P10_1030","P10_1040","P10_1050","P10_1060","P10_1070","P10_1080","P10_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PCH_1000","PCH_1010","PCH_1020","PCH_1030","PCH_1040","PCH_1050","PCH_1060","PCH_1070","PCH_1080","PCH_1090","PCH_1100","PCH_1110","PCH_1120","PCH_1130","PCH_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","IAQ_1000","IAQ_1010","IAQ_1020","IAQ_1030","IAQ_1040","IAQ_1050","IAQ_1060","IAQ_1070","IAQ_1080","IAQ_1090","IAQ_1100","IAQ_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","REGIMEN_PERIOD_1","REGIMEN_PERIOD_2","REGIMEN_PERIOD_3","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","REG_1080","REG_1150","AGE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SER_1010","SER_1020","SER_1030","SER_1040","SER_1050","SER_1060","SER_1070","SER_1100","SER_1110","SER_1120","SER_1130","SER_1140","SER_1150","SER_1160","SER_1170","SER_1180","SER_1190","SER_1200","SER_1210","SER_1220","SER_1240","SER_1250","SER_1000","SER_1080","SER_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PSE_1060","PSE_1070","PSE_1080","PSE_1090","PSE_1100","PSE_1110","PSE_1130","PSE_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SIG_1000","SIG_1010","SIG_1020","SIG_1030","SIG_1040","SIG_1050","SIG_1060","SIG_1080","SIG_1090","SIG_1100","SIG_1110","SIG_1120","SIG_1130","SIG_1140","SIG_1150","SIG_1160","SIG_1170","SIG_1180","SIG_1190","SIG_1200","SIG_1210","SIG_1220","SIG_1230","SIG_1240","SIG_1250","SIG_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SFU_1000","SFU_1010","SFU_1020","SFU_1030","SFU_1040","SFU_1050","SFU_1060","SFU_1070","SFU_1080","SFU_1090","SFU_1100","SFU_1110","SFU_1120","SFU_1130","SFU_1140","SFU_1150","SFU_1160","SFU_1170","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SNQ_1000","SNQ_1010","SNQ_1020","SNQ_1030","SNQ_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SPI_1010","SPI_1020","SPI_1040","SPI_1050","SPI_1060","SPI_1070","VNUM_C","VNUM","VDATE","SPI_1030","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","AM_TRIAL_TYPE","AM_FVC","AM_FEV1","AM_PEF","AM_FEF2575","AM_FET","DRY_1","DRY_2","DRY_3","DRY_4","DRY_5","DRY_6","DRY_7","DRY_8","DRY_9","VNUM_C","VNUM","VDATE","DDATE","PM_TRIAL_TYPE","PM_FVC","PM_FEV1","PM_PEF","PM_FEF2575","PM_FET","DRY_10","DRY_11","DRY_12","DRY_13","DRY_14","DRY_15","DRY_16","DRY_17","DRY_18","DRY_19","DRY_20","DRY_21","DRY_22","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SLA_1010","SLA_1020","SLA_1030","SLA_1000","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SRE_1010","SRE_1020","SRE_1030","SRE_1040","SRE_1050","SRE_1060","SRE_1070","SRE_1080","SRE_1090","SRE_1100","SRE_1000","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SPU_1000","SPU_1010","SPU_1020","SPU_1030","SPU_1040","SPU_1050","SPU_1060","SPU_1070","SPU_1080","SPU_1090","SPU_1100","SPU_1110","SPU_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SAD_1000","SAD_1010","SAD_1020","SAD_1030","SAD_1040","SAD_1050","SAD_1060","SAD_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SCH_1000","SCH_1010","SCH_1020","SCH_1030","SCH_1040","SCH_1050","SCH_1055","SCH_1060","SCH_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TER_1000","TER_1010","TER_1020","TER_1040","TER_1050","TER_1070","TER_1080","TER_1090","TER_1100","TER_1110","TER_1120","TER_1130","TER_1140","TER_1150","TER_1160","TER_1170","TER_1180","TER_1185","TER_1200","TER_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TXF_1000","TXF_1010","TXF_1020","TXF_1030","TXF_1040","TXF_1050","TXF_1060","TXF_1070","TXF_1080","TXF_1090","TXF_1100","TXF_1110","TXF_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TCH_1000","TCH_1010","TCH_1020","TCH_1030","TCH_1040","TCH_1050","TCH_1060","TCH_1080","TCH_1090","TCH_1100","TCH_1110","TCH_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","WSH_1000","WSH_1010","WSH_1020","WSH_1040","WSH_1050","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","WPA_1000","WPA_1010","WPA_1020","WPA_1030","WPA_1040","WPA_1050","WPA_1060","WPA_1070","WPA_1080","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","HOW_SICK_TODAY","RUNNY_NOSE","PLUGGED_NOSE","SNEEZING","SORE_THROAT","SCRATCHY_THROAT","COUGH","HOARSENESS","HEAD_CONGESTION","CHEST_CONGESTION","FEELING_TIRED","THINK_CLEARLY","SLEEP_WELL","BREATH_EASILY","WALK","DAILY_ACTIVITIES","WORK_OUTSIDE_HOME","WORK_INSIDE_HOME","INTERACT","LIVE_LIFE","MY_COLD_IS","VNUM_C","VNUM","VDATE","WDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER"],["Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Char","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Char","Char","Num","Num","Num","Num","Char","Char","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Char","Num","Char","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Char","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num"],["In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, school, or at home? 0=None of the time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All of the time","During the past 3 days, how often have you had asthma symptoms? 0=Not at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how often have you used your rescue inhaler or nebulizer medication ? 0=Not at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how many total times did your asthma symptoms wake you up from sleep? 0=Not at all, 1=1 time in the last 3 days, 2=2-3 times in the last 3 days, 3=4-5 times in the last 3 days, 4=&gt;=6 times in the last 3 days","How would you rate the amount of impairment you have experienced due to your asthma in the past 3 days? 0=No impairment, 1=Mild impairment, 2=Moderate impairment, 3=Severe impairment, 4=Very severe impairment","How stressed or frightened were you by your asthma symptoms in the past 3 days? 0=Not at all, 1=Mildly, 2=Moderately, 3=Severely, 4=Very severely","Why do you think your asthma was worse in the past 3 days compared to what is normal for you? 0=I have not been worse over the past 3 days. My asthma symptoms have been usual., 1=Common cold, 2=Allergies, 3=Pollution or chemical irritant, 4=Too little asth","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Are you currently retired? 1=Yes,0=No","Are you retired because of asthma? 1=Yes,0=No","Are you currently unemployed? 1=Yes,0=No","Are you unemployed because of asthma? 1=Yes,0=No","Do you get paid to do work? 1=Yes,0=No","How much does your asthma bother you at your paid work? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Overall, how much does your asthma bother you when you do jobs around the house? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Overall, how much does your asthma bother your social life? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Overall, how much does your asthma bother your personal life? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Are you involved in leisure activities, such as: walking for pleasure, sports, exercise, travelling, taking vacations? 1=Yes,0=No","When involved in leisure activities, how much does your asthma bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Would you say that you can't do some of these sorts of things because of asthma? 1=Yes,0=No","How much does your asthma bother you when you sleep? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much does the cost of your asthma medicines bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Do you get free prescriptions? 1=Yes,0=No","How much does the inconvenience or embarrassment of taking your asthma medicines bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much do coughs and colds bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Feeling upset is also a bother.  Does your asthma make you feel anxious, depressed, tired, or helpless? 1=Yes,0=No","Feeling upset is also a bother.  Does your asthma make you feel anxious, depressed, tired, or helpless? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will have an asthma attack when visiting a new place? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will catch a cold? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will let others down? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that your health may get worse in the future? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you won't be able to cope with an asthma attack? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past 4 weeks, how much of the time did your asthma keep you from getting as much done at work, school or at home? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often have your had shortness of breath? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often did your asthma symptoms wake you up at night or earlier than usual in the morning? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often have you used your rescue inhaler or nebulizer medication? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","How would you rate your asthma control during the past 4 weeks? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","ICD-9 category","Description of Adverse Event","ICD9 Code","Ongoing at current visit","Type (1=Intermitent; 2=Continuous)","Severity (1=Mild; 2=Moderate; 3=Severe)","Serious (0=No, 1=Yes)","Likelyhood of Relationship to Study Drug(s) (1=None; 2=Unlikely (remote); 3=Possible; 4=Probable)","Change in Study Drugs (1=Unchanged; 2=Altered)","Outcome (1=Completely recovered; 2=Recovered, but with lasting effects; 3=Death)","Treatment Required (1=None; 2=Medication; 3=Hospitalization; 4=Other)","Ongoing at final visit","Date Started","Date Stopped","Visit Number (character)","Visit Number (numeric)","Long ICD-9 description","Short ICD-9 description","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","ASTHMA HISTORY1.Approximately how old were you when chest symptoms suggesting asthma first appeared?","Did these symptoms appear immediately after or as a result of:1a. a respiratory infection such as a cold or pneumonia? 1=Yes, 0=No, 8=Don't Know","an occupational or job change? 1=Yes, 0=No, 8=Don't Know","a household move? 1=Yes, 0=No, 8=Don't Know","a pregnancy? 1=Yes, 0=No, 8=Don't Know","a hormonal change ? 1=Yes, 0=No, 8=Don't Know","How old were you when a doctor first diagnosed you with asthma?","Have any of your immediate blood relatives been told by a physician that they have asthma? 1=Yes, 0=No, 8=Don't Know","Father  (1=Yes, 0=No, 8=Don't Know)","Brother(s) or Sister(s)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Child(ren)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","ASTHMA SYMPTOMS1.How do you categorize your asthma symptoms throughout the course of the year? 1=Relatively the same all year, 2=Vary by season","Winter? 1=Yes,0=No","Spring? 1=Yes,0=No","Summer? 1=Yes,0=No","Fall? 1=Yes,0=No","In the last 12 months, how many...  5a.Asthma episodes have you had that required emergency care or an unscheduled office visit?","Overnight hospitalizations have you had due to asthma?","Courses of systemic corticosteroid therapy  for asthma have you taken?","Days of work, school, or housework have you missed due to asthma?","In the past 3 months, how many days of work, school, or housework have you missed due to asthma?","Have you ever been admitted to an intensive care unit for asthma? 1=Yes,0=No","How many times have you been admitted to an intensive care unit for asthma?","Have you ever had invasive mechanical ventilation? 1=Yes, 0=No, 8=Don't Know","Have you ever had non-invasive mechanical ventilation? 1=Yes, 0=No, 8=Don't Know","Do any of the following currently provoke your asthma? 1=Yes, 0=No, 8=Don't Know","Menstrual cycle(If participant is male or a postmenopausal female, leave blank.) (1=Yes, 0=No, 8=Don't Know)","Aspirin or non-steroidal anti-inflammatory drugs (e.g., Aleve, Motrin) (1=Yes, 0=No, 8=Don't Know)","Respiratory infections (e.g., colds)  (1=Yes, 0=No, 8=Don't Know)","Irritants (e.g., pollution, odors, perfumes, chemicals, household cleaners) (1=Yes, 0=No, 8=Don't Know)","Weather conditions (e.g., change in weather, humidity)  (1=Yes, 0=No, 8=Don't Know)","Exposure to cold air  (1=Yes, 0=No, 8=Don't Know)","Emotional factors (e.g., stress, laughing)  (1=Yes, 0=No, 8=Don't Know)","Tobacco smoke  (1=Yes, 0=No, 8=Don't Know)","Food additives/preservatives (e.g., MSG, sulfites)  (1=Yes, 0=No, 8=Don't Know)","Allergies (e.g., dust, animals, pollens)(1=Yes, 0=No, 8=Don't Know)","Other (1=Yes,0=No)","To which of the following did a doctor or other health practitioner say you were allergic? 1=Yes, 0=No, 8=Don't Know","Foods  (1=Yes, 0=No, 8=Don't Know)","Things you breathe in or are exposed to (e.g., dust, pollens, molds, animal fur, feathers, dander)  (1=Yes, 0=No, 8=Don't Know)","Stinging insects such as bees or wasps (1=Yes, 0=No, 8=Don't Know)","Latex  (1=Yes, 0=No, 8=Don't Know)","Other  (1=Yes,0=No)","Have you ever had eczema / atopic dermatitis ? 1=Yes, 0=No, 8=Don't Know","If YES, was your eczema diagnosed by a doctor? 1=Yes,0=No","Have any of your immediate blood relatives been told by a physician that they have allergies/eczema/hay fever? 1=Yes, 0=No, 8=Don't Know","Father  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Brother(s) or Sister(s)(1=Yes, 0=No, 8=Don't Know, 9=N/A)","Child(ren)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Did you grow up in a household where you were exposed to tobacco smoke? 1=Yes,0=No","Do you currently smoke? 1=Yes,0=No","Record smoking history in pack-years*. __ __ . __ pack-years","Were you ever a smoker? 1=Yes,0=No","Record smoking history in pack-years*. __ __ . __ pack-years","Do you currently live in a household where you are exposed to tobacco smoke? 1=Yes,0=No","Do you currently vape  or use a hookah ? 1=Yes,0=No","How frequently do you vape or use a hookah? 1=Infrequently , 2=Occasionally , 3=Weekly , 4=Daily","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many days a week","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many times a day","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many years have y","Have you ever vaped or used a hookah in the past? 1=Yes,0=No","Have you ever vaped or used a hookah in the past?  If NO, skip to Q17. Approximately how many years did you vape or use a hookah?  (1-99 years)","Do you currently live in a household where you are exposed to others vaping or using a hookah? 1=Yes,0=No","Do you spend time in social settings  where you are exposed to others vaping or using a hookah? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","When was the last time that you vaped or used a hookah? (number of days from Visit 1)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","How many days were you bothered by coughing during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe was your coughing during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by wheezing during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe was your wheezing during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by shortness of breath during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","Over average, how severe was your shortness of breath during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you awakened at night during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how much of a problem was being awakened at night during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by side effects of your asthma medication during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe were the side effects of your asthma medication during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Participants baseline peak flow (PEF) value (0-999 L/M) First visit (Visit 0A, or Visit 1): PEF (FEF Max) from prebronchodilator (baseline) spirometry at first visit (convert to L/M)","Participants baseline rescue use value (0-99 puffs/day) First visit (Visit 0A, or Visit 1): Self-reported average daily use of albuterol during the 14 days prior to first visit.","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","body mass index computed from height and weight measured at Visit 1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Page Number","Reason for change in scheduled medications status (1=Adverse Event, 2=Other (specify))","Reason for change in scheduled medications status Related adverse event number (0-99)","Current status of participants scheduled medications (1=Discontinued, 2=Resumed)","Date change took effect","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Code to identify a drug unit of measure; (Units)","Code to identify how frequently a drug is taken or administered. (Frequency)","Code to identify the route used to administer a drug. (Route)","Name of Medication","This number corresponds to the ID number of the AHFS monograph. (Medication Code)","Related Event","Medication Dose","Ongoing at current visit","Ongoing at final visit","Start Date","Stop Date","Visit Number (character)","Visit Number (numeric)","GEN_DRUG_NAME","CLASS_TEXT","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Diary and Peak Flow Compliance: Number of full days since the last visit (0-999 days)","Diary and Peak Flow Compliance: Number of days where AM and PM scheduled sessions are complete (AM and PM PEF and all diary questions for AM and PM answered) (0-999 days)","Diary and Peak Flow Compliance: Percent compliance If the compliance value in Q1c is less than 75Pct, re-emphasize the importance of completing scheduled diary assessments and peak flows. (0-150.0Pct)","Scheduled Respimat Compliance (Visits 2-9 only): Number of scheduled puffs since the last visit (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Number of remaining puffs on scheduled Respimat Reference all returned Respimats (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Number of puffs taken (60 x (# used Respimats)) - Q2b (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Percent compliance=Q2c/Q2a x 100 If the participant took less than 75Pct of the scheduled Respimat puffs, re-emphasize the importance of maintaining the daily dosing schedule. (0-150.0Pct)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of scheduled puffs since the last visit (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of remaining puffs reflected on scheduled Twisthaler counters Total the values reflected on all counters for all returned Twisthalers (i.e., out of their pouches). (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of puffs taken (60 x (# used Twisthalers)) - Q3b (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Percent compliance=Q3c/Q3a x 100 If the participant took less than 75Pct of the scheduled Twisthaler puffs, re-emphasize the importance of maintaining the daily dosing schedule. (0-150.0Pct)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","1=mometasone, 2=placebo","Blinded Scheduled Twisthaler Contents - How sure are you about your answer in Q1? 1=Absolutely sure - I know what the Twisthaler/MDI contains, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","1=tiotropium, 2=placebo","Blinded Scheduled Respimat Contents - How sure are you about your answer in Q4? 1=Absolutely sure - I know what the Respimat capsules contain, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has the participant or parent/legal guardian signed the SIENA Informed Consent document? 1=Yes,0=No","Ages 12-17 Only: Has the participant signed and dated the assent form or, if the participant is less than the local age of assent, has the participant given verbal assent? 1=Yes,0=No","Has the participant used an inhaled corticosteroid in the past 6 weeks? 1=Yes,0=No","Will the participant be 12 years of age, or older, as of Visit 1? 1=Yes,0=No","Has the participant used an inhaled corticosteroid or combination inhaled corticosteroid/LABA in the past 3 weeks? 1=Yes,0=No","If NO, has the participant used a leukotriene modifier in the past 3 weeks? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used inhaled corticosteroid therapy?","Over the past 3 months, on average, how many days per week has the participant used inhaled corticosteroid therapy: &lt; 5? 1=Yes,0=No","Over the past 3 months, has the participant used greater than the equivalent of 80-240 mcg of inhaled beclomethasone daily? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used combination inhaled corticosteroid/LABA therapy?","Over the past 3 months, on average, how many days per week has the participant used combination inhaled corticosteroid/LABA therapy,  &lt; 5 (1=Yes,0=No)","Over the past 3 months, on average, how many days per week has the participant had daytime asthma symptoms?","Over the past 3 months, on average, how many days per week has the participant had daytime asthma symptoms,  &lt;=2? 1=Yes,0=No","Over the past 3 months, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms?","Over the past 3 months, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms, &lt;=2? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used his/her short acting betaagonist  for relief of symptoms?","Over the past 3 months, on average, how many days per week has the participant used his/her short acting betaagonist  for relief of symptoms, &lt; 2? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 80Pct? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 70Pct? 1=Yes,0=No","Does the participant have current evidence of any of the conditions listed on the Exclusionary Medical Conditions for SIENA  reference card, or any chronic diseases that would prevent participation in the trial or put the participant at risk by participati","Is the participant currently taking any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Is the participant able to go off these medications for the required washout period prior to Visit 1 and for the duration of the study? 1=Yes,0=No","Is the participant able to use the spirotel e-diary/PEF meter correctly, as evidenced by achieving a score of 13 on the spirotel Performance Checklist ? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Has the participant or parent/legal guardian signed the SIENA Informed Consent document?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past month, on average, how many days per week has the participant had daytime symptoms?","In the past month, on average, how many days per week has the participant had daytime symptoms, &lt;=2? 1=Yes,0=No","In the past month, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms?","In the past month, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms,  &lt;=2? 1=Yes,0=No","In the past month, on average, how many days per week has the participant used his/her short acting beta-agonist  for relief of symptoms?","In the past month, on average, how many days per week has the participant used his/her short acting beta-agonist  for relief of symptoms, &lt; 2? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 80Pct? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 70Pct? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant or parent/legal guardian sign the SIENA Informed Consent document? 1=Yes,0=No","Ages 12-17 Only: Has the participant signed and dated the assent form or, if the participant is less than the local age of assent, has the participant given verbal assent? 1=Yes,0=No","Is the participant 12 years of age, or older? 1=Yes,0=No","Does the participant plan to move away from the clinical site in the upcoming 11 months such that his/her ability to complete the study will be jeopardized? 1=Yes,0=No","Has the participant used investigative drugs and/or enrolled in an intervention trial in the past 30 days, or have plans to enroll in such a trial during the SIENA study? 1=Yes,0=No","Has the participant had a respiratory infection within the past 6 weeks? 1=Yes,0=No","Has the participant had a respiratory infection within the past 4 weeks? 1=Yes,0=No","During the past 4 weeks, on average, how many days per week has the participant had daytime asthma symptoms?","During the past 4 weeks, on average, how many days per week has the participant had daytime asthma symptoms  &gt; 2? 1=Yes,0=No","During the past 4 weeks, how many nighttime awakenings due to asthma symptoms has the participant had?","During the past 4 weeks, how many nighttime awakenings due to asthma symptoms has the participant had,  &gt; 2? 1=Yes,0=No","During the past 4 weeks, on average, how many days per week has the participant used his/her short-acting beta-agonist  for relief of symptoms?","During the past 4 weeks, on average, how many days per week has the participant used his/her short-acting beta-agonist  for relief of symptoms, &gt; 2? 1=Yes,0=No","Is Q7a, Q8a, or Q9a checked YES? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Did the participant or parent/legal guardian sign the SIENA Informed Consent document?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Does the participant have current evidence of any of the conditions listed on the Exclusionary Medical Conditions for SIENA  reference card, or any chronic diseases  that would prevent participation in the trial or put the participant at risk by participat","Does the participant have a history of bladder-neck obstruction, urinary retention, benign prostatic hyperplasia , or a clinically relevant urologic disorder that precludes study participation? 1=Yes,0=No","Does the participant have a history of narrow angle glaucoma? 1=Yes,0=No","Does the participant have a history of significant cardiovascular disorders or arrhythmias? 1=Yes,0=No","Has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card within the specified time periods? 1=Yes,0=No","Is the participant currently taking prescription or OTC medication other than those listed on the Allowed Medications  reference card? 1=Yes,0=No","Based on input from the participant and the study physician, will the participant need to use intranasal steroids at any time during the study? 1=Yes,0=No","Is the participant willing to use a single intranasal steroid at a stable dose continuously for the duration of the study, starting at Visit 1? 1=Yes,0=No","Is the participant currently receiving allergen immunotherapy  other than an established maintenance regimen implemented continuously for a minimum of 3 months? 1=Yes,0=No","Has the participant taken omalizumab within the past 3 months? 1=Yes,0=No","Has the participant used any smokeless tobacco products  in the past year? 1=Yes,0=No","Has the participant smoked cigarettes, a pipe, cigar, marijuana, electronic cigarettes, or any other substance in the past year? 1=Yes,0=No","Ages 18+ Only: Does the participant have a smoking history of greater than 10 pack-years? 1=Yes,0=No","Ages 12-17 Only: Does the participant have a smoking history of greater than 5 pack-years? 1=Yes,0=No","Has the participant received a physician diagnosis of asthma at least 12 months ago? 1=Yes,0=No","Has the participant received a physician diagnosis of asthma at least 12 months ago or had a history consistent with asthma for the previous 12 months? 1=Yes,0=No","Has the participant experienced a life-threatening asthma exacerbation requiring treatment with intubation or mechanical ventilation in the past 5 years? 1=Yes,0=No","Has the participant had an asthma exacerbation requiring systemic corticosteroid treatment in the past 6 weeks? 1=Yes,0=No","Has the participant used an oral or inhaled corticosteroid, or leukotriene modifier in the past 6 weeks? 1=Yes,0=No","Has the participant used an oral corticosteroid in the past 6 weeks? 1=Yes,0=No","Has the participant used an inhaled corticosteroid or leukotriene modifier in the past 3 weeks? 1=Yes,0=No","Is the participant potentially able to bear children? 1=Yes, 0=No, 9=N/A","Is the participant currently pregnant or lactating? 1=Yes,0=No","Does the participant agree to use one of the approved methods indicated on the Birth Control Methods  reference card for the duration of the study? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was the participants prebronchodilator FEV1 &gt;=70Pct predicted? 1=Yes,0=No","Did the participants FEV1 improve &gt;=12Pct and &gt;=200 ml in response to four puffs of albuterol? 1=Yes,0=No","Does the participant have valid source documentation within the past 6 months for an overread AsthmaNet methacholine challenge  with a PC20 &lt;=16 mg/ml? 1=Yes,0=No","Did the participants FEV1 improve &gt;=12Pct and &gt;=200 ml in response to four puffs of albuterol?","Was the participants methacholine PC20 &lt;=16 mg/ml? 1=Yes,0=No","Is the participant able to use the spirotel e-diary/PEF meter correctly, as evidenced by achieving a score of 13 on the Spirotel Performance Checklist ? 1=Yes, 0=No, 9=N/A","Has the participant used the Twisthaler previously? 1=Yes,0=No","Is the participant able to use a Twisthaler properly, as evidenced by achieving a score of 12 on the Twisthaler Inhalation Technique Checklist ? 1=Yes,0=No","Is the participant able to use a metered dose inhaler properly, as evidenced by achieving a score of 11 on the MDI Inhalation Technique Checklist  ? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Does the participant have valid source documentation within the past 6 months for an overread AsthmaNet methacholine challenge  with a PC20 &lt;=16 mg/ml?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Since Visit 1, has the participant experienced a treatment failure event as defined in the protocol? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Since Visit 1, has the participant experienced one or more asthma exacerbations as defined in the protocol? 1=Yes,0=No","Since Visit 1, has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant provide two acceptable sputum induction samples during the run-in? 1=Yes,0=No","Since Visit 1, has the participant experienced a treatment failure event as defined in the protocol? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Did the participant complete Asmanex220 mcg treatment at least 6 weeks ago? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Since Visit 1, has the participant experienced one or more asthma exacerbations as defined in the protocol? 1=Yes,0=No","Since Visit 1, has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Has the participant experienced a respiratory tract infection in the past 6 weeks? 1=Yes,0=No","Has the participant experienced a respiratory tract infection in the past 4 weeks? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report daytime asthma symptoms daily during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report night-time awakenings due to asthma symptoms more than once per week during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report night-time awakenings due to asthma symptoms more than twice per week during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report daily albuterol use for asthma symptom control during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant record at least 75Pct of AM and PM peak flow measurements and symptoms on his or her spirotel during the run-in? 1=Yes,0=No","Did the participant take at least 75Pct of the required puffs from his or her Respimat during the run-in? 1=Yes,0=No","Was the participants prebronchodilator FEV1 &gt;=70Pct predicted? 1=Yes,0=No","Has the participant used the Respimat previously? 1=Yes,0=No","Is the participant able to use a Respimat inhaler properly, as evidenced by achieving a score of 15 on the Respimat Inhalation Technique Checklist? 1=Yes,0=No","Does the participant wish to withdraw consent from the study? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has QC procedure been performed on the NIOX MINO today? 1=Yes,0=No","Did the participant eat or drink within the past hour? 1=Yes,0=No","Did the participant take part in strenuous activity/exercise within the past hour? 1=Yes,0=No","Time eNO started (based on a 24-hour clock) (0000-2359)","ENO Measurement","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","How long have you lived in the current house? (Estimate if uncertain.) (1010) years","How long have you lived in the current house? (Estimate if uncertain.)(1020) months","Does your house use a wood burning stove as a primary source of heat? 1=Yes, 0=No, 8=Don't Know","Does your house use an air conditioner? 1=Yes, 0=No, 8=Don't Know","Does your house use an evaporative cooler ? 1=Yes, 0=No, 8=Don't Know","Does your house use a humidifier? 1=Yes, 0=No, 8=Don't Know","Does your house use a dehumidifier? 1=Yes, 0=No, 8=Don't Know","Has there been water damage to your house, basement, or its contents during the past 12 months? 1=Yes, 0=No, 8=Don't Know","Has there been any mold or mildew, on any surfaces, inside your house in the past 12 months? 1=Yes, 0=No, 8=Don't Know","Which rooms have or have had mold or mildew?10a.Bathroom(s)  (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew?10b.Basement or attic (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10c.Kitchen (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10d.Your bedroom (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10e.Other bedrooms (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10f.Living or family room (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10g.Other (1 = Yes, 0 = No)","Do you ever see cockroaches in your house? 1=Yes,0=No","In which room(s) have you seen cockroaches?12a.Kitchen (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches?  12b.Basement or attic (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12c.Bathroom(s) (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12d.Living or family room (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12e.Your bedroom  (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12f.Other bedrooms (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12g.Garage (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12h.Other (1 = Yes, 0 = No)","Do you ever see rodents  or rodent droppings in your house? 1=Yes,0=No","14.In which room(s) have you seen rodents or rodent droppings? 14a. Kitchen (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14b. Basement or attic (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14c. Bathroom(s) (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings?14d. Living or family room (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14e. Your bedroom(1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14f. Other bedrooms (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14g. Garage (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14h. Other (1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15a. Barns(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property?15b. Hay(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15c. Woodsheds(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15d. Firewood(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property?15e. Chicken coops (1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15f. Corral (1 = Yes, 0 = No)","CHARACTERISTICS OF THE PARTICIPANT'S BEDROOM16.What is the floor covering in your bedroom? 1=Rug/carpet, 2=Vinyl tile or linoleum, 3=Wood, 4=Ceramic tile, 5=Other , 9=Don't know","What type of mattress is on your bed? 1=None, 2=Inner spring mattress, 3=Foam mattress, 4=Waterbed, 5=Air mattress, 6=Other , 9=Don't know","Is the mattress completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","Does your bed have a box spring? 1=Yes,0=No","Is the box spring completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","What type of pillow do you usually sleep with? 1=None, 2=Feather/down, 3=Foam/Dacron/synthetic, 5=Other , 9=Don't know","Is the pillow completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","Does your household have any pets? 1=Yes,0=No","Enter the number of pets that the household has. (Enter '00' if none. If none to Q24a - Q24d, skip to the next question.)24a.Cat","Cat  (1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has. Dog","Dog(1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has 24c.Rabbit, guinea pig, hamster, gerbil, or mouse","Rabbit, guinea pig, hamster, gerbil, or mouse (1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has 24d.Bird","Bird (1=Indoor, 2=Outdoor, 3=Both)","In general, and on a regular basis, are you exposed to any of the following animals?  25a.Cat(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25b.Dog(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25c.Rabbit, guinea pig, hamster, gerbil, or mouse(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25d.Bird (1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25e.Farm animals (1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals?  25f.Other (1 = Yes, 0 = No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other (specify","Highest education of any household member (1=No High School diploma/GED, 2=High Schol grad, 3=Technical training, 4=Some college, no degree, 5=Assoc degree, 6=Bachelor degree, 7=Masters degree/MD/PhD/JD/PharmD, 9=Decline to answer, 10=Don't know)","Category best describes the combined annual income, before taxes, of all members of your household for the last year. (1=Less than $25,000, 2=$25,000 - $49,999, 3=$50,000 - $99,999, 4=$100,000 or more, 9=Decline to answer, 10=Don't know)","How many people  are supported by this income reported in Q3?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Total IgE","Cat (E1)","Dog (E5)","Mouse (E72)","Rat (E74)","Mold mix (MX1)","Cockroach, German (I6)","Grass mix (GX2)","Tree mix (TX4)","Tree mix (TX6)","Weed mix (WX1)","Weed (W3)","Mite (D2)","Mite (D1)","Number or allergen specific (Cat, Dog, Mouse, etc) that were either POS or greater than 0.35","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","CBC with differential cell count: Eosinophils (absolute count) (0-9999/uL)","CBC with differential cell count: WBC (0-999.9 K/uL)","CBC with differential cell count: Differential Lymphocytes (0-99.9Pct)","CBC with differential cell count: Differential Monocytes (0-99.9Pct)","CBC with differential cell count: Differential Basophils (0-99.9Pct)","CBC with differential cell count: Differential Neutrophils (0-99.9Pct)","CBC with differential cell count: Differential Eosinophils (0-99.9Pct)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","PARENTAL HEIGHT - First study visit only or until both are completed 1.Biological mother's height (complete height or check unknown) __ feet","PARENTAL HEIGHT - First study visit only or until both are completed1.Biological mother's height (complete height or check unknown) __ __ inches","PARENTAL HEIGHT - First study visit only or until both are completed 1.Biological mother's height (complete height or check unknown) (9=Don't Know)","Biological father's height (complete height or check unknown) __ feet","Biological father's height (complete height or check unknown) __ __ inches","Biological father's height (complete height or check unknown) (9=Don't Know)","PARTICIPANT MEASUREMENTS - Complete at all applicable study visits3. What type of height measurement was obtained? 1=Standing height, 2=Length","First measurement __ __ __ . __ cm","Second measurement __ __ __ . __ cm","Third measurement __ __ __ . __ cm","Average height or length measurement __ __ __ . __ cm","In your judgment, was the participant's height or length measurement acceptable? 1=Yes,0=No","Weight (shoes off, light clothing) __ __ __ . __ kg","ORAL CANDIDIASIS5.Does the participant have evidence of oral candidiasis? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Post Diluent FEV1","Did the participant drop20Pct at the diluent stage? 1=Yes,0=No","Last concentration of methacholine administered __ __ . __ __ __ __ mg/ml","FEV1 after last concentration of methacholine administered","Did the participant achieve a PC20? 1=Yes,0=No","PC20 __ __ . __ __ mg/ml","Time methacholine challenge ended (based on 24-hour clock)","FEV1","Time of FEV1 in Q7a (based on 24-hour clock)","Was the FEV1 from Q7athe methacholine reversal reference value  in the gray box above? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was an additional treatment used in the first hour? 1=Yes,0=No","Additional albuterol by MDI (1=Yes,0=No)","Number of additional puffs of albuterol administered (1=2, 2=4, 3=&gt;4)","Nebulized Beta-agonist(1=Yes,0=No)","Subcutaneous epinephrine(1=Yes,0=No)","Implementation of clinic emergency protocol or algorithm (1=Yes,0=No)","Other  (1=Yes,0=No)","Participant's FEV1 after additional treatment within first hour.2a.FEV1","Time of FEV1 in Q2a (based on 24-hour clock)","Was the FEV1 from Q2a &gt;the methacholine reversal reference value  in the gray box on the Methacholine Challenge Testing  form? 1=Yes,0=No","Was additional treatment used after one hour? 1=Yes,0=No","Additional albuterol by MDI (1=Yes,0=No)","Number of additional puffs of albuterol administered (1=2, 2=4, 3=&gt;4)","Nebulized Beta-agonist  (1=Yes,0=No)","Subcutaneous epinephrine(1=Yes,0=No)","Implementation of clinic emergency protocol or algorithm  (1=Yes,0=No)","Treatment in the emergency room  (1=Yes,0=No)","Overnight hospitalization (1=Yes,0=No)","Other (1=Yes,0=No)","Participant's final FEV1 after methacholine challenge4a.FEV1","Time of FEV1 in Q4a (based on 24-hour clock)","Was the FEV1 from Q4a ? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Exclusions and Confounders1.Has the participant had any severe acute illness in the past 4 weeks? 1=Yes,0=No","If YES, has the participant received permission from the supervising physician to proceed with the methacholine challenge testing? 1=Yes,0=No","Has the participant used an oral or injectable steroid  in the last 4 weeks? 1=Yes,0=No","Does the participant have a baseline  FEV1 less than 55Pct of predicted or less than 1.0 L? 1=Yes,0=No","Pregnancy test results:  (Check N/A if the participant is male, or is female and is post-menopausal, had a hysterectomy or tubal ligation.) (1=Positive, 0=Negative, 9=N/A)","Is the participant's systolic blood pressure &gt; 200 mm Hg or diastolic blood pressure &gt; 100 mm Hg? 1=Yes,0=No","Is there any other reason the participant should not proceed with the methacholine challenge testing? 1=Yes,0=No","Is the participant eligible to proceed with the diluent  pulmonary function testing for the methacholine challenge? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Exclusions and Confounders:  1.Has the participant had any severe acute illness in the past 4 weeks? 1=Yes,0=No","If YES, has the participant received permission from the supervising physician to proceed with the methacholine challenge testing? 1=Yes,0=No","During the past 4 weeks, has the participant had any respiratory infections, colds, or bronchitis ? 1=Yes,0=No","If YES, during the past 2 weeks, has the participant had any respiratory infections, colds, or bronchitis ? 1=Yes,0=No","Has the participant used an oral or injectable steroid  in the last 4 weeks? 1=Yes,0=No","Does the participant have a baseline  FEV1 less than 70Pct of predicted? 1=Yes,0=No","Pregnancy test results(Check N/A if the participant is male, or is female and has not started menses.)(1=Positive, 0=Negative, 9=N/A)","If participant's age is ? 1=Yes,0=No","If participant's age is &lt; 12 years: Is the participant's systolic blood pressure &gt; 180 mm Hg or diastolic blood pressure &gt; 90 mm Hg? 1=Yes,0=No","Is there any other reason the participant should not proceed with the methacholine challenge testing? 1=Yes,0=No","Is the participant eligible to proceed with the diluent  pulmonary function testing for the methacholine challenge? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time albuterol administered (based on 24-hour clock)","Time post-albuterol spirometry started (based on 24-hour clock)","Highest FVC","Highest FEV1","Highest FEV1 (Pct predicted)","FEF Max","FEF25-75","In your judgment, was the subject's spirometry technique acceptable? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Please check the box next to the treatment that you believe you received over the past 12 weeks. 1=mometasone, 2=placebo","Blinded Scheduled Twisthaler Contents: How sure are you about your answer to Question 2? 1=Absolutely sure -  I know what the MDI contains, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","Please comment with respect to any observations you made that helped you make your choice in Question 2 . 1=I have no comments, 2=I noticed the following:","Please check the box next to the treatment that you believe you received over the past 12 weeks. 1=tiotropium, 2=placebo","How sure are you about your answer to Question 5? 1=Absolutely sure - I know what the Respimat capsules contain, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","1=I have no comments, 2=I noticed the following:","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Visit Number (numeric)","Level of periostin in peripheral blood (ng/ml)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time ipratropium administered (based on 24-hour clock) (0000-2359)","Time post-ipratropium spirometry started (based on 24-hour clock) (0000-2359)","Highest FVC (0-9.99 L)","Highest FEV1 (0-9.99 L)","Highest FEV1 (Pct predicted) (0-999 Pct predicted)","FEF Max (0-99.99 L/S)","FEF25-75 (0-9.99 L/S)","In your judgment, was the participants spirometry technique acceptable? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Is the participant unable to bear children due to any of the following reasons? 1=Yes,0=No","Post-menopausal (at least one year since last menses) (1=Yes,0=No)","Hysterectomy(1=Yes,0=No)","Tubal ligation(1=Yes,0=No)","Pregnancy test results  (1=Positive, 0=Negative)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Have you had any diseases, illnesses, conditions, or surgeries related Blood, Lymph, or Immune Systems? 1=Yes,0=No","Eyes(1=Yes,0=No)","Breasts (1=Yes,0=No)","Endocrine Systems (1=Yes,0=No)","Heart and Blood Vessels  (1=Yes,0=No)","Liver or Pancreas (1=Yes,0=No)","Kidneys or Urinary Tract System  (1=Yes,0=No)","Reproductive System (1=Yes,0=No)","Muscles or Bones (1=Yes,0=No)","Nervous System (1=Yes,0=No)","Psychiatric(1=Yes,0=No)","Drug Allergies (1=Yes,0=No)","Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Skin (1=Yes,0=No)","Have you ever had allergic rhinitis ? 1=Yes, 0=No, 9=Don't know","Have you ever had nasal polyps? 1=Yes, 0=No, 9=Don't know","Do you have chronic or recurrent sinusitis ? 1=Yes, 0=No, 9=Don't know","Have you ever been diagnosed with vocal cord dysfunction? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the ear, nose, or throat? 1=Yes,0=No","Have you ever had pneumonia? 1=Yes, 0=No, 9=Don't know","If YES, were you diagnosed by chest x-ray? 1=Yes, 0=No, 9=Don't know","If YES, were you treated with antibiotics? 1=Yes, 0=No, 9=Don't know","Have you ever had bronchitis? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the lungs ? 1=Yes, 0=No, 9=Don't know","Do you have gastroesophageal reflux disease ? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the stomach or intestines? 1=Yes,0=No","Have you been diagnosed with sleep disordered breathing ? 1=Yes,0=No","If YES, are you being treated with CPAP or BiPAP? 1=Yes,0=No","Have you ever had other sleep disorders? 1=Yes,0=No","Have you ever had other conditions that have not been mentioned on this form? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Short-acting Inhaled Beta-Agonists by Inhaler (e.g., albuterol, Primatene Mist, Maxair, ProAir, Proventil, Ventolin, Xopenex)  (1=Yes, 0=No, 9=Don't Know)","If YES, indicate average weekly puffs in the past month (Enter '000' if none used) weekly puffs","Rescue treatment via a Nebulizer Machine (e.g., albuterol, ipratropium, Combivent, Xopenex, levalbuterol)  (1=Yes, 0=No, 9=Don't Know)","Long-acting Inhaled Beta-Agonists (e.g., Serevent, Foradil, salmeterol, formoterol) Do not consider combination medications. (1=Yes, 0=No, 9=Don't Know)","Oral Beta-Agonists (e.g., albuterol, Brethine, Bricanyl, metaproterenol, Proventil, Ventolin, Repetabs, Volmax)  (1=Yes, 0=No, 9=Don't Know)","Oral Theophylline (short-acting or sustained release)(e.g., Aminophylline, Slo-Phyllin, Slo-bid, Theo-Dur, Uniphyl)(1=Yes, 0=No, 9=Don't Know)","Inhaled Anticholinergic by Inhaler (e.g., Atrovent, Combivent, Spiriva) (1=Yes, 0=No, 9=Don't Know)","Leukotriene Antagonist / 5LO Inhibitors (e.g., Accolate, Zyflo, Singulair) (1=Yes, 0=No, 9=Don't Know)","IgE Blocker (e.g., Xolair)(1=Yes, 0=No, 9=Don't Know)","Oral Steroids FOR ASTHMA (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron,  dexamethasone)(1=Yes, 0=No, 9=Don't Know)","If YES, in the past 12 months, how many courses of steroids by mouth have you taken FOR ASTHMA? 1=1 course, 2=2 courses, 3=3 courses, 4=4 courses, 5=5 courses, 6=More than 5 courses","Injectable Steroids FOR ASTHMA (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV)(1=Yes, 0=No, 9=Don't Know)","Steroids by Inhaler (e.g., Asmanex Twisthaler, QVAR, Flovent, Pulmicort Flexhaler)(1=Yes, 0=No, 9=Don't Know)","Indicate most recent type of inhaled steroid taken (refer to PRIOR_TRT_CARD reference card) code","Indicate number of daily puffs used daily puffs","Indicate the total number of months that you used the inhaled steroid out of the past 12 months months","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide) (1=Yes, 0=No, 9=Don't Know)","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide). If YES, complete Q13a - Q13c. Indicate most recent type of nebulized steroid taken (refer to PRIOR_TRT_CARD reference card)","Indicate number of daily treatments used daily treatments","Indicate the total number of months that you used the nebulized steroid out of the past 12 months months","Long-Acting Beta-Agonist and Inhaled Steroid Combination Medications (e.g., Advair Diskus, Symbicort MDI, Dulera MDI)(1=Yes, 0=No, 9=Don't Know)","Indicate most recent type of combination medication taken (refer to PRIOR_TRT_CARD reference card) code","Indicate number of daily puffs used daily puffs","Indicate the total number of months that you used the combination medication out of the past 12 months months","Nasal Steroids (e.g., Beconase, Vancenase, Flonase, Nasacort, Nasalide, Nasarel, Omnaris, Rhinocort, Nasonex)(1=Yes, 0=No, 9=Don't Know)","Non-steroidal Anti-allergic Nasal Medications (e.g., Nasalcrom, Astelin, Astepro, ipratropium)(1=Yes, 0=No, 9=Don't Know)","Anti-allergic Oral Medications (e.g., fexofenadine, loratadine, cetirizine, chlorpheniramine) (1=Yes, 0=No, 9=Don't Know)","Topical Steroids - Prescription (e.g., Synalar, Lidex, Dermacin, Fluocinonide) (1=Yes, 0=No, 9=Don't Know)","Topical Steroids - OTC (e.g., Hydrocortisone - multiple strengths and products) (1=Yes, 0=No, 9=Don't Know)","Other Medication FOR ASTHMA OR ALLERGIES (1=Yes, 0=No, 9=Don't Know)","Oral Steroids for Conditions Other Than Asthma (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) (1=Yes, 0=No, 9=Don't Know)","Injectable Steroids for Conditions Other Than Asthma (e.g., Medrol, Solumedrol, Decadron, dexamethasone,  triamcinolone, Kenalog, hydrocortisone IV)(1=Yes, 0=No, 9=Don't Know)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Short-acting Inhaled Beta-Agonists by Inhaler (e.g., albuterol, Primatene Mist, Maxair, ProAir, Proventil, Ventolin, Xopenex) Date","Rescue treatment via a Nebulizer Machine (e.g., albuterol, ipratropium, Combivent, Xopenex, levalbuterol) Date","Long-acting Inhaled Beta-Agonists (e.g., Serevent, Foradil, salmeterol, formoterol) Do not consider combination medications. Date","Oral Beta-Agonists (e.g., albuterol, Brethine, Bricanyl, metaproterenol, Proventil, Ventolin, Repetabs, Volmax) Date","Inhaled Anticholinergic by Inhaler (e.g., Atrovent, Combivent, Spiriva) Date","Leukotriene Antagonist / 5LO Inhibitors (e.g., Accolate, Zyflo, Singulair) Date","IgE Blocker (e.g., Xolair) Date","Oral Steroids FOR ASTHMA (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) Date","Injectable Steroids FOR ASTHMA (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV) Date","Steroids by Inhaler (e.g., Asmanex Twisthaler, QVAR, Flovent, Pulmicort Flexhaler) Date","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide) Date","Long-Acting Beta-Agonist and Inhaled Steroid Combination Medications (e.g., Advair Diskus, Symbicort MDI, Dulera MDI) Date","Nasal Steroids (e.g., Beconase, Vancenase, Flonase, Nasacort, Nasalide, Nasarel, Omnaris, Rhinocort, Nasonex) Date","Non-steroidal Anti-allergic Nasal Medications (e.g., Nasalcrom, Astelin, Astepro, ipratropium) Date","Anti-allergic Oral Medications (e.g., fexofenadine, loratadine, cetirizine, chlorpheniramine) Date","Topical Steroids - Prescription (e.g., Synalar, Lidex, Dermacin, Fluocinonide) Date","Topical Steroids - OTC (e.g., Hydrocortisone - multiple strengths and products) Date","Other Medication FOR ASTHMA OR ALLERGIES Date","Oral Steroids for Conditions Other Than Asthma (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) Date","Injectable Steroids for Conditions Other Than Asthma (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV) Date","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the last month, how often have you been upset because of something that happened unexpectedly? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that you were unable to control the important things in your life? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt nervous and stressed? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt confident about being able to handle your personal problems? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that things were going your way? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you found that you could not cope with all the things that you had to do? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you been able to control irritations in your life? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that you were on top of things? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you been angered because of things that happened that were outside of your control? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that your difficulties were piling up so high that you could not overcome them? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Have you consumed caffeine in the past 4 hours? 1=Yes,0=No","Have you used medications with caffeine in the past 4 hours? 1=Yes,0=No","Have you used any weight loss medications in the past 4 hours? 1=Yes,0=No","Have you consumed any food containing alcohol or beverages containing alcohol in the past 4 hours? 1=Yes,0=No","Have you used a rescue intermediate-acting inhaled betaagonist in the past 6 hours? 1=Yes,0=No","Have you used a short-acting anticholinergic in the past 6 hours? 1=Yes,0=No","Have you used any ophthalmic antihistamines in the past 6 hours? 1=Yes,0=No","Have you used any nasal antihistamines in the past 6 hours? 1=Yes,0=No","Have you used any nasal decongestants in the past 6 hours? 1=Yes,0=No","Have you used any oral antihistamines in the past 48 hours? 1=Yes,0=No","Have you used any oral decongestants or cold remedies in the past 48 hours? 1=Yes,0=No","Have you used any smokeless tobacco products today? 1=Yes,0=No","At this time, is your asthma worse because of recent exposure to triggers? 1=Yes,0=No","Is there any other reason you should not proceed with spirometry testing? 1=Yes,0=No","Is the participant eligible to proceed with the spirometry testing? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past 4 weeks, I worried about the long-term effects of asthma on my health (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I had to worry about asthma triggers (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, my asthma was on my mind (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, it was hard to get a good nights sleep because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt like I couldnt enjoy life because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt that asthma was controlling my life (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt frustrated that I couldnt make plans in advance because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, everyday activities were a struggle (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, asthma placed stress on my relationships with family, friends, significant others, or co-workers (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, I felt frustrated that I have to do things differently than people who dont have asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt like I missed out on doing things with others because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, I had to do a lot of planning to make sure I always had an inhaler ready (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Randomized treatment assignment period 1: (Mometasone, Tiotropium or Placebo)","Randomized treatment assignment period 2: (Mometasone, Tiotropium or Placebo)","Randomized treatment assignment period 3: (Mometasone, Tiotropium or Placebo)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Sex (1=Male, 2=Female)","Primary Racial Identification for Spirometry (1=Amer Indian or Alaskan Native, 2=Asian or Pacific Islander, 3=Black, 4=White, 5=Hispanic or Latino, 6=Other)","Participant's Age at Visit 1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Description of Adverse Event (ICD9 Code)","Is the participant currently taking study drug? 1=Yes,0=No","Time interval between the last administration of the study drug and the Adverse Event","What was the unit of time for the interval in Question #4? 1=Second, 2=Minute, 3=Hour, 4=Day","Fatal event(1=Yes,0=No)","Life-threatening event (1=Yes,0=No)","Inpatient hospitalization required(1=Yes,0=No)","Hospitalization prolonged  (1=Yes,0=No)","Disabling or incapacitating  (1=Yes,0=No)","Overdose (1=Yes,0=No)","Cancer (1=Yes,0=No)","Congenital anomaly(1=Yes,0=No)","Serious laboratory abnormality with clinical symptoms(1=Yes,0=No)","Height failure (per protocol MOP)(1=Yes,0=No)","Pregnancy(1=Yes,0=No)","Other (1=Yes,0=No)","What in your opinion caused the event? 1=Yes,0=No","Withdrawal of study drug(s)(1=Yes,0=No)","Concurrent medication  (1=Yes,0=No)","Other condition or event  (1=Yes,0=No)","Was the event expected or unexpected? 1=Expected, 2=Unexpected","Was the event possibly, probably, or definitely related to study participation? 1=Yes,0=No","Date of Adverse Event","Admission date","Discharge date","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","What type of height measurement was obtained? 1=Standing height, 2=Length","First measurement __ __ __ . __ cm","Second measurement __ __ __ . __ cm","Third measurement __ __ __ . __ cm","Average height or length measurement __ __ __ . __ cm","In your judgment, was the participant's height or length measurement acceptable? 1=Yes,0=No","Weight (shoes off, light clothing) __ __ __ . __ kg","Does the participant have evidence of oral candidiasis? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant fail to respond within 48 hours to the treatment failure rescue algorithm? 1=Yes,0=No","Did the participant use at least 16 puffs as needed albuterol per 24 hours for a period of 48 hours? 1=Yes,0=No","Did the participant experience prebronchodilator FEV1 values &lt; 50Pct of the baseline prebronchodilator value obtained at Visit 1 on two consecutive spirometric determinations made on different days? 1=Yes, 0=No, 9=Not evaluated","Did the participant experience prebronchodilator FEV1 values &lt; 40Pct of predicted on two consecutive spirometric determinations made on different days? 1=Yes, 0=No, 9=Not evaluated","Did the study or treating physician prescribe the participant oral/parenteral corticosteroids for the treatment of his/her asthma? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation in the opinion of the study investigator or personal physician? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation? 1=Yes,0=No","Did the participant seek care for significant asthma exacerbation conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Was the participant hospitalized? 1=Yes,0=No","Duration of hospital stay (0-99.9 days)","Was intubation or ventilation assistance required? 1=Yes,0=No","Was the participant admitted to the intensive care unit? 1=Yes,0=No","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Inhaled corticosteroids (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Nebulized bronchodilator (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Oral corticosteroids (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, IM or IV steroids (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Antibiotics (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Other (1=Yes, 0=No )","Why do you think the participant experienced a significant asthma exacerbation? 1=Common cold, 2=Allergies, 3=Pollution or chemical irritant, 4=Too little asthma maintenance medication, 5=Exercise, 6=Other","Date exacerbation conditions were met","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Which phone contact following Asthma Exacerbation visit is being performed? 1=Phone Contact #1 , 2=Phone Contact #2 , 3=Phone Contact #3","Since the most recent Asthma Exacerbation contact, has the participant sought additional care for significant asthma exacerbation conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Was the participant hospitalized? 1=Yes,0=No","Duration of hospital stay (0-99.9 days)","Was intubation or ventilation assistance required? 1=Yes,0=No","Was the participant admitted to the intensive care unit? 1=Yes,0=No","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Inhaled corticosteroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Nebulized bronchodilator (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Oral corticosteroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, IM or IV steroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Antibiotics (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Runny Nose  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Post nasal drip  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Need to blow your nose  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Facial pain/pressure(0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Nasal obstruction  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time spirometry started (based on 24-hour clock)","The reported FEV1 and FVC are the best volumes of all acceptable maneuvers. 1.Highest FVC","Highest FEV1 (Pct predicted)","The reported flow rates correspond to the maneuver where FEV1 + FVC is maximized. FEF Max","FEF25-75","In your judgment, was the subject's spirometry technique acceptable? 0=No, 1=Yes","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Highest FEV1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","TRIAL_TYPE","FVC","FEV1","PEF","FEF2575","FET","Number of times the participant woke up last night due to asthma symptoms # Times woke up due to asthma (0-9)","Number of puffs the participant will take from the WHITE study inhaler this morning # Puffs from WHITE inhaler in AM (0-9)","only) Number of puffs the participant will take from the BLUE study inhaler this morning # Puffs from BLUE inhaler in AM (0-9)","Has the participant taken any puffs from his/her RED RESCUE albuterol inhaler in the past 4 hours? 1=yes, 0=no","Shortness of Breath Score Shortness of Breath Score overnight (0,1,2,3)","Chest tightness score Chest Tightness Score overnight (0,1,2,3)","Wheezing score Wheezing Score overnight (0,1,2,3)","Coughing score Coughing Score overnight (0,1,2,3)","Phlegm/Mucus score Phlegm/Mucus Score overnight (0,1,2,3)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","date 'fixed' for analysis so that associated AM and PM always have same ddate","TRIAL_TYPE","FVC","FEV1","PEF","FEF2575","FET","Number of puffs the participant will take from the WHITE study inhaler tonight # Puffs from WHITE inhaler in PM (0-9)","Has the participant taken any puffs from his/her RED RESCUE albuterol inhaler during the past 4 hours? 1=yes, 0=no","Shortness of Breath Score Shortness of Breath Score since waking this morning (0,1,2,3)","Chest tightness score Chest Tightness Score since waking this morning (0,1,2,3)","Wheezing score Wheezing Score since waking this morning (0,1,2,3)","Coughing score Coughing Score since waking this morning (0,1,2,3)","Phlegm/Mucus score Phlegm/Mucus Score since waking this morning (0,1,2,3)","Did your regular exercise cause unusually severe shortness of breath during the past 24 hours? 1=Yes, 0=No, 9=N/A","Number of RED RESCUE albuterol puffs taken in the past 24 hours to prevent symptoms# Preventive  RED RESCUE puffs  in past 24 hours?","Number of RED RESCUE albuterol puffs taken for asthma symptoms or low peak flow during past 24 hours  # RED RESCUE  puffs for symptoms  in past 24 hours?","Number of times used RED RESCUE albuterol inhaler for asthma symptoms during past 24 hours  # Times used  RED RESCUE for symptoms in past 24 hours?","Did RED RESCUE albuterol relieve symptoms for less than 4 hours after treatment during past 24 hours? 1=Yes,0=No","Was the participant seen by a healthcare provider  for an unscheduled visit in the past 24 hours due to asthma symptoms? 1=Yes,0=No","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Processing: Time processing started (based on 24-hour clock) __ __ __ __ (0 - 2359)","Processing Sample: Total Cell Count: __ __ __ __. __x 10^4 cells/ml (0 - 9999.9)","Was the participant's sputum sample processed within 4 hours after collection? 1=Yes,0=No","Processing Sample: Processing Date __ __ / __ __ / 20 __ __","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Rate slide's quality:  (1=Very good, 2=Good, 3=Acceptable, 4=Poor but readable, 5=Not readable)","Record the number on the slide(s) that was (were) read (0 - 9)","Record the number on the slide(s) that was (were) read: These are numbers that were assigned to the slides at each site. (0 -9)","Total Cell Count: __ __ __ __ . __ x 10^4 cells/ml (0 - 9999.9)","Differential Cell Counts: Squamous Cells __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Epithelial Cells __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Macrophages __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Neutrophils __ __ . __ Pct (0 -99.9)","Differential Cell Counts: Eosinophils __ __ . __ Pct (0 - 99.9)","Differential Cell Counts: Lymphocytes __ __ . __ Pct (0 -99.9)","Date of Read __ __ / __ __ / 20 __ __","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","For this protocol, what was the duration of sputum induction the first time the participant's sample was processed within 4 hours after collection?","Sputum induction start time (based on 24-hour clock)","Sputum induction stop time (based on 24-hour clock)","Duration of sputum induction collection phase at this visit -- minutes (0 - 99.9)","Was the duration ? 1=Yes,0=No","Volume of sputum sample at this visit -- ml (0.0 - 999.9)","Is the volume adequate for processing? 1=Yes,0=No","Is the sample adequate for laboratory analysis? 1=Yes,0=No","Participant's FEV1 immediately after completion of sputum induction: -- (0.0 -9.99 L)","FEV1 (Pct predicted) (0 - 999)","Time of FEV1 in Q7a (based on 24-hour clock)","Percent difference in FEV1 ((Reference - Q7a) / Reference) X100 (-99.9 - 99.9 Pct)","Did the participant's FEV1 drop &gt; 10Pct from reference FEV1 as indicated in Q7d? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Participant's FEV1 after initial 2 puffs of albuterol: 1a.FEV1 __ . __ __ L (0 -9.99)","Participant's FEV1 after initial 2 puffs of albuterol: 1b. FEV1 (Pct predicted)__ __ __ Pct predicted (0 - 999)","Participant's FEV1 after initial 2 puffs of albuterol: 1c. Time of FEV1 from Q1a (based on 24-hour clock) __ __ __ __ (0 - 2359)","Participant's FEV1 after initial 2 puffs of albuterol: 1d. Was the FEV1 from Q1a &gt;=the sputum induction reversal reference value in the gray box above? 1=Yes,0=No","Participant's FEV1 after 2 additional puffs of albuterol: 2a. FEV1 __ . __ __ L (0 - 9.99)","Participant's FEV1 after 2 additional puffs of albuterol: 2b. FEV1 (Pct predicted) __ __ __ Pct predicted (0 - 999)","Participant's FEV1 after 2 additional puffs of albuterol: 2c. Time of FEV1 from Q2a (based on 24-hour clock)__ __ __ __ (0 - 2359)","Participant's FEV1 after 2 additional puffs of albuterol: 2d.Was the FEV1 from Q2a ? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was the participant's FEV1 after reversal from the methacholine challenge greater yhan 90Pct of the baseline FEV1 (1=Yes,0=No)","If NO, has the participant received permission from the supervising physician to proceed with sputum induction testing? 1=Yes,0=No","Physician's Signature:_________ (1=signature present, NULL=signature missing)","Participant's FEV1 used for assessment of eligibility for sputum induction __ . __ __ L (0 - 9.99)","Participant's FEV1 (Pct predicted) used for assessment of eligibility for sputum induction __ __ __ Pct predicted (O-999)","Was the participant's FEV1  from Q3 ? 1=Yes,0=No","Has the participant used any smokeless tobacco products  today? 1=Yes,0=No","Is there any other reason the participant should not proceed with sputum induction? 1=Yes,0=No","Is the participant eligible for sputum induction? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has the participant completed the study through Visit 9? 1=Yes,0=No","Who initiated termination of the participant? 1=Participant, 2=Clinical Staff","1=no longer interested in participating*, 2=no longer willing to follow protocol*, 3=difficult access to clinic , 4=unable to make visits during clinic hours, 5=moving out of the area, 6=unable to continue due to personal constraints*, 7=unable to continue","Did clinical staff terminate the participant due to pregnancy? 1=Yes, 0=No, 9=N/A","Did clinical staff terminate the participant due to loss to follow-up? 1=Yes,0=No","Did clinical staff terminate the participant due to loss to follow-up, type of contact (1=In-person visit, 2=Phone call)","Did clinical staff terminate the participant due to an asthma-related adverse event? 1=Yes,0=No","Did clinical staff terminate the participant due to a medication-related adverse event? 1=Yes,0=No","Did clinical staff terminate the participant due to an adverse event not related to asthma or medications? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility during the Supervised Washout ? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with medication dosing? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with diary completion? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with visit attendance? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with peak flow monitoring? 1=Yes,0=No","Did clinical staff terminate the participant due to significant asthma exacerbation or two treatment failures during run-in ? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility during the run-in period  for reasons other than compliance or exacerbation/treatment failure or high sputum eosinophils ? 1=Yes,0=No","Did clinical staff terminate the participant due to other reason? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility due to high sputum eosinophils in the run-in ? 1=Yes,0=No","Did clinical staff terminate the participant due to... Indicate the letter corresponding to the primary reason the participant was terminated. (a-n)","Did clinical staff terminate the participant due to loss to follow-up, date of last contact with participant","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant seek care for treatment failure conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact, other than call to notify clinic as instructed in spirotel","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Inhaled corticosteroids (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Nebulized bronchodilator (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started: Oral corticosteroids form.  (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  IM or IV steroids form.  (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started: Antibiotics (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant awaken from asthma three or more times in a two-week period or on two consecutive nights? 1=Yes,0=No","Did the participant use albuterol for relief of symptoms four or more times/day for two or more consecutive days? 1=Yes,0=No","Did albuterol relieve symptoms for less than four hours after treatment? 1=Yes,0=No","Did the participant use albuterol for relief of symptoms daily for seven days, and this use exceeded two times the weekly use of albuterol in the baseline period? 1=Yes,0=No","Did regular exercise cause unusually severe shortness of breath on two or more days during a seven day period? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation? 1=Yes,0=No","Did the participant experience a treatment failure? 1=Yes,0=No","Has the participant experienced two treatment failures in the current treatment period ? 1=Yes,0=No","Does the study physician feel the study treatment contributed directly to the treatment failure, such that it is in the participants best interest to discontinue the current blinded treatment? 1=Yes,0=No","Did the participant begin daily treatment with high-dose ICS for 10 days? 1=Yes,0=No","Did the participant begin daily treatment with high-dose ICS for 10 days? 1=Did not think symptoms werebad enough, 2=Other","Date treatment failure conditions were met","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","During this last treatment period , has the participant experienced a treatment failure or asthma exacerbation? 1=Yes,0=No","Has the participant experienced two or more treatment failures in this last treatment period, or does the study physician feel the study treatment contributed directly to the treatment failure such that it is in the participants best interest to discontinu","As of today, have 2 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","As of today, have 6 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","As of today, have 3 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Are you currently employed ? 1=Yes,0=No","In general, how many hours per week do you usually work?","During the past seven days, how many hours did you miss from work because of problems associated with your asthma?","During the past seven days, how much did asthma affect your productivity while you were working?","Do you currently attend classes in an academic setting ? 1=Yes,0=No","In general, how many hours per week do you usually attend classes?","During the past seven days, how many hours did you miss from class or school because of problems associated with your asthma?","During the past seven days, how much did asthma affect your productivity while in school or attending classes in an academic setting?","During the past seven days, how much did your asthma affect your ability to do your regular daily activities, other than work at a job or attend classes?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","How sick today? 0=Not Sick, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely","Runny Nose:(0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","PLUGGED_NOSE  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SNEEZING  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SORE_THROAT  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SCRATCHY_THROAT  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","COUGH (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","HOARSENESS  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","HEAD_CONGESTION  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","CHEST_CONGESTION  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","FEELING_TIRED (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","THINK_CLEARLY  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SLEEP_WELL  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","BREATH_EASILY  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WALK  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","DAILY_ACTIVITIES  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WORK_OUTSIDE_HOME  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WORK_INSIDE_HOME  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","INTERACT  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","LIVE_LIFE  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","MY_COLD_IS  (1=Very much better, 2=Somewhat better, 3=A little better, 4=The same, 5=A little worse, 6=Somewhat worse, 7=Very much worse)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Date (number of days from Visit 1)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Num<\/th>\n      <th>Variable<\/th>\n      <th>Type<\/th>\n      <th>Label<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":1},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Num","targets":1},{"name":"Variable","targets":2},{"name":"Type","targets":3},{"name":"Label","targets":4}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-b55be07e8cea91a90454" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-b55be07e8cea91a90454">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80","81","82","83","84","85","86","87","88","89","90","91","92","93","94","95","96","97","98","99","100","101","102","103","104","105","106","107","108","109","110","111","112","113","114","115","116","117","118","119","120","121","122","123","124","125","126","127","128","129","130","131","132","133","134","135","136","137","138","139","140","141","142","143","144","145","146","147","148","149","150","151","152","153","154","155","156","157","158","159","160","161","162","163","164","165","166","167","168","169","170","171","172","173","174","175","176","177","178","179","180","181","182","183","184","185","186","187","188","189","190","191","192","193","194","195","196","197","198","199","200","201","202","203","204","205","206","207","208","209","210","211","212","213","214","215","216","217","218","219","220","221","222","223","224","225","226","227","228","229","230","231","232","233","234","235","236","237","238","239","240","241","242","243","244","245","246","247","248","249","250","251","252","253","254","255","256","257","258","259","260","261","262","263","264","265","266","267","268","269","270","271","272","273","274","275","276","277","278","279","280","281","282","283","284","285","286","287","288","289","290","291","292","293","294","295","296","297","298","299","300","301","302","303","304","305","306","307","308","309","310","311","312","313","314","315","316","317","318","319","320","321","322","323","324","325","326","327","328","329","330","331","332","333","334","335","336","337","338","339","340","341","342","343","344","345","346","347","348","349","350","351","352","353","354","355","356","357","358","359","360","361","362","363","364","365","366","367","368","369","370","371","372","373","374","375","376","377","378","379","380","381","382","383","384","385","386","387","388","389","390","391","392","393","394","395","396","397","398","399","400","401","402","403","404","405","406","407","408","409","410","411","412","413","414","415","416","417","418","419","420","421","422","423","424","425","426","427","428","429","430","431","432","433","434","435","436","437","438","439","440","441","442","443","444","445","446","447","448","449","450","451","452","453","454","455","456","457","458","459","460","461","462","463","464","465","466","467","468","469","470","471","472","473","474","475","476","477","478","479","480","481","482","483","484","485","486","487","488","489","490","491","492","493","494","495","496","497","498","499","500","501","502","503","504","505","506","507","508","509","510","511","512","513","514","515","516","517","518","519","520","521","522","523","524","525","526","527","528","529","530","531","532","533","534","535","536","537","538","539","540","541","542","543","544","545","546","547","548","549","550","551","552","553","554","555","556","557","558","559","560","561","562","563","564","565","566","567","568","569","570","571","572","573","574","575","576","577","578","579","580","581","582","583","584","585","586","587","588","589","590","591","592","593","594","595","596","597","598","599","600","601","602","603","604","605","606","607","608","609","610","611","612","613","614","615","616","617","618","619","620","621","622","623","624","625","626","627","628","629","630","631","632","633","634","635","636","637","638","639","640","641","642","643","644","645","646","647","648","649","650","651","652","653","654","655","656","657","658","659","660","661","662","663","664","665","666","667","668","669","670","671","672","673","674","675","676","677","678","679","680","681","682","683","684","685","686","687","688","689","690","691","692","693","694","695","696","697","698","699","700","701","702","703","704","705","706","707","708","709","710","711","712","713","714","715","716","717","718","719","720","721","722","723","724","725","726","727","728","729","730","731","732","733","734","735","736","737","738","739","740","741","742","743","744","745","746","747","748","749","750","751","752","753","754","755","756","757","758","759","760","761","762","763","764","765","766","767","768","769","770","771","772","773","774","775","776","777","778","779","780","781","782","783","784","785","786","787","788","789","790","791","792","793","794","795","796","797","798","799","800","801","802","803","804","805","806","807","808","809","810","811","812","813","814","815","816","817","818","819","820","821","822","823","824","825","826","827","828","829","830","831","832","833","834","835","836","837","838","839","840","841","842","843","844","845","846","847","848","849","850","851","852","853","854","855","856","857","858","859","860","861","862","863","864","865","866","867","868","869","870","871","872","873","874","875","876","877","878","879","880","881","882","883","884","885","886","887","888","889","890","891","892","893","894","895","896","897","898","899","900","901","902","903","904","905","906","907","908","909","910","911","912","913","914","915","916","917","918","919","920","921","922","923","924","925","926","927","928","929","930","931","932","933","934","935","936","937","938","939","940","941","942","943","944","945","946","947","948","949","950","951","952","953","954","955","956","957","958","959","960","961","962","963","964","965","966","967","968","969","970","971","972","973","974","975","976","977","978","979","980","981","982","983","984","985","986","987","988","989","990","991","992","993","994","995","996","997","998","999","1000","1001","1002","1003","1004","1005","1006","1007","1008","1009","1010","1011","1012","1013","1014","1015","1016","1017","1018","1019","1020","1021","1022","1023","1024","1025","1026","1027","1028","1029","1030","1031","1032","1033","1034","1035","1036","1037","1038","1039","1040","1041","1042","1043","1044","1045","1046","1047","1048","1049","1050","1051","1052","1053","1054","1055","1056","1057","1058","1059","1060","1061","1062","1063","1064","1065","1066","1067","1068","1069","1070","1071","1072","1073","1074","1075","1076","1077","1078","1079","1080","1081","1082","1083","1084","1085","1086","1087","1088","1089","1090","1091","1092","1093","1094","1095","1096","1097","1098","1099","1100","1101","1102","1103","1104","1105","1106","1107","1108","1109","1110","1111","1112","1113","1114","1115","1116","1117","1118","1119","1120","1121","1122","1123","1124","1125","1126","1127","1128","1129","1130","1131","1132","1133","1134","1135","1136","1137","1138","1139","1140","1141","1142","1143","1144","1145","1146","1147","1148","1149","1150","1151","1152","1153","1154","1155","1156","1157","1158","1159","1160","1161","1162","1163","1164","1165","1166","1167","1168","1169","1170","1171","1172","1173","1174","1175","1176","1177","1178","1179","1180","1181","1182","1183","1184","1185","1186","1187","1188","1189","1190"],[1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,1,2,3,4,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,1,2,3,4,5,6,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28],["AAQ_1000","AAQ_1010","AAQ_1020","AAQ_1030","AAQ_1040","AAQ_1050","AAQ_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ABP_1000","ABP_1010","ABP_1020","ABP_1030","ABP_1040","ABP_1050","ABP_1060","ABP_1070","ABP_1080","ABP_1090","ABP_1100","ABP_1110","ABP_1120","ABP_1130","ABP_1140","ABP_1150","ABP_1160","ABP_1170","ABP_1180","ABP_1190","ABP_1200","ABP_1210","ABP_1220","ABP_1230","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ACT_1","ACT_2","ACT_3","ACT_4","ACT_5","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ICD9_CAT","AEC_1000","AEC_1010","AEC_1040","AEC_1050","AEC_1060","AEC_1070","AEC_1080","AEC_1090","AEC_1100","AEC_1110","AEC_1120","AEC_1020","AEC_1030","VNUM_C","VNUM","ICD9LONG","ICD9SHRT","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","AHA_1000","AHA_1020","AHA_1030","AHA_1040","AHA_1050","AHA_1060","AHA_1070","AHA_1090","AHA_1100","AHA_1110","AHA_1120","AHA_1130","AHA_1140","AHA_1150","AHA_1160","AHA_1170","AHA_1180","AHA_1190","AHA_1200","AHA_1210","AHA_1220","AHA_1250","AHA_1260","AHA_1270","AHA_1280","AHA_1290","AHA_1300","AHA_1310","AHA_1320","AHA_1330","AHA_1340","AHA_1350","AHA_1360","AHA_1370","AHA_1380","AHA_1390","AHA_1400","AHA_1410","AHA_1420","AHA_1430","AHA_1440","AHA_1450","AHA_1460","AHA_1470","AHA_1500","AHA_1570","AHA_1580","AHA_1590","AHA_1600","AHA_1730","AHA_1740","AHA_1750","AHA_1760","AHA_1770","AHA_1780","AHA_1790","AHA_1800","AHA_1810","AHA_1820","AHA_1830","AHA_1840","AHA_1850","AHA_1890","AHA_1900","VNUM_C","VNUM","VDATE","DATE_VAPE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ASU_1000","ASU_1010","ASU_1020","ASU_1030","ASU_1040","ASU_1050","ASU_1060","ASU_1070","ASU_1080","ASU_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","BAS_1000","BAS_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","BMI","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAGE","CHG_1000","CHG_1010","CHG_1020","CHG_1030","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","CME_1040","CME_1050","CME_1055","CME_1000","CME_1010","CME_1020","CME_1030","CME_1080","CME_1090","CME_1060","CME_1070","VNUM_C","VNUM","GEN_NAME","CLASS","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","COM_1000","COM_1010","COM_1020","COM_1030","COM_1040","COM_1050","COM_1060","COM_1070","COM_1080","COM_1090","COM_1100","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","CTX_1000","CTX_1010","CTX_1030","CTX_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E0A_1000","E0A_1020","E0A_1030","E0A_1040","E0A_1043","E0A_1045","E0A_1050","E0A_1060","E0A_1070","E0A_1073","E0A_1075","E0A_1080","E0A_1090","E0A_1100","E0A_1110","E0A_1120","E0A_1130","E0A_1140","E0A_1145","E0A_1150","E0A_1160","E0A_1170","E0A_1180","E0A_1190","E0A_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E0B_1000","E0B_1010","E0B_1020","E0B_1030","E0B_1040","E0B_1050","E0B_1060","E0B_1065","E0B_1070","E0B_1080","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E1_1000","E1_1020","E1_1030","E1_1040","E1_1050","E1_1060","E1_1065","E1_1070","E1_1080","E1_1090","E1_1100","E1_1110","E1_1120","E1_1130","E1_1140","E1_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E2_1000","E2_1010","E2_1020","E2_1030","E2_1040","E2_1050","E2_1060","E2_1070","E2_1080","E2_1090","E2_1100","E2_1110","E2_1120","E2_1130","E2_1140","E2_1145","E2_1150","E2_1160","E2_1170","E2_1173","E2_1175","E2_1180","E2_1190","E2_1200","E2_1210","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E3_1000","E3_1010","E3_1020","E3_1030","E3_1070","E3_1080","E3_1090","E3_1100","E3_1110","E3_1120","E3_1130","E3_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E4_1000","E4_1010","E4_1020","E4_1030","E4_1050","E4_1060","E4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E5_1000","E5_1010","E5_1020","E5_1030","E5_1035","E5_1040","E5_1050","E5_1060","E5_1065","E5_1070","E5_1080","E5_1085","E5_1090","E5_1100","E5_1110","E5_1120","E5_1130","E5_1140","E5_1150","E5_1160","E5_1170","E5_1180","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ENO_1000","ENO_1010","ENO_1020","ENO_1040","ENO_1050","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","HEQ_1000","HEQ_1010","HEQ_1020","HEQ_1030","HEQ_1040","HEQ_1050","HEQ_1060","HEQ_1070","HEQ_1080","HEQ_1090","HEQ_1100","HEQ_1110","HEQ_1120","HEQ_1130","HEQ_1140","HEQ_1150","HEQ_1160","HEQ_1170","HEQ_1180","HEQ_1190","HEQ_1200","HEQ_1210","HEQ_1220","HEQ_1230","HEQ_1240","HEQ_1250","HEQ_1260","HEQ_1270","HEQ_1280","HEQ_1290","HEQ_1300","HEQ_1310","HEQ_1320","HEQ_1330","HEQ_1340","HEQ_1350","HEQ_1360","HEQ_1370","HEQ_1380","HEQ_1390","HEQ_1400","HEQ_1410","HEQ_1420","HEQ_1430","HEQ_1440","HEQ_1450","HEQ_1460","HEQ_1470","HEQ_1480","HEQ_1490","HEQ_1500","HEQ_1510","HEQ_1520","HEQ_1530","HEQ_1540","HEQ_1550","HEQ_1560","HEQ_1570","HEQ_1580","HEQ_1590","HEQ_1600","HEQ_1610","HEQ_1620","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SEI_1000","SEI_1010","SEI_1020","SEI_1030","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","IGE","RE1","RE5","RE72","RE74","RMX1","RI6","RGX2","RTX4","RTX6","RWX1","RW3","RD2","RD1","N_ALLERGENPOS","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","LAB_1000","LAB_1010","LAB_1020","LAB_1030","LAB_1040","LAB_1050","LAB_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PLE_1000","PLE_1010","PLE_1020","PLE_1030","PLE_1040","PLE_1050","PLE_1060","PLE_1070","PLE_1080","PLE_1090","PLE_1100","PLE_1110","PLE_1130","PLE_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MTH_1000","MTH_1010","MTH_1020","MTH_1030","MTH_1040","MTH_1050","MTH_1060","MTH_1070","MTH_1080","MTH_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MAD_1000","MAD_1010","MAD_1020","MAD_1030","MAD_1040","MAD_1050","MAD_1060","MAD_1070","MAD_1090","MAD_1100","MAD_1110","MAD_1120","MAD_1130","MAD_1140","MAD_1150","MAD_1160","MAD_1170","MAD_1180","MAD_1190","MAD_1200","MAD_1220","MAD_1230","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MCA_1000","MCA_1010","MCA_1050","MCA_1060","MCA_1070","MCA_1080","MCA_1100","MCA_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MCP_1000","MCP_1010","MCP_1030","MCP_1040","MCP_1050","MCP_1060","MCP_1070","MCP_1080","MCP_1090","MCP_1100","MCP_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PA4_1000","PA4_1010","PA4_1020","PA4_1030","PA4_1040","PA4_1050","PA4_1060","PA4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PTX_1000","PTX_1010","PTX_1020","PTX_1030","PTX_1040","PTX_1050","PTX_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","VNUM","PERIOSTIN","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PI4_1000","PI4_1010","PI4_1020","PI4_1030","PI4_1040","PI4_1050","PI4_1060","PI4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PRG_1000","PRG_1010","PRG_1020","PRG_1030","PRG_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAD_1000","PAD_1010","PAD_1020","PAD_1030","PAD_1040","PAD_1050","PAD_1060","PAD_1070","PAD_1080","PAD_1090","PAD_1100","PAD_1110","PAD_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAL_1000","PAL_1010","PAL_1020","PAL_1030","PAL_1040","PAL_1050","PAL_1060","PAL_1070","PAL_1080","PAL_1090","PAL_1100","PAL_1110","PAL_1120","PAL_1130","PAL_1150","PAL_1160","PAL_1170","PAL_1180","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PTR_1000","PTR_1010","PTR_1050","PTR_1060","PTR_1100","PTR_1140","PTR_1180","PTR_1220","PTR_1260","PTR_1300","PTR_1340","PTR_1380","PTR_1390","PTR_1430","PTR_1470","PTR_1480","PTR_1490","PTR_1500","PTR_1535","PTR_1540","PTR_1550","PTR_1560","PTR_1600","PTR_1610","PTR_1620","PTR_1630","PTR_1670","PTR_1710","PTR_1750","PTR_1790","PTR_1830","PTR_1870","PTR_1910","VNUM_C","VNUM","VDATE","DATE_Q02","DATE_Q03","DATE_Q04","DATE_Q05","DATE_Q07","DATE_Q08","DATE_Q09","DATE_Q10","DATE_Q11","DATE_Q12","DATE_Q13","DATE_Q14","DATE_Q15","DATE_Q16","DATE_Q17","DATE_Q18","DATE_Q19","DATE_Q20","DATE_Q21","DATE_Q22","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","P10_1000","P10_1010","P10_1020","P10_1030","P10_1040","P10_1050","P10_1060","P10_1070","P10_1080","P10_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PCH_1000","PCH_1010","PCH_1020","PCH_1030","PCH_1040","PCH_1050","PCH_1060","PCH_1070","PCH_1080","PCH_1090","PCH_1100","PCH_1110","PCH_1120","PCH_1130","PCH_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","IAQ_1000","IAQ_1010","IAQ_1020","IAQ_1030","IAQ_1040","IAQ_1050","IAQ_1060","IAQ_1070","IAQ_1080","IAQ_1090","IAQ_1100","IAQ_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","REGIMEN_PERIOD_1","REGIMEN_PERIOD_2","REGIMEN_PERIOD_3","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","REG_1080","REG_1150","AGE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SER_1010","SER_1020","SER_1030","SER_1040","SER_1050","SER_1060","SER_1070","SER_1100","SER_1110","SER_1120","SER_1130","SER_1140","SER_1150","SER_1160","SER_1170","SER_1180","SER_1190","SER_1200","SER_1210","SER_1220","SER_1240","SER_1250","SER_1000","SER_1080","SER_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PSE_1060","PSE_1070","PSE_1080","PSE_1090","PSE_1100","PSE_1110","PSE_1130","PSE_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SIG_1000","SIG_1010","SIG_1020","SIG_1030","SIG_1040","SIG_1050","SIG_1060","SIG_1080","SIG_1090","SIG_1100","SIG_1110","SIG_1120","SIG_1130","SIG_1140","SIG_1150","SIG_1160","SIG_1170","SIG_1180","SIG_1190","SIG_1200","SIG_1210","SIG_1220","SIG_1230","SIG_1240","SIG_1250","SIG_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SFU_1000","SFU_1010","SFU_1020","SFU_1030","SFU_1040","SFU_1050","SFU_1060","SFU_1070","SFU_1080","SFU_1090","SFU_1100","SFU_1110","SFU_1120","SFU_1130","SFU_1140","SFU_1150","SFU_1160","SFU_1170","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SNQ_1000","SNQ_1010","SNQ_1020","SNQ_1030","SNQ_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SPI_1010","SPI_1020","SPI_1040","SPI_1050","SPI_1060","SPI_1070","VNUM_C","VNUM","VDATE","SPI_1030","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","AM_TRIAL_TYPE","AM_FVC","AM_FEV1","AM_PEF","AM_FEF2575","AM_FET","DRY_1","DRY_2","DRY_3","DRY_4","DRY_5","DRY_6","DRY_7","DRY_8","DRY_9","VNUM_C","VNUM","VDATE","DDATE","PM_TRIAL_TYPE","PM_FVC","PM_FEV1","PM_PEF","PM_FEF2575","PM_FET","DRY_10","DRY_11","DRY_12","DRY_13","DRY_14","DRY_15","DRY_16","DRY_17","DRY_18","DRY_19","DRY_20","DRY_21","DRY_22","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SLA_1010","SLA_1020","SLA_1030","SLA_1000","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SRE_1010","SRE_1020","SRE_1030","SRE_1040","SRE_1050","SRE_1060","SRE_1070","SRE_1080","SRE_1090","SRE_1100","SRE_1000","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SPU_1000","SPU_1010","SPU_1020","SPU_1030","SPU_1040","SPU_1050","SPU_1060","SPU_1070","SPU_1080","SPU_1090","SPU_1100","SPU_1110","SPU_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SAD_1000","SAD_1010","SAD_1020","SAD_1030","SAD_1040","SAD_1050","SAD_1060","SAD_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SCH_1000","SCH_1010","SCH_1020","SCH_1030","SCH_1040","SCH_1050","SCH_1055","SCH_1060","SCH_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TER_1000","TER_1010","TER_1020","TER_1040","TER_1050","TER_1070","TER_1080","TER_1090","TER_1100","TER_1110","TER_1120","TER_1130","TER_1140","TER_1150","TER_1160","TER_1170","TER_1180","TER_1185","TER_1200","TER_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TXF_1000","TXF_1010","TXF_1020","TXF_1030","TXF_1040","TXF_1050","TXF_1060","TXF_1070","TXF_1080","TXF_1090","TXF_1100","TXF_1110","TXF_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TCH_1000","TCH_1010","TCH_1020","TCH_1030","TCH_1040","TCH_1050","TCH_1060","TCH_1080","TCH_1090","TCH_1100","TCH_1110","TCH_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","WSH_1000","WSH_1010","WSH_1020","WSH_1040","WSH_1050","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","WPA_1000","WPA_1010","WPA_1020","WPA_1030","WPA_1040","WPA_1050","WPA_1060","WPA_1070","WPA_1080","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","HOW_SICK_TODAY","RUNNY_NOSE","PLUGGED_NOSE","SNEEZING","SORE_THROAT","SCRATCHY_THROAT","COUGH","HOARSENESS","HEAD_CONGESTION","CHEST_CONGESTION","FEELING_TIRED","THINK_CLEARLY","SLEEP_WELL","BREATH_EASILY","WALK","DAILY_ACTIVITIES","WORK_OUTSIDE_HOME","WORK_INSIDE_HOME","INTERACT","LIVE_LIFE","MY_COLD_IS","VNUM_C","VNUM","VDATE","WDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER"],["Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Char","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Char","Char","Num","Num","Num","Num","Char","Char","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Char","Num","Char","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Char","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num"],["In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, school, or at home? 0=None of the time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All of the time","During the past 3 days, how often have you had asthma symptoms? 0=Not at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how often have you used your rescue inhaler or nebulizer medication ? 0=Not at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how many total times did your asthma symptoms wake you up from sleep? 0=Not at all, 1=1 time in the last 3 days, 2=2-3 times in the last 3 days, 3=4-5 times in the last 3 days, 4=&gt;=6 times in the last 3 days","How would you rate the amount of impairment you have experienced due to your asthma in the past 3 days? 0=No impairment, 1=Mild impairment, 2=Moderate impairment, 3=Severe impairment, 4=Very severe impairment","How stressed or frightened were you by your asthma symptoms in the past 3 days? 0=Not at all, 1=Mildly, 2=Moderately, 3=Severely, 4=Very severely","Why do you think your asthma was worse in the past 3 days compared to what is normal for you? 0=I have not been worse over the past 3 days. My asthma symptoms have been usual., 1=Common cold, 2=Allergies, 3=Pollution or chemical irritant, 4=Too little asth","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Are you currently retired? 1=Yes,0=No","Are you retired because of asthma? 1=Yes,0=No","Are you currently unemployed? 1=Yes,0=No","Are you unemployed because of asthma? 1=Yes,0=No","Do you get paid to do work? 1=Yes,0=No","How much does your asthma bother you at your paid work? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Overall, how much does your asthma bother you when you do jobs around the house? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Overall, how much does your asthma bother your social life? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Overall, how much does your asthma bother your personal life? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Are you involved in leisure activities, such as: walking for pleasure, sports, exercise, travelling, taking vacations? 1=Yes,0=No","When involved in leisure activities, how much does your asthma bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Would you say that you can't do some of these sorts of things because of asthma? 1=Yes,0=No","How much does your asthma bother you when you sleep? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much does the cost of your asthma medicines bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Do you get free prescriptions? 1=Yes,0=No","How much does the inconvenience or embarrassment of taking your asthma medicines bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much do coughs and colds bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Feeling upset is also a bother.  Does your asthma make you feel anxious, depressed, tired, or helpless? 1=Yes,0=No","Feeling upset is also a bother.  Does your asthma make you feel anxious, depressed, tired, or helpless? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will have an asthma attack when visiting a new place? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will catch a cold? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will let others down? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that your health may get worse in the future? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you won't be able to cope with an asthma attack? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past 4 weeks, how much of the time did your asthma keep you from getting as much done at work, school or at home? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often have your had shortness of breath? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often did your asthma symptoms wake you up at night or earlier than usual in the morning? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often have you used your rescue inhaler or nebulizer medication? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","How would you rate your asthma control during the past 4 weeks? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","ICD-9 category","Description of Adverse Event","ICD9 Code","Ongoing at current visit","Type (1=Intermitent; 2=Continuous)","Severity (1=Mild; 2=Moderate; 3=Severe)","Serious (0=No, 1=Yes)","Likelyhood of Relationship to Study Drug(s) (1=None; 2=Unlikely (remote); 3=Possible; 4=Probable)","Change in Study Drugs (1=Unchanged; 2=Altered)","Outcome (1=Completely recovered; 2=Recovered, but with lasting effects; 3=Death)","Treatment Required (1=None; 2=Medication; 3=Hospitalization; 4=Other)","Ongoing at final visit","Date Started","Date Stopped","Visit Number (character)","Visit Number (numeric)","Long ICD-9 description","Short ICD-9 description","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","ASTHMA HISTORY1.Approximately how old were you when chest symptoms suggesting asthma first appeared?","Did these symptoms appear immediately after or as a result of:1a. a respiratory infection such as a cold or pneumonia? 1=Yes, 0=No, 8=Don't Know","an occupational or job change? 1=Yes, 0=No, 8=Don't Know","a household move? 1=Yes, 0=No, 8=Don't Know","a pregnancy? 1=Yes, 0=No, 8=Don't Know","a hormonal change ? 1=Yes, 0=No, 8=Don't Know","How old were you when a doctor first diagnosed you with asthma?","Have any of your immediate blood relatives been told by a physician that they have asthma? 1=Yes, 0=No, 8=Don't Know","Father  (1=Yes, 0=No, 8=Don't Know)","Brother(s) or Sister(s)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Child(ren)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","ASTHMA SYMPTOMS1.How do you categorize your asthma symptoms throughout the course of the year? 1=Relatively the same all year, 2=Vary by season","Winter? 1=Yes,0=No","Spring? 1=Yes,0=No","Summer? 1=Yes,0=No","Fall? 1=Yes,0=No","In the last 12 months, how many...  5a.Asthma episodes have you had that required emergency care or an unscheduled office visit?","Overnight hospitalizations have you had due to asthma?","Courses of systemic corticosteroid therapy  for asthma have you taken?","Days of work, school, or housework have you missed due to asthma?","In the past 3 months, how many days of work, school, or housework have you missed due to asthma?","Have you ever been admitted to an intensive care unit for asthma? 1=Yes,0=No","How many times have you been admitted to an intensive care unit for asthma?","Have you ever had invasive mechanical ventilation? 1=Yes, 0=No, 8=Don't Know","Have you ever had non-invasive mechanical ventilation? 1=Yes, 0=No, 8=Don't Know","Do any of the following currently provoke your asthma? 1=Yes, 0=No, 8=Don't Know","Menstrual cycle(If participant is male or a postmenopausal female, leave blank.) (1=Yes, 0=No, 8=Don't Know)","Aspirin or non-steroidal anti-inflammatory drugs (e.g., Aleve, Motrin) (1=Yes, 0=No, 8=Don't Know)","Respiratory infections (e.g., colds)  (1=Yes, 0=No, 8=Don't Know)","Irritants (e.g., pollution, odors, perfumes, chemicals, household cleaners) (1=Yes, 0=No, 8=Don't Know)","Weather conditions (e.g., change in weather, humidity)  (1=Yes, 0=No, 8=Don't Know)","Exposure to cold air  (1=Yes, 0=No, 8=Don't Know)","Emotional factors (e.g., stress, laughing)  (1=Yes, 0=No, 8=Don't Know)","Tobacco smoke  (1=Yes, 0=No, 8=Don't Know)","Food additives/preservatives (e.g., MSG, sulfites)  (1=Yes, 0=No, 8=Don't Know)","Allergies (e.g., dust, animals, pollens)(1=Yes, 0=No, 8=Don't Know)","Other (1=Yes,0=No)","To which of the following did a doctor or other health practitioner say you were allergic? 1=Yes, 0=No, 8=Don't Know","Foods  (1=Yes, 0=No, 8=Don't Know)","Things you breathe in or are exposed to (e.g., dust, pollens, molds, animal fur, feathers, dander)  (1=Yes, 0=No, 8=Don't Know)","Stinging insects such as bees or wasps (1=Yes, 0=No, 8=Don't Know)","Latex  (1=Yes, 0=No, 8=Don't Know)","Other  (1=Yes,0=No)","Have you ever had eczema / atopic dermatitis ? 1=Yes, 0=No, 8=Don't Know","If YES, was your eczema diagnosed by a doctor? 1=Yes,0=No","Have any of your immediate blood relatives been told by a physician that they have allergies/eczema/hay fever? 1=Yes, 0=No, 8=Don't Know","Father  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Brother(s) or Sister(s)(1=Yes, 0=No, 8=Don't Know, 9=N/A)","Child(ren)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Did you grow up in a household where you were exposed to tobacco smoke? 1=Yes,0=No","Do you currently smoke? 1=Yes,0=No","Record smoking history in pack-years*. __ __ . __ pack-years","Were you ever a smoker? 1=Yes,0=No","Record smoking history in pack-years*. __ __ . __ pack-years","Do you currently live in a household where you are exposed to tobacco smoke? 1=Yes,0=No","Do you currently vape  or use a hookah ? 1=Yes,0=No","How frequently do you vape or use a hookah? 1=Infrequently , 2=Occasionally , 3=Weekly , 4=Daily","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many days a week","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many times a day","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many years have y","Have you ever vaped or used a hookah in the past? 1=Yes,0=No","Have you ever vaped or used a hookah in the past?  If NO, skip to Q17. Approximately how many years did you vape or use a hookah?  (1-99 years)","Do you currently live in a household where you are exposed to others vaping or using a hookah? 1=Yes,0=No","Do you spend time in social settings  where you are exposed to others vaping or using a hookah? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","When was the last time that you vaped or used a hookah? (number of days from Visit 1)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","How many days were you bothered by coughing during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe was your coughing during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by wheezing during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe was your wheezing during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by shortness of breath during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","Over average, how severe was your shortness of breath during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you awakened at night during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how much of a problem was being awakened at night during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by side effects of your asthma medication during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe were the side effects of your asthma medication during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Participants baseline peak flow (PEF) value (0-999 L/M) First visit (Visit 0A, or Visit 1): PEF (FEF Max) from prebronchodilator (baseline) spirometry at first visit (convert to L/M)","Participants baseline rescue use value (0-99 puffs/day) First visit (Visit 0A, or Visit 1): Self-reported average daily use of albuterol during the 14 days prior to first visit.","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","body mass index computed from height and weight measured at Visit 1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Page Number","Reason for change in scheduled medications status (1=Adverse Event, 2=Other (specify))","Reason for change in scheduled medications status Related adverse event number (0-99)","Current status of participants scheduled medications (1=Discontinued, 2=Resumed)","Date change took effect","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Code to identify a drug unit of measure; (Units)","Code to identify how frequently a drug is taken or administered. (Frequency)","Code to identify the route used to administer a drug. (Route)","Name of Medication","This number corresponds to the ID number of the AHFS monograph. (Medication Code)","Related Event","Medication Dose","Ongoing at current visit","Ongoing at final visit","Start Date","Stop Date","Visit Number (character)","Visit Number (numeric)","GEN_DRUG_NAME","CLASS_TEXT","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Diary and Peak Flow Compliance: Number of full days since the last visit (0-999 days)","Diary and Peak Flow Compliance: Number of days where AM and PM scheduled sessions are complete (AM and PM PEF and all diary questions for AM and PM answered) (0-999 days)","Diary and Peak Flow Compliance: Percent compliance If the compliance value in Q1c is less than 75Pct, re-emphasize the importance of completing scheduled diary assessments and peak flows. (0-150.0Pct)","Scheduled Respimat Compliance (Visits 2-9 only): Number of scheduled puffs since the last visit (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Number of remaining puffs on scheduled Respimat Reference all returned Respimats (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Number of puffs taken (60 x (# used Respimats)) - Q2b (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Percent compliance=Q2c/Q2a x 100 If the participant took less than 75Pct of the scheduled Respimat puffs, re-emphasize the importance of maintaining the daily dosing schedule. (0-150.0Pct)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of scheduled puffs since the last visit (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of remaining puffs reflected on scheduled Twisthaler counters Total the values reflected on all counters for all returned Twisthalers (i.e., out of their pouches). (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of puffs taken (60 x (# used Twisthalers)) - Q3b (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Percent compliance=Q3c/Q3a x 100 If the participant took less than 75Pct of the scheduled Twisthaler puffs, re-emphasize the importance of maintaining the daily dosing schedule. (0-150.0Pct)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","1=mometasone, 2=placebo","Blinded Scheduled Twisthaler Contents - How sure are you about your answer in Q1? 1=Absolutely sure - I know what the Twisthaler/MDI contains, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","1=tiotropium, 2=placebo","Blinded Scheduled Respimat Contents - How sure are you about your answer in Q4? 1=Absolutely sure - I know what the Respimat capsules contain, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has the participant or parent/legal guardian signed the SIENA Informed Consent document? 1=Yes,0=No","Ages 12-17 Only: Has the participant signed and dated the assent form or, if the participant is less than the local age of assent, has the participant given verbal assent? 1=Yes,0=No","Has the participant used an inhaled corticosteroid in the past 6 weeks? 1=Yes,0=No","Will the participant be 12 years of age, or older, as of Visit 1? 1=Yes,0=No","Has the participant used an inhaled corticosteroid or combination inhaled corticosteroid/LABA in the past 3 weeks? 1=Yes,0=No","If NO, has the participant used a leukotriene modifier in the past 3 weeks? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used inhaled corticosteroid therapy?","Over the past 3 months, on average, how many days per week has the participant used inhaled corticosteroid therapy: &lt; 5? 1=Yes,0=No","Over the past 3 months, has the participant used greater than the equivalent of 80-240 mcg of inhaled beclomethasone daily? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used combination inhaled corticosteroid/LABA therapy?","Over the past 3 months, on average, how many days per week has the participant used combination inhaled corticosteroid/LABA therapy,  &lt; 5 (1=Yes,0=No)","Over the past 3 months, on average, how many days per week has the participant had daytime asthma symptoms?","Over the past 3 months, on average, how many days per week has the participant had daytime asthma symptoms,  &lt;=2? 1=Yes,0=No","Over the past 3 months, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms?","Over the past 3 months, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms, &lt;=2? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used his/her short acting betaagonist  for relief of symptoms?","Over the past 3 months, on average, how many days per week has the participant used his/her short acting betaagonist  for relief of symptoms, &lt; 2? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 80Pct? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 70Pct? 1=Yes,0=No","Does the participant have current evidence of any of the conditions listed on the Exclusionary Medical Conditions for SIENA  reference card, or any chronic diseases that would prevent participation in the trial or put the participant at risk by participati","Is the participant currently taking any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Is the participant able to go off these medications for the required washout period prior to Visit 1 and for the duration of the study? 1=Yes,0=No","Is the participant able to use the spirotel e-diary/PEF meter correctly, as evidenced by achieving a score of 13 on the spirotel Performance Checklist ? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Has the participant or parent/legal guardian signed the SIENA Informed Consent document?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past month, on average, how many days per week has the participant had daytime symptoms?","In the past month, on average, how many days per week has the participant had daytime symptoms, &lt;=2? 1=Yes,0=No","In the past month, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms?","In the past month, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms,  &lt;=2? 1=Yes,0=No","In the past month, on average, how many days per week has the participant used his/her short acting beta-agonist  for relief of symptoms?","In the past month, on average, how many days per week has the participant used his/her short acting beta-agonist  for relief of symptoms, &lt; 2? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 80Pct? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 70Pct? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant or parent/legal guardian sign the SIENA Informed Consent document? 1=Yes,0=No","Ages 12-17 Only: Has the participant signed and dated the assent form or, if the participant is less than the local age of assent, has the participant given verbal assent? 1=Yes,0=No","Is the participant 12 years of age, or older? 1=Yes,0=No","Does the participant plan to move away from the clinical site in the upcoming 11 months such that his/her ability to complete the study will be jeopardized? 1=Yes,0=No","Has the participant used investigative drugs and/or enrolled in an intervention trial in the past 30 days, or have plans to enroll in such a trial during the SIENA study? 1=Yes,0=No","Has the participant had a respiratory infection within the past 6 weeks? 1=Yes,0=No","Has the participant had a respiratory infection within the past 4 weeks? 1=Yes,0=No","During the past 4 weeks, on average, how many days per week has the participant had daytime asthma symptoms?","During the past 4 weeks, on average, how many days per week has the participant had daytime asthma symptoms  &gt; 2? 1=Yes,0=No","During the past 4 weeks, how many nighttime awakenings due to asthma symptoms has the participant had?","During the past 4 weeks, how many nighttime awakenings due to asthma symptoms has the participant had,  &gt; 2? 1=Yes,0=No","During the past 4 weeks, on average, how many days per week has the participant used his/her short-acting beta-agonist  for relief of symptoms?","During the past 4 weeks, on average, how many days per week has the participant used his/her short-acting beta-agonist  for relief of symptoms, &gt; 2? 1=Yes,0=No","Is Q7a, Q8a, or Q9a checked YES? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Did the participant or parent/legal guardian sign the SIENA Informed Consent document?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Does the participant have current evidence of any of the conditions listed on the Exclusionary Medical Conditions for SIENA  reference card, or any chronic diseases  that would prevent participation in the trial or put the participant at risk by participat","Does the participant have a history of bladder-neck obstruction, urinary retention, benign prostatic hyperplasia , or a clinically relevant urologic disorder that precludes study participation? 1=Yes,0=No","Does the participant have a history of narrow angle glaucoma? 1=Yes,0=No","Does the participant have a history of significant cardiovascular disorders or arrhythmias? 1=Yes,0=No","Has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card within the specified time periods? 1=Yes,0=No","Is the participant currently taking prescription or OTC medication other than those listed on the Allowed Medications  reference card? 1=Yes,0=No","Based on input from the participant and the study physician, will the participant need to use intranasal steroids at any time during the study? 1=Yes,0=No","Is the participant willing to use a single intranasal steroid at a stable dose continuously for the duration of the study, starting at Visit 1? 1=Yes,0=No","Is the participant currently receiving allergen immunotherapy  other than an established maintenance regimen implemented continuously for a minimum of 3 months? 1=Yes,0=No","Has the participant taken omalizumab within the past 3 months? 1=Yes,0=No","Has the participant used any smokeless tobacco products  in the past year? 1=Yes,0=No","Has the participant smoked cigarettes, a pipe, cigar, marijuana, electronic cigarettes, or any other substance in the past year? 1=Yes,0=No","Ages 18+ Only: Does the participant have a smoking history of greater than 10 pack-years? 1=Yes,0=No","Ages 12-17 Only: Does the participant have a smoking history of greater than 5 pack-years? 1=Yes,0=No","Has the participant received a physician diagnosis of asthma at least 12 months ago? 1=Yes,0=No","Has the participant received a physician diagnosis of asthma at least 12 months ago or had a history consistent with asthma for the previous 12 months? 1=Yes,0=No","Has the participant experienced a life-threatening asthma exacerbation requiring treatment with intubation or mechanical ventilation in the past 5 years? 1=Yes,0=No","Has the participant had an asthma exacerbation requiring systemic corticosteroid treatment in the past 6 weeks? 1=Yes,0=No","Has the participant used an oral or inhaled corticosteroid, or leukotriene modifier in the past 6 weeks? 1=Yes,0=No","Has the participant used an oral corticosteroid in the past 6 weeks? 1=Yes,0=No","Has the participant used an inhaled corticosteroid or leukotriene modifier in the past 3 weeks? 1=Yes,0=No","Is the participant potentially able to bear children? 1=Yes, 0=No, 9=N/A","Is the participant currently pregnant or lactating? 1=Yes,0=No","Does the participant agree to use one of the approved methods indicated on the Birth Control Methods  reference card for the duration of the study? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was the participants prebronchodilator FEV1 &gt;=70Pct predicted? 1=Yes,0=No","Did the participants FEV1 improve &gt;=12Pct and &gt;=200 ml in response to four puffs of albuterol? 1=Yes,0=No","Does the participant have valid source documentation within the past 6 months for an overread AsthmaNet methacholine challenge  with a PC20 &lt;=16 mg/ml? 1=Yes,0=No","Did the participants FEV1 improve &gt;=12Pct and &gt;=200 ml in response to four puffs of albuterol?","Was the participants methacholine PC20 &lt;=16 mg/ml? 1=Yes,0=No","Is the participant able to use the spirotel e-diary/PEF meter correctly, as evidenced by achieving a score of 13 on the Spirotel Performance Checklist ? 1=Yes, 0=No, 9=N/A","Has the participant used the Twisthaler previously? 1=Yes,0=No","Is the participant able to use a Twisthaler properly, as evidenced by achieving a score of 12 on the Twisthaler Inhalation Technique Checklist ? 1=Yes,0=No","Is the participant able to use a metered dose inhaler properly, as evidenced by achieving a score of 11 on the MDI Inhalation Technique Checklist  ? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Does the participant have valid source documentation within the past 6 months for an overread AsthmaNet methacholine challenge  with a PC20 &lt;=16 mg/ml?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Since Visit 1, has the participant experienced a treatment failure event as defined in the protocol? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Since Visit 1, has the participant experienced one or more asthma exacerbations as defined in the protocol? 1=Yes,0=No","Since Visit 1, has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant provide two acceptable sputum induction samples during the run-in? 1=Yes,0=No","Since Visit 1, has the participant experienced a treatment failure event as defined in the protocol? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Did the participant complete Asmanex220 mcg treatment at least 6 weeks ago? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Since Visit 1, has the participant experienced one or more asthma exacerbations as defined in the protocol? 1=Yes,0=No","Since Visit 1, has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Has the participant experienced a respiratory tract infection in the past 6 weeks? 1=Yes,0=No","Has the participant experienced a respiratory tract infection in the past 4 weeks? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report daytime asthma symptoms daily during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report night-time awakenings due to asthma symptoms more than once per week during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report night-time awakenings due to asthma symptoms more than twice per week during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report daily albuterol use for asthma symptom control during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant record at least 75Pct of AM and PM peak flow measurements and symptoms on his or her spirotel during the run-in? 1=Yes,0=No","Did the participant take at least 75Pct of the required puffs from his or her Respimat during the run-in? 1=Yes,0=No","Was the participants prebronchodilator FEV1 &gt;=70Pct predicted? 1=Yes,0=No","Has the participant used the Respimat previously? 1=Yes,0=No","Is the participant able to use a Respimat inhaler properly, as evidenced by achieving a score of 15 on the Respimat Inhalation Technique Checklist? 1=Yes,0=No","Does the participant wish to withdraw consent from the study? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has QC procedure been performed on the NIOX MINO today? 1=Yes,0=No","Did the participant eat or drink within the past hour? 1=Yes,0=No","Did the participant take part in strenuous activity/exercise within the past hour? 1=Yes,0=No","Time eNO started (based on a 24-hour clock) (0000-2359)","ENO Measurement","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","How long have you lived in the current house? (Estimate if uncertain.) (1010) years","How long have you lived in the current house? (Estimate if uncertain.)(1020) months","Does your house use a wood burning stove as a primary source of heat? 1=Yes, 0=No, 8=Don't Know","Does your house use an air conditioner? 1=Yes, 0=No, 8=Don't Know","Does your house use an evaporative cooler ? 1=Yes, 0=No, 8=Don't Know","Does your house use a humidifier? 1=Yes, 0=No, 8=Don't Know","Does your house use a dehumidifier? 1=Yes, 0=No, 8=Don't Know","Has there been water damage to your house, basement, or its contents during the past 12 months? 1=Yes, 0=No, 8=Don't Know","Has there been any mold or mildew, on any surfaces, inside your house in the past 12 months? 1=Yes, 0=No, 8=Don't Know","Which rooms have or have had mold or mildew?10a.Bathroom(s)  (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew?10b.Basement or attic (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10c.Kitchen (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10d.Your bedroom (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10e.Other bedrooms (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10f.Living or family room (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10g.Other (1 = Yes, 0 = No)","Do you ever see cockroaches in your house? 1=Yes,0=No","In which room(s) have you seen cockroaches?12a.Kitchen (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches?  12b.Basement or attic (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12c.Bathroom(s) (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12d.Living or family room (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12e.Your bedroom  (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12f.Other bedrooms (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12g.Garage (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12h.Other (1 = Yes, 0 = No)","Do you ever see rodents  or rodent droppings in your house? 1=Yes,0=No","14.In which room(s) have you seen rodents or rodent droppings? 14a. Kitchen (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14b. Basement or attic (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14c. Bathroom(s) (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings?14d. Living or family room (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14e. Your bedroom(1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14f. Other bedrooms (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14g. Garage (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14h. Other (1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15a. Barns(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property?15b. Hay(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15c. Woodsheds(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15d. Firewood(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property?15e. Chicken coops (1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15f. Corral (1 = Yes, 0 = No)","CHARACTERISTICS OF THE PARTICIPANT'S BEDROOM16.What is the floor covering in your bedroom? 1=Rug/carpet, 2=Vinyl tile or linoleum, 3=Wood, 4=Ceramic tile, 5=Other , 9=Don't know","What type of mattress is on your bed? 1=None, 2=Inner spring mattress, 3=Foam mattress, 4=Waterbed, 5=Air mattress, 6=Other , 9=Don't know","Is the mattress completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","Does your bed have a box spring? 1=Yes,0=No","Is the box spring completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","What type of pillow do you usually sleep with? 1=None, 2=Feather/down, 3=Foam/Dacron/synthetic, 5=Other , 9=Don't know","Is the pillow completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","Does your household have any pets? 1=Yes,0=No","Enter the number of pets that the household has. (Enter '00' if none. If none to Q24a - Q24d, skip to the next question.)24a.Cat","Cat  (1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has. Dog","Dog(1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has 24c.Rabbit, guinea pig, hamster, gerbil, or mouse","Rabbit, guinea pig, hamster, gerbil, or mouse (1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has 24d.Bird","Bird (1=Indoor, 2=Outdoor, 3=Both)","In general, and on a regular basis, are you exposed to any of the following animals?  25a.Cat(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25b.Dog(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25c.Rabbit, guinea pig, hamster, gerbil, or mouse(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25d.Bird (1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25e.Farm animals (1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals?  25f.Other (1 = Yes, 0 = No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other (specify","Highest education of any household member (1=No High School diploma/GED, 2=High Schol grad, 3=Technical training, 4=Some college, no degree, 5=Assoc degree, 6=Bachelor degree, 7=Masters degree/MD/PhD/JD/PharmD, 9=Decline to answer, 10=Don't know)","Category best describes the combined annual income, before taxes, of all members of your household for the last year. (1=Less than $25,000, 2=$25,000 - $49,999, 3=$50,000 - $99,999, 4=$100,000 or more, 9=Decline to answer, 10=Don't know)","How many people  are supported by this income reported in Q3?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Total IgE","Cat (E1)","Dog (E5)","Mouse (E72)","Rat (E74)","Mold mix (MX1)","Cockroach, German (I6)","Grass mix (GX2)","Tree mix (TX4)","Tree mix (TX6)","Weed mix (WX1)","Weed (W3)","Mite (D2)","Mite (D1)","Number or allergen specific (Cat, Dog, Mouse, etc) that were either POS or greater than 0.35","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","CBC with differential cell count: Eosinophils (absolute count) (0-9999/uL)","CBC with differential cell count: WBC (0-999.9 K/uL)","CBC with differential cell count: Differential Lymphocytes (0-99.9Pct)","CBC with differential cell count: Differential Monocytes (0-99.9Pct)","CBC with differential cell count: Differential Basophils (0-99.9Pct)","CBC with differential cell count: Differential Neutrophils (0-99.9Pct)","CBC with differential cell count: Differential Eosinophils (0-99.9Pct)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","PARENTAL HEIGHT - First study visit only or until both are completed 1.Biological mother's height (complete height or check unknown) __ feet","PARENTAL HEIGHT - First study visit only or until both are completed1.Biological mother's height (complete height or check unknown) __ __ inches","PARENTAL HEIGHT - First study visit only or until both are completed 1.Biological mother's height (complete height or check unknown) (9=Don't Know)","Biological father's height (complete height or check unknown) __ feet","Biological father's height (complete height or check unknown) __ __ inches","Biological father's height (complete height or check unknown) (9=Don't Know)","PARTICIPANT MEASUREMENTS - Complete at all applicable study visits3. What type of height measurement was obtained? 1=Standing height, 2=Length","First measurement __ __ __ . __ cm","Second measurement __ __ __ . __ cm","Third measurement __ __ __ . __ cm","Average height or length measurement __ __ __ . __ cm","In your judgment, was the participant's height or length measurement acceptable? 1=Yes,0=No","Weight (shoes off, light clothing) __ __ __ . __ kg","ORAL CANDIDIASIS5.Does the participant have evidence of oral candidiasis? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Post Diluent FEV1","Did the participant drop20Pct at the diluent stage? 1=Yes,0=No","Last concentration of methacholine administered __ __ . __ __ __ __ mg/ml","FEV1 after last concentration of methacholine administered","Did the participant achieve a PC20? 1=Yes,0=No","PC20 __ __ . __ __ mg/ml","Time methacholine challenge ended (based on 24-hour clock)","FEV1","Time of FEV1 in Q7a (based on 24-hour clock)","Was the FEV1 from Q7athe methacholine reversal reference value  in the gray box above? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was an additional treatment used in the first hour? 1=Yes,0=No","Additional albuterol by MDI (1=Yes,0=No)","Number of additional puffs of albuterol administered (1=2, 2=4, 3=&gt;4)","Nebulized Beta-agonist(1=Yes,0=No)","Subcutaneous epinephrine(1=Yes,0=No)","Implementation of clinic emergency protocol or algorithm (1=Yes,0=No)","Other  (1=Yes,0=No)","Participant's FEV1 after additional treatment within first hour.2a.FEV1","Time of FEV1 in Q2a (based on 24-hour clock)","Was the FEV1 from Q2a &gt;the methacholine reversal reference value  in the gray box on the Methacholine Challenge Testing  form? 1=Yes,0=No","Was additional treatment used after one hour? 1=Yes,0=No","Additional albuterol by MDI (1=Yes,0=No)","Number of additional puffs of albuterol administered (1=2, 2=4, 3=&gt;4)","Nebulized Beta-agonist  (1=Yes,0=No)","Subcutaneous epinephrine(1=Yes,0=No)","Implementation of clinic emergency protocol or algorithm  (1=Yes,0=No)","Treatment in the emergency room  (1=Yes,0=No)","Overnight hospitalization (1=Yes,0=No)","Other (1=Yes,0=No)","Participant's final FEV1 after methacholine challenge4a.FEV1","Time of FEV1 in Q4a (based on 24-hour clock)","Was the FEV1 from Q4a ? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Exclusions and Confounders1.Has the participant had any severe acute illness in the past 4 weeks? 1=Yes,0=No","If YES, has the participant received permission from the supervising physician to proceed with the methacholine challenge testing? 1=Yes,0=No","Has the participant used an oral or injectable steroid  in the last 4 weeks? 1=Yes,0=No","Does the participant have a baseline  FEV1 less than 55Pct of predicted or less than 1.0 L? 1=Yes,0=No","Pregnancy test results:  (Check N/A if the participant is male, or is female and is post-menopausal, had a hysterectomy or tubal ligation.) (1=Positive, 0=Negative, 9=N/A)","Is the participant's systolic blood pressure &gt; 200 mm Hg or diastolic blood pressure &gt; 100 mm Hg? 1=Yes,0=No","Is there any other reason the participant should not proceed with the methacholine challenge testing? 1=Yes,0=No","Is the participant eligible to proceed with the diluent  pulmonary function testing for the methacholine challenge? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Exclusions and Confounders:  1.Has the participant had any severe acute illness in the past 4 weeks? 1=Yes,0=No","If YES, has the participant received permission from the supervising physician to proceed with the methacholine challenge testing? 1=Yes,0=No","During the past 4 weeks, has the participant had any respiratory infections, colds, or bronchitis ? 1=Yes,0=No","If YES, during the past 2 weeks, has the participant had any respiratory infections, colds, or bronchitis ? 1=Yes,0=No","Has the participant used an oral or injectable steroid  in the last 4 weeks? 1=Yes,0=No","Does the participant have a baseline  FEV1 less than 70Pct of predicted? 1=Yes,0=No","Pregnancy test results(Check N/A if the participant is male, or is female and has not started menses.)(1=Positive, 0=Negative, 9=N/A)","If participant's age is ? 1=Yes,0=No","If participant's age is &lt; 12 years: Is the participant's systolic blood pressure &gt; 180 mm Hg or diastolic blood pressure &gt; 90 mm Hg? 1=Yes,0=No","Is there any other reason the participant should not proceed with the methacholine challenge testing? 1=Yes,0=No","Is the participant eligible to proceed with the diluent  pulmonary function testing for the methacholine challenge? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time albuterol administered (based on 24-hour clock)","Time post-albuterol spirometry started (based on 24-hour clock)","Highest FVC","Highest FEV1","Highest FEV1 (Pct predicted)","FEF Max","FEF25-75","In your judgment, was the subject's spirometry technique acceptable? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Please check the box next to the treatment that you believe you received over the past 12 weeks. 1=mometasone, 2=placebo","Blinded Scheduled Twisthaler Contents: How sure are you about your answer to Question 2? 1=Absolutely sure -  I know what the MDI contains, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","Please comment with respect to any observations you made that helped you make your choice in Question 2 . 1=I have no comments, 2=I noticed the following:","Please check the box next to the treatment that you believe you received over the past 12 weeks. 1=tiotropium, 2=placebo","How sure are you about your answer to Question 5? 1=Absolutely sure - I know what the Respimat capsules contain, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","1=I have no comments, 2=I noticed the following:","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Visit Number (numeric)","Level of periostin in peripheral blood (ng/ml)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time ipratropium administered (based on 24-hour clock) (0000-2359)","Time post-ipratropium spirometry started (based on 24-hour clock) (0000-2359)","Highest FVC (0-9.99 L)","Highest FEV1 (0-9.99 L)","Highest FEV1 (Pct predicted) (0-999 Pct predicted)","FEF Max (0-99.99 L/S)","FEF25-75 (0-9.99 L/S)","In your judgment, was the participants spirometry technique acceptable? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Is the participant unable to bear children due to any of the following reasons? 1=Yes,0=No","Post-menopausal (at least one year since last menses) (1=Yes,0=No)","Hysterectomy(1=Yes,0=No)","Tubal ligation(1=Yes,0=No)","Pregnancy test results  (1=Positive, 0=Negative)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Have you had any diseases, illnesses, conditions, or surgeries related Blood, Lymph, or Immune Systems? 1=Yes,0=No","Eyes(1=Yes,0=No)","Breasts (1=Yes,0=No)","Endocrine Systems (1=Yes,0=No)","Heart and Blood Vessels  (1=Yes,0=No)","Liver or Pancreas (1=Yes,0=No)","Kidneys or Urinary Tract System  (1=Yes,0=No)","Reproductive System (1=Yes,0=No)","Muscles or Bones (1=Yes,0=No)","Nervous System (1=Yes,0=No)","Psychiatric(1=Yes,0=No)","Drug Allergies (1=Yes,0=No)","Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Skin (1=Yes,0=No)","Have you ever had allergic rhinitis ? 1=Yes, 0=No, 9=Don't know","Have you ever had nasal polyps? 1=Yes, 0=No, 9=Don't know","Do you have chronic or recurrent sinusitis ? 1=Yes, 0=No, 9=Don't know","Have you ever been diagnosed with vocal cord dysfunction? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the ear, nose, or throat? 1=Yes,0=No","Have you ever had pneumonia? 1=Yes, 0=No, 9=Don't know","If YES, were you diagnosed by chest x-ray? 1=Yes, 0=No, 9=Don't know","If YES, were you treated with antibiotics? 1=Yes, 0=No, 9=Don't know","Have you ever had bronchitis? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the lungs ? 1=Yes, 0=No, 9=Don't know","Do you have gastroesophageal reflux disease ? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the stomach or intestines? 1=Yes,0=No","Have you been diagnosed with sleep disordered breathing ? 1=Yes,0=No","If YES, are you being treated with CPAP or BiPAP? 1=Yes,0=No","Have you ever had other sleep disorders? 1=Yes,0=No","Have you ever had other conditions that have not been mentioned on this form? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Short-acting Inhaled Beta-Agonists by Inhaler (e.g., albuterol, Primatene Mist, Maxair, ProAir, Proventil, Ventolin, Xopenex)  (1=Yes, 0=No, 9=Don't Know)","If YES, indicate average weekly puffs in the past month (Enter '000' if none used) weekly puffs","Rescue treatment via a Nebulizer Machine (e.g., albuterol, ipratropium, Combivent, Xopenex, levalbuterol)  (1=Yes, 0=No, 9=Don't Know)","Long-acting Inhaled Beta-Agonists (e.g., Serevent, Foradil, salmeterol, formoterol) Do not consider combination medications. (1=Yes, 0=No, 9=Don't Know)","Oral Beta-Agonists (e.g., albuterol, Brethine, Bricanyl, metaproterenol, Proventil, Ventolin, Repetabs, Volmax)  (1=Yes, 0=No, 9=Don't Know)","Oral Theophylline (short-acting or sustained release)(e.g., Aminophylline, Slo-Phyllin, Slo-bid, Theo-Dur, Uniphyl)(1=Yes, 0=No, 9=Don't Know)","Inhaled Anticholinergic by Inhaler (e.g., Atrovent, Combivent, Spiriva) (1=Yes, 0=No, 9=Don't Know)","Leukotriene Antagonist / 5LO Inhibitors (e.g., Accolate, Zyflo, Singulair) (1=Yes, 0=No, 9=Don't Know)","IgE Blocker (e.g., Xolair)(1=Yes, 0=No, 9=Don't Know)","Oral Steroids FOR ASTHMA (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron,  dexamethasone)(1=Yes, 0=No, 9=Don't Know)","If YES, in the past 12 months, how many courses of steroids by mouth have you taken FOR ASTHMA? 1=1 course, 2=2 courses, 3=3 courses, 4=4 courses, 5=5 courses, 6=More than 5 courses","Injectable Steroids FOR ASTHMA (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV)(1=Yes, 0=No, 9=Don't Know)","Steroids by Inhaler (e.g., Asmanex Twisthaler, QVAR, Flovent, Pulmicort Flexhaler)(1=Yes, 0=No, 9=Don't Know)","Indicate most recent type of inhaled steroid taken (refer to PRIOR_TRT_CARD reference card) code","Indicate number of daily puffs used daily puffs","Indicate the total number of months that you used the inhaled steroid out of the past 12 months months","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide) (1=Yes, 0=No, 9=Don't Know)","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide). If YES, complete Q13a - Q13c. Indicate most recent type of nebulized steroid taken (refer to PRIOR_TRT_CARD reference card)","Indicate number of daily treatments used daily treatments","Indicate the total number of months that you used the nebulized steroid out of the past 12 months months","Long-Acting Beta-Agonist and Inhaled Steroid Combination Medications (e.g., Advair Diskus, Symbicort MDI, Dulera MDI)(1=Yes, 0=No, 9=Don't Know)","Indicate most recent type of combination medication taken (refer to PRIOR_TRT_CARD reference card) code","Indicate number of daily puffs used daily puffs","Indicate the total number of months that you used the combination medication out of the past 12 months months","Nasal Steroids (e.g., Beconase, Vancenase, Flonase, Nasacort, Nasalide, Nasarel, Omnaris, Rhinocort, Nasonex)(1=Yes, 0=No, 9=Don't Know)","Non-steroidal Anti-allergic Nasal Medications (e.g., Nasalcrom, Astelin, Astepro, ipratropium)(1=Yes, 0=No, 9=Don't Know)","Anti-allergic Oral Medications (e.g., fexofenadine, loratadine, cetirizine, chlorpheniramine) (1=Yes, 0=No, 9=Don't Know)","Topical Steroids - Prescription (e.g., Synalar, Lidex, Dermacin, Fluocinonide) (1=Yes, 0=No, 9=Don't Know)","Topical Steroids - OTC (e.g., Hydrocortisone - multiple strengths and products) (1=Yes, 0=No, 9=Don't Know)","Other Medication FOR ASTHMA OR ALLERGIES (1=Yes, 0=No, 9=Don't Know)","Oral Steroids for Conditions Other Than Asthma (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) (1=Yes, 0=No, 9=Don't Know)","Injectable Steroids for Conditions Other Than Asthma (e.g., Medrol, Solumedrol, Decadron, dexamethasone,  triamcinolone, Kenalog, hydrocortisone IV)(1=Yes, 0=No, 9=Don't Know)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Short-acting Inhaled Beta-Agonists by Inhaler (e.g., albuterol, Primatene Mist, Maxair, ProAir, Proventil, Ventolin, Xopenex) Date","Rescue treatment via a Nebulizer Machine (e.g., albuterol, ipratropium, Combivent, Xopenex, levalbuterol) Date","Long-acting Inhaled Beta-Agonists (e.g., Serevent, Foradil, salmeterol, formoterol) Do not consider combination medications. Date","Oral Beta-Agonists (e.g., albuterol, Brethine, Bricanyl, metaproterenol, Proventil, Ventolin, Repetabs, Volmax) Date","Inhaled Anticholinergic by Inhaler (e.g., Atrovent, Combivent, Spiriva) Date","Leukotriene Antagonist / 5LO Inhibitors (e.g., Accolate, Zyflo, Singulair) Date","IgE Blocker (e.g., Xolair) Date","Oral Steroids FOR ASTHMA (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) Date","Injectable Steroids FOR ASTHMA (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV) Date","Steroids by Inhaler (e.g., Asmanex Twisthaler, QVAR, Flovent, Pulmicort Flexhaler) Date","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide) Date","Long-Acting Beta-Agonist and Inhaled Steroid Combination Medications (e.g., Advair Diskus, Symbicort MDI, Dulera MDI) Date","Nasal Steroids (e.g., Beconase, Vancenase, Flonase, Nasacort, Nasalide, Nasarel, Omnaris, Rhinocort, Nasonex) Date","Non-steroidal Anti-allergic Nasal Medications (e.g., Nasalcrom, Astelin, Astepro, ipratropium) Date","Anti-allergic Oral Medications (e.g., fexofenadine, loratadine, cetirizine, chlorpheniramine) Date","Topical Steroids - Prescription (e.g., Synalar, Lidex, Dermacin, Fluocinonide) Date","Topical Steroids - OTC (e.g., Hydrocortisone - multiple strengths and products) Date","Other Medication FOR ASTHMA OR ALLERGIES Date","Oral Steroids for Conditions Other Than Asthma (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) Date","Injectable Steroids for Conditions Other Than Asthma (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV) Date","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the last month, how often have you been upset because of something that happened unexpectedly? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that you were unable to control the important things in your life? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt nervous and stressed? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt confident about being able to handle your personal problems? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that things were going your way? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you found that you could not cope with all the things that you had to do? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you been able to control irritations in your life? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that you were on top of things? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you been angered because of things that happened that were outside of your control? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that your difficulties were piling up so high that you could not overcome them? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Have you consumed caffeine in the past 4 hours? 1=Yes,0=No","Have you used medications with caffeine in the past 4 hours? 1=Yes,0=No","Have you used any weight loss medications in the past 4 hours? 1=Yes,0=No","Have you consumed any food containing alcohol or beverages containing alcohol in the past 4 hours? 1=Yes,0=No","Have you used a rescue intermediate-acting inhaled betaagonist in the past 6 hours? 1=Yes,0=No","Have you used a short-acting anticholinergic in the past 6 hours? 1=Yes,0=No","Have you used any ophthalmic antihistamines in the past 6 hours? 1=Yes,0=No","Have you used any nasal antihistamines in the past 6 hours? 1=Yes,0=No","Have you used any nasal decongestants in the past 6 hours? 1=Yes,0=No","Have you used any oral antihistamines in the past 48 hours? 1=Yes,0=No","Have you used any oral decongestants or cold remedies in the past 48 hours? 1=Yes,0=No","Have you used any smokeless tobacco products today? 1=Yes,0=No","At this time, is your asthma worse because of recent exposure to triggers? 1=Yes,0=No","Is there any other reason you should not proceed with spirometry testing? 1=Yes,0=No","Is the participant eligible to proceed with the spirometry testing? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past 4 weeks, I worried about the long-term effects of asthma on my health (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I had to worry about asthma triggers (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, my asthma was on my mind (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, it was hard to get a good nights sleep because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt like I couldnt enjoy life because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt that asthma was controlling my life (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt frustrated that I couldnt make plans in advance because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, everyday activities were a struggle (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, asthma placed stress on my relationships with family, friends, significant others, or co-workers (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, I felt frustrated that I have to do things differently than people who dont have asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt like I missed out on doing things with others because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, I had to do a lot of planning to make sure I always had an inhaler ready (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Randomized treatment assignment period 1: (Mometasone, Tiotropium or Placebo)","Randomized treatment assignment period 2: (Mometasone, Tiotropium or Placebo)","Randomized treatment assignment period 3: (Mometasone, Tiotropium or Placebo)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Sex (1=Male, 2=Female)","Primary Racial Identification for Spirometry (1=Amer Indian or Alaskan Native, 2=Asian or Pacific Islander, 3=Black, 4=White, 5=Hispanic or Latino, 6=Other)","Participant's Age at Visit 1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Description of Adverse Event (ICD9 Code)","Is the participant currently taking study drug? 1=Yes,0=No","Time interval between the last administration of the study drug and the Adverse Event","What was the unit of time for the interval in Question #4? 1=Second, 2=Minute, 3=Hour, 4=Day","Fatal event(1=Yes,0=No)","Life-threatening event (1=Yes,0=No)","Inpatient hospitalization required(1=Yes,0=No)","Hospitalization prolonged  (1=Yes,0=No)","Disabling or incapacitating  (1=Yes,0=No)","Overdose (1=Yes,0=No)","Cancer (1=Yes,0=No)","Congenital anomaly(1=Yes,0=No)","Serious laboratory abnormality with clinical symptoms(1=Yes,0=No)","Height failure (per protocol MOP)(1=Yes,0=No)","Pregnancy(1=Yes,0=No)","Other (1=Yes,0=No)","What in your opinion caused the event? 1=Yes,0=No","Withdrawal of study drug(s)(1=Yes,0=No)","Concurrent medication  (1=Yes,0=No)","Other condition or event  (1=Yes,0=No)","Was the event expected or unexpected? 1=Expected, 2=Unexpected","Was the event possibly, probably, or definitely related to study participation? 1=Yes,0=No","Date of Adverse Event","Admission date","Discharge date","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","What type of height measurement was obtained? 1=Standing height, 2=Length","First measurement __ __ __ . __ cm","Second measurement __ __ __ . __ cm","Third measurement __ __ __ . __ cm","Average height or length measurement __ __ __ . __ cm","In your judgment, was the participant's height or length measurement acceptable? 1=Yes,0=No","Weight (shoes off, light clothing) __ __ __ . __ kg","Does the participant have evidence of oral candidiasis? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant fail to respond within 48 hours to the treatment failure rescue algorithm? 1=Yes,0=No","Did the participant use at least 16 puffs as needed albuterol per 24 hours for a period of 48 hours? 1=Yes,0=No","Did the participant experience prebronchodilator FEV1 values &lt; 50Pct of the baseline prebronchodilator value obtained at Visit 1 on two consecutive spirometric determinations made on different days? 1=Yes, 0=No, 9=Not evaluated","Did the participant experience prebronchodilator FEV1 values &lt; 40Pct of predicted on two consecutive spirometric determinations made on different days? 1=Yes, 0=No, 9=Not evaluated","Did the study or treating physician prescribe the participant oral/parenteral corticosteroids for the treatment of his/her asthma? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation in the opinion of the study investigator or personal physician? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation? 1=Yes,0=No","Did the participant seek care for significant asthma exacerbation conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Was the participant hospitalized? 1=Yes,0=No","Duration of hospital stay (0-99.9 days)","Was intubation or ventilation assistance required? 1=Yes,0=No","Was the participant admitted to the intensive care unit? 1=Yes,0=No","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Inhaled corticosteroids (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Nebulized bronchodilator (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Oral corticosteroids (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, IM or IV steroids (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Antibiotics (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Other (1=Yes, 0=No )","Why do you think the participant experienced a significant asthma exacerbation? 1=Common cold, 2=Allergies, 3=Pollution or chemical irritant, 4=Too little asthma maintenance medication, 5=Exercise, 6=Other","Date exacerbation conditions were met","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Which phone contact following Asthma Exacerbation visit is being performed? 1=Phone Contact #1 , 2=Phone Contact #2 , 3=Phone Contact #3","Since the most recent Asthma Exacerbation contact, has the participant sought additional care for significant asthma exacerbation conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Was the participant hospitalized? 1=Yes,0=No","Duration of hospital stay (0-99.9 days)","Was intubation or ventilation assistance required? 1=Yes,0=No","Was the participant admitted to the intensive care unit? 1=Yes,0=No","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Inhaled corticosteroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Nebulized bronchodilator (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Oral corticosteroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, IM or IV steroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Antibiotics (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Runny Nose  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Post nasal drip  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Need to blow your nose  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Facial pain/pressure(0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Nasal obstruction  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time spirometry started (based on 24-hour clock)","The reported FEV1 and FVC are the best volumes of all acceptable maneuvers. 1.Highest FVC","Highest FEV1 (Pct predicted)","The reported flow rates correspond to the maneuver where FEV1 + FVC is maximized. FEF Max","FEF25-75","In your judgment, was the subject's spirometry technique acceptable? 0=No, 1=Yes","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Highest FEV1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","TRIAL_TYPE","FVC","FEV1","PEF","FEF2575","FET","Number of times the participant woke up last night due to asthma symptoms # Times woke up due to asthma (0-9)","Number of puffs the participant will take from the WHITE study inhaler this morning # Puffs from WHITE inhaler in AM (0-9)","only) Number of puffs the participant will take from the BLUE study inhaler this morning # Puffs from BLUE inhaler in AM (0-9)","Has the participant taken any puffs from his/her RED RESCUE albuterol inhaler in the past 4 hours? 1=yes, 0=no","Shortness of Breath Score Shortness of Breath Score overnight (0,1,2,3)","Chest tightness score Chest Tightness Score overnight (0,1,2,3)","Wheezing score Wheezing Score overnight (0,1,2,3)","Coughing score Coughing Score overnight (0,1,2,3)","Phlegm/Mucus score Phlegm/Mucus Score overnight (0,1,2,3)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","date 'fixed' for analysis so that associated AM and PM always have same ddate","TRIAL_TYPE","FVC","FEV1","PEF","FEF2575","FET","Number of puffs the participant will take from the WHITE study inhaler tonight # Puffs from WHITE inhaler in PM (0-9)","Has the participant taken any puffs from his/her RED RESCUE albuterol inhaler during the past 4 hours? 1=yes, 0=no","Shortness of Breath Score Shortness of Breath Score since waking this morning (0,1,2,3)","Chest tightness score Chest Tightness Score since waking this morning (0,1,2,3)","Wheezing score Wheezing Score since waking this morning (0,1,2,3)","Coughing score Coughing Score since waking this morning (0,1,2,3)","Phlegm/Mucus score Phlegm/Mucus Score since waking this morning (0,1,2,3)","Did your regular exercise cause unusually severe shortness of breath during the past 24 hours? 1=Yes, 0=No, 9=N/A","Number of RED RESCUE albuterol puffs taken in the past 24 hours to prevent symptoms# Preventive  RED RESCUE puffs  in past 24 hours?","Number of RED RESCUE albuterol puffs taken for asthma symptoms or low peak flow during past 24 hours  # RED RESCUE  puffs for symptoms  in past 24 hours?","Number of times used RED RESCUE albuterol inhaler for asthma symptoms during past 24 hours  # Times used  RED RESCUE for symptoms in past 24 hours?","Did RED RESCUE albuterol relieve symptoms for less than 4 hours after treatment during past 24 hours? 1=Yes,0=No","Was the participant seen by a healthcare provider  for an unscheduled visit in the past 24 hours due to asthma symptoms? 1=Yes,0=No","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Processing: Time processing started (based on 24-hour clock) __ __ __ __ (0 - 2359)","Processing Sample: Total Cell Count: __ __ __ __. __x 10^4 cells/ml (0 - 9999.9)","Was the participant's sputum sample processed within 4 hours after collection? 1=Yes,0=No","Processing Sample: Processing Date __ __ / __ __ / 20 __ __","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Rate slide's quality:  (1=Very good, 2=Good, 3=Acceptable, 4=Poor but readable, 5=Not readable)","Record the number on the slide(s) that was (were) read (0 - 9)","Record the number on the slide(s) that was (were) read: These are numbers that were assigned to the slides at each site. (0 -9)","Total Cell Count: __ __ __ __ . __ x 10^4 cells/ml (0 - 9999.9)","Differential Cell Counts: Squamous Cells __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Epithelial Cells __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Macrophages __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Neutrophils __ __ . __ Pct (0 -99.9)","Differential Cell Counts: Eosinophils __ __ . __ Pct (0 - 99.9)","Differential Cell Counts: Lymphocytes __ __ . __ Pct (0 -99.9)","Date of Read __ __ / __ __ / 20 __ __","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","For this protocol, what was the duration of sputum induction the first time the participant's sample was processed within 4 hours after collection?","Sputum induction start time (based on 24-hour clock)","Sputum induction stop time (based on 24-hour clock)","Duration of sputum induction collection phase at this visit -- minutes (0 - 99.9)","Was the duration ? 1=Yes,0=No","Volume of sputum sample at this visit -- ml (0.0 - 999.9)","Is the volume adequate for processing? 1=Yes,0=No","Is the sample adequate for laboratory analysis? 1=Yes,0=No","Participant's FEV1 immediately after completion of sputum induction: -- (0.0 -9.99 L)","FEV1 (Pct predicted) (0 - 999)","Time of FEV1 in Q7a (based on 24-hour clock)","Percent difference in FEV1 ((Reference - Q7a) / Reference) X100 (-99.9 - 99.9 Pct)","Did the participant's FEV1 drop &gt; 10Pct from reference FEV1 as indicated in Q7d? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Participant's FEV1 after initial 2 puffs of albuterol: 1a.FEV1 __ . __ __ L (0 -9.99)","Participant's FEV1 after initial 2 puffs of albuterol: 1b. FEV1 (Pct predicted)__ __ __ Pct predicted (0 - 999)","Participant's FEV1 after initial 2 puffs of albuterol: 1c. Time of FEV1 from Q1a (based on 24-hour clock) __ __ __ __ (0 - 2359)","Participant's FEV1 after initial 2 puffs of albuterol: 1d. Was the FEV1 from Q1a &gt;=the sputum induction reversal reference value in the gray box above? 1=Yes,0=No","Participant's FEV1 after 2 additional puffs of albuterol: 2a. FEV1 __ . __ __ L (0 - 9.99)","Participant's FEV1 after 2 additional puffs of albuterol: 2b. FEV1 (Pct predicted) __ __ __ Pct predicted (0 - 999)","Participant's FEV1 after 2 additional puffs of albuterol: 2c. Time of FEV1 from Q2a (based on 24-hour clock)__ __ __ __ (0 - 2359)","Participant's FEV1 after 2 additional puffs of albuterol: 2d.Was the FEV1 from Q2a ? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was the participant's FEV1 after reversal from the methacholine challenge greater yhan 90Pct of the baseline FEV1 (1=Yes,0=No)","If NO, has the participant received permission from the supervising physician to proceed with sputum induction testing? 1=Yes,0=No","Physician's Signature:_________ (1=signature present, NULL=signature missing)","Participant's FEV1 used for assessment of eligibility for sputum induction __ . __ __ L (0 - 9.99)","Participant's FEV1 (Pct predicted) used for assessment of eligibility for sputum induction __ __ __ Pct predicted (O-999)","Was the participant's FEV1  from Q3 ? 1=Yes,0=No","Has the participant used any smokeless tobacco products  today? 1=Yes,0=No","Is there any other reason the participant should not proceed with sputum induction? 1=Yes,0=No","Is the participant eligible for sputum induction? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has the participant completed the study through Visit 9? 1=Yes,0=No","Who initiated termination of the participant? 1=Participant, 2=Clinical Staff","1=no longer interested in participating*, 2=no longer willing to follow protocol*, 3=difficult access to clinic , 4=unable to make visits during clinic hours, 5=moving out of the area, 6=unable to continue due to personal constraints*, 7=unable to continue","Did clinical staff terminate the participant due to pregnancy? 1=Yes, 0=No, 9=N/A","Did clinical staff terminate the participant due to loss to follow-up? 1=Yes,0=No","Did clinical staff terminate the participant due to loss to follow-up, type of contact (1=In-person visit, 2=Phone call)","Did clinical staff terminate the participant due to an asthma-related adverse event? 1=Yes,0=No","Did clinical staff terminate the participant due to a medication-related adverse event? 1=Yes,0=No","Did clinical staff terminate the participant due to an adverse event not related to asthma or medications? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility during the Supervised Washout ? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with medication dosing? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with diary completion? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with visit attendance? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with peak flow monitoring? 1=Yes,0=No","Did clinical staff terminate the participant due to significant asthma exacerbation or two treatment failures during run-in ? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility during the run-in period  for reasons other than compliance or exacerbation/treatment failure or high sputum eosinophils ? 1=Yes,0=No","Did clinical staff terminate the participant due to other reason? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility due to high sputum eosinophils in the run-in ? 1=Yes,0=No","Did clinical staff terminate the participant due to... Indicate the letter corresponding to the primary reason the participant was terminated. (a-n)","Did clinical staff terminate the participant due to loss to follow-up, date of last contact with participant","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant seek care for treatment failure conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact, other than call to notify clinic as instructed in spirotel","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Inhaled corticosteroids (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Nebulized bronchodilator (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started: Oral corticosteroids form.  (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  IM or IV steroids form.  (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started: Antibiotics (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant awaken from asthma three or more times in a two-week period or on two consecutive nights? 1=Yes,0=No","Did the participant use albuterol for relief of symptoms four or more times/day for two or more consecutive days? 1=Yes,0=No","Did albuterol relieve symptoms for less than four hours after treatment? 1=Yes,0=No","Did the participant use albuterol for relief of symptoms daily for seven days, and this use exceeded two times the weekly use of albuterol in the baseline period? 1=Yes,0=No","Did regular exercise cause unusually severe shortness of breath on two or more days during a seven day period? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation? 1=Yes,0=No","Did the participant experience a treatment failure? 1=Yes,0=No","Has the participant experienced two treatment failures in the current treatment period ? 1=Yes,0=No","Does the study physician feel the study treatment contributed directly to the treatment failure, such that it is in the participants best interest to discontinue the current blinded treatment? 1=Yes,0=No","Did the participant begin daily treatment with high-dose ICS for 10 days? 1=Yes,0=No","Did the participant begin daily treatment with high-dose ICS for 10 days? 1=Did not think symptoms werebad enough, 2=Other","Date treatment failure conditions were met","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","During this last treatment period , has the participant experienced a treatment failure or asthma exacerbation? 1=Yes,0=No","Has the participant experienced two or more treatment failures in this last treatment period, or does the study physician feel the study treatment contributed directly to the treatment failure such that it is in the participants best interest to discontinu","As of today, have 2 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","As of today, have 6 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","As of today, have 3 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Are you currently employed ? 1=Yes,0=No","In general, how many hours per week do you usually work?","During the past seven days, how many hours did you miss from work because of problems associated with your asthma?","During the past seven days, how much did asthma affect your productivity while you were working?","Do you currently attend classes in an academic setting ? 1=Yes,0=No","In general, how many hours per week do you usually attend classes?","During the past seven days, how many hours did you miss from class or school because of problems associated with your asthma?","During the past seven days, how much did asthma affect your productivity while in school or attending classes in an academic setting?","During the past seven days, how much did your asthma affect your ability to do your regular daily activities, other than work at a job or attend classes?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","How sick today? 0=Not Sick, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely","Runny Nose:(0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","PLUGGED_NOSE  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SNEEZING  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SORE_THROAT  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SCRATCHY_THROAT  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","COUGH (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","HOARSENESS  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","HEAD_CONGESTION  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","CHEST_CONGESTION  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","FEELING_TIRED (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","THINK_CLEARLY  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SLEEP_WELL  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","BREATH_EASILY  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WALK  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","DAILY_ACTIVITIES  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WORK_OUTSIDE_HOME  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WORK_INSIDE_HOME  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","INTERACT  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","LIVE_LIFE  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","MY_COLD_IS  (1=Very much better, 2=Somewhat better, 3=A little better, 4=The same, 5=A little worse, 6=Somewhat worse, 7=Very much worse)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Date (number of days from Visit 1)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Num<\/th>\n      <th>Variable<\/th>\n      <th>Type<\/th>\n      <th>Label<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":1},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Num","targets":1},{"name":"Variable","targets":2},{"name":"Type","targets":3},{"name":"Label","targets":4}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 </div>
 <div id="adding-dataset-names" class="section level3 hasAnchor">
 <h3>Adding Dataset Names<a href="data-import-from-pdf-sources.html#adding-dataset-names" class="anchor-section" aria-label="Anchor link to header"></a></h3>
@@ -891,8 +900,8 @@ <h3>Adding Dataset Names<a href="data-import-from-pdf-sources.html#adding-datase
 <span id="cb396-6"><a href="data-import-from-pdf-sources.html#cb396-6" tabindex="-1"></a></span>
 <span id="cb396-7"><a href="data-import-from-pdf-sources.html#cb396-7" tabindex="-1"></a><span class="co"># View dataframe</span></span>
 <span id="cb396-8"><a href="data-import-from-pdf-sources.html#cb396-8" tabindex="-1"></a><span class="fu">datatable</span>(dataset_name_df_var1)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-c2cad6a139279b5368ef" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-c2cad6a139279b5368ef">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36"],["                                                                                    10:13 Tuesday, February 11, 2020 1","","","Data Set Name: aaaq.sas7bdat","","Num Variable        Type Len Format Informat Label","   1 AAQ_1000       Num        8 2.    2.    In the past 3 days, how much of the time did your asthma keep you from","                                             doing your usual activities at work, school, or at home? 0=None of the","                                             time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All","                                             of the time","   2 AAQ_1010       Num        8 2.    2.    During the past 3 days, how often have you had asthma symptoms? 0=Not","                                             at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or","                                             more times per day","   3 AAQ_1020       Num        8 2.    2.    During the past 3 days, how often have you used your rescue inhaler or","                                             nebulizer medication ? 0=Not at all, 1=Once per day, 2=2-3 times per day,","                                             3=4-5 times per day, 4=6 or more times per day","   4 AAQ_1030       Num        8 2.    2.    During the past 3 days, how many total times did your asthma symptoms","                                             wake you up from sleep? 0=Not at all, 1=1 time in the last 3 days, 2=2-3","                                             times in the last 3 days, 3=4-5 times in the last 3 days, 4=&gt;=6 times in the","                                             last 3 days","   5 AAQ_1040       Num        8 2.    2.    How would you rate the amount of impairment you have experienced due","                                             to your asthma in the past 3 days? 0=No impairment, 1=Mild impairment,","                                             2=Moderate impairment, 3=Severe impairment, 4=Very severe impairment","   6 AAQ_1050       Num        8 2.    2.    How stressed or frightened were you by your asthma symptoms in the past","                                             3 days? 0=Not at all, 1=Mildly, 2=Moderately, 3=Severely, 4=Very","                                             severely","   7 AAQ_1060       Num        8 2.    2.    Why do you think your asthma was worse in the past 3 days compared to","                                             what is normal for you? 0=I have not been worse over the past 3 days. My","                                             asthma symptoms have been usual., 1=Common cold, 2=Allergies,","                                             3=Pollution or chemical irritant, 4=Too little asth","   8 VNUM_C         Char       3 $3.   $3.   Visit Number (character)","   9 VNUM           Num        8             Visit Number (numeric)","  10 VDATE          Num        8             Number of days from Visit 1 to this visit","  11 RAND_ID        Char       6             Randomized Master ID","  12 ENROLL_TYPE    Char   15                Enrollment Type (Screen Fail, Randomized, Healthy Control)","  13 ENROLL_ORDER Num          8             Enrollment Order Number"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Text<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Text","targets":1}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-596d2d88d07b2bfd00f3" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-596d2d88d07b2bfd00f3">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36"],["                                                                                    10:13 Tuesday, February 11, 2020 1","","","Data Set Name: aaaq.sas7bdat","","Num Variable        Type Len Format Informat Label","   1 AAQ_1000       Num        8 2.    2.    In the past 3 days, how much of the time did your asthma keep you from","                                             doing your usual activities at work, school, or at home? 0=None of the","                                             time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All","                                             of the time","   2 AAQ_1010       Num        8 2.    2.    During the past 3 days, how often have you had asthma symptoms? 0=Not","                                             at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or","                                             more times per day","   3 AAQ_1020       Num        8 2.    2.    During the past 3 days, how often have you used your rescue inhaler or","                                             nebulizer medication ? 0=Not at all, 1=Once per day, 2=2-3 times per day,","                                             3=4-5 times per day, 4=6 or more times per day","   4 AAQ_1030       Num        8 2.    2.    During the past 3 days, how many total times did your asthma symptoms","                                             wake you up from sleep? 0=Not at all, 1=1 time in the last 3 days, 2=2-3","                                             times in the last 3 days, 3=4-5 times in the last 3 days, 4=&gt;=6 times in the","                                             last 3 days","   5 AAQ_1040       Num        8 2.    2.    How would you rate the amount of impairment you have experienced due","                                             to your asthma in the past 3 days? 0=No impairment, 1=Mild impairment,","                                             2=Moderate impairment, 3=Severe impairment, 4=Very severe impairment","   6 AAQ_1050       Num        8 2.    2.    How stressed or frightened were you by your asthma symptoms in the past","                                             3 days? 0=Not at all, 1=Mildly, 2=Moderately, 3=Severely, 4=Very","                                             severely","   7 AAQ_1060       Num        8 2.    2.    Why do you think your asthma was worse in the past 3 days compared to","                                             what is normal for you? 0=I have not been worse over the past 3 days. My","                                             asthma symptoms have been usual., 1=Common cold, 2=Allergies,","                                             3=Pollution or chemical irritant, 4=Too little asth","   8 VNUM_C         Char       3 $3.   $3.   Visit Number (character)","   9 VNUM           Num        8             Visit Number (numeric)","  10 VDATE          Num        8             Number of days from Visit 1 to this visit","  11 RAND_ID        Char       6             Randomized Master ID","  12 ENROLL_TYPE    Char   15                Enrollment Type (Screen Fail, Randomized, Healthy Control)","  13 ENROLL_ORDER Num          8             Enrollment Order Number"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Text<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Text","targets":1}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>Next, we will extract the dataset name using the same approach used in extracting values from the nanoparticle tracking example above and assign the name to a variable. We filter by the string “Data Set Name” because this is the start of the text string in the row where our dataset name is stored and is the same across all of our datasets.</p>
 <div class="sourceCode" id="cb397"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb397-1"><a href="data-import-from-pdf-sources.html#cb397-1" tabindex="-1"></a><span class="co"># Create dataframe</span></span>
 <span id="cb397-2"><a href="data-import-from-pdf-sources.html#cb397-2" tabindex="-1"></a>dataset_name_df_var1 <span class="ot">&lt;-</span> dataset_name_df_var1 <span class="sc">%&gt;%</span></span>
@@ -935,8 +944,8 @@ <h3>Adding Dataset Names<a href="data-import-from-pdf-sources.html#adding-datase
 <span id="cb401-6"><a href="data-import-from-pdf-sources.html#cb401-6" tabindex="-1"></a></span>
 <span id="cb401-7"><a href="data-import-from-pdf-sources.html#cb401-7" tabindex="-1"></a><span class="co"># View data farme</span></span>
 <span id="cb401-8"><a href="data-import-from-pdf-sources.html#cb401-8" tabindex="-1"></a><span class="fu">datatable</span>(dataset_name_var1)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-ef5ff6c82915cb259d5c" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-ef5ff6c82915cb259d5c">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13"],["aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>dataset_name<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"dataset_name","targets":1}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-e46788784fbc5f4263df" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-e46788784fbc5f4263df">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13"],["aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>dataset_name<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"dataset_name","targets":1}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>We now have a dataframe that can be joined with our <code>variables</code> dataframe for the first table. We can apply this approach to each table in our original PDF using a <code>for</code> loop.</p>
 <div class="sourceCode" id="cb402"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb402-1"><a href="data-import-from-pdf-sources.html#cb402-1" tabindex="-1"></a><span class="co"># Make dataframe to store dataset names</span></span>
 <span id="cb402-2"><a href="data-import-from-pdf-sources.html#cb402-2" tabindex="-1"></a>dataset_names <span class="ot">&lt;-</span> <span class="fu">data.frame</span>()</span>
@@ -979,8 +988,8 @@ <h3>Adding Dataset Names<a href="data-import-from-pdf-sources.html#adding-datase
 <span id="cb402-39"><a href="data-import-from-pdf-sources.html#cb402-39" tabindex="-1"></a></span>
 <span id="cb402-40"><a href="data-import-from-pdf-sources.html#cb402-40" tabindex="-1"></a><span class="co"># View </span></span>
 <span id="cb402-41"><a href="data-import-from-pdf-sources.html#cb402-41" tabindex="-1"></a><span class="fu">datatable</span>(dataset_names)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-e27e55a39994c29e92f2" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-e27e55a39994c29e92f2">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80","81","82","83","84","85","86","87","88","89","90","91","92","93","94","95","96","97","98","99","100","101","102","103","104","105","106","107","108","109","110","111","112","113","114","115","116","117","118","119","120","121","122","123","124","125","126","127","128","129","130","131","132","133","134","135","136","137","138","139","140","141","142","143","144","145","146","147","148","149","150","151","152","153","154","155","156","157","158","159","160","161","162","163","164","165","166","167","168","169","170","171","172","173","174","175","176","177","178","179","180","181","182","183","184","185","186","187","188","189","190","191","192","193","194","195","196","197","198","199","200","201","202","203","204","205","206","207","208","209","210","211","212","213","214","215","216","217","218","219","220","221","222","223","224","225","226","227","228","229","230","231","232","233","234","235","236","237","238","239","240","241","242","243","244","245","246","247","248","249","250","251","252","253","254","255","256","257","258","259","260","261","262","263","264","265","266","267","268","269","270","271","272","273","274","275","276","277","278","279","280","281","282","283","284","285","286","287","288","289","290","291","292","293","294","295","296","297","298","299","300","301","302","303","304","305","306","307","308","309","310","311","312","313","314","315","316","317","318","319","320","321","322","323","324","325","326","327","328","329","330","331","332","333","334","335","336","337","338","339","340","341","342","343","344","345","346","347","348","349","350","351","352","353","354","355","356","357","358","359","360","361","362","363","364","365","366","367","368","369","370","371","372","373","374","375","376","377","378","379","380","381","382","383","384","385","386","387","388","389","390","391","392","393","394","395","396","397","398","399","400","401","402","403","404","405","406","407","408","409","410","411","412","413","414","415","416","417","418","419","420","421","422","423","424","425","426","427","428","429","430","431","432","433","434","435","436","437","438","439","440","441","442","443","444","445","446","447","448","449","450","451","452","453","454","455","456","457","458","459","460","461","462","463","464","465","466","467","468","469","470","471","472","473","474","475","476","477","478","479","480","481","482","483","484","485","486","487","488","489","490","491","492","493","494","495","496","497","498","499","500","501","502","503","504","505","506","507","508","509","510","511","512","513","514","515","516","517","518","519","520","521","522","523","524","525","526","527","528","529","530","531","532","533","534","535","536","537","538","539","540","541","542","543","544","545","546","547","548","549","550","551","552","553","554","555","556","557","558","559","560","561","562","563","564","565","566","567","568","569","570","571","572","573","574","575","576","577","578","579","580","581","582","583","584","585","586","587","588","589","590","591","592","593","594","595","596","597","598","599","600","601","602","603","604","605","606","607","608","609","610","611","612","613","614","615","616","617","618","619","620","621","622","623","624","625","626","627","628","629","630","631","632","633","634","635","636","637","638","639","640","641","642","643","644","645","646","647","648","649","650","651","652","653","654","655","656","657","658","659","660","661","662","663","664","665","666","667","668","669","670","671","672","673","674","675","676","677","678","679","680","681","682","683","684","685","686","687","688","689","690","691","692","693","694","695","696","697","698","699","700","701","702","703","704","705","706","707","708","709","710","711","712","713","714","715","716","717","718","719","720","721","722","723","724","725","726","727","728","729","730","731","732","733","734","735","736","737","738","739","740","741","742","743","744","745","746","747","748","749","750","751","752","753","754","755","756","757","758","759","760","761","762","763","764","765","766","767","768","769","770","771","772","773","774","775","776","777","778","779","780","781","782","783","784","785","786","787","788","789","790","791","792","793","794","795","796","797","798","799","800","801","802","803","804","805","806","807","808","809","810","811","812","813","814","815","816","817","818","819","820","821","822","823","824","825","826","827","828","829","830","831","832","833","834","835","836","837","838","839","840","841","842","843","844","845","846","847","848","849","850","851","852","853","854","855","856","857","858","859","860","861","862","863","864","865","866","867","868","869","870","871","872","873","874","875","876","877","878","879","880","881","882","883","884","885","886","887","888","889","890","891","892","893","894","895","896","897","898","899","900","901","902","903","904","905","906","907","908","909","910","911","912","913","914","915","916","917","918","919","920","921","922","923","924","925","926","927","928","929","930","931","932","933","934","935","936","937","938","939","940","941","942","943","944","945","946","947","948","949","950","951","952","953","954","955","956","957","958","959","960","961","962","963","964","965","966","967","968","969","970","971","972","973","974","975","976","977","978","979","980","981","982","983","984","985","986","987","988","989","990","991","992","993","994","995","996","997","998","999","1000","1001","1002","1003","1004","1005","1006","1007","1008","1009","1010","1011","1012","1013","1014","1015","1016","1017","1018","1019","1020","1021","1022","1023","1024","1025","1026","1027","1028","1029","1030","1031","1032","1033","1034","1035","1036","1037","1038","1039","1040","1041","1042","1043","1044","1045","1046","1047","1048","1049","1050","1051","1052","1053","1054","1055","1056","1057","1058","1059","1060","1061","1062","1063","1064","1065","1066","1067","1068","1069","1070","1071","1072","1073","1074","1075","1076","1077","1078","1079","1080","1081","1082","1083","1084","1085","1086","1087","1088","1089","1090","1091","1092","1093","1094","1095","1096","1097","1098","1099","1100","1101","1102","1103","1104","1105","1106","1107","1108","1109","1110","1111","1112","1113","1114","1115","1116","1117","1118","1119","1120","1121","1122","1123","1124","1125","1126","1127","1128","1129","1130","1131","1132","1133","1134","1135","1136","1137","1138","1139","1140","1141","1142","1143","1144","1145","1146","1147","1148","1149","1150","1151","1152","1153","1154","1155","1156","1157","1158","1159","1160","1161","1162","1163","1164","1165","1166","1167","1168","1169","1170","1171","1172","1173","1174","1175","1176","1177","1178","1179","1180","1181","1182","1183","1184","1185","1186","1187","1188","1189","1190"],["aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","bmi.sas7bdat","bmi.sas7bdat","bmi.sas7bdat","bmi.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Data Set Name<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Data Set Name","targets":1}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-d47e6c002911e0fb75d4" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-d47e6c002911e0fb75d4">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80","81","82","83","84","85","86","87","88","89","90","91","92","93","94","95","96","97","98","99","100","101","102","103","104","105","106","107","108","109","110","111","112","113","114","115","116","117","118","119","120","121","122","123","124","125","126","127","128","129","130","131","132","133","134","135","136","137","138","139","140","141","142","143","144","145","146","147","148","149","150","151","152","153","154","155","156","157","158","159","160","161","162","163","164","165","166","167","168","169","170","171","172","173","174","175","176","177","178","179","180","181","182","183","184","185","186","187","188","189","190","191","192","193","194","195","196","197","198","199","200","201","202","203","204","205","206","207","208","209","210","211","212","213","214","215","216","217","218","219","220","221","222","223","224","225","226","227","228","229","230","231","232","233","234","235","236","237","238","239","240","241","242","243","244","245","246","247","248","249","250","251","252","253","254","255","256","257","258","259","260","261","262","263","264","265","266","267","268","269","270","271","272","273","274","275","276","277","278","279","280","281","282","283","284","285","286","287","288","289","290","291","292","293","294","295","296","297","298","299","300","301","302","303","304","305","306","307","308","309","310","311","312","313","314","315","316","317","318","319","320","321","322","323","324","325","326","327","328","329","330","331","332","333","334","335","336","337","338","339","340","341","342","343","344","345","346","347","348","349","350","351","352","353","354","355","356","357","358","359","360","361","362","363","364","365","366","367","368","369","370","371","372","373","374","375","376","377","378","379","380","381","382","383","384","385","386","387","388","389","390","391","392","393","394","395","396","397","398","399","400","401","402","403","404","405","406","407","408","409","410","411","412","413","414","415","416","417","418","419","420","421","422","423","424","425","426","427","428","429","430","431","432","433","434","435","436","437","438","439","440","441","442","443","444","445","446","447","448","449","450","451","452","453","454","455","456","457","458","459","460","461","462","463","464","465","466","467","468","469","470","471","472","473","474","475","476","477","478","479","480","481","482","483","484","485","486","487","488","489","490","491","492","493","494","495","496","497","498","499","500","501","502","503","504","505","506","507","508","509","510","511","512","513","514","515","516","517","518","519","520","521","522","523","524","525","526","527","528","529","530","531","532","533","534","535","536","537","538","539","540","541","542","543","544","545","546","547","548","549","550","551","552","553","554","555","556","557","558","559","560","561","562","563","564","565","566","567","568","569","570","571","572","573","574","575","576","577","578","579","580","581","582","583","584","585","586","587","588","589","590","591","592","593","594","595","596","597","598","599","600","601","602","603","604","605","606","607","608","609","610","611","612","613","614","615","616","617","618","619","620","621","622","623","624","625","626","627","628","629","630","631","632","633","634","635","636","637","638","639","640","641","642","643","644","645","646","647","648","649","650","651","652","653","654","655","656","657","658","659","660","661","662","663","664","665","666","667","668","669","670","671","672","673","674","675","676","677","678","679","680","681","682","683","684","685","686","687","688","689","690","691","692","693","694","695","696","697","698","699","700","701","702","703","704","705","706","707","708","709","710","711","712","713","714","715","716","717","718","719","720","721","722","723","724","725","726","727","728","729","730","731","732","733","734","735","736","737","738","739","740","741","742","743","744","745","746","747","748","749","750","751","752","753","754","755","756","757","758","759","760","761","762","763","764","765","766","767","768","769","770","771","772","773","774","775","776","777","778","779","780","781","782","783","784","785","786","787","788","789","790","791","792","793","794","795","796","797","798","799","800","801","802","803","804","805","806","807","808","809","810","811","812","813","814","815","816","817","818","819","820","821","822","823","824","825","826","827","828","829","830","831","832","833","834","835","836","837","838","839","840","841","842","843","844","845","846","847","848","849","850","851","852","853","854","855","856","857","858","859","860","861","862","863","864","865","866","867","868","869","870","871","872","873","874","875","876","877","878","879","880","881","882","883","884","885","886","887","888","889","890","891","892","893","894","895","896","897","898","899","900","901","902","903","904","905","906","907","908","909","910","911","912","913","914","915","916","917","918","919","920","921","922","923","924","925","926","927","928","929","930","931","932","933","934","935","936","937","938","939","940","941","942","943","944","945","946","947","948","949","950","951","952","953","954","955","956","957","958","959","960","961","962","963","964","965","966","967","968","969","970","971","972","973","974","975","976","977","978","979","980","981","982","983","984","985","986","987","988","989","990","991","992","993","994","995","996","997","998","999","1000","1001","1002","1003","1004","1005","1006","1007","1008","1009","1010","1011","1012","1013","1014","1015","1016","1017","1018","1019","1020","1021","1022","1023","1024","1025","1026","1027","1028","1029","1030","1031","1032","1033","1034","1035","1036","1037","1038","1039","1040","1041","1042","1043","1044","1045","1046","1047","1048","1049","1050","1051","1052","1053","1054","1055","1056","1057","1058","1059","1060","1061","1062","1063","1064","1065","1066","1067","1068","1069","1070","1071","1072","1073","1074","1075","1076","1077","1078","1079","1080","1081","1082","1083","1084","1085","1086","1087","1088","1089","1090","1091","1092","1093","1094","1095","1096","1097","1098","1099","1100","1101","1102","1103","1104","1105","1106","1107","1108","1109","1110","1111","1112","1113","1114","1115","1116","1117","1118","1119","1120","1121","1122","1123","1124","1125","1126","1127","1128","1129","1130","1131","1132","1133","1134","1135","1136","1137","1138","1139","1140","1141","1142","1143","1144","1145","1146","1147","1148","1149","1150","1151","1152","1153","1154","1155","1156","1157","1158","1159","1160","1161","1162","1163","1164","1165","1166","1167","1168","1169","1170","1171","1172","1173","1174","1175","1176","1177","1178","1179","1180","1181","1182","1183","1184","1185","1186","1187","1188","1189","1190"],["aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","bmi.sas7bdat","bmi.sas7bdat","bmi.sas7bdat","bmi.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Data Set Name<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Data Set Name","targets":1}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 </div>
 <div id="combining-dataset-names-and-variable-information" class="section level3 hasAnchor">
 <h3>Combining Dataset Names and Variable Information<a href="data-import-from-pdf-sources.html#combining-dataset-names-and-variable-information" class="anchor-section" aria-label="Anchor link to header"></a></h3>
@@ -991,8 +1000,8 @@ <h3>Combining Dataset Names and Variable Information<a href="data-import-from-pd
 <span id="cb403-4"><a href="data-import-from-pdf-sources.html#cb403-4" tabindex="-1"></a>  <span class="fu">clean_names</span>()</span>
 <span id="cb403-5"><a href="data-import-from-pdf-sources.html#cb403-5" tabindex="-1"></a></span>
 <span id="cb403-6"><a href="data-import-from-pdf-sources.html#cb403-6" tabindex="-1"></a><span class="fu">datatable</span>(final_variable_df)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-d0560e119fd7663c6719" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-d0560e119fd7663c6719">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80","81","82","83","84","85","86","87","88","89","90","91","92","93","94","95","96","97","98","99","100","101","102","103","104","105","106","107","108","109","110","111","112","113","114","115","116","117","118","119","120","121","122","123","124","125","126","127","128","129","130","131","132","133","134","135","136","137","138","139","140","141","142","143","144","145","146","147","148","149","150","151","152","153","154","155","156","157","158","159","160","161","162","163","164","165","166","167","168","169","170","171","172","173","174","175","176","177","178","179","180","181","182","183","184","185","186","187","188","189","190","191","192","193","194","195","196","197","198","199","200","201","202","203","204","205","206","207","208","209","210","211","212","213","214","215","216","217","218","219","220","221","222","223","224","225","226","227","228","229","230","231","232","233","234","235","236","237","238","239","240","241","242","243","244","245","246","247","248","249","250","251","252","253","254","255","256","257","258","259","260","261","262","263","264","265","266","267","268","269","270","271","272","273","274","275","276","277","278","279","280","281","282","283","284","285","286","287","288","289","290","291","292","293","294","295","296","297","298","299","300","301","302","303","304","305","306","307","308","309","310","311","312","313","314","315","316","317","318","319","320","321","322","323","324","325","326","327","328","329","330","331","332","333","334","335","336","337","338","339","340","341","342","343","344","345","346","347","348","349","350","351","352","353","354","355","356","357","358","359","360","361","362","363","364","365","366","367","368","369","370","371","372","373","374","375","376","377","378","379","380","381","382","383","384","385","386","387","388","389","390","391","392","393","394","395","396","397","398","399","400","401","402","403","404","405","406","407","408","409","410","411","412","413","414","415","416","417","418","419","420","421","422","423","424","425","426","427","428","429","430","431","432","433","434","435","436","437","438","439","440","441","442","443","444","445","446","447","448","449","450","451","452","453","454","455","456","457","458","459","460","461","462","463","464","465","466","467","468","469","470","471","472","473","474","475","476","477","478","479","480","481","482","483","484","485","486","487","488","489","490","491","492","493","494","495","496","497","498","499","500","501","502","503","504","505","506","507","508","509","510","511","512","513","514","515","516","517","518","519","520","521","522","523","524","525","526","527","528","529","530","531","532","533","534","535","536","537","538","539","540","541","542","543","544","545","546","547","548","549","550","551","552","553","554","555","556","557","558","559","560","561","562","563","564","565","566","567","568","569","570","571","572","573","574","575","576","577","578","579","580","581","582","583","584","585","586","587","588","589","590","591","592","593","594","595","596","597","598","599","600","601","602","603","604","605","606","607","608","609","610","611","612","613","614","615","616","617","618","619","620","621","622","623","624","625","626","627","628","629","630","631","632","633","634","635","636","637","638","639","640","641","642","643","644","645","646","647","648","649","650","651","652","653","654","655","656","657","658","659","660","661","662","663","664","665","666","667","668","669","670","671","672","673","674","675","676","677","678","679","680","681","682","683","684","685","686","687","688","689","690","691","692","693","694","695","696","697","698","699","700","701","702","703","704","705","706","707","708","709","710","711","712","713","714","715","716","717","718","719","720","721","722","723","724","725","726","727","728","729","730","731","732","733","734","735","736","737","738","739","740","741","742","743","744","745","746","747","748","749","750","751","752","753","754","755","756","757","758","759","760","761","762","763","764","765","766","767","768","769","770","771","772","773","774","775","776","777","778","779","780","781","782","783","784","785","786","787","788","789","790","791","792","793","794","795","796","797","798","799","800","801","802","803","804","805","806","807","808","809","810","811","812","813","814","815","816","817","818","819","820","821","822","823","824","825","826","827","828","829","830","831","832","833","834","835","836","837","838","839","840","841","842","843","844","845","846","847","848","849","850","851","852","853","854","855","856","857","858","859","860","861","862","863","864","865","866","867","868","869","870","871","872","873","874","875","876","877","878","879","880","881","882","883","884","885","886","887","888","889","890","891","892","893","894","895","896","897","898","899","900","901","902","903","904","905","906","907","908","909","910","911","912","913","914","915","916","917","918","919","920","921","922","923","924","925","926","927","928","929","930","931","932","933","934","935","936","937","938","939","940","941","942","943","944","945","946","947","948","949","950","951","952","953","954","955","956","957","958","959","960","961","962","963","964","965","966","967","968","969","970","971","972","973","974","975","976","977","978","979","980","981","982","983","984","985","986","987","988","989","990","991","992","993","994","995","996","997","998","999","1000","1001","1002","1003","1004","1005","1006","1007","1008","1009","1010","1011","1012","1013","1014","1015","1016","1017","1018","1019","1020","1021","1022","1023","1024","1025","1026","1027","1028","1029","1030","1031","1032","1033","1034","1035","1036","1037","1038","1039","1040","1041","1042","1043","1044","1045","1046","1047","1048","1049","1050","1051","1052","1053","1054","1055","1056","1057","1058","1059","1060","1061","1062","1063","1064","1065","1066","1067","1068","1069","1070","1071","1072","1073","1074","1075","1076","1077","1078","1079","1080","1081","1082","1083","1084","1085","1086","1087","1088","1089","1090","1091","1092","1093","1094","1095","1096","1097","1098","1099","1100","1101","1102","1103","1104","1105","1106","1107","1108","1109","1110","1111","1112","1113","1114","1115","1116","1117","1118","1119","1120","1121","1122","1123","1124","1125","1126","1127","1128","1129","1130","1131","1132","1133","1134","1135","1136","1137","1138","1139","1140","1141","1142","1143","1144","1145","1146","1147","1148","1149","1150","1151","1152","1153","1154","1155","1156","1157","1158","1159","1160","1161","1162","1163","1164","1165","1166","1167","1168","1169","1170","1171","1172","1173","1174","1175","1176","1177","1178","1179","1180","1181","1182","1183","1184","1185","1186","1187","1188","1189","1190"],["aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","bmi.sas7bdat","bmi.sas7bdat","bmi.sas7bdat","bmi.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat"],[1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,1,2,3,4,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,1,2,3,4,5,6,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28],["AAQ_1000","AAQ_1010","AAQ_1020","AAQ_1030","AAQ_1040","AAQ_1050","AAQ_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ABP_1000","ABP_1010","ABP_1020","ABP_1030","ABP_1040","ABP_1050","ABP_1060","ABP_1070","ABP_1080","ABP_1090","ABP_1100","ABP_1110","ABP_1120","ABP_1130","ABP_1140","ABP_1150","ABP_1160","ABP_1170","ABP_1180","ABP_1190","ABP_1200","ABP_1210","ABP_1220","ABP_1230","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ACT_1","ACT_2","ACT_3","ACT_4","ACT_5","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ICD9_CAT","AEC_1000","AEC_1010","AEC_1040","AEC_1050","AEC_1060","AEC_1070","AEC_1080","AEC_1090","AEC_1100","AEC_1110","AEC_1120","AEC_1020","AEC_1030","VNUM_C","VNUM","ICD9LONG","ICD9SHRT","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","AHA_1000","AHA_1020","AHA_1030","AHA_1040","AHA_1050","AHA_1060","AHA_1070","AHA_1090","AHA_1100","AHA_1110","AHA_1120","AHA_1130","AHA_1140","AHA_1150","AHA_1160","AHA_1170","AHA_1180","AHA_1190","AHA_1200","AHA_1210","AHA_1220","AHA_1250","AHA_1260","AHA_1270","AHA_1280","AHA_1290","AHA_1300","AHA_1310","AHA_1320","AHA_1330","AHA_1340","AHA_1350","AHA_1360","AHA_1370","AHA_1380","AHA_1390","AHA_1400","AHA_1410","AHA_1420","AHA_1430","AHA_1440","AHA_1450","AHA_1460","AHA_1470","AHA_1500","AHA_1570","AHA_1580","AHA_1590","AHA_1600","AHA_1730","AHA_1740","AHA_1750","AHA_1760","AHA_1770","AHA_1780","AHA_1790","AHA_1800","AHA_1810","AHA_1820","AHA_1830","AHA_1840","AHA_1850","AHA_1890","AHA_1900","VNUM_C","VNUM","VDATE","DATE_VAPE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ASU_1000","ASU_1010","ASU_1020","ASU_1030","ASU_1040","ASU_1050","ASU_1060","ASU_1070","ASU_1080","ASU_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","BAS_1000","BAS_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","BMI","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAGE","CHG_1000","CHG_1010","CHG_1020","CHG_1030","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","CME_1040","CME_1050","CME_1055","CME_1000","CME_1010","CME_1020","CME_1030","CME_1080","CME_1090","CME_1060","CME_1070","VNUM_C","VNUM","GEN_NAME","CLASS","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","COM_1000","COM_1010","COM_1020","COM_1030","COM_1040","COM_1050","COM_1060","COM_1070","COM_1080","COM_1090","COM_1100","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","CTX_1000","CTX_1010","CTX_1030","CTX_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E0A_1000","E0A_1020","E0A_1030","E0A_1040","E0A_1043","E0A_1045","E0A_1050","E0A_1060","E0A_1070","E0A_1073","E0A_1075","E0A_1080","E0A_1090","E0A_1100","E0A_1110","E0A_1120","E0A_1130","E0A_1140","E0A_1145","E0A_1150","E0A_1160","E0A_1170","E0A_1180","E0A_1190","E0A_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E0B_1000","E0B_1010","E0B_1020","E0B_1030","E0B_1040","E0B_1050","E0B_1060","E0B_1065","E0B_1070","E0B_1080","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E1_1000","E1_1020","E1_1030","E1_1040","E1_1050","E1_1060","E1_1065","E1_1070","E1_1080","E1_1090","E1_1100","E1_1110","E1_1120","E1_1130","E1_1140","E1_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E2_1000","E2_1010","E2_1020","E2_1030","E2_1040","E2_1050","E2_1060","E2_1070","E2_1080","E2_1090","E2_1100","E2_1110","E2_1120","E2_1130","E2_1140","E2_1145","E2_1150","E2_1160","E2_1170","E2_1173","E2_1175","E2_1180","E2_1190","E2_1200","E2_1210","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E3_1000","E3_1010","E3_1020","E3_1030","E3_1070","E3_1080","E3_1090","E3_1100","E3_1110","E3_1120","E3_1130","E3_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E4_1000","E4_1010","E4_1020","E4_1030","E4_1050","E4_1060","E4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E5_1000","E5_1010","E5_1020","E5_1030","E5_1035","E5_1040","E5_1050","E5_1060","E5_1065","E5_1070","E5_1080","E5_1085","E5_1090","E5_1100","E5_1110","E5_1120","E5_1130","E5_1140","E5_1150","E5_1160","E5_1170","E5_1180","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ENO_1000","ENO_1010","ENO_1020","ENO_1040","ENO_1050","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","HEQ_1000","HEQ_1010","HEQ_1020","HEQ_1030","HEQ_1040","HEQ_1050","HEQ_1060","HEQ_1070","HEQ_1080","HEQ_1090","HEQ_1100","HEQ_1110","HEQ_1120","HEQ_1130","HEQ_1140","HEQ_1150","HEQ_1160","HEQ_1170","HEQ_1180","HEQ_1190","HEQ_1200","HEQ_1210","HEQ_1220","HEQ_1230","HEQ_1240","HEQ_1250","HEQ_1260","HEQ_1270","HEQ_1280","HEQ_1290","HEQ_1300","HEQ_1310","HEQ_1320","HEQ_1330","HEQ_1340","HEQ_1350","HEQ_1360","HEQ_1370","HEQ_1380","HEQ_1390","HEQ_1400","HEQ_1410","HEQ_1420","HEQ_1430","HEQ_1440","HEQ_1450","HEQ_1460","HEQ_1470","HEQ_1480","HEQ_1490","HEQ_1500","HEQ_1510","HEQ_1520","HEQ_1530","HEQ_1540","HEQ_1550","HEQ_1560","HEQ_1570","HEQ_1580","HEQ_1590","HEQ_1600","HEQ_1610","HEQ_1620","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SEI_1000","SEI_1010","SEI_1020","SEI_1030","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","IGE","RE1","RE5","RE72","RE74","RMX1","RI6","RGX2","RTX4","RTX6","RWX1","RW3","RD2","RD1","N_ALLERGENPOS","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","LAB_1000","LAB_1010","LAB_1020","LAB_1030","LAB_1040","LAB_1050","LAB_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PLE_1000","PLE_1010","PLE_1020","PLE_1030","PLE_1040","PLE_1050","PLE_1060","PLE_1070","PLE_1080","PLE_1090","PLE_1100","PLE_1110","PLE_1130","PLE_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MTH_1000","MTH_1010","MTH_1020","MTH_1030","MTH_1040","MTH_1050","MTH_1060","MTH_1070","MTH_1080","MTH_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MAD_1000","MAD_1010","MAD_1020","MAD_1030","MAD_1040","MAD_1050","MAD_1060","MAD_1070","MAD_1090","MAD_1100","MAD_1110","MAD_1120","MAD_1130","MAD_1140","MAD_1150","MAD_1160","MAD_1170","MAD_1180","MAD_1190","MAD_1200","MAD_1220","MAD_1230","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MCA_1000","MCA_1010","MCA_1050","MCA_1060","MCA_1070","MCA_1080","MCA_1100","MCA_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MCP_1000","MCP_1010","MCP_1030","MCP_1040","MCP_1050","MCP_1060","MCP_1070","MCP_1080","MCP_1090","MCP_1100","MCP_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PA4_1000","PA4_1010","PA4_1020","PA4_1030","PA4_1040","PA4_1050","PA4_1060","PA4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PTX_1000","PTX_1010","PTX_1020","PTX_1030","PTX_1040","PTX_1050","PTX_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","VNUM","PERIOSTIN","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PI4_1000","PI4_1010","PI4_1020","PI4_1030","PI4_1040","PI4_1050","PI4_1060","PI4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PRG_1000","PRG_1010","PRG_1020","PRG_1030","PRG_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAD_1000","PAD_1010","PAD_1020","PAD_1030","PAD_1040","PAD_1050","PAD_1060","PAD_1070","PAD_1080","PAD_1090","PAD_1100","PAD_1110","PAD_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAL_1000","PAL_1010","PAL_1020","PAL_1030","PAL_1040","PAL_1050","PAL_1060","PAL_1070","PAL_1080","PAL_1090","PAL_1100","PAL_1110","PAL_1120","PAL_1130","PAL_1150","PAL_1160","PAL_1170","PAL_1180","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PTR_1000","PTR_1010","PTR_1050","PTR_1060","PTR_1100","PTR_1140","PTR_1180","PTR_1220","PTR_1260","PTR_1300","PTR_1340","PTR_1380","PTR_1390","PTR_1430","PTR_1470","PTR_1480","PTR_1490","PTR_1500","PTR_1535","PTR_1540","PTR_1550","PTR_1560","PTR_1600","PTR_1610","PTR_1620","PTR_1630","PTR_1670","PTR_1710","PTR_1750","PTR_1790","PTR_1830","PTR_1870","PTR_1910","VNUM_C","VNUM","VDATE","DATE_Q02","DATE_Q03","DATE_Q04","DATE_Q05","DATE_Q07","DATE_Q08","DATE_Q09","DATE_Q10","DATE_Q11","DATE_Q12","DATE_Q13","DATE_Q14","DATE_Q15","DATE_Q16","DATE_Q17","DATE_Q18","DATE_Q19","DATE_Q20","DATE_Q21","DATE_Q22","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","P10_1000","P10_1010","P10_1020","P10_1030","P10_1040","P10_1050","P10_1060","P10_1070","P10_1080","P10_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PCH_1000","PCH_1010","PCH_1020","PCH_1030","PCH_1040","PCH_1050","PCH_1060","PCH_1070","PCH_1080","PCH_1090","PCH_1100","PCH_1110","PCH_1120","PCH_1130","PCH_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","IAQ_1000","IAQ_1010","IAQ_1020","IAQ_1030","IAQ_1040","IAQ_1050","IAQ_1060","IAQ_1070","IAQ_1080","IAQ_1090","IAQ_1100","IAQ_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","REGIMEN_PERIOD_1","REGIMEN_PERIOD_2","REGIMEN_PERIOD_3","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","REG_1080","REG_1150","AGE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SER_1010","SER_1020","SER_1030","SER_1040","SER_1050","SER_1060","SER_1070","SER_1100","SER_1110","SER_1120","SER_1130","SER_1140","SER_1150","SER_1160","SER_1170","SER_1180","SER_1190","SER_1200","SER_1210","SER_1220","SER_1240","SER_1250","SER_1000","SER_1080","SER_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PSE_1060","PSE_1070","PSE_1080","PSE_1090","PSE_1100","PSE_1110","PSE_1130","PSE_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SIG_1000","SIG_1010","SIG_1020","SIG_1030","SIG_1040","SIG_1050","SIG_1060","SIG_1080","SIG_1090","SIG_1100","SIG_1110","SIG_1120","SIG_1130","SIG_1140","SIG_1150","SIG_1160","SIG_1170","SIG_1180","SIG_1190","SIG_1200","SIG_1210","SIG_1220","SIG_1230","SIG_1240","SIG_1250","SIG_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SFU_1000","SFU_1010","SFU_1020","SFU_1030","SFU_1040","SFU_1050","SFU_1060","SFU_1070","SFU_1080","SFU_1090","SFU_1100","SFU_1110","SFU_1120","SFU_1130","SFU_1140","SFU_1150","SFU_1160","SFU_1170","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SNQ_1000","SNQ_1010","SNQ_1020","SNQ_1030","SNQ_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SPI_1010","SPI_1020","SPI_1040","SPI_1050","SPI_1060","SPI_1070","VNUM_C","VNUM","VDATE","SPI_1030","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","AM_TRIAL_TYPE","AM_FVC","AM_FEV1","AM_PEF","AM_FEF2575","AM_FET","DRY_1","DRY_2","DRY_3","DRY_4","DRY_5","DRY_6","DRY_7","DRY_8","DRY_9","VNUM_C","VNUM","VDATE","DDATE","PM_TRIAL_TYPE","PM_FVC","PM_FEV1","PM_PEF","PM_FEF2575","PM_FET","DRY_10","DRY_11","DRY_12","DRY_13","DRY_14","DRY_15","DRY_16","DRY_17","DRY_18","DRY_19","DRY_20","DRY_21","DRY_22","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SLA_1010","SLA_1020","SLA_1030","SLA_1000","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SRE_1010","SRE_1020","SRE_1030","SRE_1040","SRE_1050","SRE_1060","SRE_1070","SRE_1080","SRE_1090","SRE_1100","SRE_1000","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SPU_1000","SPU_1010","SPU_1020","SPU_1030","SPU_1040","SPU_1050","SPU_1060","SPU_1070","SPU_1080","SPU_1090","SPU_1100","SPU_1110","SPU_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SAD_1000","SAD_1010","SAD_1020","SAD_1030","SAD_1040","SAD_1050","SAD_1060","SAD_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SCH_1000","SCH_1010","SCH_1020","SCH_1030","SCH_1040","SCH_1050","SCH_1055","SCH_1060","SCH_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TER_1000","TER_1010","TER_1020","TER_1040","TER_1050","TER_1070","TER_1080","TER_1090","TER_1100","TER_1110","TER_1120","TER_1130","TER_1140","TER_1150","TER_1160","TER_1170","TER_1180","TER_1185","TER_1200","TER_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TXF_1000","TXF_1010","TXF_1020","TXF_1030","TXF_1040","TXF_1050","TXF_1060","TXF_1070","TXF_1080","TXF_1090","TXF_1100","TXF_1110","TXF_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TCH_1000","TCH_1010","TCH_1020","TCH_1030","TCH_1040","TCH_1050","TCH_1060","TCH_1080","TCH_1090","TCH_1100","TCH_1110","TCH_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","WSH_1000","WSH_1010","WSH_1020","WSH_1040","WSH_1050","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","WPA_1000","WPA_1010","WPA_1020","WPA_1030","WPA_1040","WPA_1050","WPA_1060","WPA_1070","WPA_1080","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","HOW_SICK_TODAY","RUNNY_NOSE","PLUGGED_NOSE","SNEEZING","SORE_THROAT","SCRATCHY_THROAT","COUGH","HOARSENESS","HEAD_CONGESTION","CHEST_CONGESTION","FEELING_TIRED","THINK_CLEARLY","SLEEP_WELL","BREATH_EASILY","WALK","DAILY_ACTIVITIES","WORK_OUTSIDE_HOME","WORK_INSIDE_HOME","INTERACT","LIVE_LIFE","MY_COLD_IS","VNUM_C","VNUM","VDATE","WDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER"],["Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Char","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Char","Char","Num","Num","Num","Num","Char","Char","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Char","Num","Char","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Char","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num"],["In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, school, or at home? 0=None of the time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All of the time","During the past 3 days, how often have you had asthma symptoms? 0=Not at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how often have you used your rescue inhaler or nebulizer medication ? 0=Not at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how many total times did your asthma symptoms wake you up from sleep? 0=Not at all, 1=1 time in the last 3 days, 2=2-3 times in the last 3 days, 3=4-5 times in the last 3 days, 4=&gt;=6 times in the last 3 days","How would you rate the amount of impairment you have experienced due to your asthma in the past 3 days? 0=No impairment, 1=Mild impairment, 2=Moderate impairment, 3=Severe impairment, 4=Very severe impairment","How stressed or frightened were you by your asthma symptoms in the past 3 days? 0=Not at all, 1=Mildly, 2=Moderately, 3=Severely, 4=Very severely","Why do you think your asthma was worse in the past 3 days compared to what is normal for you? 0=I have not been worse over the past 3 days. My asthma symptoms have been usual., 1=Common cold, 2=Allergies, 3=Pollution or chemical irritant, 4=Too little asth","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Are you currently retired? 1=Yes,0=No","Are you retired because of asthma? 1=Yes,0=No","Are you currently unemployed? 1=Yes,0=No","Are you unemployed because of asthma? 1=Yes,0=No","Do you get paid to do work? 1=Yes,0=No","How much does your asthma bother you at your paid work? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Overall, how much does your asthma bother you when you do jobs around the house? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Overall, how much does your asthma bother your social life? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Overall, how much does your asthma bother your personal life? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Are you involved in leisure activities, such as: walking for pleasure, sports, exercise, travelling, taking vacations? 1=Yes,0=No","When involved in leisure activities, how much does your asthma bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Would you say that you can't do some of these sorts of things because of asthma? 1=Yes,0=No","How much does your asthma bother you when you sleep? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much does the cost of your asthma medicines bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Do you get free prescriptions? 1=Yes,0=No","How much does the inconvenience or embarrassment of taking your asthma medicines bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much do coughs and colds bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Feeling upset is also a bother.  Does your asthma make you feel anxious, depressed, tired, or helpless? 1=Yes,0=No","Feeling upset is also a bother.  Does your asthma make you feel anxious, depressed, tired, or helpless? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will have an asthma attack when visiting a new place? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will catch a cold? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will let others down? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that your health may get worse in the future? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you won't be able to cope with an asthma attack? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past 4 weeks, how much of the time did your asthma keep you from getting as much done at work, school or at home? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often have your had shortness of breath? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often did your asthma symptoms wake you up at night or earlier than usual in the morning? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often have you used your rescue inhaler or nebulizer medication? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","How would you rate your asthma control during the past 4 weeks? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","ICD-9 category","Description of Adverse Event","ICD9 Code","Ongoing at current visit","Type (1=Intermitent; 2=Continuous)","Severity (1=Mild; 2=Moderate; 3=Severe)","Serious (0=No, 1=Yes)","Likelyhood of Relationship to Study Drug(s) (1=None; 2=Unlikely (remote); 3=Possible; 4=Probable)","Change in Study Drugs (1=Unchanged; 2=Altered)","Outcome (1=Completely recovered; 2=Recovered, but with lasting effects; 3=Death)","Treatment Required (1=None; 2=Medication; 3=Hospitalization; 4=Other)","Ongoing at final visit","Date Started","Date Stopped","Visit Number (character)","Visit Number (numeric)","Long ICD-9 description","Short ICD-9 description","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","ASTHMA HISTORY1.Approximately how old were you when chest symptoms suggesting asthma first appeared?","Did these symptoms appear immediately after or as a result of:1a. a respiratory infection such as a cold or pneumonia? 1=Yes, 0=No, 8=Don't Know","an occupational or job change? 1=Yes, 0=No, 8=Don't Know","a household move? 1=Yes, 0=No, 8=Don't Know","a pregnancy? 1=Yes, 0=No, 8=Don't Know","a hormonal change ? 1=Yes, 0=No, 8=Don't Know","How old were you when a doctor first diagnosed you with asthma?","Have any of your immediate blood relatives been told by a physician that they have asthma? 1=Yes, 0=No, 8=Don't Know","Father  (1=Yes, 0=No, 8=Don't Know)","Brother(s) or Sister(s)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Child(ren)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","ASTHMA SYMPTOMS1.How do you categorize your asthma symptoms throughout the course of the year? 1=Relatively the same all year, 2=Vary by season","Winter? 1=Yes,0=No","Spring? 1=Yes,0=No","Summer? 1=Yes,0=No","Fall? 1=Yes,0=No","In the last 12 months, how many...  5a.Asthma episodes have you had that required emergency care or an unscheduled office visit?","Overnight hospitalizations have you had due to asthma?","Courses of systemic corticosteroid therapy  for asthma have you taken?","Days of work, school, or housework have you missed due to asthma?","In the past 3 months, how many days of work, school, or housework have you missed due to asthma?","Have you ever been admitted to an intensive care unit for asthma? 1=Yes,0=No","How many times have you been admitted to an intensive care unit for asthma?","Have you ever had invasive mechanical ventilation? 1=Yes, 0=No, 8=Don't Know","Have you ever had non-invasive mechanical ventilation? 1=Yes, 0=No, 8=Don't Know","Do any of the following currently provoke your asthma? 1=Yes, 0=No, 8=Don't Know","Menstrual cycle(If participant is male or a postmenopausal female, leave blank.) (1=Yes, 0=No, 8=Don't Know)","Aspirin or non-steroidal anti-inflammatory drugs (e.g., Aleve, Motrin) (1=Yes, 0=No, 8=Don't Know)","Respiratory infections (e.g., colds)  (1=Yes, 0=No, 8=Don't Know)","Irritants (e.g., pollution, odors, perfumes, chemicals, household cleaners) (1=Yes, 0=No, 8=Don't Know)","Weather conditions (e.g., change in weather, humidity)  (1=Yes, 0=No, 8=Don't Know)","Exposure to cold air  (1=Yes, 0=No, 8=Don't Know)","Emotional factors (e.g., stress, laughing)  (1=Yes, 0=No, 8=Don't Know)","Tobacco smoke  (1=Yes, 0=No, 8=Don't Know)","Food additives/preservatives (e.g., MSG, sulfites)  (1=Yes, 0=No, 8=Don't Know)","Allergies (e.g., dust, animals, pollens)(1=Yes, 0=No, 8=Don't Know)","Other (1=Yes,0=No)","To which of the following did a doctor or other health practitioner say you were allergic? 1=Yes, 0=No, 8=Don't Know","Foods  (1=Yes, 0=No, 8=Don't Know)","Things you breathe in or are exposed to (e.g., dust, pollens, molds, animal fur, feathers, dander)  (1=Yes, 0=No, 8=Don't Know)","Stinging insects such as bees or wasps (1=Yes, 0=No, 8=Don't Know)","Latex  (1=Yes, 0=No, 8=Don't Know)","Other  (1=Yes,0=No)","Have you ever had eczema / atopic dermatitis ? 1=Yes, 0=No, 8=Don't Know","If YES, was your eczema diagnosed by a doctor? 1=Yes,0=No","Have any of your immediate blood relatives been told by a physician that they have allergies/eczema/hay fever? 1=Yes, 0=No, 8=Don't Know","Father  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Brother(s) or Sister(s)(1=Yes, 0=No, 8=Don't Know, 9=N/A)","Child(ren)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Did you grow up in a household where you were exposed to tobacco smoke? 1=Yes,0=No","Do you currently smoke? 1=Yes,0=No","Record smoking history in pack-years*. __ __ . __ pack-years","Were you ever a smoker? 1=Yes,0=No","Record smoking history in pack-years*. __ __ . __ pack-years","Do you currently live in a household where you are exposed to tobacco smoke? 1=Yes,0=No","Do you currently vape  or use a hookah ? 1=Yes,0=No","How frequently do you vape or use a hookah? 1=Infrequently , 2=Occasionally , 3=Weekly , 4=Daily","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many days a week","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many times a day","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many years have y","Have you ever vaped or used a hookah in the past? 1=Yes,0=No","Have you ever vaped or used a hookah in the past?  If NO, skip to Q17. Approximately how many years did you vape or use a hookah?  (1-99 years)","Do you currently live in a household where you are exposed to others vaping or using a hookah? 1=Yes,0=No","Do you spend time in social settings  where you are exposed to others vaping or using a hookah? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","When was the last time that you vaped or used a hookah? (number of days from Visit 1)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","How many days were you bothered by coughing during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe was your coughing during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by wheezing during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe was your wheezing during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by shortness of breath during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","Over average, how severe was your shortness of breath during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you awakened at night during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how much of a problem was being awakened at night during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by side effects of your asthma medication during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe were the side effects of your asthma medication during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Participants baseline peak flow (PEF) value (0-999 L/M) First visit (Visit 0A, or Visit 1): PEF (FEF Max) from prebronchodilator (baseline) spirometry at first visit (convert to L/M)","Participants baseline rescue use value (0-99 puffs/day) First visit (Visit 0A, or Visit 1): Self-reported average daily use of albuterol during the 14 days prior to first visit.","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","body mass index computed from height and weight measured at Visit 1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Page Number","Reason for change in scheduled medications status (1=Adverse Event, 2=Other (specify))","Reason for change in scheduled medications status Related adverse event number (0-99)","Current status of participants scheduled medications (1=Discontinued, 2=Resumed)","Date change took effect","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Code to identify a drug unit of measure; (Units)","Code to identify how frequently a drug is taken or administered. (Frequency)","Code to identify the route used to administer a drug. (Route)","Name of Medication","This number corresponds to the ID number of the AHFS monograph. (Medication Code)","Related Event","Medication Dose","Ongoing at current visit","Ongoing at final visit","Start Date","Stop Date","Visit Number (character)","Visit Number (numeric)","GEN_DRUG_NAME","CLASS_TEXT","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Diary and Peak Flow Compliance: Number of full days since the last visit (0-999 days)","Diary and Peak Flow Compliance: Number of days where AM and PM scheduled sessions are complete (AM and PM PEF and all diary questions for AM and PM answered) (0-999 days)","Diary and Peak Flow Compliance: Percent compliance If the compliance value in Q1c is less than 75Pct, re-emphasize the importance of completing scheduled diary assessments and peak flows. (0-150.0Pct)","Scheduled Respimat Compliance (Visits 2-9 only): Number of scheduled puffs since the last visit (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Number of remaining puffs on scheduled Respimat Reference all returned Respimats (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Number of puffs taken (60 x (# used Respimats)) - Q2b (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Percent compliance=Q2c/Q2a x 100 If the participant took less than 75Pct of the scheduled Respimat puffs, re-emphasize the importance of maintaining the daily dosing schedule. (0-150.0Pct)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of scheduled puffs since the last visit (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of remaining puffs reflected on scheduled Twisthaler counters Total the values reflected on all counters for all returned Twisthalers (i.e., out of their pouches). (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of puffs taken (60 x (# used Twisthalers)) - Q3b (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Percent compliance=Q3c/Q3a x 100 If the participant took less than 75Pct of the scheduled Twisthaler puffs, re-emphasize the importance of maintaining the daily dosing schedule. (0-150.0Pct)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","1=mometasone, 2=placebo","Blinded Scheduled Twisthaler Contents - How sure are you about your answer in Q1? 1=Absolutely sure - I know what the Twisthaler/MDI contains, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","1=tiotropium, 2=placebo","Blinded Scheduled Respimat Contents - How sure are you about your answer in Q4? 1=Absolutely sure - I know what the Respimat capsules contain, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has the participant or parent/legal guardian signed the SIENA Informed Consent document? 1=Yes,0=No","Ages 12-17 Only: Has the participant signed and dated the assent form or, if the participant is less than the local age of assent, has the participant given verbal assent? 1=Yes,0=No","Has the participant used an inhaled corticosteroid in the past 6 weeks? 1=Yes,0=No","Will the participant be 12 years of age, or older, as of Visit 1? 1=Yes,0=No","Has the participant used an inhaled corticosteroid or combination inhaled corticosteroid/LABA in the past 3 weeks? 1=Yes,0=No","If NO, has the participant used a leukotriene modifier in the past 3 weeks? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used inhaled corticosteroid therapy?","Over the past 3 months, on average, how many days per week has the participant used inhaled corticosteroid therapy: &lt; 5? 1=Yes,0=No","Over the past 3 months, has the participant used greater than the equivalent of 80-240 mcg of inhaled beclomethasone daily? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used combination inhaled corticosteroid/LABA therapy?","Over the past 3 months, on average, how many days per week has the participant used combination inhaled corticosteroid/LABA therapy,  &lt; 5 (1=Yes,0=No)","Over the past 3 months, on average, how many days per week has the participant had daytime asthma symptoms?","Over the past 3 months, on average, how many days per week has the participant had daytime asthma symptoms,  &lt;=2? 1=Yes,0=No","Over the past 3 months, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms?","Over the past 3 months, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms, &lt;=2? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used his/her short acting betaagonist  for relief of symptoms?","Over the past 3 months, on average, how many days per week has the participant used his/her short acting betaagonist  for relief of symptoms, &lt; 2? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 80Pct? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 70Pct? 1=Yes,0=No","Does the participant have current evidence of any of the conditions listed on the Exclusionary Medical Conditions for SIENA  reference card, or any chronic diseases that would prevent participation in the trial or put the participant at risk by participati","Is the participant currently taking any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Is the participant able to go off these medications for the required washout period prior to Visit 1 and for the duration of the study? 1=Yes,0=No","Is the participant able to use the spirotel e-diary/PEF meter correctly, as evidenced by achieving a score of 13 on the spirotel Performance Checklist ? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Has the participant or parent/legal guardian signed the SIENA Informed Consent document?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past month, on average, how many days per week has the participant had daytime symptoms?","In the past month, on average, how many days per week has the participant had daytime symptoms, &lt;=2? 1=Yes,0=No","In the past month, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms?","In the past month, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms,  &lt;=2? 1=Yes,0=No","In the past month, on average, how many days per week has the participant used his/her short acting beta-agonist  for relief of symptoms?","In the past month, on average, how many days per week has the participant used his/her short acting beta-agonist  for relief of symptoms, &lt; 2? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 80Pct? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 70Pct? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant or parent/legal guardian sign the SIENA Informed Consent document? 1=Yes,0=No","Ages 12-17 Only: Has the participant signed and dated the assent form or, if the participant is less than the local age of assent, has the participant given verbal assent? 1=Yes,0=No","Is the participant 12 years of age, or older? 1=Yes,0=No","Does the participant plan to move away from the clinical site in the upcoming 11 months such that his/her ability to complete the study will be jeopardized? 1=Yes,0=No","Has the participant used investigative drugs and/or enrolled in an intervention trial in the past 30 days, or have plans to enroll in such a trial during the SIENA study? 1=Yes,0=No","Has the participant had a respiratory infection within the past 6 weeks? 1=Yes,0=No","Has the participant had a respiratory infection within the past 4 weeks? 1=Yes,0=No","During the past 4 weeks, on average, how many days per week has the participant had daytime asthma symptoms?","During the past 4 weeks, on average, how many days per week has the participant had daytime asthma symptoms  &gt; 2? 1=Yes,0=No","During the past 4 weeks, how many nighttime awakenings due to asthma symptoms has the participant had?","During the past 4 weeks, how many nighttime awakenings due to asthma symptoms has the participant had,  &gt; 2? 1=Yes,0=No","During the past 4 weeks, on average, how many days per week has the participant used his/her short-acting beta-agonist  for relief of symptoms?","During the past 4 weeks, on average, how many days per week has the participant used his/her short-acting beta-agonist  for relief of symptoms, &gt; 2? 1=Yes,0=No","Is Q7a, Q8a, or Q9a checked YES? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Did the participant or parent/legal guardian sign the SIENA Informed Consent document?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Does the participant have current evidence of any of the conditions listed on the Exclusionary Medical Conditions for SIENA  reference card, or any chronic diseases  that would prevent participation in the trial or put the participant at risk by participat","Does the participant have a history of bladder-neck obstruction, urinary retention, benign prostatic hyperplasia , or a clinically relevant urologic disorder that precludes study participation? 1=Yes,0=No","Does the participant have a history of narrow angle glaucoma? 1=Yes,0=No","Does the participant have a history of significant cardiovascular disorders or arrhythmias? 1=Yes,0=No","Has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card within the specified time periods? 1=Yes,0=No","Is the participant currently taking prescription or OTC medication other than those listed on the Allowed Medications  reference card? 1=Yes,0=No","Based on input from the participant and the study physician, will the participant need to use intranasal steroids at any time during the study? 1=Yes,0=No","Is the participant willing to use a single intranasal steroid at a stable dose continuously for the duration of the study, starting at Visit 1? 1=Yes,0=No","Is the participant currently receiving allergen immunotherapy  other than an established maintenance regimen implemented continuously for a minimum of 3 months? 1=Yes,0=No","Has the participant taken omalizumab within the past 3 months? 1=Yes,0=No","Has the participant used any smokeless tobacco products  in the past year? 1=Yes,0=No","Has the participant smoked cigarettes, a pipe, cigar, marijuana, electronic cigarettes, or any other substance in the past year? 1=Yes,0=No","Ages 18+ Only: Does the participant have a smoking history of greater than 10 pack-years? 1=Yes,0=No","Ages 12-17 Only: Does the participant have a smoking history of greater than 5 pack-years? 1=Yes,0=No","Has the participant received a physician diagnosis of asthma at least 12 months ago? 1=Yes,0=No","Has the participant received a physician diagnosis of asthma at least 12 months ago or had a history consistent with asthma for the previous 12 months? 1=Yes,0=No","Has the participant experienced a life-threatening asthma exacerbation requiring treatment with intubation or mechanical ventilation in the past 5 years? 1=Yes,0=No","Has the participant had an asthma exacerbation requiring systemic corticosteroid treatment in the past 6 weeks? 1=Yes,0=No","Has the participant used an oral or inhaled corticosteroid, or leukotriene modifier in the past 6 weeks? 1=Yes,0=No","Has the participant used an oral corticosteroid in the past 6 weeks? 1=Yes,0=No","Has the participant used an inhaled corticosteroid or leukotriene modifier in the past 3 weeks? 1=Yes,0=No","Is the participant potentially able to bear children? 1=Yes, 0=No, 9=N/A","Is the participant currently pregnant or lactating? 1=Yes,0=No","Does the participant agree to use one of the approved methods indicated on the Birth Control Methods  reference card for the duration of the study? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was the participants prebronchodilator FEV1 &gt;=70Pct predicted? 1=Yes,0=No","Did the participants FEV1 improve &gt;=12Pct and &gt;=200 ml in response to four puffs of albuterol? 1=Yes,0=No","Does the participant have valid source documentation within the past 6 months for an overread AsthmaNet methacholine challenge  with a PC20 &lt;=16 mg/ml? 1=Yes,0=No","Did the participants FEV1 improve &gt;=12Pct and &gt;=200 ml in response to four puffs of albuterol?","Was the participants methacholine PC20 &lt;=16 mg/ml? 1=Yes,0=No","Is the participant able to use the spirotel e-diary/PEF meter correctly, as evidenced by achieving a score of 13 on the Spirotel Performance Checklist ? 1=Yes, 0=No, 9=N/A","Has the participant used the Twisthaler previously? 1=Yes,0=No","Is the participant able to use a Twisthaler properly, as evidenced by achieving a score of 12 on the Twisthaler Inhalation Technique Checklist ? 1=Yes,0=No","Is the participant able to use a metered dose inhaler properly, as evidenced by achieving a score of 11 on the MDI Inhalation Technique Checklist  ? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Does the participant have valid source documentation within the past 6 months for an overread AsthmaNet methacholine challenge  with a PC20 &lt;=16 mg/ml?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Since Visit 1, has the participant experienced a treatment failure event as defined in the protocol? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Since Visit 1, has the participant experienced one or more asthma exacerbations as defined in the protocol? 1=Yes,0=No","Since Visit 1, has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant provide two acceptable sputum induction samples during the run-in? 1=Yes,0=No","Since Visit 1, has the participant experienced a treatment failure event as defined in the protocol? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Did the participant complete Asmanex220 mcg treatment at least 6 weeks ago? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Since Visit 1, has the participant experienced one or more asthma exacerbations as defined in the protocol? 1=Yes,0=No","Since Visit 1, has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Has the participant experienced a respiratory tract infection in the past 6 weeks? 1=Yes,0=No","Has the participant experienced a respiratory tract infection in the past 4 weeks? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report daytime asthma symptoms daily during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report night-time awakenings due to asthma symptoms more than once per week during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report night-time awakenings due to asthma symptoms more than twice per week during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report daily albuterol use for asthma symptom control during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant record at least 75Pct of AM and PM peak flow measurements and symptoms on his or her spirotel during the run-in? 1=Yes,0=No","Did the participant take at least 75Pct of the required puffs from his or her Respimat during the run-in? 1=Yes,0=No","Was the participants prebronchodilator FEV1 &gt;=70Pct predicted? 1=Yes,0=No","Has the participant used the Respimat previously? 1=Yes,0=No","Is the participant able to use a Respimat inhaler properly, as evidenced by achieving a score of 15 on the Respimat Inhalation Technique Checklist? 1=Yes,0=No","Does the participant wish to withdraw consent from the study? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has QC procedure been performed on the NIOX MINO today? 1=Yes,0=No","Did the participant eat or drink within the past hour? 1=Yes,0=No","Did the participant take part in strenuous activity/exercise within the past hour? 1=Yes,0=No","Time eNO started (based on a 24-hour clock) (0000-2359)","ENO Measurement","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","How long have you lived in the current house? (Estimate if uncertain.) (1010) years","How long have you lived in the current house? (Estimate if uncertain.)(1020) months","Does your house use a wood burning stove as a primary source of heat? 1=Yes, 0=No, 8=Don't Know","Does your house use an air conditioner? 1=Yes, 0=No, 8=Don't Know","Does your house use an evaporative cooler ? 1=Yes, 0=No, 8=Don't Know","Does your house use a humidifier? 1=Yes, 0=No, 8=Don't Know","Does your house use a dehumidifier? 1=Yes, 0=No, 8=Don't Know","Has there been water damage to your house, basement, or its contents during the past 12 months? 1=Yes, 0=No, 8=Don't Know","Has there been any mold or mildew, on any surfaces, inside your house in the past 12 months? 1=Yes, 0=No, 8=Don't Know","Which rooms have or have had mold or mildew?10a.Bathroom(s)  (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew?10b.Basement or attic (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10c.Kitchen (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10d.Your bedroom (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10e.Other bedrooms (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10f.Living or family room (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10g.Other (1 = Yes, 0 = No)","Do you ever see cockroaches in your house? 1=Yes,0=No","In which room(s) have you seen cockroaches?12a.Kitchen (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches?  12b.Basement or attic (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12c.Bathroom(s) (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12d.Living or family room (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12e.Your bedroom  (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12f.Other bedrooms (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12g.Garage (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12h.Other (1 = Yes, 0 = No)","Do you ever see rodents  or rodent droppings in your house? 1=Yes,0=No","14.In which room(s) have you seen rodents or rodent droppings? 14a. Kitchen (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14b. Basement or attic (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14c. Bathroom(s) (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings?14d. Living or family room (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14e. Your bedroom(1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14f. Other bedrooms (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14g. Garage (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14h. Other (1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15a. Barns(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property?15b. Hay(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15c. Woodsheds(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15d. Firewood(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property?15e. Chicken coops (1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15f. Corral (1 = Yes, 0 = No)","CHARACTERISTICS OF THE PARTICIPANT'S BEDROOM16.What is the floor covering in your bedroom? 1=Rug/carpet, 2=Vinyl tile or linoleum, 3=Wood, 4=Ceramic tile, 5=Other , 9=Don't know","What type of mattress is on your bed? 1=None, 2=Inner spring mattress, 3=Foam mattress, 4=Waterbed, 5=Air mattress, 6=Other , 9=Don't know","Is the mattress completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","Does your bed have a box spring? 1=Yes,0=No","Is the box spring completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","What type of pillow do you usually sleep with? 1=None, 2=Feather/down, 3=Foam/Dacron/synthetic, 5=Other , 9=Don't know","Is the pillow completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","Does your household have any pets? 1=Yes,0=No","Enter the number of pets that the household has. (Enter '00' if none. If none to Q24a - Q24d, skip to the next question.)24a.Cat","Cat  (1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has. Dog","Dog(1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has 24c.Rabbit, guinea pig, hamster, gerbil, or mouse","Rabbit, guinea pig, hamster, gerbil, or mouse (1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has 24d.Bird","Bird (1=Indoor, 2=Outdoor, 3=Both)","In general, and on a regular basis, are you exposed to any of the following animals?  25a.Cat(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25b.Dog(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25c.Rabbit, guinea pig, hamster, gerbil, or mouse(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25d.Bird (1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25e.Farm animals (1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals?  25f.Other (1 = Yes, 0 = No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other (specify","Highest education of any household member (1=No High School diploma/GED, 2=High Schol grad, 3=Technical training, 4=Some college, no degree, 5=Assoc degree, 6=Bachelor degree, 7=Masters degree/MD/PhD/JD/PharmD, 9=Decline to answer, 10=Don't know)","Category best describes the combined annual income, before taxes, of all members of your household for the last year. (1=Less than $25,000, 2=$25,000 - $49,999, 3=$50,000 - $99,999, 4=$100,000 or more, 9=Decline to answer, 10=Don't know)","How many people  are supported by this income reported in Q3?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Total IgE","Cat (E1)","Dog (E5)","Mouse (E72)","Rat (E74)","Mold mix (MX1)","Cockroach, German (I6)","Grass mix (GX2)","Tree mix (TX4)","Tree mix (TX6)","Weed mix (WX1)","Weed (W3)","Mite (D2)","Mite (D1)","Number or allergen specific (Cat, Dog, Mouse, etc) that were either POS or greater than 0.35","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","CBC with differential cell count: Eosinophils (absolute count) (0-9999/uL)","CBC with differential cell count: WBC (0-999.9 K/uL)","CBC with differential cell count: Differential Lymphocytes (0-99.9Pct)","CBC with differential cell count: Differential Monocytes (0-99.9Pct)","CBC with differential cell count: Differential Basophils (0-99.9Pct)","CBC with differential cell count: Differential Neutrophils (0-99.9Pct)","CBC with differential cell count: Differential Eosinophils (0-99.9Pct)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","PARENTAL HEIGHT - First study visit only or until both are completed 1.Biological mother's height (complete height or check unknown) __ feet","PARENTAL HEIGHT - First study visit only or until both are completed1.Biological mother's height (complete height or check unknown) __ __ inches","PARENTAL HEIGHT - First study visit only or until both are completed 1.Biological mother's height (complete height or check unknown) (9=Don't Know)","Biological father's height (complete height or check unknown) __ feet","Biological father's height (complete height or check unknown) __ __ inches","Biological father's height (complete height or check unknown) (9=Don't Know)","PARTICIPANT MEASUREMENTS - Complete at all applicable study visits3. What type of height measurement was obtained? 1=Standing height, 2=Length","First measurement __ __ __ . __ cm","Second measurement __ __ __ . __ cm","Third measurement __ __ __ . __ cm","Average height or length measurement __ __ __ . __ cm","In your judgment, was the participant's height or length measurement acceptable? 1=Yes,0=No","Weight (shoes off, light clothing) __ __ __ . __ kg","ORAL CANDIDIASIS5.Does the participant have evidence of oral candidiasis? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Post Diluent FEV1","Did the participant drop20Pct at the diluent stage? 1=Yes,0=No","Last concentration of methacholine administered __ __ . __ __ __ __ mg/ml","FEV1 after last concentration of methacholine administered","Did the participant achieve a PC20? 1=Yes,0=No","PC20 __ __ . __ __ mg/ml","Time methacholine challenge ended (based on 24-hour clock)","FEV1","Time of FEV1 in Q7a (based on 24-hour clock)","Was the FEV1 from Q7athe methacholine reversal reference value  in the gray box above? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was an additional treatment used in the first hour? 1=Yes,0=No","Additional albuterol by MDI (1=Yes,0=No)","Number of additional puffs of albuterol administered (1=2, 2=4, 3=&gt;4)","Nebulized Beta-agonist(1=Yes,0=No)","Subcutaneous epinephrine(1=Yes,0=No)","Implementation of clinic emergency protocol or algorithm (1=Yes,0=No)","Other  (1=Yes,0=No)","Participant's FEV1 after additional treatment within first hour.2a.FEV1","Time of FEV1 in Q2a (based on 24-hour clock)","Was the FEV1 from Q2a &gt;the methacholine reversal reference value  in the gray box on the Methacholine Challenge Testing  form? 1=Yes,0=No","Was additional treatment used after one hour? 1=Yes,0=No","Additional albuterol by MDI (1=Yes,0=No)","Number of additional puffs of albuterol administered (1=2, 2=4, 3=&gt;4)","Nebulized Beta-agonist  (1=Yes,0=No)","Subcutaneous epinephrine(1=Yes,0=No)","Implementation of clinic emergency protocol or algorithm  (1=Yes,0=No)","Treatment in the emergency room  (1=Yes,0=No)","Overnight hospitalization (1=Yes,0=No)","Other (1=Yes,0=No)","Participant's final FEV1 after methacholine challenge4a.FEV1","Time of FEV1 in Q4a (based on 24-hour clock)","Was the FEV1 from Q4a ? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Exclusions and Confounders1.Has the participant had any severe acute illness in the past 4 weeks? 1=Yes,0=No","If YES, has the participant received permission from the supervising physician to proceed with the methacholine challenge testing? 1=Yes,0=No","Has the participant used an oral or injectable steroid  in the last 4 weeks? 1=Yes,0=No","Does the participant have a baseline  FEV1 less than 55Pct of predicted or less than 1.0 L? 1=Yes,0=No","Pregnancy test results:  (Check N/A if the participant is male, or is female and is post-menopausal, had a hysterectomy or tubal ligation.) (1=Positive, 0=Negative, 9=N/A)","Is the participant's systolic blood pressure &gt; 200 mm Hg or diastolic blood pressure &gt; 100 mm Hg? 1=Yes,0=No","Is there any other reason the participant should not proceed with the methacholine challenge testing? 1=Yes,0=No","Is the participant eligible to proceed with the diluent  pulmonary function testing for the methacholine challenge? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Exclusions and Confounders:  1.Has the participant had any severe acute illness in the past 4 weeks? 1=Yes,0=No","If YES, has the participant received permission from the supervising physician to proceed with the methacholine challenge testing? 1=Yes,0=No","During the past 4 weeks, has the participant had any respiratory infections, colds, or bronchitis ? 1=Yes,0=No","If YES, during the past 2 weeks, has the participant had any respiratory infections, colds, or bronchitis ? 1=Yes,0=No","Has the participant used an oral or injectable steroid  in the last 4 weeks? 1=Yes,0=No","Does the participant have a baseline  FEV1 less than 70Pct of predicted? 1=Yes,0=No","Pregnancy test results(Check N/A if the participant is male, or is female and has not started menses.)(1=Positive, 0=Negative, 9=N/A)","If participant's age is ? 1=Yes,0=No","If participant's age is &lt; 12 years: Is the participant's systolic blood pressure &gt; 180 mm Hg or diastolic blood pressure &gt; 90 mm Hg? 1=Yes,0=No","Is there any other reason the participant should not proceed with the methacholine challenge testing? 1=Yes,0=No","Is the participant eligible to proceed with the diluent  pulmonary function testing for the methacholine challenge? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time albuterol administered (based on 24-hour clock)","Time post-albuterol spirometry started (based on 24-hour clock)","Highest FVC","Highest FEV1","Highest FEV1 (Pct predicted)","FEF Max","FEF25-75","In your judgment, was the subject's spirometry technique acceptable? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Please check the box next to the treatment that you believe you received over the past 12 weeks. 1=mometasone, 2=placebo","Blinded Scheduled Twisthaler Contents: How sure are you about your answer to Question 2? 1=Absolutely sure -  I know what the MDI contains, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","Please comment with respect to any observations you made that helped you make your choice in Question 2 . 1=I have no comments, 2=I noticed the following:","Please check the box next to the treatment that you believe you received over the past 12 weeks. 1=tiotropium, 2=placebo","How sure are you about your answer to Question 5? 1=Absolutely sure - I know what the Respimat capsules contain, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","1=I have no comments, 2=I noticed the following:","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Visit Number (numeric)","Level of periostin in peripheral blood (ng/ml)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time ipratropium administered (based on 24-hour clock) (0000-2359)","Time post-ipratropium spirometry started (based on 24-hour clock) (0000-2359)","Highest FVC (0-9.99 L)","Highest FEV1 (0-9.99 L)","Highest FEV1 (Pct predicted) (0-999 Pct predicted)","FEF Max (0-99.99 L/S)","FEF25-75 (0-9.99 L/S)","In your judgment, was the participants spirometry technique acceptable? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Is the participant unable to bear children due to any of the following reasons? 1=Yes,0=No","Post-menopausal (at least one year since last menses) (1=Yes,0=No)","Hysterectomy(1=Yes,0=No)","Tubal ligation(1=Yes,0=No)","Pregnancy test results  (1=Positive, 0=Negative)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Have you had any diseases, illnesses, conditions, or surgeries related Blood, Lymph, or Immune Systems? 1=Yes,0=No","Eyes(1=Yes,0=No)","Breasts (1=Yes,0=No)","Endocrine Systems (1=Yes,0=No)","Heart and Blood Vessels  (1=Yes,0=No)","Liver or Pancreas (1=Yes,0=No)","Kidneys or Urinary Tract System  (1=Yes,0=No)","Reproductive System (1=Yes,0=No)","Muscles or Bones (1=Yes,0=No)","Nervous System (1=Yes,0=No)","Psychiatric(1=Yes,0=No)","Drug Allergies (1=Yes,0=No)","Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Skin (1=Yes,0=No)","Have you ever had allergic rhinitis ? 1=Yes, 0=No, 9=Don't know","Have you ever had nasal polyps? 1=Yes, 0=No, 9=Don't know","Do you have chronic or recurrent sinusitis ? 1=Yes, 0=No, 9=Don't know","Have you ever been diagnosed with vocal cord dysfunction? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the ear, nose, or throat? 1=Yes,0=No","Have you ever had pneumonia? 1=Yes, 0=No, 9=Don't know","If YES, were you diagnosed by chest x-ray? 1=Yes, 0=No, 9=Don't know","If YES, were you treated with antibiotics? 1=Yes, 0=No, 9=Don't know","Have you ever had bronchitis? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the lungs ? 1=Yes, 0=No, 9=Don't know","Do you have gastroesophageal reflux disease ? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the stomach or intestines? 1=Yes,0=No","Have you been diagnosed with sleep disordered breathing ? 1=Yes,0=No","If YES, are you being treated with CPAP or BiPAP? 1=Yes,0=No","Have you ever had other sleep disorders? 1=Yes,0=No","Have you ever had other conditions that have not been mentioned on this form? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Short-acting Inhaled Beta-Agonists by Inhaler (e.g., albuterol, Primatene Mist, Maxair, ProAir, Proventil, Ventolin, Xopenex)  (1=Yes, 0=No, 9=Don't Know)","If YES, indicate average weekly puffs in the past month (Enter '000' if none used) weekly puffs","Rescue treatment via a Nebulizer Machine (e.g., albuterol, ipratropium, Combivent, Xopenex, levalbuterol)  (1=Yes, 0=No, 9=Don't Know)","Long-acting Inhaled Beta-Agonists (e.g., Serevent, Foradil, salmeterol, formoterol) Do not consider combination medications. (1=Yes, 0=No, 9=Don't Know)","Oral Beta-Agonists (e.g., albuterol, Brethine, Bricanyl, metaproterenol, Proventil, Ventolin, Repetabs, Volmax)  (1=Yes, 0=No, 9=Don't Know)","Oral Theophylline (short-acting or sustained release)(e.g., Aminophylline, Slo-Phyllin, Slo-bid, Theo-Dur, Uniphyl)(1=Yes, 0=No, 9=Don't Know)","Inhaled Anticholinergic by Inhaler (e.g., Atrovent, Combivent, Spiriva) (1=Yes, 0=No, 9=Don't Know)","Leukotriene Antagonist / 5LO Inhibitors (e.g., Accolate, Zyflo, Singulair) (1=Yes, 0=No, 9=Don't Know)","IgE Blocker (e.g., Xolair)(1=Yes, 0=No, 9=Don't Know)","Oral Steroids FOR ASTHMA (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron,  dexamethasone)(1=Yes, 0=No, 9=Don't Know)","If YES, in the past 12 months, how many courses of steroids by mouth have you taken FOR ASTHMA? 1=1 course, 2=2 courses, 3=3 courses, 4=4 courses, 5=5 courses, 6=More than 5 courses","Injectable Steroids FOR ASTHMA (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV)(1=Yes, 0=No, 9=Don't Know)","Steroids by Inhaler (e.g., Asmanex Twisthaler, QVAR, Flovent, Pulmicort Flexhaler)(1=Yes, 0=No, 9=Don't Know)","Indicate most recent type of inhaled steroid taken (refer to PRIOR_TRT_CARD reference card) code","Indicate number of daily puffs used daily puffs","Indicate the total number of months that you used the inhaled steroid out of the past 12 months months","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide) (1=Yes, 0=No, 9=Don't Know)","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide). If YES, complete Q13a - Q13c. Indicate most recent type of nebulized steroid taken (refer to PRIOR_TRT_CARD reference card)","Indicate number of daily treatments used daily treatments","Indicate the total number of months that you used the nebulized steroid out of the past 12 months months","Long-Acting Beta-Agonist and Inhaled Steroid Combination Medications (e.g., Advair Diskus, Symbicort MDI, Dulera MDI)(1=Yes, 0=No, 9=Don't Know)","Indicate most recent type of combination medication taken (refer to PRIOR_TRT_CARD reference card) code","Indicate number of daily puffs used daily puffs","Indicate the total number of months that you used the combination medication out of the past 12 months months","Nasal Steroids (e.g., Beconase, Vancenase, Flonase, Nasacort, Nasalide, Nasarel, Omnaris, Rhinocort, Nasonex)(1=Yes, 0=No, 9=Don't Know)","Non-steroidal Anti-allergic Nasal Medications (e.g., Nasalcrom, Astelin, Astepro, ipratropium)(1=Yes, 0=No, 9=Don't Know)","Anti-allergic Oral Medications (e.g., fexofenadine, loratadine, cetirizine, chlorpheniramine) (1=Yes, 0=No, 9=Don't Know)","Topical Steroids - Prescription (e.g., Synalar, Lidex, Dermacin, Fluocinonide) (1=Yes, 0=No, 9=Don't Know)","Topical Steroids - OTC (e.g., Hydrocortisone - multiple strengths and products) (1=Yes, 0=No, 9=Don't Know)","Other Medication FOR ASTHMA OR ALLERGIES (1=Yes, 0=No, 9=Don't Know)","Oral Steroids for Conditions Other Than Asthma (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) (1=Yes, 0=No, 9=Don't Know)","Injectable Steroids for Conditions Other Than Asthma (e.g., Medrol, Solumedrol, Decadron, dexamethasone,  triamcinolone, Kenalog, hydrocortisone IV)(1=Yes, 0=No, 9=Don't Know)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Short-acting Inhaled Beta-Agonists by Inhaler (e.g., albuterol, Primatene Mist, Maxair, ProAir, Proventil, Ventolin, Xopenex) Date","Rescue treatment via a Nebulizer Machine (e.g., albuterol, ipratropium, Combivent, Xopenex, levalbuterol) Date","Long-acting Inhaled Beta-Agonists (e.g., Serevent, Foradil, salmeterol, formoterol) Do not consider combination medications. Date","Oral Beta-Agonists (e.g., albuterol, Brethine, Bricanyl, metaproterenol, Proventil, Ventolin, Repetabs, Volmax) Date","Inhaled Anticholinergic by Inhaler (e.g., Atrovent, Combivent, Spiriva) Date","Leukotriene Antagonist / 5LO Inhibitors (e.g., Accolate, Zyflo, Singulair) Date","IgE Blocker (e.g., Xolair) Date","Oral Steroids FOR ASTHMA (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) Date","Injectable Steroids FOR ASTHMA (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV) Date","Steroids by Inhaler (e.g., Asmanex Twisthaler, QVAR, Flovent, Pulmicort Flexhaler) Date","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide) Date","Long-Acting Beta-Agonist and Inhaled Steroid Combination Medications (e.g., Advair Diskus, Symbicort MDI, Dulera MDI) Date","Nasal Steroids (e.g., Beconase, Vancenase, Flonase, Nasacort, Nasalide, Nasarel, Omnaris, Rhinocort, Nasonex) Date","Non-steroidal Anti-allergic Nasal Medications (e.g., Nasalcrom, Astelin, Astepro, ipratropium) Date","Anti-allergic Oral Medications (e.g., fexofenadine, loratadine, cetirizine, chlorpheniramine) Date","Topical Steroids - Prescription (e.g., Synalar, Lidex, Dermacin, Fluocinonide) Date","Topical Steroids - OTC (e.g., Hydrocortisone - multiple strengths and products) Date","Other Medication FOR ASTHMA OR ALLERGIES Date","Oral Steroids for Conditions Other Than Asthma (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) Date","Injectable Steroids for Conditions Other Than Asthma (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV) Date","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the last month, how often have you been upset because of something that happened unexpectedly? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that you were unable to control the important things in your life? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt nervous and stressed? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt confident about being able to handle your personal problems? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that things were going your way? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you found that you could not cope with all the things that you had to do? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you been able to control irritations in your life? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that you were on top of things? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you been angered because of things that happened that were outside of your control? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that your difficulties were piling up so high that you could not overcome them? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Have you consumed caffeine in the past 4 hours? 1=Yes,0=No","Have you used medications with caffeine in the past 4 hours? 1=Yes,0=No","Have you used any weight loss medications in the past 4 hours? 1=Yes,0=No","Have you consumed any food containing alcohol or beverages containing alcohol in the past 4 hours? 1=Yes,0=No","Have you used a rescue intermediate-acting inhaled betaagonist in the past 6 hours? 1=Yes,0=No","Have you used a short-acting anticholinergic in the past 6 hours? 1=Yes,0=No","Have you used any ophthalmic antihistamines in the past 6 hours? 1=Yes,0=No","Have you used any nasal antihistamines in the past 6 hours? 1=Yes,0=No","Have you used any nasal decongestants in the past 6 hours? 1=Yes,0=No","Have you used any oral antihistamines in the past 48 hours? 1=Yes,0=No","Have you used any oral decongestants or cold remedies in the past 48 hours? 1=Yes,0=No","Have you used any smokeless tobacco products today? 1=Yes,0=No","At this time, is your asthma worse because of recent exposure to triggers? 1=Yes,0=No","Is there any other reason you should not proceed with spirometry testing? 1=Yes,0=No","Is the participant eligible to proceed with the spirometry testing? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past 4 weeks, I worried about the long-term effects of asthma on my health (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I had to worry about asthma triggers (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, my asthma was on my mind (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, it was hard to get a good nights sleep because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt like I couldnt enjoy life because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt that asthma was controlling my life (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt frustrated that I couldnt make plans in advance because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, everyday activities were a struggle (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, asthma placed stress on my relationships with family, friends, significant others, or co-workers (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, I felt frustrated that I have to do things differently than people who dont have asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt like I missed out on doing things with others because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, I had to do a lot of planning to make sure I always had an inhaler ready (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Randomized treatment assignment period 1: (Mometasone, Tiotropium or Placebo)","Randomized treatment assignment period 2: (Mometasone, Tiotropium or Placebo)","Randomized treatment assignment period 3: (Mometasone, Tiotropium or Placebo)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Sex (1=Male, 2=Female)","Primary Racial Identification for Spirometry (1=Amer Indian or Alaskan Native, 2=Asian or Pacific Islander, 3=Black, 4=White, 5=Hispanic or Latino, 6=Other)","Participant's Age at Visit 1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Description of Adverse Event (ICD9 Code)","Is the participant currently taking study drug? 1=Yes,0=No","Time interval between the last administration of the study drug and the Adverse Event","What was the unit of time for the interval in Question #4? 1=Second, 2=Minute, 3=Hour, 4=Day","Fatal event(1=Yes,0=No)","Life-threatening event (1=Yes,0=No)","Inpatient hospitalization required(1=Yes,0=No)","Hospitalization prolonged  (1=Yes,0=No)","Disabling or incapacitating  (1=Yes,0=No)","Overdose (1=Yes,0=No)","Cancer (1=Yes,0=No)","Congenital anomaly(1=Yes,0=No)","Serious laboratory abnormality with clinical symptoms(1=Yes,0=No)","Height failure (per protocol MOP)(1=Yes,0=No)","Pregnancy(1=Yes,0=No)","Other (1=Yes,0=No)","What in your opinion caused the event? 1=Yes,0=No","Withdrawal of study drug(s)(1=Yes,0=No)","Concurrent medication  (1=Yes,0=No)","Other condition or event  (1=Yes,0=No)","Was the event expected or unexpected? 1=Expected, 2=Unexpected","Was the event possibly, probably, or definitely related to study participation? 1=Yes,0=No","Date of Adverse Event","Admission date","Discharge date","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","What type of height measurement was obtained? 1=Standing height, 2=Length","First measurement __ __ __ . __ cm","Second measurement __ __ __ . __ cm","Third measurement __ __ __ . __ cm","Average height or length measurement __ __ __ . __ cm","In your judgment, was the participant's height or length measurement acceptable? 1=Yes,0=No","Weight (shoes off, light clothing) __ __ __ . __ kg","Does the participant have evidence of oral candidiasis? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant fail to respond within 48 hours to the treatment failure rescue algorithm? 1=Yes,0=No","Did the participant use at least 16 puffs as needed albuterol per 24 hours for a period of 48 hours? 1=Yes,0=No","Did the participant experience prebronchodilator FEV1 values &lt; 50Pct of the baseline prebronchodilator value obtained at Visit 1 on two consecutive spirometric determinations made on different days? 1=Yes, 0=No, 9=Not evaluated","Did the participant experience prebronchodilator FEV1 values &lt; 40Pct of predicted on two consecutive spirometric determinations made on different days? 1=Yes, 0=No, 9=Not evaluated","Did the study or treating physician prescribe the participant oral/parenteral corticosteroids for the treatment of his/her asthma? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation in the opinion of the study investigator or personal physician? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation? 1=Yes,0=No","Did the participant seek care for significant asthma exacerbation conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Was the participant hospitalized? 1=Yes,0=No","Duration of hospital stay (0-99.9 days)","Was intubation or ventilation assistance required? 1=Yes,0=No","Was the participant admitted to the intensive care unit? 1=Yes,0=No","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Inhaled corticosteroids (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Nebulized bronchodilator (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Oral corticosteroids (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, IM or IV steroids (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Antibiotics (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Other (1=Yes, 0=No )","Why do you think the participant experienced a significant asthma exacerbation? 1=Common cold, 2=Allergies, 3=Pollution or chemical irritant, 4=Too little asthma maintenance medication, 5=Exercise, 6=Other","Date exacerbation conditions were met","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Which phone contact following Asthma Exacerbation visit is being performed? 1=Phone Contact #1 , 2=Phone Contact #2 , 3=Phone Contact #3","Since the most recent Asthma Exacerbation contact, has the participant sought additional care for significant asthma exacerbation conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Was the participant hospitalized? 1=Yes,0=No","Duration of hospital stay (0-99.9 days)","Was intubation or ventilation assistance required? 1=Yes,0=No","Was the participant admitted to the intensive care unit? 1=Yes,0=No","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Inhaled corticosteroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Nebulized bronchodilator (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Oral corticosteroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, IM or IV steroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Antibiotics (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Runny Nose  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Post nasal drip  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Need to blow your nose  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Facial pain/pressure(0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Nasal obstruction  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time spirometry started (based on 24-hour clock)","The reported FEV1 and FVC are the best volumes of all acceptable maneuvers. 1.Highest FVC","Highest FEV1 (Pct predicted)","The reported flow rates correspond to the maneuver where FEV1 + FVC is maximized. FEF Max","FEF25-75","In your judgment, was the subject's spirometry technique acceptable? 0=No, 1=Yes","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Highest FEV1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","TRIAL_TYPE","FVC","FEV1","PEF","FEF2575","FET","Number of times the participant woke up last night due to asthma symptoms # Times woke up due to asthma (0-9)","Number of puffs the participant will take from the WHITE study inhaler this morning # Puffs from WHITE inhaler in AM (0-9)","only) Number of puffs the participant will take from the BLUE study inhaler this morning # Puffs from BLUE inhaler in AM (0-9)","Has the participant taken any puffs from his/her RED RESCUE albuterol inhaler in the past 4 hours? 1=yes, 0=no","Shortness of Breath Score Shortness of Breath Score overnight (0,1,2,3)","Chest tightness score Chest Tightness Score overnight (0,1,2,3)","Wheezing score Wheezing Score overnight (0,1,2,3)","Coughing score Coughing Score overnight (0,1,2,3)","Phlegm/Mucus score Phlegm/Mucus Score overnight (0,1,2,3)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","date 'fixed' for analysis so that associated AM and PM always have same ddate","TRIAL_TYPE","FVC","FEV1","PEF","FEF2575","FET","Number of puffs the participant will take from the WHITE study inhaler tonight # Puffs from WHITE inhaler in PM (0-9)","Has the participant taken any puffs from his/her RED RESCUE albuterol inhaler during the past 4 hours? 1=yes, 0=no","Shortness of Breath Score Shortness of Breath Score since waking this morning (0,1,2,3)","Chest tightness score Chest Tightness Score since waking this morning (0,1,2,3)","Wheezing score Wheezing Score since waking this morning (0,1,2,3)","Coughing score Coughing Score since waking this morning (0,1,2,3)","Phlegm/Mucus score Phlegm/Mucus Score since waking this morning (0,1,2,3)","Did your regular exercise cause unusually severe shortness of breath during the past 24 hours? 1=Yes, 0=No, 9=N/A","Number of RED RESCUE albuterol puffs taken in the past 24 hours to prevent symptoms# Preventive  RED RESCUE puffs  in past 24 hours?","Number of RED RESCUE albuterol puffs taken for asthma symptoms or low peak flow during past 24 hours  # RED RESCUE  puffs for symptoms  in past 24 hours?","Number of times used RED RESCUE albuterol inhaler for asthma symptoms during past 24 hours  # Times used  RED RESCUE for symptoms in past 24 hours?","Did RED RESCUE albuterol relieve symptoms for less than 4 hours after treatment during past 24 hours? 1=Yes,0=No","Was the participant seen by a healthcare provider  for an unscheduled visit in the past 24 hours due to asthma symptoms? 1=Yes,0=No","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Processing: Time processing started (based on 24-hour clock) __ __ __ __ (0 - 2359)","Processing Sample: Total Cell Count: __ __ __ __. __x 10^4 cells/ml (0 - 9999.9)","Was the participant's sputum sample processed within 4 hours after collection? 1=Yes,0=No","Processing Sample: Processing Date __ __ / __ __ / 20 __ __","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Rate slide's quality:  (1=Very good, 2=Good, 3=Acceptable, 4=Poor but readable, 5=Not readable)","Record the number on the slide(s) that was (were) read (0 - 9)","Record the number on the slide(s) that was (were) read: These are numbers that were assigned to the slides at each site. (0 -9)","Total Cell Count: __ __ __ __ . __ x 10^4 cells/ml (0 - 9999.9)","Differential Cell Counts: Squamous Cells __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Epithelial Cells __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Macrophages __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Neutrophils __ __ . __ Pct (0 -99.9)","Differential Cell Counts: Eosinophils __ __ . __ Pct (0 - 99.9)","Differential Cell Counts: Lymphocytes __ __ . __ Pct (0 -99.9)","Date of Read __ __ / __ __ / 20 __ __","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","For this protocol, what was the duration of sputum induction the first time the participant's sample was processed within 4 hours after collection?","Sputum induction start time (based on 24-hour clock)","Sputum induction stop time (based on 24-hour clock)","Duration of sputum induction collection phase at this visit -- minutes (0 - 99.9)","Was the duration ? 1=Yes,0=No","Volume of sputum sample at this visit -- ml (0.0 - 999.9)","Is the volume adequate for processing? 1=Yes,0=No","Is the sample adequate for laboratory analysis? 1=Yes,0=No","Participant's FEV1 immediately after completion of sputum induction: -- (0.0 -9.99 L)","FEV1 (Pct predicted) (0 - 999)","Time of FEV1 in Q7a (based on 24-hour clock)","Percent difference in FEV1 ((Reference - Q7a) / Reference) X100 (-99.9 - 99.9 Pct)","Did the participant's FEV1 drop &gt; 10Pct from reference FEV1 as indicated in Q7d? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Participant's FEV1 after initial 2 puffs of albuterol: 1a.FEV1 __ . __ __ L (0 -9.99)","Participant's FEV1 after initial 2 puffs of albuterol: 1b. FEV1 (Pct predicted)__ __ __ Pct predicted (0 - 999)","Participant's FEV1 after initial 2 puffs of albuterol: 1c. Time of FEV1 from Q1a (based on 24-hour clock) __ __ __ __ (0 - 2359)","Participant's FEV1 after initial 2 puffs of albuterol: 1d. Was the FEV1 from Q1a &gt;=the sputum induction reversal reference value in the gray box above? 1=Yes,0=No","Participant's FEV1 after 2 additional puffs of albuterol: 2a. FEV1 __ . __ __ L (0 - 9.99)","Participant's FEV1 after 2 additional puffs of albuterol: 2b. FEV1 (Pct predicted) __ __ __ Pct predicted (0 - 999)","Participant's FEV1 after 2 additional puffs of albuterol: 2c. Time of FEV1 from Q2a (based on 24-hour clock)__ __ __ __ (0 - 2359)","Participant's FEV1 after 2 additional puffs of albuterol: 2d.Was the FEV1 from Q2a ? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was the participant's FEV1 after reversal from the methacholine challenge greater yhan 90Pct of the baseline FEV1 (1=Yes,0=No)","If NO, has the participant received permission from the supervising physician to proceed with sputum induction testing? 1=Yes,0=No","Physician's Signature:_________ (1=signature present, NULL=signature missing)","Participant's FEV1 used for assessment of eligibility for sputum induction __ . __ __ L (0 - 9.99)","Participant's FEV1 (Pct predicted) used for assessment of eligibility for sputum induction __ __ __ Pct predicted (O-999)","Was the participant's FEV1  from Q3 ? 1=Yes,0=No","Has the participant used any smokeless tobacco products  today? 1=Yes,0=No","Is there any other reason the participant should not proceed with sputum induction? 1=Yes,0=No","Is the participant eligible for sputum induction? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has the participant completed the study through Visit 9? 1=Yes,0=No","Who initiated termination of the participant? 1=Participant, 2=Clinical Staff","1=no longer interested in participating*, 2=no longer willing to follow protocol*, 3=difficult access to clinic , 4=unable to make visits during clinic hours, 5=moving out of the area, 6=unable to continue due to personal constraints*, 7=unable to continue","Did clinical staff terminate the participant due to pregnancy? 1=Yes, 0=No, 9=N/A","Did clinical staff terminate the participant due to loss to follow-up? 1=Yes,0=No","Did clinical staff terminate the participant due to loss to follow-up, type of contact (1=In-person visit, 2=Phone call)","Did clinical staff terminate the participant due to an asthma-related adverse event? 1=Yes,0=No","Did clinical staff terminate the participant due to a medication-related adverse event? 1=Yes,0=No","Did clinical staff terminate the participant due to an adverse event not related to asthma or medications? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility during the Supervised Washout ? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with medication dosing? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with diary completion? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with visit attendance? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with peak flow monitoring? 1=Yes,0=No","Did clinical staff terminate the participant due to significant asthma exacerbation or two treatment failures during run-in ? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility during the run-in period  for reasons other than compliance or exacerbation/treatment failure or high sputum eosinophils ? 1=Yes,0=No","Did clinical staff terminate the participant due to other reason? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility due to high sputum eosinophils in the run-in ? 1=Yes,0=No","Did clinical staff terminate the participant due to... Indicate the letter corresponding to the primary reason the participant was terminated. (a-n)","Did clinical staff terminate the participant due to loss to follow-up, date of last contact with participant","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant seek care for treatment failure conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact, other than call to notify clinic as instructed in spirotel","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Inhaled corticosteroids (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Nebulized bronchodilator (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started: Oral corticosteroids form.  (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  IM or IV steroids form.  (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started: Antibiotics (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant awaken from asthma three or more times in a two-week period or on two consecutive nights? 1=Yes,0=No","Did the participant use albuterol for relief of symptoms four or more times/day for two or more consecutive days? 1=Yes,0=No","Did albuterol relieve symptoms for less than four hours after treatment? 1=Yes,0=No","Did the participant use albuterol for relief of symptoms daily for seven days, and this use exceeded two times the weekly use of albuterol in the baseline period? 1=Yes,0=No","Did regular exercise cause unusually severe shortness of breath on two or more days during a seven day period? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation? 1=Yes,0=No","Did the participant experience a treatment failure? 1=Yes,0=No","Has the participant experienced two treatment failures in the current treatment period ? 1=Yes,0=No","Does the study physician feel the study treatment contributed directly to the treatment failure, such that it is in the participants best interest to discontinue the current blinded treatment? 1=Yes,0=No","Did the participant begin daily treatment with high-dose ICS for 10 days? 1=Yes,0=No","Did the participant begin daily treatment with high-dose ICS for 10 days? 1=Did not think symptoms werebad enough, 2=Other","Date treatment failure conditions were met","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","During this last treatment period , has the participant experienced a treatment failure or asthma exacerbation? 1=Yes,0=No","Has the participant experienced two or more treatment failures in this last treatment period, or does the study physician feel the study treatment contributed directly to the treatment failure such that it is in the participants best interest to discontinu","As of today, have 2 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","As of today, have 6 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","As of today, have 3 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Are you currently employed ? 1=Yes,0=No","In general, how many hours per week do you usually work?","During the past seven days, how many hours did you miss from work because of problems associated with your asthma?","During the past seven days, how much did asthma affect your productivity while you were working?","Do you currently attend classes in an academic setting ? 1=Yes,0=No","In general, how many hours per week do you usually attend classes?","During the past seven days, how many hours did you miss from class or school because of problems associated with your asthma?","During the past seven days, how much did asthma affect your productivity while in school or attending classes in an academic setting?","During the past seven days, how much did your asthma affect your ability to do your regular daily activities, other than work at a job or attend classes?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","How sick today? 0=Not Sick, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely","Runny Nose:(0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","PLUGGED_NOSE  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SNEEZING  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SORE_THROAT  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SCRATCHY_THROAT  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","COUGH (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","HOARSENESS  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","HEAD_CONGESTION  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","CHEST_CONGESTION  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","FEELING_TIRED (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","THINK_CLEARLY  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SLEEP_WELL  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","BREATH_EASILY  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WALK  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","DAILY_ACTIVITIES  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WORK_OUTSIDE_HOME  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WORK_INSIDE_HOME  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","INTERACT  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","LIVE_LIFE  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","MY_COLD_IS  (1=Very much better, 2=Somewhat better, 3=A little better, 4=The same, 5=A little worse, 6=Somewhat worse, 7=Very much worse)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Date (number of days from Visit 1)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>data_set_name<\/th>\n      <th>variable_number_within_dataset<\/th>\n      <th>variable<\/th>\n      <th>type<\/th>\n      <th>variable_description<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":2},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"data_set_name","targets":1},{"name":"variable_number_within_dataset","targets":2},{"name":"variable","targets":3},{"name":"type","targets":4},{"name":"variable_description","targets":5}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-45c6e86f681d8078437e" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-45c6e86f681d8078437e">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80","81","82","83","84","85","86","87","88","89","90","91","92","93","94","95","96","97","98","99","100","101","102","103","104","105","106","107","108","109","110","111","112","113","114","115","116","117","118","119","120","121","122","123","124","125","126","127","128","129","130","131","132","133","134","135","136","137","138","139","140","141","142","143","144","145","146","147","148","149","150","151","152","153","154","155","156","157","158","159","160","161","162","163","164","165","166","167","168","169","170","171","172","173","174","175","176","177","178","179","180","181","182","183","184","185","186","187","188","189","190","191","192","193","194","195","196","197","198","199","200","201","202","203","204","205","206","207","208","209","210","211","212","213","214","215","216","217","218","219","220","221","222","223","224","225","226","227","228","229","230","231","232","233","234","235","236","237","238","239","240","241","242","243","244","245","246","247","248","249","250","251","252","253","254","255","256","257","258","259","260","261","262","263","264","265","266","267","268","269","270","271","272","273","274","275","276","277","278","279","280","281","282","283","284","285","286","287","288","289","290","291","292","293","294","295","296","297","298","299","300","301","302","303","304","305","306","307","308","309","310","311","312","313","314","315","316","317","318","319","320","321","322","323","324","325","326","327","328","329","330","331","332","333","334","335","336","337","338","339","340","341","342","343","344","345","346","347","348","349","350","351","352","353","354","355","356","357","358","359","360","361","362","363","364","365","366","367","368","369","370","371","372","373","374","375","376","377","378","379","380","381","382","383","384","385","386","387","388","389","390","391","392","393","394","395","396","397","398","399","400","401","402","403","404","405","406","407","408","409","410","411","412","413","414","415","416","417","418","419","420","421","422","423","424","425","426","427","428","429","430","431","432","433","434","435","436","437","438","439","440","441","442","443","444","445","446","447","448","449","450","451","452","453","454","455","456","457","458","459","460","461","462","463","464","465","466","467","468","469","470","471","472","473","474","475","476","477","478","479","480","481","482","483","484","485","486","487","488","489","490","491","492","493","494","495","496","497","498","499","500","501","502","503","504","505","506","507","508","509","510","511","512","513","514","515","516","517","518","519","520","521","522","523","524","525","526","527","528","529","530","531","532","533","534","535","536","537","538","539","540","541","542","543","544","545","546","547","548","549","550","551","552","553","554","555","556","557","558","559","560","561","562","563","564","565","566","567","568","569","570","571","572","573","574","575","576","577","578","579","580","581","582","583","584","585","586","587","588","589","590","591","592","593","594","595","596","597","598","599","600","601","602","603","604","605","606","607","608","609","610","611","612","613","614","615","616","617","618","619","620","621","622","623","624","625","626","627","628","629","630","631","632","633","634","635","636","637","638","639","640","641","642","643","644","645","646","647","648","649","650","651","652","653","654","655","656","657","658","659","660","661","662","663","664","665","666","667","668","669","670","671","672","673","674","675","676","677","678","679","680","681","682","683","684","685","686","687","688","689","690","691","692","693","694","695","696","697","698","699","700","701","702","703","704","705","706","707","708","709","710","711","712","713","714","715","716","717","718","719","720","721","722","723","724","725","726","727","728","729","730","731","732","733","734","735","736","737","738","739","740","741","742","743","744","745","746","747","748","749","750","751","752","753","754","755","756","757","758","759","760","761","762","763","764","765","766","767","768","769","770","771","772","773","774","775","776","777","778","779","780","781","782","783","784","785","786","787","788","789","790","791","792","793","794","795","796","797","798","799","800","801","802","803","804","805","806","807","808","809","810","811","812","813","814","815","816","817","818","819","820","821","822","823","824","825","826","827","828","829","830","831","832","833","834","835","836","837","838","839","840","841","842","843","844","845","846","847","848","849","850","851","852","853","854","855","856","857","858","859","860","861","862","863","864","865","866","867","868","869","870","871","872","873","874","875","876","877","878","879","880","881","882","883","884","885","886","887","888","889","890","891","892","893","894","895","896","897","898","899","900","901","902","903","904","905","906","907","908","909","910","911","912","913","914","915","916","917","918","919","920","921","922","923","924","925","926","927","928","929","930","931","932","933","934","935","936","937","938","939","940","941","942","943","944","945","946","947","948","949","950","951","952","953","954","955","956","957","958","959","960","961","962","963","964","965","966","967","968","969","970","971","972","973","974","975","976","977","978","979","980","981","982","983","984","985","986","987","988","989","990","991","992","993","994","995","996","997","998","999","1000","1001","1002","1003","1004","1005","1006","1007","1008","1009","1010","1011","1012","1013","1014","1015","1016","1017","1018","1019","1020","1021","1022","1023","1024","1025","1026","1027","1028","1029","1030","1031","1032","1033","1034","1035","1036","1037","1038","1039","1040","1041","1042","1043","1044","1045","1046","1047","1048","1049","1050","1051","1052","1053","1054","1055","1056","1057","1058","1059","1060","1061","1062","1063","1064","1065","1066","1067","1068","1069","1070","1071","1072","1073","1074","1075","1076","1077","1078","1079","1080","1081","1082","1083","1084","1085","1086","1087","1088","1089","1090","1091","1092","1093","1094","1095","1096","1097","1098","1099","1100","1101","1102","1103","1104","1105","1106","1107","1108","1109","1110","1111","1112","1113","1114","1115","1116","1117","1118","1119","1120","1121","1122","1123","1124","1125","1126","1127","1128","1129","1130","1131","1132","1133","1134","1135","1136","1137","1138","1139","1140","1141","1142","1143","1144","1145","1146","1147","1148","1149","1150","1151","1152","1153","1154","1155","1156","1157","1158","1159","1160","1161","1162","1163","1164","1165","1166","1167","1168","1169","1170","1171","1172","1173","1174","1175","1176","1177","1178","1179","1180","1181","1182","1183","1184","1185","1186","1187","1188","1189","1190"],["aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","aaaq.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","abp.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","act.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","aeclin.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asthma_hx_adult.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","asui.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","baseline.sas7bdat","bmi.sas7bdat","bmi.sas7bdat","bmi.sas7bdat","bmi.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","change_meds.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","cmed.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","comply.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","ctxqx.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0a.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig0b.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig1.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig2.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig3.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig4.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","elig5.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","eno.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","heq.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","household_sei.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","ige.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lab.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","lexam_ped.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","metha_add_trt.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_adult.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","methachk_ped.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","palb4_spiro.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","parttxqx.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","periostin.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","pipra4_spiro.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","preg_test.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_adult.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_cond_all.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","prior_trt.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pss_10.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","pulmonarychk.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","rand_iaql_12.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","regimen.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","registry.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","serious.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sexam_ped.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","sigex_follow.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","snq.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spiro.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","spirotel.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputlab.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputread.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputum_add_trt.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","sputumchk.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","term.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","txfail_chk.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","washout.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wpai_asthma.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat","wurss_21.sas7bdat"],[1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,1,2,3,4,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,1,2,3,4,5,6,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,1,2,3,4,5,6,7,8,9,10,11,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28],["AAQ_1000","AAQ_1010","AAQ_1020","AAQ_1030","AAQ_1040","AAQ_1050","AAQ_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ABP_1000","ABP_1010","ABP_1020","ABP_1030","ABP_1040","ABP_1050","ABP_1060","ABP_1070","ABP_1080","ABP_1090","ABP_1100","ABP_1110","ABP_1120","ABP_1130","ABP_1140","ABP_1150","ABP_1160","ABP_1170","ABP_1180","ABP_1190","ABP_1200","ABP_1210","ABP_1220","ABP_1230","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ACT_1","ACT_2","ACT_3","ACT_4","ACT_5","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ICD9_CAT","AEC_1000","AEC_1010","AEC_1040","AEC_1050","AEC_1060","AEC_1070","AEC_1080","AEC_1090","AEC_1100","AEC_1110","AEC_1120","AEC_1020","AEC_1030","VNUM_C","VNUM","ICD9LONG","ICD9SHRT","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","AHA_1000","AHA_1020","AHA_1030","AHA_1040","AHA_1050","AHA_1060","AHA_1070","AHA_1090","AHA_1100","AHA_1110","AHA_1120","AHA_1130","AHA_1140","AHA_1150","AHA_1160","AHA_1170","AHA_1180","AHA_1190","AHA_1200","AHA_1210","AHA_1220","AHA_1250","AHA_1260","AHA_1270","AHA_1280","AHA_1290","AHA_1300","AHA_1310","AHA_1320","AHA_1330","AHA_1340","AHA_1350","AHA_1360","AHA_1370","AHA_1380","AHA_1390","AHA_1400","AHA_1410","AHA_1420","AHA_1430","AHA_1440","AHA_1450","AHA_1460","AHA_1470","AHA_1500","AHA_1570","AHA_1580","AHA_1590","AHA_1600","AHA_1730","AHA_1740","AHA_1750","AHA_1760","AHA_1770","AHA_1780","AHA_1790","AHA_1800","AHA_1810","AHA_1820","AHA_1830","AHA_1840","AHA_1850","AHA_1890","AHA_1900","VNUM_C","VNUM","VDATE","DATE_VAPE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ASU_1000","ASU_1010","ASU_1020","ASU_1030","ASU_1040","ASU_1050","ASU_1060","ASU_1070","ASU_1080","ASU_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","BAS_1000","BAS_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","BMI","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAGE","CHG_1000","CHG_1010","CHG_1020","CHG_1030","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","CME_1040","CME_1050","CME_1055","CME_1000","CME_1010","CME_1020","CME_1030","CME_1080","CME_1090","CME_1060","CME_1070","VNUM_C","VNUM","GEN_NAME","CLASS","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","COM_1000","COM_1010","COM_1020","COM_1030","COM_1040","COM_1050","COM_1060","COM_1070","COM_1080","COM_1090","COM_1100","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","CTX_1000","CTX_1010","CTX_1030","CTX_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E0A_1000","E0A_1020","E0A_1030","E0A_1040","E0A_1043","E0A_1045","E0A_1050","E0A_1060","E0A_1070","E0A_1073","E0A_1075","E0A_1080","E0A_1090","E0A_1100","E0A_1110","E0A_1120","E0A_1130","E0A_1140","E0A_1145","E0A_1150","E0A_1160","E0A_1170","E0A_1180","E0A_1190","E0A_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E0B_1000","E0B_1010","E0B_1020","E0B_1030","E0B_1040","E0B_1050","E0B_1060","E0B_1065","E0B_1070","E0B_1080","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E1_1000","E1_1020","E1_1030","E1_1040","E1_1050","E1_1060","E1_1065","E1_1070","E1_1080","E1_1090","E1_1100","E1_1110","E1_1120","E1_1130","E1_1140","E1_1010","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E2_1000","E2_1010","E2_1020","E2_1030","E2_1040","E2_1050","E2_1060","E2_1070","E2_1080","E2_1090","E2_1100","E2_1110","E2_1120","E2_1130","E2_1140","E2_1145","E2_1150","E2_1160","E2_1170","E2_1173","E2_1175","E2_1180","E2_1190","E2_1200","E2_1210","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E3_1000","E3_1010","E3_1020","E3_1030","E3_1070","E3_1080","E3_1090","E3_1100","E3_1110","E3_1120","E3_1130","E3_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E4_1000","E4_1010","E4_1020","E4_1030","E4_1050","E4_1060","E4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","E5_1000","E5_1010","E5_1020","E5_1030","E5_1035","E5_1040","E5_1050","E5_1060","E5_1065","E5_1070","E5_1080","E5_1085","E5_1090","E5_1100","E5_1110","E5_1120","E5_1130","E5_1140","E5_1150","E5_1160","E5_1170","E5_1180","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","ENO_1000","ENO_1010","ENO_1020","ENO_1040","ENO_1050","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","HEQ_1000","HEQ_1010","HEQ_1020","HEQ_1030","HEQ_1040","HEQ_1050","HEQ_1060","HEQ_1070","HEQ_1080","HEQ_1090","HEQ_1100","HEQ_1110","HEQ_1120","HEQ_1130","HEQ_1140","HEQ_1150","HEQ_1160","HEQ_1170","HEQ_1180","HEQ_1190","HEQ_1200","HEQ_1210","HEQ_1220","HEQ_1230","HEQ_1240","HEQ_1250","HEQ_1260","HEQ_1270","HEQ_1280","HEQ_1290","HEQ_1300","HEQ_1310","HEQ_1320","HEQ_1330","HEQ_1340","HEQ_1350","HEQ_1360","HEQ_1370","HEQ_1380","HEQ_1390","HEQ_1400","HEQ_1410","HEQ_1420","HEQ_1430","HEQ_1440","HEQ_1450","HEQ_1460","HEQ_1470","HEQ_1480","HEQ_1490","HEQ_1500","HEQ_1510","HEQ_1520","HEQ_1530","HEQ_1540","HEQ_1550","HEQ_1560","HEQ_1570","HEQ_1580","HEQ_1590","HEQ_1600","HEQ_1610","HEQ_1620","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SEI_1000","SEI_1010","SEI_1020","SEI_1030","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","IGE","RE1","RE5","RE72","RE74","RMX1","RI6","RGX2","RTX4","RTX6","RWX1","RW3","RD2","RD1","N_ALLERGENPOS","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","LAB_1000","LAB_1010","LAB_1020","LAB_1030","LAB_1040","LAB_1050","LAB_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PLE_1000","PLE_1010","PLE_1020","PLE_1030","PLE_1040","PLE_1050","PLE_1060","PLE_1070","PLE_1080","PLE_1090","PLE_1100","PLE_1110","PLE_1130","PLE_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MTH_1000","MTH_1010","MTH_1020","MTH_1030","MTH_1040","MTH_1050","MTH_1060","MTH_1070","MTH_1080","MTH_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MAD_1000","MAD_1010","MAD_1020","MAD_1030","MAD_1040","MAD_1050","MAD_1060","MAD_1070","MAD_1090","MAD_1100","MAD_1110","MAD_1120","MAD_1130","MAD_1140","MAD_1150","MAD_1160","MAD_1170","MAD_1180","MAD_1190","MAD_1200","MAD_1220","MAD_1230","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MCA_1000","MCA_1010","MCA_1050","MCA_1060","MCA_1070","MCA_1080","MCA_1100","MCA_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","MCP_1000","MCP_1010","MCP_1030","MCP_1040","MCP_1050","MCP_1060","MCP_1070","MCP_1080","MCP_1090","MCP_1100","MCP_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PA4_1000","PA4_1010","PA4_1020","PA4_1030","PA4_1040","PA4_1050","PA4_1060","PA4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PTX_1000","PTX_1010","PTX_1020","PTX_1030","PTX_1040","PTX_1050","PTX_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","VNUM","PERIOSTIN","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PI4_1000","PI4_1010","PI4_1020","PI4_1030","PI4_1040","PI4_1050","PI4_1060","PI4_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PRG_1000","PRG_1010","PRG_1020","PRG_1030","PRG_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAD_1000","PAD_1010","PAD_1020","PAD_1030","PAD_1040","PAD_1050","PAD_1060","PAD_1070","PAD_1080","PAD_1090","PAD_1100","PAD_1110","PAD_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PAL_1000","PAL_1010","PAL_1020","PAL_1030","PAL_1040","PAL_1050","PAL_1060","PAL_1070","PAL_1080","PAL_1090","PAL_1100","PAL_1110","PAL_1120","PAL_1130","PAL_1150","PAL_1160","PAL_1170","PAL_1180","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PTR_1000","PTR_1010","PTR_1050","PTR_1060","PTR_1100","PTR_1140","PTR_1180","PTR_1220","PTR_1260","PTR_1300","PTR_1340","PTR_1380","PTR_1390","PTR_1430","PTR_1470","PTR_1480","PTR_1490","PTR_1500","PTR_1535","PTR_1540","PTR_1550","PTR_1560","PTR_1600","PTR_1610","PTR_1620","PTR_1630","PTR_1670","PTR_1710","PTR_1750","PTR_1790","PTR_1830","PTR_1870","PTR_1910","VNUM_C","VNUM","VDATE","DATE_Q02","DATE_Q03","DATE_Q04","DATE_Q05","DATE_Q07","DATE_Q08","DATE_Q09","DATE_Q10","DATE_Q11","DATE_Q12","DATE_Q13","DATE_Q14","DATE_Q15","DATE_Q16","DATE_Q17","DATE_Q18","DATE_Q19","DATE_Q20","DATE_Q21","DATE_Q22","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","P10_1000","P10_1010","P10_1020","P10_1030","P10_1040","P10_1050","P10_1060","P10_1070","P10_1080","P10_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PCH_1000","PCH_1010","PCH_1020","PCH_1030","PCH_1040","PCH_1050","PCH_1060","PCH_1070","PCH_1080","PCH_1090","PCH_1100","PCH_1110","PCH_1120","PCH_1130","PCH_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","IAQ_1000","IAQ_1010","IAQ_1020","IAQ_1030","IAQ_1040","IAQ_1050","IAQ_1060","IAQ_1070","IAQ_1080","IAQ_1090","IAQ_1100","IAQ_1110","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","REGIMEN_PERIOD_1","REGIMEN_PERIOD_2","REGIMEN_PERIOD_3","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","REG_1080","REG_1150","AGE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SER_1010","SER_1020","SER_1030","SER_1040","SER_1050","SER_1060","SER_1070","SER_1100","SER_1110","SER_1120","SER_1130","SER_1140","SER_1150","SER_1160","SER_1170","SER_1180","SER_1190","SER_1200","SER_1210","SER_1220","SER_1240","SER_1250","SER_1000","SER_1080","SER_1090","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","PSE_1060","PSE_1070","PSE_1080","PSE_1090","PSE_1100","PSE_1110","PSE_1130","PSE_1140","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SIG_1000","SIG_1010","SIG_1020","SIG_1030","SIG_1040","SIG_1050","SIG_1060","SIG_1080","SIG_1090","SIG_1100","SIG_1110","SIG_1120","SIG_1130","SIG_1140","SIG_1150","SIG_1160","SIG_1170","SIG_1180","SIG_1190","SIG_1200","SIG_1210","SIG_1220","SIG_1230","SIG_1240","SIG_1250","SIG_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SFU_1000","SFU_1010","SFU_1020","SFU_1030","SFU_1040","SFU_1050","SFU_1060","SFU_1070","SFU_1080","SFU_1090","SFU_1100","SFU_1110","SFU_1120","SFU_1130","SFU_1140","SFU_1150","SFU_1160","SFU_1170","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SNQ_1000","SNQ_1010","SNQ_1020","SNQ_1030","SNQ_1040","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SPI_1010","SPI_1020","SPI_1040","SPI_1050","SPI_1060","SPI_1070","VNUM_C","VNUM","VDATE","SPI_1030","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","AM_TRIAL_TYPE","AM_FVC","AM_FEV1","AM_PEF","AM_FEF2575","AM_FET","DRY_1","DRY_2","DRY_3","DRY_4","DRY_5","DRY_6","DRY_7","DRY_8","DRY_9","VNUM_C","VNUM","VDATE","DDATE","PM_TRIAL_TYPE","PM_FVC","PM_FEV1","PM_PEF","PM_FEF2575","PM_FET","DRY_10","DRY_11","DRY_12","DRY_13","DRY_14","DRY_15","DRY_16","DRY_17","DRY_18","DRY_19","DRY_20","DRY_21","DRY_22","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SLA_1010","SLA_1020","SLA_1030","SLA_1000","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SRE_1010","SRE_1020","SRE_1030","SRE_1040","SRE_1050","SRE_1060","SRE_1070","SRE_1080","SRE_1090","SRE_1100","SRE_1000","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SPU_1000","SPU_1010","SPU_1020","SPU_1030","SPU_1040","SPU_1050","SPU_1060","SPU_1070","SPU_1080","SPU_1090","SPU_1100","SPU_1110","SPU_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SAD_1000","SAD_1010","SAD_1020","SAD_1030","SAD_1040","SAD_1050","SAD_1060","SAD_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","SCH_1000","SCH_1010","SCH_1020","SCH_1030","SCH_1040","SCH_1050","SCH_1055","SCH_1060","SCH_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TER_1000","TER_1010","TER_1020","TER_1040","TER_1050","TER_1070","TER_1080","TER_1090","TER_1100","TER_1110","TER_1120","TER_1130","TER_1140","TER_1150","TER_1160","TER_1170","TER_1180","TER_1185","TER_1200","TER_1060","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TXF_1000","TXF_1010","TXF_1020","TXF_1030","TXF_1040","TXF_1050","TXF_1060","TXF_1070","TXF_1080","TXF_1090","TXF_1100","TXF_1110","TXF_1120","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","TCH_1000","TCH_1010","TCH_1020","TCH_1030","TCH_1040","TCH_1050","TCH_1060","TCH_1080","TCH_1090","TCH_1100","TCH_1110","TCH_1070","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","WSH_1000","WSH_1010","WSH_1020","WSH_1040","WSH_1050","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","WPA_1000","WPA_1010","WPA_1020","WPA_1030","WPA_1040","WPA_1050","WPA_1060","WPA_1070","WPA_1080","VNUM_C","VNUM","VDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER","HOW_SICK_TODAY","RUNNY_NOSE","PLUGGED_NOSE","SNEEZING","SORE_THROAT","SCRATCHY_THROAT","COUGH","HOARSENESS","HEAD_CONGESTION","CHEST_CONGESTION","FEELING_TIRED","THINK_CLEARLY","SLEEP_WELL","BREATH_EASILY","WALK","DAILY_ACTIVITIES","WORK_OUTSIDE_HOME","WORK_INSIDE_HOME","INTERACT","LIVE_LIFE","MY_COLD_IS","VNUM_C","VNUM","VDATE","WDATE","RAND_ID","ENROLL_TYPE","ENROLL_ORDER"],["Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Char","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Char","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Char","Char","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Char","Char","Char","Char","Num","Num","Num","Num","Char","Char","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Char","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Char","Num","Char","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Char","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Char","Char","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Num","Char","Num","Num","Num","Char","Char","Num"],["In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, school, or at home? 0=None of the time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All of the time","During the past 3 days, how often have you had asthma symptoms? 0=Not at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how often have you used your rescue inhaler or nebulizer medication ? 0=Not at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or more times per day","During the past 3 days, how many total times did your asthma symptoms wake you up from sleep? 0=Not at all, 1=1 time in the last 3 days, 2=2-3 times in the last 3 days, 3=4-5 times in the last 3 days, 4=&gt;=6 times in the last 3 days","How would you rate the amount of impairment you have experienced due to your asthma in the past 3 days? 0=No impairment, 1=Mild impairment, 2=Moderate impairment, 3=Severe impairment, 4=Very severe impairment","How stressed or frightened were you by your asthma symptoms in the past 3 days? 0=Not at all, 1=Mildly, 2=Moderately, 3=Severely, 4=Very severely","Why do you think your asthma was worse in the past 3 days compared to what is normal for you? 0=I have not been worse over the past 3 days. My asthma symptoms have been usual., 1=Common cold, 2=Allergies, 3=Pollution or chemical irritant, 4=Too little asth","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Are you currently retired? 1=Yes,0=No","Are you retired because of asthma? 1=Yes,0=No","Are you currently unemployed? 1=Yes,0=No","Are you unemployed because of asthma? 1=Yes,0=No","Do you get paid to do work? 1=Yes,0=No","How much does your asthma bother you at your paid work? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Overall, how much does your asthma bother you when you do jobs around the house? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Overall, how much does your asthma bother your social life? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Overall, how much does your asthma bother your personal life? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Are you involved in leisure activities, such as: walking for pleasure, sports, exercise, travelling, taking vacations? 1=Yes,0=No","When involved in leisure activities, how much does your asthma bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Would you say that you can't do some of these sorts of things because of asthma? 1=Yes,0=No","How much does your asthma bother you when you sleep? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much does the cost of your asthma medicines bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Do you get free prescriptions? 1=Yes,0=No","How much does the inconvenience or embarrassment of taking your asthma medicines bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much do coughs and colds bother you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery, 0=None of these really apply to me","Feeling upset is also a bother.  Does your asthma make you feel anxious, depressed, tired, or helpless? 1=Yes,0=No","Feeling upset is also a bother.  Does your asthma make you feel anxious, depressed, tired, or helpless? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will have an asthma attack when visiting a new place? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will catch a cold? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you will let others down? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that your health may get worse in the future? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","How much bother is the worry that you won't be able to cope with an asthma attack? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past 4 weeks, how much of the time did your asthma keep you from getting as much done at work, school or at home? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often have your had shortness of breath? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often did your asthma symptoms wake you up at night or earlier than usual in the morning? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","During the past 4 weeks, how often have you used your rescue inhaler or nebulizer medication? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","How would you rate your asthma control during the past 4 weeks? 1=All of the time, 2=Most of the time, 3=Some of the time, 4=A little of the time, 5=None of the time","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","ICD-9 category","Description of Adverse Event","ICD9 Code","Ongoing at current visit","Type (1=Intermitent; 2=Continuous)","Severity (1=Mild; 2=Moderate; 3=Severe)","Serious (0=No, 1=Yes)","Likelyhood of Relationship to Study Drug(s) (1=None; 2=Unlikely (remote); 3=Possible; 4=Probable)","Change in Study Drugs (1=Unchanged; 2=Altered)","Outcome (1=Completely recovered; 2=Recovered, but with lasting effects; 3=Death)","Treatment Required (1=None; 2=Medication; 3=Hospitalization; 4=Other)","Ongoing at final visit","Date Started","Date Stopped","Visit Number (character)","Visit Number (numeric)","Long ICD-9 description","Short ICD-9 description","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","ASTHMA HISTORY1.Approximately how old were you when chest symptoms suggesting asthma first appeared?","Did these symptoms appear immediately after or as a result of:1a. a respiratory infection such as a cold or pneumonia? 1=Yes, 0=No, 8=Don't Know","an occupational or job change? 1=Yes, 0=No, 8=Don't Know","a household move? 1=Yes, 0=No, 8=Don't Know","a pregnancy? 1=Yes, 0=No, 8=Don't Know","a hormonal change ? 1=Yes, 0=No, 8=Don't Know","How old were you when a doctor first diagnosed you with asthma?","Have any of your immediate blood relatives been told by a physician that they have asthma? 1=Yes, 0=No, 8=Don't Know","Father  (1=Yes, 0=No, 8=Don't Know)","Brother(s) or Sister(s)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Child(ren)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","ASTHMA SYMPTOMS1.How do you categorize your asthma symptoms throughout the course of the year? 1=Relatively the same all year, 2=Vary by season","Winter? 1=Yes,0=No","Spring? 1=Yes,0=No","Summer? 1=Yes,0=No","Fall? 1=Yes,0=No","In the last 12 months, how many...  5a.Asthma episodes have you had that required emergency care or an unscheduled office visit?","Overnight hospitalizations have you had due to asthma?","Courses of systemic corticosteroid therapy  for asthma have you taken?","Days of work, school, or housework have you missed due to asthma?","In the past 3 months, how many days of work, school, or housework have you missed due to asthma?","Have you ever been admitted to an intensive care unit for asthma? 1=Yes,0=No","How many times have you been admitted to an intensive care unit for asthma?","Have you ever had invasive mechanical ventilation? 1=Yes, 0=No, 8=Don't Know","Have you ever had non-invasive mechanical ventilation? 1=Yes, 0=No, 8=Don't Know","Do any of the following currently provoke your asthma? 1=Yes, 0=No, 8=Don't Know","Menstrual cycle(If participant is male or a postmenopausal female, leave blank.) (1=Yes, 0=No, 8=Don't Know)","Aspirin or non-steroidal anti-inflammatory drugs (e.g., Aleve, Motrin) (1=Yes, 0=No, 8=Don't Know)","Respiratory infections (e.g., colds)  (1=Yes, 0=No, 8=Don't Know)","Irritants (e.g., pollution, odors, perfumes, chemicals, household cleaners) (1=Yes, 0=No, 8=Don't Know)","Weather conditions (e.g., change in weather, humidity)  (1=Yes, 0=No, 8=Don't Know)","Exposure to cold air  (1=Yes, 0=No, 8=Don't Know)","Emotional factors (e.g., stress, laughing)  (1=Yes, 0=No, 8=Don't Know)","Tobacco smoke  (1=Yes, 0=No, 8=Don't Know)","Food additives/preservatives (e.g., MSG, sulfites)  (1=Yes, 0=No, 8=Don't Know)","Allergies (e.g., dust, animals, pollens)(1=Yes, 0=No, 8=Don't Know)","Other (1=Yes,0=No)","To which of the following did a doctor or other health practitioner say you were allergic? 1=Yes, 0=No, 8=Don't Know","Foods  (1=Yes, 0=No, 8=Don't Know)","Things you breathe in or are exposed to (e.g., dust, pollens, molds, animal fur, feathers, dander)  (1=Yes, 0=No, 8=Don't Know)","Stinging insects such as bees or wasps (1=Yes, 0=No, 8=Don't Know)","Latex  (1=Yes, 0=No, 8=Don't Know)","Other  (1=Yes,0=No)","Have you ever had eczema / atopic dermatitis ? 1=Yes, 0=No, 8=Don't Know","If YES, was your eczema diagnosed by a doctor? 1=Yes,0=No","Have any of your immediate blood relatives been told by a physician that they have allergies/eczema/hay fever? 1=Yes, 0=No, 8=Don't Know","Father  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Brother(s) or Sister(s)(1=Yes, 0=No, 8=Don't Know, 9=N/A)","Child(ren)  (1=Yes, 0=No, 8=Don't Know, 9=N/A)","Did you grow up in a household where you were exposed to tobacco smoke? 1=Yes,0=No","Do you currently smoke? 1=Yes,0=No","Record smoking history in pack-years*. __ __ . __ pack-years","Were you ever a smoker? 1=Yes,0=No","Record smoking history in pack-years*. __ __ . __ pack-years","Do you currently live in a household where you are exposed to tobacco smoke? 1=Yes,0=No","Do you currently vape  or use a hookah ? 1=Yes,0=No","How frequently do you vape or use a hookah? 1=Infrequently , 2=Occasionally , 3=Weekly , 4=Daily","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many days a week","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many times a day","Do you currently vape (i.e., use nicotine or any other substances in an e-cigarette device) or use a hookah (waterpipe)?  If NO, skip to Q16. How frequently do you vape or use a hookah?  If Infrequently or Occasionally, skip to Q17.  How many years have y","Have you ever vaped or used a hookah in the past? 1=Yes,0=No","Have you ever vaped or used a hookah in the past?  If NO, skip to Q17. Approximately how many years did you vape or use a hookah?  (1-99 years)","Do you currently live in a household where you are exposed to others vaping or using a hookah? 1=Yes,0=No","Do you spend time in social settings  where you are exposed to others vaping or using a hookah? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","When was the last time that you vaped or used a hookah? (number of days from Visit 1)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","How many days were you bothered by coughing during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe was your coughing during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by wheezing during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe was your wheezing during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by shortness of breath during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","Over average, how severe was your shortness of breath during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you awakened at night during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how much of a problem was being awakened at night during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","How many days were you bothered by side effects of your asthma medication during the past 2 weeks? 0=Not at all, 1=1-3 days, 2=4-7 days, 3=8-14 days","On average, how severe were the side effects of your asthma medication during the past 2 weeks? 1=Mild, 2=Moderate, 3=Severe","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Participants baseline peak flow (PEF) value (0-999 L/M) First visit (Visit 0A, or Visit 1): PEF (FEF Max) from prebronchodilator (baseline) spirometry at first visit (convert to L/M)","Participants baseline rescue use value (0-99 puffs/day) First visit (Visit 0A, or Visit 1): Self-reported average daily use of albuterol during the 14 days prior to first visit.","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","body mass index computed from height and weight measured at Visit 1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Page Number","Reason for change in scheduled medications status (1=Adverse Event, 2=Other (specify))","Reason for change in scheduled medications status Related adverse event number (0-99)","Current status of participants scheduled medications (1=Discontinued, 2=Resumed)","Date change took effect","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Code to identify a drug unit of measure; (Units)","Code to identify how frequently a drug is taken or administered. (Frequency)","Code to identify the route used to administer a drug. (Route)","Name of Medication","This number corresponds to the ID number of the AHFS monograph. (Medication Code)","Related Event","Medication Dose","Ongoing at current visit","Ongoing at final visit","Start Date","Stop Date","Visit Number (character)","Visit Number (numeric)","GEN_DRUG_NAME","CLASS_TEXT","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Diary and Peak Flow Compliance: Number of full days since the last visit (0-999 days)","Diary and Peak Flow Compliance: Number of days where AM and PM scheduled sessions are complete (AM and PM PEF and all diary questions for AM and PM answered) (0-999 days)","Diary and Peak Flow Compliance: Percent compliance If the compliance value in Q1c is less than 75Pct, re-emphasize the importance of completing scheduled diary assessments and peak flows. (0-150.0Pct)","Scheduled Respimat Compliance (Visits 2-9 only): Number of scheduled puffs since the last visit (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Number of remaining puffs on scheduled Respimat Reference all returned Respimats (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Number of puffs taken (60 x (# used Respimats)) - Q2b (0-999 puffs)","Scheduled Respimat Compliance (Visits 2-9 only): Percent compliance=Q2c/Q2a x 100 If the participant took less than 75Pct of the scheduled Respimat puffs, re-emphasize the importance of maintaining the daily dosing schedule. (0-150.0Pct)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of scheduled puffs since the last visit (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of remaining puffs reflected on scheduled Twisthaler counters Total the values reflected on all counters for all returned Twisthalers (i.e., out of their pouches). (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Number of puffs taken (60 x (# used Twisthalers)) - Q3b (0-999 puffs)","Scheduled Twisthaler Compliance (Visits 4-9 only): Percent compliance=Q3c/Q3a x 100 If the participant took less than 75Pct of the scheduled Twisthaler puffs, re-emphasize the importance of maintaining the daily dosing schedule. (0-150.0Pct)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","1=mometasone, 2=placebo","Blinded Scheduled Twisthaler Contents - How sure are you about your answer in Q1? 1=Absolutely sure - I know what the Twisthaler/MDI contains, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","1=tiotropium, 2=placebo","Blinded Scheduled Respimat Contents - How sure are you about your answer in Q4? 1=Absolutely sure - I know what the Respimat capsules contain, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has the participant or parent/legal guardian signed the SIENA Informed Consent document? 1=Yes,0=No","Ages 12-17 Only: Has the participant signed and dated the assent form or, if the participant is less than the local age of assent, has the participant given verbal assent? 1=Yes,0=No","Has the participant used an inhaled corticosteroid in the past 6 weeks? 1=Yes,0=No","Will the participant be 12 years of age, or older, as of Visit 1? 1=Yes,0=No","Has the participant used an inhaled corticosteroid or combination inhaled corticosteroid/LABA in the past 3 weeks? 1=Yes,0=No","If NO, has the participant used a leukotriene modifier in the past 3 weeks? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used inhaled corticosteroid therapy?","Over the past 3 months, on average, how many days per week has the participant used inhaled corticosteroid therapy: &lt; 5? 1=Yes,0=No","Over the past 3 months, has the participant used greater than the equivalent of 80-240 mcg of inhaled beclomethasone daily? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used combination inhaled corticosteroid/LABA therapy?","Over the past 3 months, on average, how many days per week has the participant used combination inhaled corticosteroid/LABA therapy,  &lt; 5 (1=Yes,0=No)","Over the past 3 months, on average, how many days per week has the participant had daytime asthma symptoms?","Over the past 3 months, on average, how many days per week has the participant had daytime asthma symptoms,  &lt;=2? 1=Yes,0=No","Over the past 3 months, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms?","Over the past 3 months, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms, &lt;=2? 1=Yes,0=No","Over the past 3 months, on average, how many days per week has the participant used his/her short acting betaagonist  for relief of symptoms?","Over the past 3 months, on average, how many days per week has the participant used his/her short acting betaagonist  for relief of symptoms, &lt; 2? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 80Pct? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 70Pct? 1=Yes,0=No","Does the participant have current evidence of any of the conditions listed on the Exclusionary Medical Conditions for SIENA  reference card, or any chronic diseases that would prevent participation in the trial or put the participant at risk by participati","Is the participant currently taking any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Is the participant able to go off these medications for the required washout period prior to Visit 1 and for the duration of the study? 1=Yes,0=No","Is the participant able to use the spirotel e-diary/PEF meter correctly, as evidenced by achieving a score of 13 on the spirotel Performance Checklist ? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Has the participant or parent/legal guardian signed the SIENA Informed Consent document?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past month, on average, how many days per week has the participant had daytime symptoms?","In the past month, on average, how many days per week has the participant had daytime symptoms, &lt;=2? 1=Yes,0=No","In the past month, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms?","In the past month, on average, how many nighttime awakenings per month has the participant had due to asthma symptoms,  &lt;=2? 1=Yes,0=No","In the past month, on average, how many days per week has the participant used his/her short acting beta-agonist  for relief of symptoms?","In the past month, on average, how many days per week has the participant used his/her short acting beta-agonist  for relief of symptoms, &lt; 2? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 80Pct? 1=Yes,0=No","Was the participants FEV1 Pct predicted &gt; 70Pct? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant or parent/legal guardian sign the SIENA Informed Consent document? 1=Yes,0=No","Ages 12-17 Only: Has the participant signed and dated the assent form or, if the participant is less than the local age of assent, has the participant given verbal assent? 1=Yes,0=No","Is the participant 12 years of age, or older? 1=Yes,0=No","Does the participant plan to move away from the clinical site in the upcoming 11 months such that his/her ability to complete the study will be jeopardized? 1=Yes,0=No","Has the participant used investigative drugs and/or enrolled in an intervention trial in the past 30 days, or have plans to enroll in such a trial during the SIENA study? 1=Yes,0=No","Has the participant had a respiratory infection within the past 6 weeks? 1=Yes,0=No","Has the participant had a respiratory infection within the past 4 weeks? 1=Yes,0=No","During the past 4 weeks, on average, how many days per week has the participant had daytime asthma symptoms?","During the past 4 weeks, on average, how many days per week has the participant had daytime asthma symptoms  &gt; 2? 1=Yes,0=No","During the past 4 weeks, how many nighttime awakenings due to asthma symptoms has the participant had?","During the past 4 weeks, how many nighttime awakenings due to asthma symptoms has the participant had,  &gt; 2? 1=Yes,0=No","During the past 4 weeks, on average, how many days per week has the participant used his/her short-acting beta-agonist  for relief of symptoms?","During the past 4 weeks, on average, how many days per week has the participant used his/her short-acting beta-agonist  for relief of symptoms, &gt; 2? 1=Yes,0=No","Is Q7a, Q8a, or Q9a checked YES? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Did the participant or parent/legal guardian sign the SIENA Informed Consent document?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Does the participant have current evidence of any of the conditions listed on the Exclusionary Medical Conditions for SIENA  reference card, or any chronic diseases  that would prevent participation in the trial or put the participant at risk by participat","Does the participant have a history of bladder-neck obstruction, urinary retention, benign prostatic hyperplasia , or a clinically relevant urologic disorder that precludes study participation? 1=Yes,0=No","Does the participant have a history of narrow angle glaucoma? 1=Yes,0=No","Does the participant have a history of significant cardiovascular disorders or arrhythmias? 1=Yes,0=No","Has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card within the specified time periods? 1=Yes,0=No","Is the participant currently taking prescription or OTC medication other than those listed on the Allowed Medications  reference card? 1=Yes,0=No","Based on input from the participant and the study physician, will the participant need to use intranasal steroids at any time during the study? 1=Yes,0=No","Is the participant willing to use a single intranasal steroid at a stable dose continuously for the duration of the study, starting at Visit 1? 1=Yes,0=No","Is the participant currently receiving allergen immunotherapy  other than an established maintenance regimen implemented continuously for a minimum of 3 months? 1=Yes,0=No","Has the participant taken omalizumab within the past 3 months? 1=Yes,0=No","Has the participant used any smokeless tobacco products  in the past year? 1=Yes,0=No","Has the participant smoked cigarettes, a pipe, cigar, marijuana, electronic cigarettes, or any other substance in the past year? 1=Yes,0=No","Ages 18+ Only: Does the participant have a smoking history of greater than 10 pack-years? 1=Yes,0=No","Ages 12-17 Only: Does the participant have a smoking history of greater than 5 pack-years? 1=Yes,0=No","Has the participant received a physician diagnosis of asthma at least 12 months ago? 1=Yes,0=No","Has the participant received a physician diagnosis of asthma at least 12 months ago or had a history consistent with asthma for the previous 12 months? 1=Yes,0=No","Has the participant experienced a life-threatening asthma exacerbation requiring treatment with intubation or mechanical ventilation in the past 5 years? 1=Yes,0=No","Has the participant had an asthma exacerbation requiring systemic corticosteroid treatment in the past 6 weeks? 1=Yes,0=No","Has the participant used an oral or inhaled corticosteroid, or leukotriene modifier in the past 6 weeks? 1=Yes,0=No","Has the participant used an oral corticosteroid in the past 6 weeks? 1=Yes,0=No","Has the participant used an inhaled corticosteroid or leukotriene modifier in the past 3 weeks? 1=Yes,0=No","Is the participant potentially able to bear children? 1=Yes, 0=No, 9=N/A","Is the participant currently pregnant or lactating? 1=Yes,0=No","Does the participant agree to use one of the approved methods indicated on the Birth Control Methods  reference card for the duration of the study? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was the participants prebronchodilator FEV1 &gt;=70Pct predicted? 1=Yes,0=No","Did the participants FEV1 improve &gt;=12Pct and &gt;=200 ml in response to four puffs of albuterol? 1=Yes,0=No","Does the participant have valid source documentation within the past 6 months for an overread AsthmaNet methacholine challenge  with a PC20 &lt;=16 mg/ml? 1=Yes,0=No","Did the participants FEV1 improve &gt;=12Pct and &gt;=200 ml in response to four puffs of albuterol?","Was the participants methacholine PC20 &lt;=16 mg/ml? 1=Yes,0=No","Is the participant able to use the spirotel e-diary/PEF meter correctly, as evidenced by achieving a score of 13 on the Spirotel Performance Checklist ? 1=Yes, 0=No, 9=N/A","Has the participant used the Twisthaler previously? 1=Yes,0=No","Is the participant able to use a Twisthaler properly, as evidenced by achieving a score of 12 on the Twisthaler Inhalation Technique Checklist ? 1=Yes,0=No","Is the participant able to use a metered dose inhaler properly, as evidenced by achieving a score of 11 on the MDI Inhalation Technique Checklist  ? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Does the participant have valid source documentation within the past 6 months for an overread AsthmaNet methacholine challenge  with a PC20 &lt;=16 mg/ml?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Since Visit 1, has the participant experienced a treatment failure event as defined in the protocol? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Since Visit 1, has the participant experienced one or more asthma exacerbations as defined in the protocol? 1=Yes,0=No","Since Visit 1, has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant provide two acceptable sputum induction samples during the run-in? 1=Yes,0=No","Since Visit 1, has the participant experienced a treatment failure event as defined in the protocol? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Did the participant complete Asmanex220 mcg treatment at least 6 weeks ago? 1=Yes,0=No","Has the participant experienced two or more treatment failures? 1=Yes,0=No","Since Visit 1, has the participant experienced one or more asthma exacerbations as defined in the protocol? 1=Yes,0=No","Since Visit 1, has the participant taken any medications listed on the Exclusionary Drugs for SIENA  reference card? 1=Yes,0=No","Has the participant experienced a respiratory tract infection in the past 6 weeks? 1=Yes,0=No","Has the participant experienced a respiratory tract infection in the past 4 weeks? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report daytime asthma symptoms daily during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report night-time awakenings due to asthma symptoms more than once per week during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report night-time awakenings due to asthma symptoms more than twice per week during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant report daily albuterol use for asthma symptom control during the run-in? 1=Yes,0=No","According to the Spirotel SIENA Eligibility Report: Did the participant record at least 75Pct of AM and PM peak flow measurements and symptoms on his or her spirotel during the run-in? 1=Yes,0=No","Did the participant take at least 75Pct of the required puffs from his or her Respimat during the run-in? 1=Yes,0=No","Was the participants prebronchodilator FEV1 &gt;=70Pct predicted? 1=Yes,0=No","Has the participant used the Respimat previously? 1=Yes,0=No","Is the participant able to use a Respimat inhaler properly, as evidenced by achieving a score of 15 on the Respimat Inhalation Technique Checklist? 1=Yes,0=No","Does the participant wish to withdraw consent from the study? 1=Yes,0=No","Is there any new information that makes the participant ineligible according to the eligibility criteria? 1=Yes,0=No","Does the participant have any condition or issue which, in the opinion of the investigator, might interfere with study participation? 1=Yes,0=No","Is the participant eligible to proceed? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has QC procedure been performed on the NIOX MINO today? 1=Yes,0=No","Did the participant eat or drink within the past hour? 1=Yes,0=No","Did the participant take part in strenuous activity/exercise within the past hour? 1=Yes,0=No","Time eNO started (based on a 24-hour clock) (0000-2359)","ENO Measurement","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","How long have you lived in the current house? (Estimate if uncertain.) (1010) years","How long have you lived in the current house? (Estimate if uncertain.)(1020) months","Does your house use a wood burning stove as a primary source of heat? 1=Yes, 0=No, 8=Don't Know","Does your house use an air conditioner? 1=Yes, 0=No, 8=Don't Know","Does your house use an evaporative cooler ? 1=Yes, 0=No, 8=Don't Know","Does your house use a humidifier? 1=Yes, 0=No, 8=Don't Know","Does your house use a dehumidifier? 1=Yes, 0=No, 8=Don't Know","Has there been water damage to your house, basement, or its contents during the past 12 months? 1=Yes, 0=No, 8=Don't Know","Has there been any mold or mildew, on any surfaces, inside your house in the past 12 months? 1=Yes, 0=No, 8=Don't Know","Which rooms have or have had mold or mildew?10a.Bathroom(s)  (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew?10b.Basement or attic (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10c.Kitchen (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10d.Your bedroom (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10e.Other bedrooms (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10f.Living or family room (1 = Yes, 0 = No)","Which rooms have or have had mold or mildew? 10g.Other (1 = Yes, 0 = No)","Do you ever see cockroaches in your house? 1=Yes,0=No","In which room(s) have you seen cockroaches?12a.Kitchen (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches?  12b.Basement or attic (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12c.Bathroom(s) (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12d.Living or family room (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12e.Your bedroom  (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12f.Other bedrooms (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12g.Garage (1 = Yes, 0 = No)","In which room(s) have you seen cockroaches? 12h.Other (1 = Yes, 0 = No)","Do you ever see rodents  or rodent droppings in your house? 1=Yes,0=No","14.In which room(s) have you seen rodents or rodent droppings? 14a. Kitchen (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14b. Basement or attic (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14c. Bathroom(s) (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings?14d. Living or family room (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14e. Your bedroom(1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14f. Other bedrooms (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14g. Garage (1 = Yes, 0 = No)","14.In which room(s) have you seen rodents or rodent droppings? 14h. Other (1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15a. Barns(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property?15b. Hay(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15c. Woodsheds(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15d. Firewood(1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property?15e. Chicken coops (1 = Yes, 0 = No)","15.Are any of the following located on your property or next to your property? 15f. Corral (1 = Yes, 0 = No)","CHARACTERISTICS OF THE PARTICIPANT'S BEDROOM16.What is the floor covering in your bedroom? 1=Rug/carpet, 2=Vinyl tile or linoleum, 3=Wood, 4=Ceramic tile, 5=Other , 9=Don't know","What type of mattress is on your bed? 1=None, 2=Inner spring mattress, 3=Foam mattress, 4=Waterbed, 5=Air mattress, 6=Other , 9=Don't know","Is the mattress completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","Does your bed have a box spring? 1=Yes,0=No","Is the box spring completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","What type of pillow do you usually sleep with? 1=None, 2=Feather/down, 3=Foam/Dacron/synthetic, 5=Other , 9=Don't know","Is the pillow completely enclosed in an allergy-proof, encasing cover? 1=Yes,0=No","Does your household have any pets? 1=Yes,0=No","Enter the number of pets that the household has. (Enter '00' if none. If none to Q24a - Q24d, skip to the next question.)24a.Cat","Cat  (1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has. Dog","Dog(1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has 24c.Rabbit, guinea pig, hamster, gerbil, or mouse","Rabbit, guinea pig, hamster, gerbil, or mouse (1=Indoor, 2=Outdoor, 3=Both)","Enter the number of pets that the household has 24d.Bird","Bird (1=Indoor, 2=Outdoor, 3=Both)","In general, and on a regular basis, are you exposed to any of the following animals?  25a.Cat(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25b.Dog(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25c.Rabbit, guinea pig, hamster, gerbil, or mouse(1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25d.Bird (1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals? 25e.Farm animals (1 = Yes, 0 = No)","In general, and on a regular basis, are you exposed to any of the following animals?  25f.Other (1 = Yes, 0 = No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other (specify","Highest education of any household member (1=No High School diploma/GED, 2=High Schol grad, 3=Technical training, 4=Some college, no degree, 5=Assoc degree, 6=Bachelor degree, 7=Masters degree/MD/PhD/JD/PharmD, 9=Decline to answer, 10=Don't know)","Category best describes the combined annual income, before taxes, of all members of your household for the last year. (1=Less than $25,000, 2=$25,000 - $49,999, 3=$50,000 - $99,999, 4=$100,000 or more, 9=Decline to answer, 10=Don't know)","How many people  are supported by this income reported in Q3?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Total IgE","Cat (E1)","Dog (E5)","Mouse (E72)","Rat (E74)","Mold mix (MX1)","Cockroach, German (I6)","Grass mix (GX2)","Tree mix (TX4)","Tree mix (TX6)","Weed mix (WX1)","Weed (W3)","Mite (D2)","Mite (D1)","Number or allergen specific (Cat, Dog, Mouse, etc) that were either POS or greater than 0.35","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","CBC with differential cell count: Eosinophils (absolute count) (0-9999/uL)","CBC with differential cell count: WBC (0-999.9 K/uL)","CBC with differential cell count: Differential Lymphocytes (0-99.9Pct)","CBC with differential cell count: Differential Monocytes (0-99.9Pct)","CBC with differential cell count: Differential Basophils (0-99.9Pct)","CBC with differential cell count: Differential Neutrophils (0-99.9Pct)","CBC with differential cell count: Differential Eosinophils (0-99.9Pct)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","PARENTAL HEIGHT - First study visit only or until both are completed 1.Biological mother's height (complete height or check unknown) __ feet","PARENTAL HEIGHT - First study visit only or until both are completed1.Biological mother's height (complete height or check unknown) __ __ inches","PARENTAL HEIGHT - First study visit only or until both are completed 1.Biological mother's height (complete height or check unknown) (9=Don't Know)","Biological father's height (complete height or check unknown) __ feet","Biological father's height (complete height or check unknown) __ __ inches","Biological father's height (complete height or check unknown) (9=Don't Know)","PARTICIPANT MEASUREMENTS - Complete at all applicable study visits3. What type of height measurement was obtained? 1=Standing height, 2=Length","First measurement __ __ __ . __ cm","Second measurement __ __ __ . __ cm","Third measurement __ __ __ . __ cm","Average height or length measurement __ __ __ . __ cm","In your judgment, was the participant's height or length measurement acceptable? 1=Yes,0=No","Weight (shoes off, light clothing) __ __ __ . __ kg","ORAL CANDIDIASIS5.Does the participant have evidence of oral candidiasis? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Post Diluent FEV1","Did the participant drop20Pct at the diluent stage? 1=Yes,0=No","Last concentration of methacholine administered __ __ . __ __ __ __ mg/ml","FEV1 after last concentration of methacholine administered","Did the participant achieve a PC20? 1=Yes,0=No","PC20 __ __ . __ __ mg/ml","Time methacholine challenge ended (based on 24-hour clock)","FEV1","Time of FEV1 in Q7a (based on 24-hour clock)","Was the FEV1 from Q7athe methacholine reversal reference value  in the gray box above? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was an additional treatment used in the first hour? 1=Yes,0=No","Additional albuterol by MDI (1=Yes,0=No)","Number of additional puffs of albuterol administered (1=2, 2=4, 3=&gt;4)","Nebulized Beta-agonist(1=Yes,0=No)","Subcutaneous epinephrine(1=Yes,0=No)","Implementation of clinic emergency protocol or algorithm (1=Yes,0=No)","Other  (1=Yes,0=No)","Participant's FEV1 after additional treatment within first hour.2a.FEV1","Time of FEV1 in Q2a (based on 24-hour clock)","Was the FEV1 from Q2a &gt;the methacholine reversal reference value  in the gray box on the Methacholine Challenge Testing  form? 1=Yes,0=No","Was additional treatment used after one hour? 1=Yes,0=No","Additional albuterol by MDI (1=Yes,0=No)","Number of additional puffs of albuterol administered (1=2, 2=4, 3=&gt;4)","Nebulized Beta-agonist  (1=Yes,0=No)","Subcutaneous epinephrine(1=Yes,0=No)","Implementation of clinic emergency protocol or algorithm  (1=Yes,0=No)","Treatment in the emergency room  (1=Yes,0=No)","Overnight hospitalization (1=Yes,0=No)","Other (1=Yes,0=No)","Participant's final FEV1 after methacholine challenge4a.FEV1","Time of FEV1 in Q4a (based on 24-hour clock)","Was the FEV1 from Q4a ? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Exclusions and Confounders1.Has the participant had any severe acute illness in the past 4 weeks? 1=Yes,0=No","If YES, has the participant received permission from the supervising physician to proceed with the methacholine challenge testing? 1=Yes,0=No","Has the participant used an oral or injectable steroid  in the last 4 weeks? 1=Yes,0=No","Does the participant have a baseline  FEV1 less than 55Pct of predicted or less than 1.0 L? 1=Yes,0=No","Pregnancy test results:  (Check N/A if the participant is male, or is female and is post-menopausal, had a hysterectomy or tubal ligation.) (1=Positive, 0=Negative, 9=N/A)","Is the participant's systolic blood pressure &gt; 200 mm Hg or diastolic blood pressure &gt; 100 mm Hg? 1=Yes,0=No","Is there any other reason the participant should not proceed with the methacholine challenge testing? 1=Yes,0=No","Is the participant eligible to proceed with the diluent  pulmonary function testing for the methacholine challenge? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Exclusions and Confounders:  1.Has the participant had any severe acute illness in the past 4 weeks? 1=Yes,0=No","If YES, has the participant received permission from the supervising physician to proceed with the methacholine challenge testing? 1=Yes,0=No","During the past 4 weeks, has the participant had any respiratory infections, colds, or bronchitis ? 1=Yes,0=No","If YES, during the past 2 weeks, has the participant had any respiratory infections, colds, or bronchitis ? 1=Yes,0=No","Has the participant used an oral or injectable steroid  in the last 4 weeks? 1=Yes,0=No","Does the participant have a baseline  FEV1 less than 70Pct of predicted? 1=Yes,0=No","Pregnancy test results(Check N/A if the participant is male, or is female and has not started menses.)(1=Positive, 0=Negative, 9=N/A)","If participant's age is ? 1=Yes,0=No","If participant's age is &lt; 12 years: Is the participant's systolic blood pressure &gt; 180 mm Hg or diastolic blood pressure &gt; 90 mm Hg? 1=Yes,0=No","Is there any other reason the participant should not proceed with the methacholine challenge testing? 1=Yes,0=No","Is the participant eligible to proceed with the diluent  pulmonary function testing for the methacholine challenge? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time albuterol administered (based on 24-hour clock)","Time post-albuterol spirometry started (based on 24-hour clock)","Highest FVC","Highest FEV1","Highest FEV1 (Pct predicted)","FEF Max","FEF25-75","In your judgment, was the subject's spirometry technique acceptable? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Please check the box next to the treatment that you believe you received over the past 12 weeks. 1=mometasone, 2=placebo","Blinded Scheduled Twisthaler Contents: How sure are you about your answer to Question 2? 1=Absolutely sure -  I know what the MDI contains, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","Please comment with respect to any observations you made that helped you make your choice in Question 2 . 1=I have no comments, 2=I noticed the following:","Please check the box next to the treatment that you believe you received over the past 12 weeks. 1=tiotropium, 2=placebo","How sure are you about your answer to Question 5? 1=Absolutely sure - I know what the Respimat capsules contain, 2=Moderately sure, 3=Somewhat sure, 4=Not sure at all - purely a guess","1=I have no comments, 2=I noticed the following:","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Visit Number (numeric)","Level of periostin in peripheral blood (ng/ml)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time ipratropium administered (based on 24-hour clock) (0000-2359)","Time post-ipratropium spirometry started (based on 24-hour clock) (0000-2359)","Highest FVC (0-9.99 L)","Highest FEV1 (0-9.99 L)","Highest FEV1 (Pct predicted) (0-999 Pct predicted)","FEF Max (0-99.99 L/S)","FEF25-75 (0-9.99 L/S)","In your judgment, was the participants spirometry technique acceptable? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Is the participant unable to bear children due to any of the following reasons? 1=Yes,0=No","Post-menopausal (at least one year since last menses) (1=Yes,0=No)","Hysterectomy(1=Yes,0=No)","Tubal ligation(1=Yes,0=No)","Pregnancy test results  (1=Positive, 0=Negative)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Have you had any diseases, illnesses, conditions, or surgeries related Blood, Lymph, or Immune Systems? 1=Yes,0=No","Eyes(1=Yes,0=No)","Breasts (1=Yes,0=No)","Endocrine Systems (1=Yes,0=No)","Heart and Blood Vessels  (1=Yes,0=No)","Liver or Pancreas (1=Yes,0=No)","Kidneys or Urinary Tract System  (1=Yes,0=No)","Reproductive System (1=Yes,0=No)","Muscles or Bones (1=Yes,0=No)","Nervous System (1=Yes,0=No)","Psychiatric(1=Yes,0=No)","Drug Allergies (1=Yes,0=No)","Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Skin (1=Yes,0=No)","Have you ever had allergic rhinitis ? 1=Yes, 0=No, 9=Don't know","Have you ever had nasal polyps? 1=Yes, 0=No, 9=Don't know","Do you have chronic or recurrent sinusitis ? 1=Yes, 0=No, 9=Don't know","Have you ever been diagnosed with vocal cord dysfunction? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the ear, nose, or throat? 1=Yes,0=No","Have you ever had pneumonia? 1=Yes, 0=No, 9=Don't know","If YES, were you diagnosed by chest x-ray? 1=Yes, 0=No, 9=Don't know","If YES, were you treated with antibiotics? 1=Yes, 0=No, 9=Don't know","Have you ever had bronchitis? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the lungs ? 1=Yes, 0=No, 9=Don't know","Do you have gastroesophageal reflux disease ? 1=Yes, 0=No, 9=Don't know","Have you ever had other conditions related to the stomach or intestines? 1=Yes,0=No","Have you been diagnosed with sleep disordered breathing ? 1=Yes,0=No","If YES, are you being treated with CPAP or BiPAP? 1=Yes,0=No","Have you ever had other sleep disorders? 1=Yes,0=No","Have you ever had other conditions that have not been mentioned on this form? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Who is the respondent? 1=Self/Participant, 2=Parent/Guardian, 3=Other","Short-acting Inhaled Beta-Agonists by Inhaler (e.g., albuterol, Primatene Mist, Maxair, ProAir, Proventil, Ventolin, Xopenex)  (1=Yes, 0=No, 9=Don't Know)","If YES, indicate average weekly puffs in the past month (Enter '000' if none used) weekly puffs","Rescue treatment via a Nebulizer Machine (e.g., albuterol, ipratropium, Combivent, Xopenex, levalbuterol)  (1=Yes, 0=No, 9=Don't Know)","Long-acting Inhaled Beta-Agonists (e.g., Serevent, Foradil, salmeterol, formoterol) Do not consider combination medications. (1=Yes, 0=No, 9=Don't Know)","Oral Beta-Agonists (e.g., albuterol, Brethine, Bricanyl, metaproterenol, Proventil, Ventolin, Repetabs, Volmax)  (1=Yes, 0=No, 9=Don't Know)","Oral Theophylline (short-acting or sustained release)(e.g., Aminophylline, Slo-Phyllin, Slo-bid, Theo-Dur, Uniphyl)(1=Yes, 0=No, 9=Don't Know)","Inhaled Anticholinergic by Inhaler (e.g., Atrovent, Combivent, Spiriva) (1=Yes, 0=No, 9=Don't Know)","Leukotriene Antagonist / 5LO Inhibitors (e.g., Accolate, Zyflo, Singulair) (1=Yes, 0=No, 9=Don't Know)","IgE Blocker (e.g., Xolair)(1=Yes, 0=No, 9=Don't Know)","Oral Steroids FOR ASTHMA (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron,  dexamethasone)(1=Yes, 0=No, 9=Don't Know)","If YES, in the past 12 months, how many courses of steroids by mouth have you taken FOR ASTHMA? 1=1 course, 2=2 courses, 3=3 courses, 4=4 courses, 5=5 courses, 6=More than 5 courses","Injectable Steroids FOR ASTHMA (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV)(1=Yes, 0=No, 9=Don't Know)","Steroids by Inhaler (e.g., Asmanex Twisthaler, QVAR, Flovent, Pulmicort Flexhaler)(1=Yes, 0=No, 9=Don't Know)","Indicate most recent type of inhaled steroid taken (refer to PRIOR_TRT_CARD reference card) code","Indicate number of daily puffs used daily puffs","Indicate the total number of months that you used the inhaled steroid out of the past 12 months months","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide) (1=Yes, 0=No, 9=Don't Know)","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide). If YES, complete Q13a - Q13c. Indicate most recent type of nebulized steroid taken (refer to PRIOR_TRT_CARD reference card)","Indicate number of daily treatments used daily treatments","Indicate the total number of months that you used the nebulized steroid out of the past 12 months months","Long-Acting Beta-Agonist and Inhaled Steroid Combination Medications (e.g., Advair Diskus, Symbicort MDI, Dulera MDI)(1=Yes, 0=No, 9=Don't Know)","Indicate most recent type of combination medication taken (refer to PRIOR_TRT_CARD reference card) code","Indicate number of daily puffs used daily puffs","Indicate the total number of months that you used the combination medication out of the past 12 months months","Nasal Steroids (e.g., Beconase, Vancenase, Flonase, Nasacort, Nasalide, Nasarel, Omnaris, Rhinocort, Nasonex)(1=Yes, 0=No, 9=Don't Know)","Non-steroidal Anti-allergic Nasal Medications (e.g., Nasalcrom, Astelin, Astepro, ipratropium)(1=Yes, 0=No, 9=Don't Know)","Anti-allergic Oral Medications (e.g., fexofenadine, loratadine, cetirizine, chlorpheniramine) (1=Yes, 0=No, 9=Don't Know)","Topical Steroids - Prescription (e.g., Synalar, Lidex, Dermacin, Fluocinonide) (1=Yes, 0=No, 9=Don't Know)","Topical Steroids - OTC (e.g., Hydrocortisone - multiple strengths and products) (1=Yes, 0=No, 9=Don't Know)","Other Medication FOR ASTHMA OR ALLERGIES (1=Yes, 0=No, 9=Don't Know)","Oral Steroids for Conditions Other Than Asthma (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) (1=Yes, 0=No, 9=Don't Know)","Injectable Steroids for Conditions Other Than Asthma (e.g., Medrol, Solumedrol, Decadron, dexamethasone,  triamcinolone, Kenalog, hydrocortisone IV)(1=Yes, 0=No, 9=Don't Know)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Short-acting Inhaled Beta-Agonists by Inhaler (e.g., albuterol, Primatene Mist, Maxair, ProAir, Proventil, Ventolin, Xopenex) Date","Rescue treatment via a Nebulizer Machine (e.g., albuterol, ipratropium, Combivent, Xopenex, levalbuterol) Date","Long-acting Inhaled Beta-Agonists (e.g., Serevent, Foradil, salmeterol, formoterol) Do not consider combination medications. Date","Oral Beta-Agonists (e.g., albuterol, Brethine, Bricanyl, metaproterenol, Proventil, Ventolin, Repetabs, Volmax) Date","Inhaled Anticholinergic by Inhaler (e.g., Atrovent, Combivent, Spiriva) Date","Leukotriene Antagonist / 5LO Inhibitors (e.g., Accolate, Zyflo, Singulair) Date","IgE Blocker (e.g., Xolair) Date","Oral Steroids FOR ASTHMA (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) Date","Injectable Steroids FOR ASTHMA (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV) Date","Steroids by Inhaler (e.g., Asmanex Twisthaler, QVAR, Flovent, Pulmicort Flexhaler) Date","Steroids by Nebulizer (e.g., Pulmicort Respules, budesonide) Date","Long-Acting Beta-Agonist and Inhaled Steroid Combination Medications (e.g., Advair Diskus, Symbicort MDI, Dulera MDI) Date","Nasal Steroids (e.g., Beconase, Vancenase, Flonase, Nasacort, Nasalide, Nasarel, Omnaris, Rhinocort, Nasonex) Date","Non-steroidal Anti-allergic Nasal Medications (e.g., Nasalcrom, Astelin, Astepro, ipratropium) Date","Anti-allergic Oral Medications (e.g., fexofenadine, loratadine, cetirizine, chlorpheniramine) Date","Topical Steroids - Prescription (e.g., Synalar, Lidex, Dermacin, Fluocinonide) Date","Topical Steroids - OTC (e.g., Hydrocortisone - multiple strengths and products) Date","Other Medication FOR ASTHMA OR ALLERGIES Date","Oral Steroids for Conditions Other Than Asthma (e.g., Prednisone, Prelone, Pediapred, Medrol, Orapred, Decadron, dexamethasone) Date","Injectable Steroids for Conditions Other Than Asthma (e.g., Medrol, Solumedrol, Decadron, dexamethasone, triamcinolone, Kenalog, hydrocortisone IV) Date","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the last month, how often have you been upset because of something that happened unexpectedly? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that you were unable to control the important things in your life? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt nervous and stressed? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt confident about being able to handle your personal problems? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that things were going your way? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you found that you could not cope with all the things that you had to do? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you been able to control irritations in your life? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that you were on top of things? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you been angered because of things that happened that were outside of your control? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","In the last month, how often have you felt that your difficulties were piling up so high that you could not overcome them? 0=Never, 1=Almost never, 2=Sometimes, 3=Fairly often, 4=Very often","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Have you consumed caffeine in the past 4 hours? 1=Yes,0=No","Have you used medications with caffeine in the past 4 hours? 1=Yes,0=No","Have you used any weight loss medications in the past 4 hours? 1=Yes,0=No","Have you consumed any food containing alcohol or beverages containing alcohol in the past 4 hours? 1=Yes,0=No","Have you used a rescue intermediate-acting inhaled betaagonist in the past 6 hours? 1=Yes,0=No","Have you used a short-acting anticholinergic in the past 6 hours? 1=Yes,0=No","Have you used any ophthalmic antihistamines in the past 6 hours? 1=Yes,0=No","Have you used any nasal antihistamines in the past 6 hours? 1=Yes,0=No","Have you used any nasal decongestants in the past 6 hours? 1=Yes,0=No","Have you used any oral antihistamines in the past 48 hours? 1=Yes,0=No","Have you used any oral decongestants or cold remedies in the past 48 hours? 1=Yes,0=No","Have you used any smokeless tobacco products today? 1=Yes,0=No","At this time, is your asthma worse because of recent exposure to triggers? 1=Yes,0=No","Is there any other reason you should not proceed with spirometry testing? 1=Yes,0=No","Is the participant eligible to proceed with the spirometry testing? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","In the past 4 weeks, I worried about the long-term effects of asthma on my health (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I had to worry about asthma triggers (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, my asthma was on my mind (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, it was hard to get a good nights sleep because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt like I couldnt enjoy life because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt that asthma was controlling my life (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt frustrated that I couldnt make plans in advance because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, everyday activities were a struggle (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, asthma placed stress on my relationships with family, friends, significant others, or co-workers (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, I felt frustrated that I have to do things differently than people who dont have asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, I felt like I missed out on doing things with others because of my asthma (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","In the past 4 weeks, because of my asthma, I had to do a lot of planning to make sure I always had an inhaler ready (1=Not at all, 2=A little bit, 3=Somewhat, 4=Quite a bit, 5=Very much)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Randomized treatment assignment period 1: (Mometasone, Tiotropium or Placebo)","Randomized treatment assignment period 2: (Mometasone, Tiotropium or Placebo)","Randomized treatment assignment period 3: (Mometasone, Tiotropium or Placebo)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Sex (1=Male, 2=Female)","Primary Racial Identification for Spirometry (1=Amer Indian or Alaskan Native, 2=Asian or Pacific Islander, 3=Black, 4=White, 5=Hispanic or Latino, 6=Other)","Participant's Age at Visit 1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Description of Adverse Event (ICD9 Code)","Is the participant currently taking study drug? 1=Yes,0=No","Time interval between the last administration of the study drug and the Adverse Event","What was the unit of time for the interval in Question #4? 1=Second, 2=Minute, 3=Hour, 4=Day","Fatal event(1=Yes,0=No)","Life-threatening event (1=Yes,0=No)","Inpatient hospitalization required(1=Yes,0=No)","Hospitalization prolonged  (1=Yes,0=No)","Disabling or incapacitating  (1=Yes,0=No)","Overdose (1=Yes,0=No)","Cancer (1=Yes,0=No)","Congenital anomaly(1=Yes,0=No)","Serious laboratory abnormality with clinical symptoms(1=Yes,0=No)","Height failure (per protocol MOP)(1=Yes,0=No)","Pregnancy(1=Yes,0=No)","Other (1=Yes,0=No)","What in your opinion caused the event? 1=Yes,0=No","Withdrawal of study drug(s)(1=Yes,0=No)","Concurrent medication  (1=Yes,0=No)","Other condition or event  (1=Yes,0=No)","Was the event expected or unexpected? 1=Expected, 2=Unexpected","Was the event possibly, probably, or definitely related to study participation? 1=Yes,0=No","Date of Adverse Event","Admission date","Discharge date","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","What type of height measurement was obtained? 1=Standing height, 2=Length","First measurement __ __ __ . __ cm","Second measurement __ __ __ . __ cm","Third measurement __ __ __ . __ cm","Average height or length measurement __ __ __ . __ cm","In your judgment, was the participant's height or length measurement acceptable? 1=Yes,0=No","Weight (shoes off, light clothing) __ __ __ . __ kg","Does the participant have evidence of oral candidiasis? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant fail to respond within 48 hours to the treatment failure rescue algorithm? 1=Yes,0=No","Did the participant use at least 16 puffs as needed albuterol per 24 hours for a period of 48 hours? 1=Yes,0=No","Did the participant experience prebronchodilator FEV1 values &lt; 50Pct of the baseline prebronchodilator value obtained at Visit 1 on two consecutive spirometric determinations made on different days? 1=Yes, 0=No, 9=Not evaluated","Did the participant experience prebronchodilator FEV1 values &lt; 40Pct of predicted on two consecutive spirometric determinations made on different days? 1=Yes, 0=No, 9=Not evaluated","Did the study or treating physician prescribe the participant oral/parenteral corticosteroids for the treatment of his/her asthma? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation in the opinion of the study investigator or personal physician? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation? 1=Yes,0=No","Did the participant seek care for significant asthma exacerbation conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Was the participant hospitalized? 1=Yes,0=No","Duration of hospital stay (0-99.9 days)","Was intubation or ventilation assistance required? 1=Yes,0=No","Was the participant admitted to the intensive care unit? 1=Yes,0=No","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Inhaled corticosteroids (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Nebulized bronchodilator (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Oral corticosteroids (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, IM or IV steroids (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Antibiotics (1=Yes, 0=No )","Has the participant taken any of the following medications (excluding study ICS) since significant asthma exacerbation conditions started, Other (1=Yes, 0=No )","Why do you think the participant experienced a significant asthma exacerbation? 1=Common cold, 2=Allergies, 3=Pollution or chemical irritant, 4=Too little asthma maintenance medication, 5=Exercise, 6=Other","Date exacerbation conditions were met","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Which phone contact following Asthma Exacerbation visit is being performed? 1=Phone Contact #1 , 2=Phone Contact #2 , 3=Phone Contact #3","Since the most recent Asthma Exacerbation contact, has the participant sought additional care for significant asthma exacerbation conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Was the participant hospitalized? 1=Yes,0=No","Duration of hospital stay (0-99.9 days)","Was intubation or ventilation assistance required? 1=Yes,0=No","Was the participant admitted to the intensive care unit? 1=Yes,0=No","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Inhaled corticosteroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Nebulized bronchodilator (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Oral corticosteroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, IM or IV steroids (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Antibiotics (1=Yes,0=No)","Due to persistent symptoms, has the participant started taking any of the following medications since the most recent Asthma Exacerbation contact, Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Runny Nose  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Post nasal drip  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Need to blow your nose  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Facial pain/pressure(0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Nasal obstruction  (0=Never1=1-4 times per month, 2=2-6 times per week, 3=Daily)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Time spirometry started (based on 24-hour clock)","The reported FEV1 and FVC are the best volumes of all acceptable maneuvers. 1.Highest FVC","Highest FEV1 (Pct predicted)","The reported flow rates correspond to the maneuver where FEV1 + FVC is maximized. FEF Max","FEF25-75","In your judgment, was the subject's spirometry technique acceptable? 0=No, 1=Yes","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Highest FEV1","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","TRIAL_TYPE","FVC","FEV1","PEF","FEF2575","FET","Number of times the participant woke up last night due to asthma symptoms # Times woke up due to asthma (0-9)","Number of puffs the participant will take from the WHITE study inhaler this morning # Puffs from WHITE inhaler in AM (0-9)","only) Number of puffs the participant will take from the BLUE study inhaler this morning # Puffs from BLUE inhaler in AM (0-9)","Has the participant taken any puffs from his/her RED RESCUE albuterol inhaler in the past 4 hours? 1=yes, 0=no","Shortness of Breath Score Shortness of Breath Score overnight (0,1,2,3)","Chest tightness score Chest Tightness Score overnight (0,1,2,3)","Wheezing score Wheezing Score overnight (0,1,2,3)","Coughing score Coughing Score overnight (0,1,2,3)","Phlegm/Mucus score Phlegm/Mucus Score overnight (0,1,2,3)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","date 'fixed' for analysis so that associated AM and PM always have same ddate","TRIAL_TYPE","FVC","FEV1","PEF","FEF2575","FET","Number of puffs the participant will take from the WHITE study inhaler tonight # Puffs from WHITE inhaler in PM (0-9)","Has the participant taken any puffs from his/her RED RESCUE albuterol inhaler during the past 4 hours? 1=yes, 0=no","Shortness of Breath Score Shortness of Breath Score since waking this morning (0,1,2,3)","Chest tightness score Chest Tightness Score since waking this morning (0,1,2,3)","Wheezing score Wheezing Score since waking this morning (0,1,2,3)","Coughing score Coughing Score since waking this morning (0,1,2,3)","Phlegm/Mucus score Phlegm/Mucus Score since waking this morning (0,1,2,3)","Did your regular exercise cause unusually severe shortness of breath during the past 24 hours? 1=Yes, 0=No, 9=N/A","Number of RED RESCUE albuterol puffs taken in the past 24 hours to prevent symptoms# Preventive  RED RESCUE puffs  in past 24 hours?","Number of RED RESCUE albuterol puffs taken for asthma symptoms or low peak flow during past 24 hours  # RED RESCUE  puffs for symptoms  in past 24 hours?","Number of times used RED RESCUE albuterol inhaler for asthma symptoms during past 24 hours  # Times used  RED RESCUE for symptoms in past 24 hours?","Did RED RESCUE albuterol relieve symptoms for less than 4 hours after treatment during past 24 hours? 1=Yes,0=No","Was the participant seen by a healthcare provider  for an unscheduled visit in the past 24 hours due to asthma symptoms? 1=Yes,0=No","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Processing: Time processing started (based on 24-hour clock) __ __ __ __ (0 - 2359)","Processing Sample: Total Cell Count: __ __ __ __. __x 10^4 cells/ml (0 - 9999.9)","Was the participant's sputum sample processed within 4 hours after collection? 1=Yes,0=No","Processing Sample: Processing Date __ __ / __ __ / 20 __ __","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Rate slide's quality:  (1=Very good, 2=Good, 3=Acceptable, 4=Poor but readable, 5=Not readable)","Record the number on the slide(s) that was (were) read (0 - 9)","Record the number on the slide(s) that was (were) read: These are numbers that were assigned to the slides at each site. (0 -9)","Total Cell Count: __ __ __ __ . __ x 10^4 cells/ml (0 - 9999.9)","Differential Cell Counts: Squamous Cells __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Epithelial Cells __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Macrophages __ __ . __ Pct (0.0 - 99.9)","Differential Cell Counts: Neutrophils __ __ . __ Pct (0 -99.9)","Differential Cell Counts: Eosinophils __ __ . __ Pct (0 - 99.9)","Differential Cell Counts: Lymphocytes __ __ . __ Pct (0 -99.9)","Date of Read __ __ / __ __ / 20 __ __","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","For this protocol, what was the duration of sputum induction the first time the participant's sample was processed within 4 hours after collection?","Sputum induction start time (based on 24-hour clock)","Sputum induction stop time (based on 24-hour clock)","Duration of sputum induction collection phase at this visit -- minutes (0 - 99.9)","Was the duration ? 1=Yes,0=No","Volume of sputum sample at this visit -- ml (0.0 - 999.9)","Is the volume adequate for processing? 1=Yes,0=No","Is the sample adequate for laboratory analysis? 1=Yes,0=No","Participant's FEV1 immediately after completion of sputum induction: -- (0.0 -9.99 L)","FEV1 (Pct predicted) (0 - 999)","Time of FEV1 in Q7a (based on 24-hour clock)","Percent difference in FEV1 ((Reference - Q7a) / Reference) X100 (-99.9 - 99.9 Pct)","Did the participant's FEV1 drop &gt; 10Pct from reference FEV1 as indicated in Q7d? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Participant's FEV1 after initial 2 puffs of albuterol: 1a.FEV1 __ . __ __ L (0 -9.99)","Participant's FEV1 after initial 2 puffs of albuterol: 1b. FEV1 (Pct predicted)__ __ __ Pct predicted (0 - 999)","Participant's FEV1 after initial 2 puffs of albuterol: 1c. Time of FEV1 from Q1a (based on 24-hour clock) __ __ __ __ (0 - 2359)","Participant's FEV1 after initial 2 puffs of albuterol: 1d. Was the FEV1 from Q1a &gt;=the sputum induction reversal reference value in the gray box above? 1=Yes,0=No","Participant's FEV1 after 2 additional puffs of albuterol: 2a. FEV1 __ . __ __ L (0 - 9.99)","Participant's FEV1 after 2 additional puffs of albuterol: 2b. FEV1 (Pct predicted) __ __ __ Pct predicted (0 - 999)","Participant's FEV1 after 2 additional puffs of albuterol: 2c. Time of FEV1 from Q2a (based on 24-hour clock)__ __ __ __ (0 - 2359)","Participant's FEV1 after 2 additional puffs of albuterol: 2d.Was the FEV1 from Q2a ? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Was the participant's FEV1 after reversal from the methacholine challenge greater yhan 90Pct of the baseline FEV1 (1=Yes,0=No)","If NO, has the participant received permission from the supervising physician to proceed with sputum induction testing? 1=Yes,0=No","Physician's Signature:_________ (1=signature present, NULL=signature missing)","Participant's FEV1 used for assessment of eligibility for sputum induction __ . __ __ L (0 - 9.99)","Participant's FEV1 (Pct predicted) used for assessment of eligibility for sputum induction __ __ __ Pct predicted (O-999)","Was the participant's FEV1  from Q3 ? 1=Yes,0=No","Has the participant used any smokeless tobacco products  today? 1=Yes,0=No","Is there any other reason the participant should not proceed with sputum induction? 1=Yes,0=No","Is the participant eligible for sputum induction? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Has the participant completed the study through Visit 9? 1=Yes,0=No","Who initiated termination of the participant? 1=Participant, 2=Clinical Staff","1=no longer interested in participating*, 2=no longer willing to follow protocol*, 3=difficult access to clinic , 4=unable to make visits during clinic hours, 5=moving out of the area, 6=unable to continue due to personal constraints*, 7=unable to continue","Did clinical staff terminate the participant due to pregnancy? 1=Yes, 0=No, 9=N/A","Did clinical staff terminate the participant due to loss to follow-up? 1=Yes,0=No","Did clinical staff terminate the participant due to loss to follow-up, type of contact (1=In-person visit, 2=Phone call)","Did clinical staff terminate the participant due to an asthma-related adverse event? 1=Yes,0=No","Did clinical staff terminate the participant due to a medication-related adverse event? 1=Yes,0=No","Did clinical staff terminate the participant due to an adverse event not related to asthma or medications? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility during the Supervised Washout ? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with medication dosing? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with diary completion? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with visit attendance? 1=Yes,0=No","Did clinical staff terminate the participant due to non-compliance with peak flow monitoring? 1=Yes,0=No","Did clinical staff terminate the participant due to significant asthma exacerbation or two treatment failures during run-in ? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility during the run-in period  for reasons other than compliance or exacerbation/treatment failure or high sputum eosinophils ? 1=Yes,0=No","Did clinical staff terminate the participant due to other reason? 1=Yes,0=No","Did clinical staff terminate the participant due to ineligibility due to high sputum eosinophils in the run-in ? 1=Yes,0=No","Did clinical staff terminate the participant due to... Indicate the letter corresponding to the primary reason the participant was terminated. (a-n)","Did clinical staff terminate the participant due to loss to follow-up, date of last contact with participant","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant seek care for treatment failure conditions? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Yes,0=No","What type of care was sought: Study Investigator or Coordinator? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact, other than call to notify clinic as instructed in spirotel","What type of care was sought: Primary Care or Other Physician? 1=Yes,0=No","What type of care was sought: Primary Care or Other Physician? 1=Scheduled clinic visit, 2=Unscheduled clinic visit, 3=Phone contact","What type of care was sought: Emergency Department visit? 1=Yes,0=No","What type of care was sought: Urgent care visit? 1=Yes,0=No","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Inhaled corticosteroids (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Nebulized bronchodilator (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started: Oral corticosteroids form.  (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  IM or IV steroids form.  (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started: Antibiotics (1=Yes,0=No)","Has the participant taken any of the following medications (excluding study medication and high-dose ICS taken as instructed for treatment failure) since treatment failure conditions started:  Other (1=Yes,0=No)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Did the participant awaken from asthma three or more times in a two-week period or on two consecutive nights? 1=Yes,0=No","Did the participant use albuterol for relief of symptoms four or more times/day for two or more consecutive days? 1=Yes,0=No","Did albuterol relieve symptoms for less than four hours after treatment? 1=Yes,0=No","Did the participant use albuterol for relief of symptoms daily for seven days, and this use exceeded two times the weekly use of albuterol in the baseline period? 1=Yes,0=No","Did regular exercise cause unusually severe shortness of breath on two or more days during a seven day period? 1=Yes,0=No","Did the participant experience a significant asthma exacerbation? 1=Yes,0=No","Did the participant experience a treatment failure? 1=Yes,0=No","Has the participant experienced two treatment failures in the current treatment period ? 1=Yes,0=No","Does the study physician feel the study treatment contributed directly to the treatment failure, such that it is in the participants best interest to discontinue the current blinded treatment? 1=Yes,0=No","Did the participant begin daily treatment with high-dose ICS for 10 days? 1=Yes,0=No","Did the participant begin daily treatment with high-dose ICS for 10 days? 1=Did not think symptoms werebad enough, 2=Other","Date treatment failure conditions were met","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","During this last treatment period , has the participant experienced a treatment failure or asthma exacerbation? 1=Yes,0=No","Has the participant experienced two or more treatment failures in this last treatment period, or does the study physician feel the study treatment contributed directly to the treatment failure such that it is in the participants best interest to discontinu","As of today, have 2 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","As of today, have 6 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","As of today, have 3 weeks passed since the participants final dose of open-label Asmanex? 1=Yes,0=No","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","Are you currently employed ? 1=Yes,0=No","In general, how many hours per week do you usually work?","During the past seven days, how many hours did you miss from work because of problems associated with your asthma?","During the past seven days, how much did asthma affect your productivity while you were working?","Do you currently attend classes in an academic setting ? 1=Yes,0=No","In general, how many hours per week do you usually attend classes?","During the past seven days, how many hours did you miss from class or school because of problems associated with your asthma?","During the past seven days, how much did asthma affect your productivity while in school or attending classes in an academic setting?","During the past seven days, how much did your asthma affect your ability to do your regular daily activities, other than work at a job or attend classes?","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number","How sick today? 0=Not Sick, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely","Runny Nose:(0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","PLUGGED_NOSE  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SNEEZING  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SORE_THROAT  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SCRATCHY_THROAT  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","COUGH (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","HOARSENESS  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","HEAD_CONGESTION  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","CHEST_CONGESTION  (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","FEELING_TIRED (0=Do Not Have this Symptom, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","THINK_CLEARLY  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","SLEEP_WELL  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","BREATH_EASILY  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WALK  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","DAILY_ACTIVITIES  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WORK_OUTSIDE_HOME  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","WORK_INSIDE_HOME  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","INTERACT  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","LIVE_LIFE  (0=Not at all, 1=Very Mildy, 2 and 3=Mildly, 4 and 5=Moderately, 6 and 7=Severely)","MY_COLD_IS  (1=Very much better, 2=Somewhat better, 3=A little better, 4=The same, 5=A little worse, 6=Somewhat worse, 7=Very much worse)","Visit Number (character)","Visit Number (numeric)","Number of days from Visit 1 to this visit","Date (number of days from Visit 1)","Randomized Master ID","Enrollment Type (Screen Fail, Randomized, Healthy Control)","Enrollment Order Number"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>data_set_name<\/th>\n      <th>variable_number_within_dataset<\/th>\n      <th>variable<\/th>\n      <th>type<\/th>\n      <th>variable_description<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":2},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"data_set_name","targets":1},{"name":"variable_number_within_dataset","targets":2},{"name":"variable","targets":3},{"name":"type","targets":4},{"name":"variable_description","targets":5}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>We can also determine how many total variables we have, all of which are accessible via the table we just generated.</p>
 <div class="sourceCode" id="cb404"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb404-1"><a href="data-import-from-pdf-sources.html#cb404-1" tabindex="-1"></a><span class="co"># Total number of variables</span></span>
 <span id="cb404-2"><a href="data-import-from-pdf-sources.html#cb404-2" tabindex="-1"></a><span class="fu">nrow</span>(final_variable_df)</span></code></pre></div>
diff --git a/data-import-processing-and-summary-statistics.html b/data-import-processing-and-summary-statistics.html
index 44edc04..6df6908 100644
--- a/data-import-processing-and-summary-statistics.html
+++ b/data-import-processing-and-summary-statistics.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
@@ -452,12 +461,12 @@ <h4>Import example datasets<a href="data-import-processing-and-summary-statistic
 <h4>View example datasets<a href="data-import-processing-and-summary-statistics.html#view-example-datasets" class="anchor-section" aria-label="Anchor link to header"></a></h4>
 <p>First, let’s preview our example data. Using the <code>datatable()</code> function from the <em>DT</em> package allows us to interactively scroll through our biomarker data.</p>
 <div class="sourceCode" id="cb325"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb325-1"><a href="data-import-processing-and-summary-statistics.html#cb325-1" tabindex="-1"></a><span class="fu">datatable</span>(biomarker_data)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-a8e9d8778c96b174c353" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-a8e9d8778c96b174c353">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80","81","82","83","84","85","86","87","88","89","90","91","92","93","94","95","96","97","98","99","100","101","102","103","104","105","106","107","108","109","110","111","112","113","114","115","116","117","118","119","120","121","122","123","124","125","126","127","128","129","130","131","132","133","134","135","136","137","138","139","140","141","142","143","144","145","146","147","148","149","150","151","152","153","154","155","156","157","158","159","160"],["D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16","D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16","D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16","D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16","D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16","D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16","D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16","D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16","D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16","D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16"],[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4],[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2],[13.26,11.02,11.32,11.18,8.69,6.71,7.87,10.11,8.869999999999999,12.39,14.11,9.73,9.83,9.75,10.34571428571429,10.11,12.42,11.1,12.54,10.24,8,7.28,7.88,10.93,8.720000000000001,12.06,14.36,8.59,9.07,9.949999999999999,10.22428571428572,10.22428571428572,12.55,11.59,11.61,12.02,8.35,11.76,9.81,11.59,11.13,12.73,14.78,11.45,9.43,11.8,11.47142857142857,11.59,13.23,8.43,12.53,3.13,12.25,11.37,10.33,12.99,10.96,12.66,12.73,10.61,10.31,13.72,11.08928571428571,11.37,10.63,12.81,9.880000000000001,17.28,13.6,11.93,10.66,14.16,12.77,12.46,13.6,7.86,12.42,13.84,12.42142857142857,12.46,12,10.15,10.6,9.880000000000001,7.26,5.83,6.65,8.68,7.74,11.92,12.43,8.800000000000001,8.67,8.16,9.197857142857142,8.800000000000001,13.02,11.3,12.16,10.51,8.16,7.71,8.26,10.91,9.09,12.36,14.15,8.789999999999999,9.130000000000001,9.69,10.37428571428572,10.37428571428572,12.34,10.47,10.94,10.61,7.22,11.33,7.52,10.48,9.789999999999999,11.37,13.67,10.34,8.470000000000001,10.07,10.33,10.47,13.44,8.699999999999999,9.5,13.68,12.26,11.94,10.94,13.5,10.9,12.53,13.25,10.68,10.41,13.64,11.81214285714286,11.94,10.09,12.14,9.18,15.99,12.24,11.11,8.76,13.65,11.74,11.05,12.6,7.31,11.41,12.61,11.42,11.42],[1100.47,472.1,611.87,328.26,14.84,174.38,277.68,2772.77,128.48,414.77,1187.36,615.6799999999999,283.72,187.51,612.135,414.77,1298.94,637.35,828.29,321.39,15.2,114.23,264.92,3417.01,158.27,438.2,1154.57,532.6,269.57,192.66,688.8000000000001,438.2,833.9299999999999,524.11,614.99,439.01,18.2,165.19,228.98,2801.85,131.09,380.11,989.41,508.88,170.14,242.6,574.8921428571429,439.01,1044.69,381.44,772.01,641.17,40.57,123.33,161.79,1859.58,155.81,854.5700000000001,899.76,304.68,163.61,228.69,545.1214285714285,381.44,1123.37,28.83,751.36,219.63,34.57,323.07,481.13,357.62,24.57,190.44,259.75,193.53,16.79,88.91,292.3978571428572,219.63,1107.72,474.48,588.45,321.59,13.91,169.02,280.06,2720.11,127.44,405.93,1199.72,617.52,285.5,188.15,607.1142857142858,405.93,1283.64,636.29,806.98,329.85,17.58,114.66,263.26,3369.52,161.08,435.3,1130.06,526.09,273.04,190.55,681.2785714285714,435.3,811.62,525.76,585.74,434.27,17.83,166.14,235.38,2819.04,129.11,373.4,983.37,511.28,169.22,234.69,571.2035714285714,434.27,1022.94,393.13,760.5700000000001,650.22,42.54,124.06,165.67,1902.58,152.02,818.87,901.15,313.42,167.34,227.12,545.8307142857144,393.13,1146.83,27.84,756.48,195.28,35.88,333.79,478.45,377.49,25.45,182.44,263.98,197.61,15.58,91.31999999999999,294.8871428571429,197.61],[36292.57,25439.34,24654.6,23825.7,17504.9,11099.84,12415.64,18131.15,20614.12,35493.56,33949.25,15885.92,20604.15,16270.8,22298.68142857143,20614.12,46107.81,25876.96,30163.14,22211.63,15943.24,11196.96,14713.98,21742.66,18827.56,36444.95,48042.38,16594.76,19994.08,17964.85,24701.78285714286,21742.66,36166.34,null,26696.07,29626.69,17143.72,13923.51,15166.47,22525.33,null,35901.82,46541.14,null,20075.02,22585.19,26031.93636363637,24308.56318181818,42729.62,31695.42,34751.24,64650.82,38096.75,47679.53,22894.16,33252.23,26798.35,81542.88,74197.75999999999,29052.07,19997.71,36989.15,41737.69214285714,36989.15,123431.3,161270.2,144296.4,41552.38,83413.5,78688.66,137315.2,37062.96,85823.24000000001,108839.5,69392.60000000001,84240.08,102113.6,84952.55,95885.15500000003,85823.24000000001,37033.11,24499.93,24966.09,23887.29,17616.43,10569.15,12178.42,17368.88,20550.84,34456.03,31713.68,15808.44,19893.89,15527.56,21862.12428571429,20550.84,46936.91,26452.63,30843.75,22811.07,16464.42,11423.21,14581.4,22215.35,18670.46,37835.25,47679.75,17309.63,19654.98,18069.5,25067.73642857143,22215.35,36194.8,null,26424.8,29906.94,17343.9,13767.75,14775.73,22228.43,null,35274.58,45466.64,null,19170.36,21962.01,25683.26727272727,23955.84863636364,44222.37,30332.27,35141.75,65716.10000000001,39143.03,48158.27,22397.82,33115.8,26747.61,85040.21000000001,73174.89999999999,28667.48,20823.16,38032.79,42193.82571428572,38032.79,121294,159996.4,141851.8,41551.05,83766.34,80988.08,131484.6,35575.33,80646.31,107679.3,71046.63,83514.25999999999,97088.25,83421.81,94278.86857142857,83766.34],[0.32,2.89,0.21,1.19,0.48,1.48,1.04,0.88,1.39,0.22,1.35,0.78,0.43,1.94,1.042857142857143,1.04,0.08,0.15,0.07000000000000001,0.37,0.16,1.93,1.64,1.13,1.02,null,1.09,0.23,null,0.16,0.6691666666666666,0.37,0.5,0.48,0.61,1.76,0.29,0.75,2.4,0.9399999999999999,0.34,0.29,1.39,0.7,4.12,1.05,1.115714285714285,0.75,0.04,0.62,0.65,2.05,1.96,0.98,0.08,0.73,0.47,0.27,1,1.77,1.39,1.6,0.9721428571428572,0.9721428571428572,0.59,1.7,0.3,3.66,0.89,0.48,1.28,1.99,1.83,1.2,2.31,0.36,0.9,1.57,1.361428571428571,1.28,0.26,0.23,0.27,0.48,1.34,3.8,0.1,1.16,8.029999999999999,0.05,6.71,0.1,2.31,0.25,1.792142857142857,0.48,0.11,0.31,0.11,0.09,0.15,1.39,0.02,0.72,0.01,null,2.11,0.12,null,0.14,0.4399999999999999,0.14,0.26,0.62,0.3,0.29,10.6,1.23,0.18,1.18,0.31,0.16,1.55,0.58,null,0.67,1.379230769230769,0.6,1.73,0.73,0.17,0.54,0.44,1.05,0.17,0.93,0.54,0.19,0.9399999999999999,1.91,0.9,0.76,0.7857142857142857,0.76,0.16,1.6,0.2,2.53,0.72,0.25,1.24,1.96,1.73,1,1.92,0.16,0.73,1.61,1.129285714285714,1.129285714285714],[2.21,2.14,7.36,3.53,3.63,1.93,1.1,3.72,0.74,1.98,7.27,2,1.01,1.61,2.873571428571428,2.14,2.56,2.72,1.61,3.65,3.5,2.51,1.11,4.26,1.12,1.89,8.25,2.53,1.03,2.11,2.775000000000001,2.53,2.51,2.73,2.15,3.29,2.42,4.01,18.78,4.49,2.34,2.15,6.57,3.64,0.86,1.88,4.13,2.73,2.41,3.98,1.62,4.92,4,6.71,1.74,6.21,4.33,2.92,7.89,9.460000000000001,5.34,5.7,4.802142857142857,4.802142857142857,3.54,6.52,2.61,11.85,5.8,2.94,4.27,9.1,7.22,6.1,8.85,2.02,4.68,7.04,5.895714285714286,5.895714285714286,3.07,2.98,2.15,5.58,3.59,2.98,2.82,5.83,1.47,2.67,9.56,2.21,2,3.93,3.631428571428572,2.98,2.37,3.32,2.03,4.26,4.67,4.14,1.88,4.92,1.91,1.99,8.65,3,1.58,2.49,3.372142857142857,3,2.78,3.38,3.08,5.21,3.38,4.52,3.42,6.65,3.55,3.51,9.210000000000001,4.56,1.77,3.93,4.210714285714285,3.55,2.97,5.22,2.05,9.029999999999999,4.57,7.55,2.91,7.77,5.68,3.57,8.050000000000001,7.44,6.14,6.36,5.664999999999999,5.68,9.220000000000001,7.36,3.77,12.72,6.7,4.03,7.25,11.08,7.75,6.16,11.51,1.92,5.44,10.29,7.514285714285713,7.36],[840.15,1473.95,659.11,317.44,684.41,1747.14,1283.13,963.13,888.54,1403.65,523.91,688.66,607.05,1619.15,978.5299999999999,888.54,843.51,1797.04,594.4400000000001,296.94,634.76,1991.14,1316.97,911.71,1087.69,1366.48,593.3,686.04,627.85,1592.17,1024.288571428571,911.71,724.58,1756.42,533.08,393.74,789.4299999999999,2897.94,1286.39,1055.57,1376.92,1397.72,678.89,793.83,1204.1,1823.47,1193.72,1193.72,1055.99,3508.22,641.35,670.59,1398.74,2525.07,1666.68,1555.25,2306.87,1672.74,942.61,1938.45,2418.53,2049.55,1739.331428571429,1672.74,2087.9,2103.69,2113.62,511.01,1674,954.04,3401.2,677.6,1948.54,2433.39,638.5,159.97,1875.89,1953.13,1609.462857142857,1875.89,817.28,1491.14,652.09,303.76,688.6900000000001,1734.19,1295.18,978.83,889.84,1387.21,518.89,687.02,606.72,1600.92,975.1257142857141,889.84,849.7,1820.32,571.1900000000001,298.29,631.72,2050.29,1329.88,930.85,1083.1,1394.52,602.22,674.5700000000001,652.22,1613.52,1035.885,930.85,714.8099999999999,1819.75,513.11,389.76,775.8200000000001,2968.98,1367.26,1059.32,1391.71,1402.56,682.28,778.09,1171.59,1777.33,1200.883571428572,1171.59,1056.56,3572.19,674.8,692.03,1421.1,2595.99,1694.72,1625.5,2367.68,1701.4,982.65,1950.98,2484.71,2066.93,1777.66,1701.4,2061.36,2080.9,2042.7,492.12,1659.16,949.41,3399.82,673.67,1951.32,2421.46,634.86,154.07,1828.42,1958.01,1593.377142857142,1828.42]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Donor<\/th>\n      <th>Dose<\/th>\n      <th>Replicate<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[2,3,4,5,6,7,8,9]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Donor","targets":1},{"name":"Dose","targets":2},{"name":"Replicate","targets":3},{"name":"IL1B","targets":4},{"name":"IL6","targets":5},{"name":"IL8","targets":6},{"name":"IL10","targets":7},{"name":"TNFa","targets":8},{"name":"VEGF","targets":9}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-1f9d0de9c6fc6e56a7e1" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-1f9d0de9c6fc6e56a7e1">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80","81","82","83","84","85","86","87","88","89","90","91","92","93","94","95","96","97","98","99","100","101","102","103","104","105","106","107","108","109","110","111","112","113","114","115","116","117","118","119","120","121","122","123","124","125","126","127","128","129","130","131","132","133","134","135","136","137","138","139","140","141","142","143","144","145","146","147","148","149","150","151","152","153","154","155","156","157","158","159","160"],["D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16","D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16","D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16","D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16","D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16","D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16","D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16","D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16","D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16","D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16"],[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,0.6,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4],[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2],[13.26,11.02,11.32,11.18,8.69,6.71,7.87,10.11,8.869999999999999,12.39,14.11,9.73,9.83,9.75,10.34571428571429,10.11,12.42,11.1,12.54,10.24,8,7.28,7.88,10.93,8.720000000000001,12.06,14.36,8.59,9.07,9.949999999999999,10.22428571428572,10.22428571428572,12.55,11.59,11.61,12.02,8.35,11.76,9.81,11.59,11.13,12.73,14.78,11.45,9.43,11.8,11.47142857142857,11.59,13.23,8.43,12.53,3.13,12.25,11.37,10.33,12.99,10.96,12.66,12.73,10.61,10.31,13.72,11.08928571428571,11.37,10.63,12.81,9.880000000000001,17.28,13.6,11.93,10.66,14.16,12.77,12.46,13.6,7.86,12.42,13.84,12.42142857142857,12.46,12,10.15,10.6,9.880000000000001,7.26,5.83,6.65,8.68,7.74,11.92,12.43,8.800000000000001,8.67,8.16,9.197857142857142,8.800000000000001,13.02,11.3,12.16,10.51,8.16,7.71,8.26,10.91,9.09,12.36,14.15,8.789999999999999,9.130000000000001,9.69,10.37428571428572,10.37428571428572,12.34,10.47,10.94,10.61,7.22,11.33,7.52,10.48,9.789999999999999,11.37,13.67,10.34,8.470000000000001,10.07,10.33,10.47,13.44,8.699999999999999,9.5,13.68,12.26,11.94,10.94,13.5,10.9,12.53,13.25,10.68,10.41,13.64,11.81214285714286,11.94,10.09,12.14,9.18,15.99,12.24,11.11,8.76,13.65,11.74,11.05,12.6,7.31,11.41,12.61,11.42,11.42],[1100.47,472.1,611.87,328.26,14.84,174.38,277.68,2772.77,128.48,414.77,1187.36,615.6799999999999,283.72,187.51,612.135,414.77,1298.94,637.35,828.29,321.39,15.2,114.23,264.92,3417.01,158.27,438.2,1154.57,532.6,269.57,192.66,688.8000000000001,438.2,833.9299999999999,524.11,614.99,439.01,18.2,165.19,228.98,2801.85,131.09,380.11,989.41,508.88,170.14,242.6,574.8921428571429,439.01,1044.69,381.44,772.01,641.17,40.57,123.33,161.79,1859.58,155.81,854.5700000000001,899.76,304.68,163.61,228.69,545.1214285714285,381.44,1123.37,28.83,751.36,219.63,34.57,323.07,481.13,357.62,24.57,190.44,259.75,193.53,16.79,88.91,292.3978571428572,219.63,1107.72,474.48,588.45,321.59,13.91,169.02,280.06,2720.11,127.44,405.93,1199.72,617.52,285.5,188.15,607.1142857142858,405.93,1283.64,636.29,806.98,329.85,17.58,114.66,263.26,3369.52,161.08,435.3,1130.06,526.09,273.04,190.55,681.2785714285714,435.3,811.62,525.76,585.74,434.27,17.83,166.14,235.38,2819.04,129.11,373.4,983.37,511.28,169.22,234.69,571.2035714285714,434.27,1022.94,393.13,760.5700000000001,650.22,42.54,124.06,165.67,1902.58,152.02,818.87,901.15,313.42,167.34,227.12,545.8307142857144,393.13,1146.83,27.84,756.48,195.28,35.88,333.79,478.45,377.49,25.45,182.44,263.98,197.61,15.58,91.31999999999999,294.8871428571429,197.61],[36292.57,25439.34,24654.6,23825.7,17504.9,11099.84,12415.64,18131.15,20614.12,35493.56,33949.25,15885.92,20604.15,16270.8,22298.68142857143,20614.12,46107.81,25876.96,30163.14,22211.63,15943.24,11196.96,14713.98,21742.66,18827.56,36444.95,48042.38,16594.76,19994.08,17964.85,24701.78285714286,21742.66,36166.34,null,26696.07,29626.69,17143.72,13923.51,15166.47,22525.33,null,35901.82,46541.14,null,20075.02,22585.19,26031.93636363637,24308.56318181818,42729.62,31695.42,34751.24,64650.82,38096.75,47679.53,22894.16,33252.23,26798.35,81542.88,74197.75999999999,29052.07,19997.71,36989.15,41737.69214285714,36989.15,123431.3,161270.2,144296.4,41552.38,83413.5,78688.66,137315.2,37062.96,85823.24000000001,108839.5,69392.60000000001,84240.08,102113.6,84952.55,95885.15500000003,85823.24000000001,37033.11,24499.93,24966.09,23887.29,17616.43,10569.15,12178.42,17368.88,20550.84,34456.03,31713.68,15808.44,19893.89,15527.56,21862.12428571429,20550.84,46936.91,26452.63,30843.75,22811.07,16464.42,11423.21,14581.4,22215.35,18670.46,37835.25,47679.75,17309.63,19654.98,18069.5,25067.73642857143,22215.35,36194.8,null,26424.8,29906.94,17343.9,13767.75,14775.73,22228.43,null,35274.58,45466.64,null,19170.36,21962.01,25683.26727272727,23955.84863636364,44222.37,30332.27,35141.75,65716.10000000001,39143.03,48158.27,22397.82,33115.8,26747.61,85040.21000000001,73174.89999999999,28667.48,20823.16,38032.79,42193.82571428572,38032.79,121294,159996.4,141851.8,41551.05,83766.34,80988.08,131484.6,35575.33,80646.31,107679.3,71046.63,83514.25999999999,97088.25,83421.81,94278.86857142857,83766.34],[0.32,2.89,0.21,1.19,0.48,1.48,1.04,0.88,1.39,0.22,1.35,0.78,0.43,1.94,1.042857142857143,1.04,0.08,0.15,0.07000000000000001,0.37,0.16,1.93,1.64,1.13,1.02,null,1.09,0.23,null,0.16,0.6691666666666666,0.37,0.5,0.48,0.61,1.76,0.29,0.75,2.4,0.9399999999999999,0.34,0.29,1.39,0.7,4.12,1.05,1.115714285714285,0.75,0.04,0.62,0.65,2.05,1.96,0.98,0.08,0.73,0.47,0.27,1,1.77,1.39,1.6,0.9721428571428572,0.9721428571428572,0.59,1.7,0.3,3.66,0.89,0.48,1.28,1.99,1.83,1.2,2.31,0.36,0.9,1.57,1.361428571428571,1.28,0.26,0.23,0.27,0.48,1.34,3.8,0.1,1.16,8.029999999999999,0.05,6.71,0.1,2.31,0.25,1.792142857142857,0.48,0.11,0.31,0.11,0.09,0.15,1.39,0.02,0.72,0.01,null,2.11,0.12,null,0.14,0.4399999999999999,0.14,0.26,0.62,0.3,0.29,10.6,1.23,0.18,1.18,0.31,0.16,1.55,0.58,null,0.67,1.379230769230769,0.6,1.73,0.73,0.17,0.54,0.44,1.05,0.17,0.93,0.54,0.19,0.9399999999999999,1.91,0.9,0.76,0.7857142857142857,0.76,0.16,1.6,0.2,2.53,0.72,0.25,1.24,1.96,1.73,1,1.92,0.16,0.73,1.61,1.129285714285714,1.129285714285714],[2.21,2.14,7.36,3.53,3.63,1.93,1.1,3.72,0.74,1.98,7.27,2,1.01,1.61,2.873571428571428,2.14,2.56,2.72,1.61,3.65,3.5,2.51,1.11,4.26,1.12,1.89,8.25,2.53,1.03,2.11,2.775000000000001,2.53,2.51,2.73,2.15,3.29,2.42,4.01,18.78,4.49,2.34,2.15,6.57,3.64,0.86,1.88,4.13,2.73,2.41,3.98,1.62,4.92,4,6.71,1.74,6.21,4.33,2.92,7.89,9.460000000000001,5.34,5.7,4.802142857142857,4.802142857142857,3.54,6.52,2.61,11.85,5.8,2.94,4.27,9.1,7.22,6.1,8.85,2.02,4.68,7.04,5.895714285714286,5.895714285714286,3.07,2.98,2.15,5.58,3.59,2.98,2.82,5.83,1.47,2.67,9.56,2.21,2,3.93,3.631428571428572,2.98,2.37,3.32,2.03,4.26,4.67,4.14,1.88,4.92,1.91,1.99,8.65,3,1.58,2.49,3.372142857142857,3,2.78,3.38,3.08,5.21,3.38,4.52,3.42,6.65,3.55,3.51,9.210000000000001,4.56,1.77,3.93,4.210714285714285,3.55,2.97,5.22,2.05,9.029999999999999,4.57,7.55,2.91,7.77,5.68,3.57,8.050000000000001,7.44,6.14,6.36,5.664999999999999,5.68,9.220000000000001,7.36,3.77,12.72,6.7,4.03,7.25,11.08,7.75,6.16,11.51,1.92,5.44,10.29,7.514285714285713,7.36],[840.15,1473.95,659.11,317.44,684.41,1747.14,1283.13,963.13,888.54,1403.65,523.91,688.66,607.05,1619.15,978.5299999999999,888.54,843.51,1797.04,594.4400000000001,296.94,634.76,1991.14,1316.97,911.71,1087.69,1366.48,593.3,686.04,627.85,1592.17,1024.288571428571,911.71,724.58,1756.42,533.08,393.74,789.4299999999999,2897.94,1286.39,1055.57,1376.92,1397.72,678.89,793.83,1204.1,1823.47,1193.72,1193.72,1055.99,3508.22,641.35,670.59,1398.74,2525.07,1666.68,1555.25,2306.87,1672.74,942.61,1938.45,2418.53,2049.55,1739.331428571429,1672.74,2087.9,2103.69,2113.62,511.01,1674,954.04,3401.2,677.6,1948.54,2433.39,638.5,159.97,1875.89,1953.13,1609.462857142857,1875.89,817.28,1491.14,652.09,303.76,688.6900000000001,1734.19,1295.18,978.83,889.84,1387.21,518.89,687.02,606.72,1600.92,975.1257142857141,889.84,849.7,1820.32,571.1900000000001,298.29,631.72,2050.29,1329.88,930.85,1083.1,1394.52,602.22,674.5700000000001,652.22,1613.52,1035.885,930.85,714.8099999999999,1819.75,513.11,389.76,775.8200000000001,2968.98,1367.26,1059.32,1391.71,1402.56,682.28,778.09,1171.59,1777.33,1200.883571428572,1171.59,1056.56,3572.19,674.8,692.03,1421.1,2595.99,1694.72,1625.5,2367.68,1701.4,982.65,1950.98,2484.71,2066.93,1777.66,1701.4,2061.36,2080.9,2042.7,492.12,1659.16,949.41,3399.82,673.67,1951.32,2421.46,634.86,154.07,1828.42,1958.01,1593.377142857142,1828.42]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Donor<\/th>\n      <th>Dose<\/th>\n      <th>Replicate<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[2,3,4,5,6,7,8,9]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Donor","targets":1},{"name":"Dose","targets":2},{"name":"Replicate","targets":3},{"name":"IL1B","targets":4},{"name":"IL6","targets":5},{"name":"IL8","targets":6},{"name":"IL10","targets":7},{"name":"TNFa","targets":8},{"name":"VEGF","targets":9}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>We can see that our biomarker data are arranged with samples in rows and sample information and biomarker measurements in the columns.</p>
 <div class="sourceCode" id="cb326"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb326-1"><a href="data-import-processing-and-summary-statistics.html#cb326-1" tabindex="-1"></a><span class="fu">datatable</span>(demographic_data)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-375af24b79b9504d3425" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-375af24b79b9504d3425">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16"],["D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16"],[22,35,27,60,43,38,26,36,40,44,46,52,58,65,37,25],["M","F","M","F","F","M","M","F","F","M","M","M","M","M","F","F"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Donor<\/th>\n      <th>Age_yr<\/th>\n      <th>Sex<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":2},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Donor","targets":1},{"name":"Age_yr","targets":2},{"name":"Sex","targets":3}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-fbbfa6f65f0521e7d313" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-fbbfa6f65f0521e7d313">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16"],["D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16"],[22,35,27,60,43,38,26,36,40,44,46,52,58,65,37,25],["M","F","M","F","F","M","M","F","F","M","M","M","M","M","F","F"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Donor<\/th>\n      <th>Age_yr<\/th>\n      <th>Sex<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":2},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Donor","targets":1},{"name":"Age_yr","targets":2},{"name":"Sex","targets":3}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>Our demographic data provide information about the donors that our cells came from, matching to the <code>Donor</code> column in our biomarker data.</p>
 </div>
 </div>
@@ -588,8 +597,8 @@ <h2>Averaging Replicates<a href="data-import-processing-and-summary-statistics.h
 <span id="cb335-15"><a href="data-import-processing-and-summary-statistics.html#cb335-15" tabindex="-1"></a></span>
 <span id="cb335-16"><a href="data-import-processing-and-summary-statistics.html#cb335-16" tabindex="-1"></a><span class="co"># View new data frame</span></span>
 <span id="cb335-17"><a href="data-import-processing-and-summary-statistics.html#cb335-17" tabindex="-1"></a><span class="fu">datatable</span>(biomarker_data_imp_avg)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-0472720e4e4117c00c60" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-0472720e4e4117c00c60">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80"],["D10","D10","D10","D10","D10","D11","D11","D11","D11","D11","D12","D12","D12","D12","D12","D13","D13","D13","D13","D13","D14","D14","D14","D14","D14","D15","D15","D15","D15","D15","D16","D16","D16","D16","D16","D1","D1","D1","D1","D1","D2","D2","D2","D2","D2","D3","D3","D3","D3","D3","D4","D4","D4","D4","D4","D5","D5","D5","D5","D5","D6","D6","D6","D6","D6","D7","D7","D7","D7","D7","D8","D8","D8","D8","D8","D9","D9","D9","D9","D9"],["0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4"],[12.16,12.21,12.05,12.6,11.75,13.27,14.25,14.23,12.99,13.1,9.27,8.69,10.89,10.64,7.58,9.25,9.1,8.949999999999999,10.36,11.91,8.960000000000001,9.82,10.94,13.68,13.23,9.77,10.3,10.9,11.45,11.92,9.460000000000001,10.3,11.03,11.66,11.94,12.63,12.72,12.45,13.33,10.36,10.59,11.2,11.03,8.57,12.47,10.96,12.35,11.28,11.02,9.529999999999999,10.53,10.38,11.31,8.4,16.63,7.98,8.08,7.78,12.25,12.92,6.27,7.5,11.55,11.66,11.52,7.26,8.07,8.66,10.64,9.710000000000001,9.390000000000001,10.92,11.04,13.25,13.9,8.31,8.91,10.46,10.93,12.25],[410.35,436.75,376.76,836.72,186.44,1193.54,1142.31,986.39,900.45,261.86,616.6,529.34,510.08,309.05,195.57,284.61,271.3,169.68,165.48,16.18,187.83,191.6,238.65,227.9,90.12,609.62,685.04,573.05,545.48,293.64,410.35,436.75,436.64,387.29,208.62,1104.09,1291.29,822.77,1033.82,1135.1,473.29,636.8200000000001,524.9400000000001,387.29,28.34,600.16,817.64,600.36,766.29,753.92,324.92,325.62,436.64,645.7,207.45,14.38,16.39,18.01,41.55,35.22,171.7,114.45,165.66,123.69,328.43,278.87,264.09,232.18,163.73,479.79,2746.44,3393.26,2810.45,1881.08,367.56,127.96,159.68,130.1,153.92,25.01],[34974.79,37140.1,35588.2,83291.53999999999,108259.4,32831.47,47861.07,46003.89,73686.33,70219.61,15847.18,16952.19,9090.41,28859.78,83877.17,20249.02,19824.53,19622.69,20410.44,99600.92,15899.18,18017.18,22273.6,37510.97,84187.17999999999,22080.4,24884.76,25857.6,41965.76,95082.00999999999,20582.48,21979,24132.21,37510.97,84794.78999999999,36662.84,46522.36,36180.57,43476,122362.65,24969.64,26164.79,9087.049999999999,31013.84,160633.3,24810.35,30503.45,26560.43,34946.49,143074.1,23856.5,22511.35,29766.81,65183.46,41551.72,17560.67,16203.83,17243.81,38619.89,83589.92,10834.49,11310.08,13845.63,47918.9,79838.37,12297.03,14647.69,14971.1,22645.99,134399.9,17750.02,21979,22376.88,33184.02,36319.14,20582.48,18749.01,9107.459999999999,26772.98,83234.77],[0.14,-0.19,0.22,0.23,1.1,4.03,1.6,1.47,0.97,2.12,0.44,0.18,0.64,1.84,0.26,1.37,-0.19,1.96,1.14,0.8100000000000001,1.09,0.15,0.86,1.18,1.59,1.42,0.55,1.25,0.88,1.25,0.76,0.26,0.68,0.87,1.2,0.29,0.1,0.38,0.88,0.37,1.56,0.23,0.55,0.68,1.65,0.24,0.09,0.45,0.41,0.25,0.84,0.23,1.02,1.29,3.09,0.91,0.15,5.44,1.2,0.8,2.64,1.66,0.99,1.01,0.36,0.57,0.83,1.29,0.12,1.26,1.02,0.92,1.06,0.83,1.98,4.71,0.52,0.33,0.5,1.78],[2.33,1.94,2.83,3.24,6.13,8.42,8.449999999999999,7.89,7.97,10.18,2.1,2.76,4.1,8.449999999999999,1.97,1.5,1.31,1.31,5.74,5.06,2.77,2.3,2.9,6.03,8.66,3.25,3.07,4.17,5.23,6.7,2.56,2.76,3.14,5.24,6.63,2.64,2.46,2.64,2.69,6.38,2.56,3.02,3.06,4.6,6.94,4.76,1.82,2.62,1.83,3.19,4.56,3.96,4.25,6.97,12.29,3.61,4.08,2.9,4.28,6.25,2.46,3.32,4.26,7.13,3.49,1.96,1.5,11.1,2.33,5.76,4.78,4.59,5.57,6.99,10.09,1.1,1.52,2.94,5,7.48],[1395.43,1380.5,1400.14,1687.07,2427.43,521.4,597.76,680.58,962.63,636.6799999999999,687.84,680.3,785.96,1944.71,157.02,606.88,640.03,1187.85,2451.62,1852.15,1610.03,1602.84,1800.4,2058.24,1955.57,976.83,1030.09,1197.3,1758.5,1601.42,889.1900000000001,921.28,1182.65,1687.07,1852.15,828.72,846.6,719.7,1056.28,2074.63,1482.54,1808.68,1788.09,3540.2,2092.3,655.6,582.8099999999999,523.09,658.08,2078.16,310.6,297.61,391.75,681.3099999999999,501.57,686.55,633.24,782.63,1409.92,1666.58,1740.66,2020.71,2933.46,2560.53,951.72,1289.16,1323.43,1326.83,1680.7,3400.51,970.98,921.28,1057.45,1590.37,675.64,889.1900000000001,1085.39,1384.32,2337.27,1949.93]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Donor<\/th>\n      <th>Dose<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[3,4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Donor","targets":1},{"name":"Dose","targets":2},{"name":"IL1B","targets":3},{"name":"IL6","targets":4},{"name":"IL8","targets":5},{"name":"IL10","targets":6},{"name":"TNFa","targets":7},{"name":"VEGF","targets":8}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-514e6343c83a7705b404" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-514e6343c83a7705b404">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80"],["D10","D10","D10","D10","D10","D11","D11","D11","D11","D11","D12","D12","D12","D12","D12","D13","D13","D13","D13","D13","D14","D14","D14","D14","D14","D15","D15","D15","D15","D15","D16","D16","D16","D16","D16","D1","D1","D1","D1","D1","D2","D2","D2","D2","D2","D3","D3","D3","D3","D3","D4","D4","D4","D4","D4","D5","D5","D5","D5","D5","D6","D6","D6","D6","D6","D7","D7","D7","D7","D7","D8","D8","D8","D8","D8","D9","D9","D9","D9","D9"],["0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4"],[12.16,12.21,12.05,12.6,11.75,13.27,14.25,14.23,12.99,13.1,9.27,8.69,10.89,10.64,7.58,9.25,9.1,8.949999999999999,10.36,11.91,8.960000000000001,9.82,10.94,13.68,13.23,9.77,10.3,10.9,11.45,11.92,9.460000000000001,10.3,11.03,11.66,11.94,12.63,12.72,12.45,13.33,10.36,10.59,11.2,11.03,8.57,12.47,10.96,12.35,11.28,11.02,9.529999999999999,10.53,10.38,11.31,8.4,16.63,7.98,8.08,7.78,12.25,12.92,6.27,7.5,11.55,11.66,11.52,7.26,8.07,8.66,10.64,9.710000000000001,9.390000000000001,10.92,11.04,13.25,13.9,8.31,8.91,10.46,10.93,12.25],[410.35,436.75,376.76,836.72,186.44,1193.54,1142.31,986.39,900.45,261.86,616.6,529.34,510.08,309.05,195.57,284.61,271.3,169.68,165.48,16.18,187.83,191.6,238.65,227.9,90.12,609.62,685.04,573.05,545.48,293.64,410.35,436.75,436.64,387.29,208.62,1104.09,1291.29,822.77,1033.82,1135.1,473.29,636.8200000000001,524.9400000000001,387.29,28.34,600.16,817.64,600.36,766.29,753.92,324.92,325.62,436.64,645.7,207.45,14.38,16.39,18.01,41.55,35.22,171.7,114.45,165.66,123.69,328.43,278.87,264.09,232.18,163.73,479.79,2746.44,3393.26,2810.45,1881.08,367.56,127.96,159.68,130.1,153.92,25.01],[34974.79,37140.1,35588.2,83291.53999999999,108259.4,32831.47,47861.07,46003.89,73686.33,70219.61,15847.18,16952.19,9090.41,28859.78,83877.17,20249.02,19824.53,19622.69,20410.44,99600.92,15899.18,18017.18,22273.6,37510.97,84187.17999999999,22080.4,24884.76,25857.6,41965.76,95082.00999999999,20582.48,21979,24132.21,37510.97,84794.78999999999,36662.84,46522.36,36180.57,43476,122362.65,24969.64,26164.79,9087.049999999999,31013.84,160633.3,24810.35,30503.45,26560.43,34946.49,143074.1,23856.5,22511.35,29766.81,65183.46,41551.72,17560.67,16203.83,17243.81,38619.89,83589.92,10834.49,11310.08,13845.63,47918.9,79838.37,12297.03,14647.69,14971.1,22645.99,134399.9,17750.02,21979,22376.88,33184.02,36319.14,20582.48,18749.01,9107.459999999999,26772.98,83234.77],[0.14,-0.19,0.22,0.23,1.1,4.03,1.6,1.47,0.97,2.12,0.44,0.18,0.64,1.84,0.26,1.37,-0.19,1.96,1.14,0.8100000000000001,1.09,0.15,0.86,1.18,1.59,1.42,0.55,1.25,0.88,1.25,0.76,0.26,0.68,0.87,1.2,0.29,0.1,0.38,0.88,0.37,1.56,0.23,0.55,0.68,1.65,0.24,0.09,0.45,0.41,0.25,0.84,0.23,1.02,1.29,3.09,0.91,0.15,5.44,1.2,0.8,2.64,1.66,0.99,1.01,0.36,0.57,0.83,1.29,0.12,1.26,1.02,0.92,1.06,0.83,1.98,4.71,0.52,0.33,0.5,1.78],[2.33,1.94,2.83,3.24,6.13,8.42,8.449999999999999,7.89,7.97,10.18,2.1,2.76,4.1,8.449999999999999,1.97,1.5,1.31,1.31,5.74,5.06,2.77,2.3,2.9,6.03,8.66,3.25,3.07,4.17,5.23,6.7,2.56,2.76,3.14,5.24,6.63,2.64,2.46,2.64,2.69,6.38,2.56,3.02,3.06,4.6,6.94,4.76,1.82,2.62,1.83,3.19,4.56,3.96,4.25,6.97,12.29,3.61,4.08,2.9,4.28,6.25,2.46,3.32,4.26,7.13,3.49,1.96,1.5,11.1,2.33,5.76,4.78,4.59,5.57,6.99,10.09,1.1,1.52,2.94,5,7.48],[1395.43,1380.5,1400.14,1687.07,2427.43,521.4,597.76,680.58,962.63,636.6799999999999,687.84,680.3,785.96,1944.71,157.02,606.88,640.03,1187.85,2451.62,1852.15,1610.03,1602.84,1800.4,2058.24,1955.57,976.83,1030.09,1197.3,1758.5,1601.42,889.1900000000001,921.28,1182.65,1687.07,1852.15,828.72,846.6,719.7,1056.28,2074.63,1482.54,1808.68,1788.09,3540.2,2092.3,655.6,582.8099999999999,523.09,658.08,2078.16,310.6,297.61,391.75,681.3099999999999,501.57,686.55,633.24,782.63,1409.92,1666.58,1740.66,2020.71,2933.46,2560.53,951.72,1289.16,1323.43,1326.83,1680.7,3400.51,970.98,921.28,1057.45,1590.37,675.64,889.1900000000001,1085.39,1384.32,2337.27,1949.93]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Donor<\/th>\n      <th>Dose<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[3,4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Donor","targets":1},{"name":"Dose","targets":2},{"name":"IL1B","targets":3},{"name":"IL6","targets":4},{"name":"IL8","targets":5},{"name":"IL10","targets":6},{"name":"TNFa","targets":7},{"name":"VEGF","targets":8}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 </div>
 <div id="descriptive-statistics" class="section level2 hasAnchor">
 <h2>Descriptive Statistics<a href="data-import-processing-and-summary-statistics.html#descriptive-statistics" class="anchor-section" aria-label="Anchor link to header"></a></h2>
@@ -604,8 +613,8 @@ <h3>Method #1 - Tidyverse and Basic Functions<a href="data-import-processing-and
 <span id="cb336-5"><a href="data-import-processing-and-summary-statistics.html#cb336-5" tabindex="-1"></a></span>
 <span id="cb336-6"><a href="data-import-processing-and-summary-statistics.html#cb336-6" tabindex="-1"></a><span class="co"># View data</span></span>
 <span id="cb336-7"><a href="data-import-processing-and-summary-statistics.html#cb336-7" tabindex="-1"></a><span class="fu">datatable</span>(biomarker_group_means)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-7009825df4c4c600926d" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-7009825df4c4c600926d">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["0","0.6","1","2","4"],[9.75375,10.3,10.909375,11.464375,11.92],[597.1693749999999,669.520625,564.5225,535.59,288.328125],[21986.78375,24703.149375,22606.77125,41687.335,94439.059375],[1.376875,0.443125,1.161875,0.876875,1.241875],[3.21,3.05375,4.105,5.2325,6.7],[971.35,1023.284375,1196.3875,1754.03125,1617.09125]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Dose<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[2,3,4,5,6,7]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Dose","targets":1},{"name":"IL1B","targets":2},{"name":"IL6","targets":3},{"name":"IL8","targets":4},{"name":"IL10","targets":5},{"name":"TNFa","targets":6},{"name":"VEGF","targets":7}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-165bcd5f636712a97cb7" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-165bcd5f636712a97cb7">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["0","0.6","1","2","4"],[9.75375,10.3,10.909375,11.464375,11.92],[597.1693749999999,669.520625,564.5225,535.59,288.328125],[21986.78375,24703.149375,22606.77125,41687.335,94439.059375],[1.376875,0.443125,1.161875,0.876875,1.241875],[3.21,3.05375,4.105,5.2325,6.7],[971.35,1023.284375,1196.3875,1754.03125,1617.09125]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Dose<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[2,3,4,5,6,7]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Dose","targets":1},{"name":"IL1B","targets":2},{"name":"IL6","targets":3},{"name":"IL8","targets":4},{"name":"IL10","targets":5},{"name":"TNFa","targets":6},{"name":"VEGF","targets":7}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>You’ll notice that there are a lot of decimal places in our calculated means, while in our original data, there are only two decimal places. We can add a step to round the data to our above code chunk to produce cleaner results.</p>
 <div class="sourceCode" id="cb337"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb337-1"><a href="data-import-processing-and-summary-statistics.html#cb337-1" tabindex="-1"></a><span class="co"># Calculate means</span></span>
 <span id="cb337-2"><a href="data-import-processing-and-summary-statistics.html#cb337-2" tabindex="-1"></a>biomarker_group_means <span class="ot">&lt;-</span> biomarker_data_imp_avg <span class="sc">%&gt;%</span></span>
@@ -615,8 +624,8 @@ <h3>Method #1 - Tidyverse and Basic Functions<a href="data-import-processing-and
 <span id="cb337-6"><a href="data-import-processing-and-summary-statistics.html#cb337-6" tabindex="-1"></a></span>
 <span id="cb337-7"><a href="data-import-processing-and-summary-statistics.html#cb337-7" tabindex="-1"></a><span class="co"># View data</span></span>
 <span id="cb337-8"><a href="data-import-processing-and-summary-statistics.html#cb337-8" tabindex="-1"></a><span class="fu">datatable</span>(biomarker_group_means)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-26d885b2aacfe626998a" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-26d885b2aacfe626998a">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["0","0.6","1","2","4"],[9.75,10.3,10.91,11.46,11.92],[597.17,669.52,564.52,535.59,288.33],[21986.78,24703.15,22606.77,41687.34,94439.06],[1.38,0.44,1.16,0.88,1.24],[3.21,3.05,4.1,5.23,6.7],[971.35,1023.28,1196.39,1754.03,1617.09]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Dose<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[2,3,4,5,6,7]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Dose","targets":1},{"name":"IL1B","targets":2},{"name":"IL6","targets":3},{"name":"IL8","targets":4},{"name":"IL10","targets":5},{"name":"TNFa","targets":6},{"name":"VEGF","targets":7}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-c6e33fd5ea101a5cdeb2" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-c6e33fd5ea101a5cdeb2">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["0","0.6","1","2","4"],[9.75,10.3,10.91,11.46,11.92],[597.17,669.52,564.52,535.59,288.33],[21986.78,24703.15,22606.77,41687.34,94439.06],[1.38,0.44,1.16,0.88,1.24],[3.21,3.05,4.1,5.23,6.7],[971.35,1023.28,1196.39,1754.03,1617.09]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Dose<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[2,3,4,5,6,7]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Dose","targets":1},{"name":"IL1B","targets":2},{"name":"IL6","targets":3},{"name":"IL8","targets":4},{"name":"IL10","targets":5},{"name":"TNFa","targets":6},{"name":"VEGF","targets":7}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <div class="question">
 <p><i>With this, we can answer <strong>Environmental Health Question 1</strong>:</i></p>
 <p>What is the mean concentration of each inflammatory biomarker by acrolein concentration?</p>
@@ -633,8 +642,8 @@ <h3>Method #1 - Tidyverse and Basic Functions<a href="data-import-processing-and
 <span id="cb338-6"><a href="data-import-processing-and-summary-statistics.html#cb338-6" tabindex="-1"></a></span>
 <span id="cb338-7"><a href="data-import-processing-and-summary-statistics.html#cb338-7" tabindex="-1"></a><span class="co"># View data</span></span>
 <span id="cb338-8"><a href="data-import-processing-and-summary-statistics.html#cb338-8" tabindex="-1"></a><span class="fu">datatable</span>(biomarker_group_sds)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-1e45c852785430c572b0" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-1e45c852785430c572b0">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["0","0.6","1","2","4"],[1.9,1.9,1.5,1.6,2],[658.2,810.1,652.3,474.2,297.2],[7570,10768,10647.6,17939.2,33535.4],[1.3,0.6,1.2,0.4,0.8],[1.8,1.7,2.4,2,2.7],[417,485.8,620.7,749.3,836.3]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Dose<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[2,3,4,5,6,7]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Dose","targets":1},{"name":"IL1B","targets":2},{"name":"IL6","targets":3},{"name":"IL8","targets":4},{"name":"IL10","targets":5},{"name":"TNFa","targets":6},{"name":"VEGF","targets":7}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-21a2c1156fd1d00ea2b6" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-21a2c1156fd1d00ea2b6">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["0","0.6","1","2","4"],[1.9,1.9,1.5,1.6,2],[658.2,810.1,652.3,474.2,297.2],[7570,10768,10647.6,17939.2,33535.4],[1.3,0.6,1.2,0.4,0.8],[1.8,1.7,2.4,2,2.7],[417,485.8,620.7,749.3,836.3]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Dose<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[2,3,4,5,6,7]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Dose","targets":1},{"name":"IL1B","targets":2},{"name":"IL6","targets":3},{"name":"IL8","targets":4},{"name":"IL10","targets":5},{"name":"TNFa","targets":6},{"name":"VEGF","targets":7}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>Now we’ve calculated both the means and standard deviations! However, these are typically presented as mean ± standard deviation. We can merge these dataframes by executing the following steps:</p>
 <ol style="list-style-type: decimal">
 <li>Pivot each dataframe to a long format, with each row containing the value for one biomarker at one dose.</li>
@@ -656,8 +665,8 @@ <h3>Method #1 - Tidyverse and Basic Functions<a href="data-import-processing-and
 <span id="cb339-10"><a href="data-import-processing-and-summary-statistics.html#cb339-10" tabindex="-1"></a></span>
 <span id="cb339-11"><a href="data-import-processing-and-summary-statistics.html#cb339-11" tabindex="-1"></a><span class="co"># Preview what dataframe looks like</span></span>
 <span id="cb339-12"><a href="data-import-processing-and-summary-statistics.html#cb339-12" tabindex="-1"></a><span class="fu">datatable</span>(biomarker_group_means_long)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-c06531efbab4020196e0" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-c06531efbab4020196e0">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30"],["0_IL1B","0_IL6","0_IL8","0_IL10","0_TNFa","0_VEGF","0.6_IL1B","0.6_IL6","0.6_IL8","0.6_IL10","0.6_TNFa","0.6_VEGF","1_IL1B","1_IL6","1_IL8","1_IL10","1_TNFa","1_VEGF","2_IL1B","2_IL6","2_IL8","2_IL10","2_TNFa","2_VEGF","4_IL1B","4_IL6","4_IL8","4_IL10","4_TNFa","4_VEGF"],["0","0","0","0","0","0","0.6","0.6","0.6","0.6","0.6","0.6","1","1","1","1","1","1","2","2","2","2","2","2","4","4","4","4","4","4"],["IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF"],[9.75,597.17,21986.78,1.38,3.21,971.35,10.3,669.52,24703.15,0.44,3.05,1023.28,10.91,564.52,22606.77,1.16,4.1,1196.39,11.46,535.59,41687.34,0.88,5.23,1754.03,11.92,288.33,94439.06,1.24,6.7,1617.09]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Dose_variable<\/th>\n      <th>Dose<\/th>\n      <th>variable<\/th>\n      <th>mean<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":4},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Dose_variable","targets":1},{"name":"Dose","targets":2},{"name":"variable","targets":3},{"name":"mean","targets":4}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-a7415faeca81c0deb6de" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-a7415faeca81c0deb6de">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30"],["0_IL1B","0_IL6","0_IL8","0_IL10","0_TNFa","0_VEGF","0.6_IL1B","0.6_IL6","0.6_IL8","0.6_IL10","0.6_TNFa","0.6_VEGF","1_IL1B","1_IL6","1_IL8","1_IL10","1_TNFa","1_VEGF","2_IL1B","2_IL6","2_IL8","2_IL10","2_TNFa","2_VEGF","4_IL1B","4_IL6","4_IL8","4_IL10","4_TNFa","4_VEGF"],["0","0","0","0","0","0","0.6","0.6","0.6","0.6","0.6","0.6","1","1","1","1","1","1","2","2","2","2","2","2","4","4","4","4","4","4"],["IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF"],[9.75,597.17,21986.78,1.38,3.21,971.35,10.3,669.52,24703.15,0.44,3.05,1023.28,10.91,564.52,22606.77,1.16,4.1,1196.39,11.46,535.59,41687.34,0.88,5.23,1754.03,11.92,288.33,94439.06,1.24,6.7,1617.09]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Dose_variable<\/th>\n      <th>Dose<\/th>\n      <th>variable<\/th>\n      <th>mean<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":4},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Dose_variable","targets":1},{"name":"Dose","targets":2},{"name":"variable","targets":3},{"name":"mean","targets":4}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>Next, we will join the mean and standard deviation datasets. Notice that we are only joining the <code>Dose_variable</code> and <code>sd</code> columns from the standard deviation data frame to prevent duplicate columns (<code>Dose</code>, <code>variable</code>) from being included.</p>
 <div class="sourceCode" id="cb340"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb340-1"><a href="data-import-processing-and-summary-statistics.html#cb340-1" tabindex="-1"></a><span class="co"># Merge the dataframes by row</span></span>
 <span id="cb340-2"><a href="data-import-processing-and-summary-statistics.html#cb340-2" tabindex="-1"></a>biomarker_group_summstats <span class="ot">&lt;-</span> <span class="fu">left_join</span>(biomarker_group_means_long, </span>
@@ -666,8 +675,8 @@ <h3>Method #1 - Tidyverse and Basic Functions<a href="data-import-processing-and
 <span id="cb340-5"><a href="data-import-processing-and-summary-statistics.html#cb340-5" tabindex="-1"></a></span>
 <span id="cb340-6"><a href="data-import-processing-and-summary-statistics.html#cb340-6" tabindex="-1"></a><span class="co"># Preview the new dataframe</span></span>
 <span id="cb340-7"><a href="data-import-processing-and-summary-statistics.html#cb340-7" tabindex="-1"></a><span class="fu">datatable</span>(biomarker_group_summstats)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-e0a95ba43a437e34c69e" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-e0a95ba43a437e34c69e">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30"],["0_IL1B","0_IL6","0_IL8","0_IL10","0_TNFa","0_VEGF","0.6_IL1B","0.6_IL6","0.6_IL8","0.6_IL10","0.6_TNFa","0.6_VEGF","1_IL1B","1_IL6","1_IL8","1_IL10","1_TNFa","1_VEGF","2_IL1B","2_IL6","2_IL8","2_IL10","2_TNFa","2_VEGF","4_IL1B","4_IL6","4_IL8","4_IL10","4_TNFa","4_VEGF"],["0","0","0","0","0","0","0.6","0.6","0.6","0.6","0.6","0.6","1","1","1","1","1","1","2","2","2","2","2","2","4","4","4","4","4","4"],["IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF"],[9.75,597.17,21986.78,1.38,3.21,971.35,10.3,669.52,24703.15,0.44,3.05,1023.28,10.91,564.52,22606.77,1.16,4.1,1196.39,11.46,535.59,41687.34,0.88,5.23,1754.03,11.92,288.33,94439.06,1.24,6.7,1617.09],[9.75,597.17,21986.78,1.38,3.21,971.35,10.3,669.52,24703.15,0.44,3.05,1023.28,10.91,564.52,22606.77,1.16,4.1,1196.39,11.46,535.59,41687.34,0.88,5.23,1754.03,11.92,288.33,94439.06,1.24,6.7,1617.09]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Dose_variable<\/th>\n      <th>Dose<\/th>\n      <th>variable<\/th>\n      <th>mean<\/th>\n      <th>sd<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[4,5]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Dose_variable","targets":1},{"name":"Dose","targets":2},{"name":"variable","targets":3},{"name":"mean","targets":4},{"name":"sd","targets":5}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-3d8273ec94c86caa9fb9" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-3d8273ec94c86caa9fb9">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30"],["0_IL1B","0_IL6","0_IL8","0_IL10","0_TNFa","0_VEGF","0.6_IL1B","0.6_IL6","0.6_IL8","0.6_IL10","0.6_TNFa","0.6_VEGF","1_IL1B","1_IL6","1_IL8","1_IL10","1_TNFa","1_VEGF","2_IL1B","2_IL6","2_IL8","2_IL10","2_TNFa","2_VEGF","4_IL1B","4_IL6","4_IL8","4_IL10","4_TNFa","4_VEGF"],["0","0","0","0","0","0","0.6","0.6","0.6","0.6","0.6","0.6","1","1","1","1","1","1","2","2","2","2","2","2","4","4","4","4","4","4"],["IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF"],[9.75,597.17,21986.78,1.38,3.21,971.35,10.3,669.52,24703.15,0.44,3.05,1023.28,10.91,564.52,22606.77,1.16,4.1,1196.39,11.46,535.59,41687.34,0.88,5.23,1754.03,11.92,288.33,94439.06,1.24,6.7,1617.09],[9.75,597.17,21986.78,1.38,3.21,971.35,10.3,669.52,24703.15,0.44,3.05,1023.28,10.91,564.52,22606.77,1.16,4.1,1196.39,11.46,535.59,41687.34,0.88,5.23,1754.03,11.92,288.33,94439.06,1.24,6.7,1617.09]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Dose_variable<\/th>\n      <th>Dose<\/th>\n      <th>variable<\/th>\n      <th>mean<\/th>\n      <th>sd<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[4,5]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Dose_variable","targets":1},{"name":"Dose","targets":2},{"name":"variable","targets":3},{"name":"mean","targets":4},{"name":"sd","targets":5}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>Then, we can unite the mean and standard deviation columns and add the ± symbol between them by storing that character as a variable and pasting that variable in our <code>paste()</code> function.</p>
 <div class="sourceCode" id="cb341"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb341-1"><a href="data-import-processing-and-summary-statistics.html#cb341-1" tabindex="-1"></a><span class="co"># Store plus/minus character</span></span>
 <span id="cb341-2"><a href="data-import-processing-and-summary-statistics.html#cb341-2" tabindex="-1"></a>plusminus <span class="ot">&lt;-</span><span class="st">&quot;\u00b1&quot;</span></span>
@@ -679,8 +688,8 @@ <h3>Method #1 - Tidyverse and Basic Functions<a href="data-import-processing-and
 <span id="cb341-8"><a href="data-import-processing-and-summary-statistics.html#cb341-8" tabindex="-1"></a></span>
 <span id="cb341-9"><a href="data-import-processing-and-summary-statistics.html#cb341-9" tabindex="-1"></a><span class="co"># Preview the new dataframe</span></span>
 <span id="cb341-10"><a href="data-import-processing-and-summary-statistics.html#cb341-10" tabindex="-1"></a><span class="fu">datatable</span>(biomarker_group_summstats)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-80fce8683da3293000e9" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-80fce8683da3293000e9">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30"],["0_IL1B","0_IL6","0_IL8","0_IL10","0_TNFa","0_VEGF","0.6_IL1B","0.6_IL6","0.6_IL8","0.6_IL10","0.6_TNFa","0.6_VEGF","1_IL1B","1_IL6","1_IL8","1_IL10","1_TNFa","1_VEGF","2_IL1B","2_IL6","2_IL8","2_IL10","2_TNFa","2_VEGF","4_IL1B","4_IL6","4_IL8","4_IL10","4_TNFa","4_VEGF"],["0","0","0","0","0","0","0.6","0.6","0.6","0.6","0.6","0.6","1","1","1","1","1","1","2","2","2","2","2","2","4","4","4","4","4","4"],["IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF"],[9.75,597.17,21986.78,1.38,3.21,971.35,10.3,669.52,24703.15,0.44,3.05,1023.28,10.91,564.52,22606.77,1.16,4.1,1196.39,11.46,535.59,41687.34,0.88,5.23,1754.03,11.92,288.33,94439.06,1.24,6.7,1617.09],[9.75,597.17,21986.78,1.38,3.21,971.35,10.3,669.52,24703.15,0.44,3.05,1023.28,10.91,564.52,22606.77,1.16,4.1,1196.39,11.46,535.59,41687.34,0.88,5.23,1754.03,11.92,288.33,94439.06,1.24,6.7,1617.09],["9.75 ± 9.75","597.17 ± 597.17","21986.78 ± 21986.78","1.38 ± 1.38","3.21 ± 3.21","971.35 ± 971.35","10.3 ± 10.3","669.52 ± 669.52","24703.15 ± 24703.15","0.44 ± 0.44","3.05 ± 3.05","1023.28 ± 1023.28","10.91 ± 10.91","564.52 ± 564.52","22606.77 ± 22606.77","1.16 ± 1.16","4.1 ± 4.1","1196.39 ± 1196.39","11.46 ± 11.46","535.59 ± 535.59","41687.34 ± 41687.34","0.88 ± 0.88","5.23 ± 5.23","1754.03 ± 1754.03","11.92 ± 11.92","288.33 ± 288.33","94439.06 ± 94439.06","1.24 ± 1.24","6.7 ± 6.7","1617.09 ± 1617.09"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Dose_variable<\/th>\n      <th>Dose<\/th>\n      <th>variable<\/th>\n      <th>mean<\/th>\n      <th>sd<\/th>\n      <th>mean_sd<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[4,5]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Dose_variable","targets":1},{"name":"Dose","targets":2},{"name":"variable","targets":3},{"name":"mean","targets":4},{"name":"sd","targets":5},{"name":"mean_sd","targets":6}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-783a412746baab11cf2f" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-783a412746baab11cf2f">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30"],["0_IL1B","0_IL6","0_IL8","0_IL10","0_TNFa","0_VEGF","0.6_IL1B","0.6_IL6","0.6_IL8","0.6_IL10","0.6_TNFa","0.6_VEGF","1_IL1B","1_IL6","1_IL8","1_IL10","1_TNFa","1_VEGF","2_IL1B","2_IL6","2_IL8","2_IL10","2_TNFa","2_VEGF","4_IL1B","4_IL6","4_IL8","4_IL10","4_TNFa","4_VEGF"],["0","0","0","0","0","0","0.6","0.6","0.6","0.6","0.6","0.6","1","1","1","1","1","1","2","2","2","2","2","2","4","4","4","4","4","4"],["IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF"],[9.75,597.17,21986.78,1.38,3.21,971.35,10.3,669.52,24703.15,0.44,3.05,1023.28,10.91,564.52,22606.77,1.16,4.1,1196.39,11.46,535.59,41687.34,0.88,5.23,1754.03,11.92,288.33,94439.06,1.24,6.7,1617.09],[9.75,597.17,21986.78,1.38,3.21,971.35,10.3,669.52,24703.15,0.44,3.05,1023.28,10.91,564.52,22606.77,1.16,4.1,1196.39,11.46,535.59,41687.34,0.88,5.23,1754.03,11.92,288.33,94439.06,1.24,6.7,1617.09],["9.75 ± 9.75","597.17 ± 597.17","21986.78 ± 21986.78","1.38 ± 1.38","3.21 ± 3.21","971.35 ± 971.35","10.3 ± 10.3","669.52 ± 669.52","24703.15 ± 24703.15","0.44 ± 0.44","3.05 ± 3.05","1023.28 ± 1023.28","10.91 ± 10.91","564.52 ± 564.52","22606.77 ± 22606.77","1.16 ± 1.16","4.1 ± 4.1","1196.39 ± 1196.39","11.46 ± 11.46","535.59 ± 535.59","41687.34 ± 41687.34","0.88 ± 0.88","5.23 ± 5.23","1754.03 ± 1754.03","11.92 ± 11.92","288.33 ± 288.33","94439.06 ± 94439.06","1.24 ± 1.24","6.7 ± 6.7","1617.09 ± 1617.09"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Dose_variable<\/th>\n      <th>Dose<\/th>\n      <th>variable<\/th>\n      <th>mean<\/th>\n      <th>sd<\/th>\n      <th>mean_sd<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[4,5]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Dose_variable","targets":1},{"name":"Dose","targets":2},{"name":"variable","targets":3},{"name":"mean","targets":4},{"name":"sd","targets":5},{"name":"mean_sd","targets":6}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>Last, we can pivot the dataframe wider to revert it to its original layout, which is more digestible to read.</p>
 <div class="sourceCode" id="cb342"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb342-1"><a href="data-import-processing-and-summary-statistics.html#cb342-1" tabindex="-1"></a><span class="co"># Pivot dataframe wider</span></span>
 <span id="cb342-2"><a href="data-import-processing-and-summary-statistics.html#cb342-2" tabindex="-1"></a>biomarker_group_summstats <span class="ot">&lt;-</span> biomarker_group_summstats <span class="sc">%&gt;%</span></span>
@@ -693,8 +702,8 @@ <h3>Method #1 - Tidyverse and Basic Functions<a href="data-import-processing-and
 <span id="cb342-9"><a href="data-import-processing-and-summary-statistics.html#cb342-9" tabindex="-1"></a></span>
 <span id="cb342-10"><a href="data-import-processing-and-summary-statistics.html#cb342-10" tabindex="-1"></a><span class="co"># View final dataframe</span></span>
 <span id="cb342-11"><a href="data-import-processing-and-summary-statistics.html#cb342-11" tabindex="-1"></a><span class="fu">datatable</span>(biomarker_group_summstats)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-7daa04276e9facacab8f" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-7daa04276e9facacab8f">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["0","0.6","1","2","4"],["9.75 ± 9.75","10.3 ± 10.3","10.91 ± 10.91","11.46 ± 11.46","11.92 ± 11.92"],["597.17 ± 597.17","669.52 ± 669.52","564.52 ± 564.52","535.59 ± 535.59","288.33 ± 288.33"],["21986.78 ± 21986.78","24703.15 ± 24703.15","22606.77 ± 22606.77","41687.34 ± 41687.34","94439.06 ± 94439.06"],["1.38 ± 1.38","0.44 ± 0.44","1.16 ± 1.16","0.88 ± 0.88","1.24 ± 1.24"],["3.21 ± 3.21","3.05 ± 3.05","4.1 ± 4.1","5.23 ± 5.23","6.7 ± 6.7"],["971.35 ± 971.35","1023.28 ± 1023.28","1196.39 ± 1196.39","1754.03 ± 1754.03","1617.09 ± 1617.09"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Dose<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Dose","targets":1},{"name":"IL1B","targets":2},{"name":"IL6","targets":3},{"name":"IL8","targets":4},{"name":"IL10","targets":5},{"name":"TNFa","targets":6},{"name":"VEGF","targets":7}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-2f273678b26535d01311" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-2f273678b26535d01311">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5"],["0","0.6","1","2","4"],["9.75 ± 9.75","10.3 ± 10.3","10.91 ± 10.91","11.46 ± 11.46","11.92 ± 11.92"],["597.17 ± 597.17","669.52 ± 669.52","564.52 ± 564.52","535.59 ± 535.59","288.33 ± 288.33"],["21986.78 ± 21986.78","24703.15 ± 24703.15","22606.77 ± 22606.77","41687.34 ± 41687.34","94439.06 ± 94439.06"],["1.38 ± 1.38","0.44 ± 0.44","1.16 ± 1.16","0.88 ± 0.88","1.24 ± 1.24"],["3.21 ± 3.21","3.05 ± 3.05","4.1 ± 4.1","5.23 ± 5.23","6.7 ± 6.7"],["971.35 ± 971.35","1023.28 ± 1023.28","1196.39 ± 1196.39","1754.03 ± 1754.03","1617.09 ± 1617.09"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Dose<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Dose","targets":1},{"name":"IL1B","targets":2},{"name":"IL6","targets":3},{"name":"IL8","targets":4},{"name":"IL10","targets":5},{"name":"TNFa","targets":6},{"name":"VEGF","targets":7}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>These data are now in a publication-ready format that can be exported to a .txt, .csv., or .xlsx file for easy sharing.</p>
 </div>
 <div id="method-2---applying-a-list-of-functions" class="section level3 hasAnchor">
@@ -711,8 +720,8 @@ <h3>Method #2 - Applying a List of Functions<a href="data-import-processing-and-
 <span id="cb343-9"><a href="data-import-processing-and-summary-statistics.html#cb343-9" tabindex="-1"></a></span>
 <span id="cb343-10"><a href="data-import-processing-and-summary-statistics.html#cb343-10" tabindex="-1"></a><span class="co"># View data</span></span>
 <span id="cb343-11"><a href="data-import-processing-and-summary-statistics.html#cb343-11" tabindex="-1"></a><span class="fu">datatable</span>(biomarker_descriptive_stats_all)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-6a995626b5964ea58500" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-6a995626b5964ea58500">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25"],["min","min","min","min","min","median","median","median","median","median","mean","mean","mean","mean","mean","max","max","max","max","max","sd","sd","sd","sd","sd"],["0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4"],[6.27,7.5,7.78,8.4,7.58,9.425000000000001,10.3,11.03,11.555,11.93,9.75375,10.3,10.909375,11.464375,11.92,13.27,14.25,14.23,13.68,16.63,1.899234494912797,1.905189404407516,1.513289876395134,1.56458287838431,2.044165681478224],[14.38,16.39,18.01,41.55,16.18,410.35,436.75,436.64,387.29,208.035,597.1693749999999,669.520625,564.5225,535.59,288.328125,2746.44,3393.26,2810.45,1881.08,1135.1,658.2397037398939,810.1192296896284,652.2647653266782,474.1640196950137,297.2423973621248],[10834.49,11310.08,9087.049999999999,20410.44,36319.14,20582.48,21979,22325.24,37510.97,84490.98499999999,21986.78375,24703.149375,22606.77125,41687.335,94439.059375,36662.84,47861.07,46003.89,83291.53999999999,160633.3,7570.028454551431,10768.04243681303,10647.63009995049,17939.24569853593,33535.41840978907],[0.14,-0.19,0.22,0.12,0.25,0.9650000000000001,0.23,0.925,0.88,1.225,1.376875,0.443125,1.161875,0.876875,1.241875,4.71,1.66,5.44,1.84,3.09,1.326864568572593,0.5551662663863262,1.2325,0.4307004953174151,0.7830982803369362],[1.1,1.31,1.31,1.83,1.97,2.6,2.76,3.1,5.235,6.505,3.21,3.05375,4.105,5.2325,6.7,8.42,8.449999999999999,11.1,8.449999999999999,12.29,1.76761232552088,1.733769977053857,2.387821880570939,2.004084163235999,2.690268883711564],[310.6,297.61,391.75,658.08,157.02,889.1900000000001,921.28,1185.25,1687.07,1852.15,971.35,1023.284375,1196.3875,1754.03125,1617.09125,1740.66,2020.71,2933.46,3540.2,3400.51,417.0417453924726,485.7680272444005,620.7245347763638,749.2713471622502,836.2607857829598]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>statistic<\/th>\n      <th>Dose<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[3,4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"statistic","targets":1},{"name":"Dose","targets":2},{"name":"IL1B","targets":3},{"name":"IL6","targets":4},{"name":"IL8","targets":5},{"name":"IL10","targets":6},{"name":"TNFa","targets":7},{"name":"VEGF","targets":8}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-5f3f37973f32f0bafae5" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-5f3f37973f32f0bafae5">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25"],["min","min","min","min","min","median","median","median","median","median","mean","mean","mean","mean","mean","max","max","max","max","max","sd","sd","sd","sd","sd"],["0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4"],[6.27,7.5,7.78,8.4,7.58,9.425000000000001,10.3,11.03,11.555,11.93,9.75375,10.3,10.909375,11.464375,11.92,13.27,14.25,14.23,13.68,16.63,1.899234494912797,1.905189404407516,1.513289876395134,1.56458287838431,2.044165681478224],[14.38,16.39,18.01,41.55,16.18,410.35,436.75,436.64,387.29,208.035,597.1693749999999,669.520625,564.5225,535.59,288.328125,2746.44,3393.26,2810.45,1881.08,1135.1,658.2397037398939,810.1192296896284,652.2647653266782,474.1640196950137,297.2423973621248],[10834.49,11310.08,9087.049999999999,20410.44,36319.14,20582.48,21979,22325.24,37510.97,84490.98499999999,21986.78375,24703.149375,22606.77125,41687.335,94439.059375,36662.84,47861.07,46003.89,83291.53999999999,160633.3,7570.028454551431,10768.04243681303,10647.63009995049,17939.24569853593,33535.41840978907],[0.14,-0.19,0.22,0.12,0.25,0.9650000000000001,0.23,0.925,0.88,1.225,1.376875,0.443125,1.161875,0.876875,1.241875,4.71,1.66,5.44,1.84,3.09,1.326864568572593,0.5551662663863262,1.2325,0.4307004953174151,0.7830982803369362],[1.1,1.31,1.31,1.83,1.97,2.6,2.76,3.1,5.235,6.505,3.21,3.05375,4.105,5.2325,6.7,8.42,8.449999999999999,11.1,8.449999999999999,12.29,1.76761232552088,1.733769977053857,2.387821880570939,2.004084163235999,2.690268883711564],[310.6,297.61,391.75,658.08,157.02,889.1900000000001,921.28,1185.25,1687.07,1852.15,971.35,1023.284375,1196.3875,1754.03125,1617.09125,1740.66,2020.71,2933.46,3540.2,3400.51,417.0417453924726,485.7680272444005,620.7245347763638,749.2713471622502,836.2607857829598]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>statistic<\/th>\n      <th>Dose<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[3,4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"statistic","targets":1},{"name":"Dose","targets":2},{"name":"IL1B","targets":3},{"name":"IL6","targets":4},{"name":"IL8","targets":5},{"name":"IL10","targets":6},{"name":"TNFa","targets":7},{"name":"VEGF","targets":8}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>Depending on your final goal, descriptive statistics data can then be extracted from this dataframe and cleaned up or reformatted as needed to create a publication-ready table!</p>
 </div>
 <div id="other-methods" class="section level3 hasAnchor">
@@ -1186,8 +1195,8 @@ <h2>Normality Assessment and Data Transformation<a href="data-import-processing-
 <span id="cb352-15"><a href="data-import-processing-and-summary-statistics.html#cb352-15" tabindex="-1"></a></span>
 <span id="cb352-16"><a href="data-import-processing-and-summary-statistics.html#cb352-16" tabindex="-1"></a><span class="co"># View cleaned up data frame</span></span>
 <span id="cb352-17"><a href="data-import-processing-and-summary-statistics.html#cb352-17" tabindex="-1"></a><span class="fu">datatable</span>(shapiro_res)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-f1cc11f5fd0b5056c42b" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-f1cc11f5fd0b5056c42b">{"x":{"filter":"none","vertical":false,"data":[["IL1B","IL6","IL8","IL10","TNFa","VEGF"],[0.7908975299411561,1.061132292837358e-11,1.578095224843693e-09,1.646966900182463e-09,4.301352465611483e-05,0.0009741571784854114],[true,false,false,false,false,false]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>p.value<\/th>\n      <th>normal<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":1},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"p.value","targets":1},{"name":"normal","targets":2}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-355365551c414dbee00d" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-355365551c414dbee00d">{"x":{"filter":"none","vertical":false,"data":[["IL1B","IL6","IL8","IL10","TNFa","VEGF"],[0.7908975299411561,1.061132292837358e-11,1.578095224843693e-09,1.646966900182463e-09,4.301352465611483e-05,0.0009741571784854114],[true,false,false,false,false,false]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>p.value<\/th>\n      <th>normal<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":1},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"p.value","targets":1},{"name":"normal","targets":2}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <div class="question">
 <p><i>With this, we can answer <strong>Environmental Health Question 2</strong>:</i></p>
 <p>Are our data normally distributed?</p>
diff --git a/data-manipulation-and-reshaping.html b/data-manipulation-and-reshaping.html
index 2a7a35f..6464a5c 100644
--- a/data-manipulation-and-reshaping.html
+++ b/data-manipulation-and-reshaping.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
diff --git a/data-sharing-through-online-repositories.html b/data-sharing-through-online-repositories.html
index c80a547..ef35952 100644
--- a/data-sharing-through-online-repositories.html
+++ b/data-sharing-through-online-repositories.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
diff --git a/data-visualizations.html b/data-visualizations.html
index ddacae9..336152e 100644
--- a/data-visualizations.html
+++ b/data-visualizations.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
diff --git a/data-wrangling-in-excel.html b/data-wrangling-in-excel.html
index e6c5e8c..577c843 100644
--- a/data-wrangling-in-excel.html
+++ b/data-wrangling-in-excel.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
diff --git a/downloading-and-programming-in-r.html b/downloading-and-programming-in-r.html
index ed8ee3c..eea5b2b 100644
--- a/downloading-and-programming-in-r.html
+++ b/downloading-and-programming-in-r.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
diff --git a/fair-data-management-practices.html b/fair-data-management-practices.html
index e35eb03..4b398c2 100644
--- a/fair-data-management-practices.html
+++ b/fair-data-management-practices.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
diff --git a/file-management-using-github.html b/file-management-using-github.html
index 0549b87..14e4ddb 100644
--- a/file-management-using-github.html
+++ b/file-management-using-github.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
diff --git a/header.html b/header.html
new file mode 100644
index 0000000..5e75718
--- /dev/null
+++ b/header.html
@@ -0,0 +1,9 @@
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
\ No newline at end of file
diff --git a/improving-coding-efficiencies.html b/improving-coding-efficiencies.html
index 91c9c4b..8300233 100644
--- a/improving-coding-efficiencies.html
+++ b/improving-coding-efficiencies.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
diff --git a/improving-data-visualizations.html b/improving-data-visualizations.html
index 94807c7..8208882 100644
--- a/improving-data-visualizations.html
+++ b/improving-data-visualizations.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
diff --git a/index.html b/index.html
index 10029bb..1e8f24a 100644
--- a/index.html
+++ b/index.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
diff --git a/intoduction-to-statistical-tests.html b/intoduction-to-statistical-tests.html
index be2df8c..80fe2dd 100644
--- a/intoduction-to-statistical-tests.html
+++ b/intoduction-to-statistical-tests.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
diff --git a/introduction-to-artificial-intelligence-machine-learning-and-predictive-modeling-for-environmental-health.html b/introduction-to-artificial-intelligence-machine-learning-and-predictive-modeling-for-environmental-health.html
index 99db7de..841e412 100644
--- a/introduction-to-artificial-intelligence-machine-learning-and-predictive-modeling-for-environmental-health.html
+++ b/introduction-to-artificial-intelligence-machine-learning-and-predictive-modeling-for-environmental-health.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
diff --git a/multi-group-and-multi-variable-comparisons-and-visualizations.html b/multi-group-and-multi-variable-comparisons-and-visualizations.html
index a304043..42ff8d8 100644
--- a/multi-group-and-multi-variable-comparisons-and-visualizations.html
+++ b/multi-group-and-multi-variable-comparisons-and-visualizations.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
@@ -434,11 +443,11 @@ <h3>Workspace Preparation and Data Import<a href="multi-group-and-multi-variable
 <span id="cb430-4"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb430-4" tabindex="-1"></a></span>
 <span id="cb430-5"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb430-5" tabindex="-1"></a><span class="co"># View data</span></span>
 <span id="cb430-6"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb430-6" tabindex="-1"></a><span class="fu">datatable</span>(biomarker_data)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-12392c79d7e859a17286" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-12392c79d7e859a17286">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80"],["D10","D10","D10","D10","D10","D11","D11","D11","D11","D11","D12","D12","D12","D12","D12","D13","D13","D13","D13","D13","D14","D14","D14","D14","D14","D15","D15","D15","D15","D15","D16","D16","D16","D16","D16","D1","D1","D1","D1","D1","D2","D2","D2","D2","D2","D3","D3","D3","D3","D3","D4","D4","D4","D4","D4","D5","D5","D5","D5","D5","D6","D6","D6","D6","D6","D7","D7","D7","D7","D7","D8","D8","D8","D8","D8","D9","D9","D9","D9","D9"],["0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4"],[3.71808758396052,3.7235585614721,3.70597790168252,3.76553474636298,3.6724253419715,3.83491342970104,3.93073733756289,3.92884403671257,3.80632405739003,3.81762325751143,3.36036427654347,3.27649666564036,3.57167680997093,3.54101915313356,3.10097764772482,3.35755200461808,3.33628338786443,3.31469652565629,3.50589092972996,3.69041709553247,3.31614574229336,3.43562859405209,3.57773093149008,3.87578006306849,3.83086375675176,3.42894634476948,3.49825086752783,3.57288966842058,3.63807383718072,3.6915341649192,3.38681094647222,3.49825086752783,3.58856473740135,3.66220549965362,3.69376571221778,3.76871365703048,3.77820857639809,3.74953426766926,3.84096670448742,3.50589092972996,3.53480866123175,3.60880924267552,3.58856473740135,3.2585189247113,3.75167794568758,3.58014548442338,3.73876783680019,3.61823865559545,3.58736499093646,3.39643353125099,3.52732060787705,3.50842865253186,3.62175885670973,3.23266075679027,4.13996056954546,3.16671544496642,3.18269229751619,3.13422093976063,3.7279204545632,3.799087306074,2.86195536414487,3.08746284125034,3.64961545906341,3.66220549965362,3.64616265715789,3.04614178164472,3.1811025507538,3.27202318906105,3.54101915313356,3.42088657497553,3.37712374912949,3.57531233068744,3.58976348698498,3.83289001416474,3.8972404255748,3.21878116778407,3.30888505741176,3.51853513898218,3.5765221379205,3.7279204545632],[8.68422263328295,8.773963368433559,8.561326136453919,9.710324307072939,7.55028504908841,10.2222394489088,10.1590009172826,9.94747622373011,9.81610366284939,8.03815081026938,9.27052894238072,9.05077375519723,8.99740532549677,8.27635707925924,7.61889934790406,8.15790268278964,8.089053172156429,7.41515020542639,7.37920505996061,4.10265813136374,7.56094417806009,7.5894638929561,7.9047851297316,7.83857365266832,6.50969584193339,9.254131033541951,9.422148885880191,9.165032591502291,9.09402488708305,8.202809492125009,8.68422263328295,8.773963368433559,8.77360079528254,8.60099074240823,7.71163256182322,10.1099481540673,10.3357141430951,9.686097776967401,10.0151641270811,10.1498741117182,8.88962564051876,9.317005526387209,9.03875441393145,8.60099074240823,4.87479696589866,9.23160520814165,9.677085350559841,9.23208509872283,9.583628142323651,9.560179957764181,8.34837407519116,8.351469324267081,8.77360079528254,9.33695280016051,7.70355756192103,3.9429835981871,4.12018602753186,4.24868662687383,5.4110872267986,5.17871464117544,7.43212427264156,6.85112436288675,7.38076407498511,6.96220195703608,8.36382813184704,8.12861303870471,8.0503384381673,7.86530024272912,7.3639595068907,8.90926307937532,11.4238722598684,11.7288813640812,11.4570986742783,10.878112237463,8.52575569282949,7.01077983875324,7.32804655627074,7.03452387533439,7.27537959625362,4.70099449416827],[15.0940690235359,15.1807289234454,15.1191518828101,16.3458996676138,16.7241460980745,15.0028356680746,15.5465951732949,15.4894995973234,16.169128959283,16.0996069083821,13.9520295507825,14.0492691434224,13.1502883461063,14.8168226706095,16.3560077655342,14.30563571243,14.2750718138596,14.2603087281481,14.3170903457307,16.6038859325635,13.9567554773633,14.1371656727442,14.4431119040824,15.1950634099335,16.3613300730261,14.4305440239947,14.6030328264996,14.6583565484435,15.3569594653045,16.5368999542995,14.3291992954213,14.4239037658366,14.5587322032289,15.1950634099335,16.3717050181953,15.1620702745062,15.5056676741312,15.1429673906615,15.4079647745043,16.9008155214109,14.6079451777076,14.6753941971444,13.1497550568455,14.9206710618732,17.2934204566841,14.5987126147926,14.8967320984425,14.6970451993658,15.0929012215711,17.1264130896721,14.542155252148,14.4584290430224,14.8614654668171,15.9922404459517,15.3426552970282,14.1001424218123,13.9841362654046,14.0738746130993,15.2370937904406,16.3510586187167,13.4034767756747,13.4654490662827,13.7572472746771,15.5483372770406,16.2848127161883,13.5861396108168,13.8384840329337,13.8699889688218,14.4670316946933,17.0361832734254,14.1156143059506,14.4239037658366,14.4497857465895,15.0182445232662,15.1484821432167,14.3291992954213,14.1946037445951,13.1529914376826,14.708543993667,16.3449160294487],[0.189033824390017,-0.3040061868901,0.286881147788162,0.298658315564515,1.0703893278914,2.3305584000308,1.37851162325373,1.30451104180995,0.978195629681652,1.64154602908752,0.526068811667588,0.238786859587116,0.713695814843359,1.50589092972996,0.333423733725192,1.24488705912353,-0.3040061868901,1.56559717585423,1.09761079662642,0.855989697308481,1.06350294230616,0.20163386116965,0.895302621333307,1.1243281350022,1.37295209791183,1.27500704749987,0.632268215499513,1.16992500144231,0.910732661902913,1.16992500144231,0.8155754288625729,0.333423733725192,0.748461233004036,0.903038270112912,1.13750352374994,0.367371065648529,0.137503523749935,0.464668267003444,0.910732661902913,0.454175893185802,1.35614381022528,0.298658315564515,0.632268215499513,0.748461233004036,1.40599235967584,0.31034012061215,0.124328135002202,0.53605290024021,0.495695162624069,0.321928094887362,0.879705766282288,0.298658315564515,1.01435529297707,1.19534759832222,2.03210084316702,0.933572638261024,0.20163386116965,2.68706068833989,1.13750352374994,0.84799690655495,1.86393845042397,1.41142624572647,0.992768430768924,1.0071955014042,0.443606651475615,0.650764559116902,0.8718436485093179,1.19534759832222,0.16349873228288,1.17632277264046,1.01435529297707,0.941106310946431,1.04264433740849,0.8718436485093179,1.57531233068744,2.51349074558812,0.604071323668861,0.411426245726465,0.584962500721156,1.47508488294878],[1.73552217729654,1.55581615506164,1.93734439215023,2.08406426478847,2.83390207666916,3.23572705983806,3.24031432933371,3.15218341905505,3.16510798514454,3.48284828306847,1.63226821549951,1.91073266190291,2.35049724708413,3.24031432933371,1.57046293102604,1.32192809488736,1.20789285164133,1.20789285164133,2.75274859140713,2.59931779369823,1.91456452349394,1.72246602447109,1.96347412397489,2.81352468929781,3.27202318906105,2.08746284125034,2.02502879449152,2.37016428054021,2.63923216324928,2.94485844580754,1.83187724119167,1.91073266190291,2.0496307677246,2.64154602908752,2.93168305705981,1.86393845042397,1.790772037862,1.86393845042397,1.88362081628567,2.88362081628567,1.83187724119167,2.0071955014042,2.02147972741045,2.48542682717024,2.98913900736823,2.52606881166759,1.49569516262407,1.85598969730848,1.50080205305716,2.06695024392463,2.47508488294878,2.31034012061215,2.39231742277876,2.99457972421575,3.73226919950145,2.20476675065461,2.34482849699744,1.96347412397489,2.40053792958373,2.85798099512757,1.790772037862,2.11103131238874,2.39506279951758,3.0232553523003,2.16671544496642,1.56559717585423,1.32192809488736,3.59693514238723,1.73552217729654,2.75702324650746,2.53106949272595,2.48284828306847,2.7158933705476,2.99819550315325,3.47118746038699,1.0703893278914,1.33342373372519,1.97819562968165,2.58496250072116,3.08406426478847],[10.4475275409266,10.4320198468125,10.4523853997561,10.7211590148677,11.2458081856384,9.02901108665907,9.22583403553784,9.412739192116421,9.91233549843777,9.316688823436939,9.428025110276749,9.412146396540679,9.620146497339491,10.9260809830959,7.3039633562453,9.247642742898019,9.32424806586749,10.2153509830024,11.2601080099627,10.8557639471917,10.6537676442859,10.6473145095471,10.8149028512154,11.0078962674403,10.9341110109735,9.933439857942711,10.009954550333,10.2267734234482,10.7809497987497,10.646036617491,9.797969484059131,9.8490610025615,10.2090268302306,10.7211590148677,10.8557639471917,9.696480751820189,9.72723977748481,9.49325503512067,10.0461417816447,11.0193335777127,10.534828110721,10.8215188973143,10.8050102485124,11.7900226112658,11.031563370024,9.35887093968562,9.18935511254853,9.033670771393091,9.36430978186441,11.0217850685075,8.283551423174311,8.22211866974429,8.617467465294199,9.414283550503709,8.973180744513019,9.425320822759881,9.308885057411761,9.61402881967061,10.4624204732526,10.7035402591663,10.7662472988462,10.9813603523445,11.5188793273505,11.3227900730377,9.89590846480624,10.3333342781343,10.3711558797312,10.374854737178,10.7157046495212,11.7319596151152,9.92478281821723,9.8490610025615,10.0477374043091,10.6360535917989,9.40224465525408,9.797969484059131,10.0853263897027,10.4360035526972,11.191225812124,10.9299462991725]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Donor<\/th>\n      <th>Dose<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[3,4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Donor","targets":1},{"name":"Dose","targets":2},{"name":"IL1B","targets":3},{"name":"IL6","targets":4},{"name":"IL8","targets":5},{"name":"IL10","targets":6},{"name":"TNFa","targets":7},{"name":"VEGF","targets":8}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-daa7ded51ca711e7c055" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-daa7ded51ca711e7c055">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80"],["D10","D10","D10","D10","D10","D11","D11","D11","D11","D11","D12","D12","D12","D12","D12","D13","D13","D13","D13","D13","D14","D14","D14","D14","D14","D15","D15","D15","D15","D15","D16","D16","D16","D16","D16","D1","D1","D1","D1","D1","D2","D2","D2","D2","D2","D3","D3","D3","D3","D3","D4","D4","D4","D4","D4","D5","D5","D5","D5","D5","D6","D6","D6","D6","D6","D7","D7","D7","D7","D7","D8","D8","D8","D8","D8","D9","D9","D9","D9","D9"],["0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4"],[3.71808758396052,3.7235585614721,3.70597790168252,3.76553474636298,3.6724253419715,3.83491342970104,3.93073733756289,3.92884403671257,3.80632405739003,3.81762325751143,3.36036427654347,3.27649666564036,3.57167680997093,3.54101915313356,3.10097764772482,3.35755200461808,3.33628338786443,3.31469652565629,3.50589092972996,3.69041709553247,3.31614574229336,3.43562859405209,3.57773093149008,3.87578006306849,3.83086375675176,3.42894634476948,3.49825086752783,3.57288966842058,3.63807383718072,3.6915341649192,3.38681094647222,3.49825086752783,3.58856473740135,3.66220549965362,3.69376571221778,3.76871365703048,3.77820857639809,3.74953426766926,3.84096670448742,3.50589092972996,3.53480866123175,3.60880924267552,3.58856473740135,3.2585189247113,3.75167794568758,3.58014548442338,3.73876783680019,3.61823865559545,3.58736499093646,3.39643353125099,3.52732060787705,3.50842865253186,3.62175885670973,3.23266075679027,4.13996056954546,3.16671544496642,3.18269229751619,3.13422093976063,3.7279204545632,3.799087306074,2.86195536414487,3.08746284125034,3.64961545906341,3.66220549965362,3.64616265715789,3.04614178164472,3.1811025507538,3.27202318906105,3.54101915313356,3.42088657497553,3.37712374912949,3.57531233068744,3.58976348698498,3.83289001416474,3.8972404255748,3.21878116778407,3.30888505741176,3.51853513898218,3.5765221379205,3.7279204545632],[8.68422263328295,8.773963368433559,8.561326136453919,9.710324307072939,7.55028504908841,10.2222394489088,10.1590009172826,9.94747622373011,9.81610366284939,8.03815081026938,9.27052894238072,9.05077375519723,8.99740532549677,8.27635707925924,7.61889934790406,8.15790268278964,8.089053172156429,7.41515020542639,7.37920505996061,4.10265813136374,7.56094417806009,7.5894638929561,7.9047851297316,7.83857365266832,6.50969584193339,9.254131033541951,9.422148885880191,9.165032591502291,9.09402488708305,8.202809492125009,8.68422263328295,8.773963368433559,8.77360079528254,8.60099074240823,7.71163256182322,10.1099481540673,10.3357141430951,9.686097776967401,10.0151641270811,10.1498741117182,8.88962564051876,9.317005526387209,9.03875441393145,8.60099074240823,4.87479696589866,9.23160520814165,9.677085350559841,9.23208509872283,9.583628142323651,9.560179957764181,8.34837407519116,8.351469324267081,8.77360079528254,9.33695280016051,7.70355756192103,3.9429835981871,4.12018602753186,4.24868662687383,5.4110872267986,5.17871464117544,7.43212427264156,6.85112436288675,7.38076407498511,6.96220195703608,8.36382813184704,8.12861303870471,8.0503384381673,7.86530024272912,7.3639595068907,8.90926307937532,11.4238722598684,11.7288813640812,11.4570986742783,10.878112237463,8.52575569282949,7.01077983875324,7.32804655627074,7.03452387533439,7.27537959625362,4.70099449416827],[15.0940690235359,15.1807289234454,15.1191518828101,16.3458996676138,16.7241460980745,15.0028356680746,15.5465951732949,15.4894995973234,16.169128959283,16.0996069083821,13.9520295507825,14.0492691434224,13.1502883461063,14.8168226706095,16.3560077655342,14.30563571243,14.2750718138596,14.2603087281481,14.3170903457307,16.6038859325635,13.9567554773633,14.1371656727442,14.4431119040824,15.1950634099335,16.3613300730261,14.4305440239947,14.6030328264996,14.6583565484435,15.3569594653045,16.5368999542995,14.3291992954213,14.4239037658366,14.5587322032289,15.1950634099335,16.3717050181953,15.1620702745062,15.5056676741312,15.1429673906615,15.4079647745043,16.9008155214109,14.6079451777076,14.6753941971444,13.1497550568455,14.9206710618732,17.2934204566841,14.5987126147926,14.8967320984425,14.6970451993658,15.0929012215711,17.1264130896721,14.542155252148,14.4584290430224,14.8614654668171,15.9922404459517,15.3426552970282,14.1001424218123,13.9841362654046,14.0738746130993,15.2370937904406,16.3510586187167,13.4034767756747,13.4654490662827,13.7572472746771,15.5483372770406,16.2848127161883,13.5861396108168,13.8384840329337,13.8699889688218,14.4670316946933,17.0361832734254,14.1156143059506,14.4239037658366,14.4497857465895,15.0182445232662,15.1484821432167,14.3291992954213,14.1946037445951,13.1529914376826,14.708543993667,16.3449160294487],[0.189033824390017,-0.3040061868901,0.286881147788162,0.298658315564515,1.0703893278914,2.3305584000308,1.37851162325373,1.30451104180995,0.978195629681652,1.64154602908752,0.526068811667588,0.238786859587116,0.713695814843359,1.50589092972996,0.333423733725192,1.24488705912353,-0.3040061868901,1.56559717585423,1.09761079662642,0.855989697308481,1.06350294230616,0.20163386116965,0.895302621333307,1.1243281350022,1.37295209791183,1.27500704749987,0.632268215499513,1.16992500144231,0.910732661902913,1.16992500144231,0.8155754288625729,0.333423733725192,0.748461233004036,0.903038270112912,1.13750352374994,0.367371065648529,0.137503523749935,0.464668267003444,0.910732661902913,0.454175893185802,1.35614381022528,0.298658315564515,0.632268215499513,0.748461233004036,1.40599235967584,0.31034012061215,0.124328135002202,0.53605290024021,0.495695162624069,0.321928094887362,0.879705766282288,0.298658315564515,1.01435529297707,1.19534759832222,2.03210084316702,0.933572638261024,0.20163386116965,2.68706068833989,1.13750352374994,0.84799690655495,1.86393845042397,1.41142624572647,0.992768430768924,1.0071955014042,0.443606651475615,0.650764559116902,0.8718436485093179,1.19534759832222,0.16349873228288,1.17632277264046,1.01435529297707,0.941106310946431,1.04264433740849,0.8718436485093179,1.57531233068744,2.51349074558812,0.604071323668861,0.411426245726465,0.584962500721156,1.47508488294878],[1.73552217729654,1.55581615506164,1.93734439215023,2.08406426478847,2.83390207666916,3.23572705983806,3.24031432933371,3.15218341905505,3.16510798514454,3.48284828306847,1.63226821549951,1.91073266190291,2.35049724708413,3.24031432933371,1.57046293102604,1.32192809488736,1.20789285164133,1.20789285164133,2.75274859140713,2.59931779369823,1.91456452349394,1.72246602447109,1.96347412397489,2.81352468929781,3.27202318906105,2.08746284125034,2.02502879449152,2.37016428054021,2.63923216324928,2.94485844580754,1.83187724119167,1.91073266190291,2.0496307677246,2.64154602908752,2.93168305705981,1.86393845042397,1.790772037862,1.86393845042397,1.88362081628567,2.88362081628567,1.83187724119167,2.0071955014042,2.02147972741045,2.48542682717024,2.98913900736823,2.52606881166759,1.49569516262407,1.85598969730848,1.50080205305716,2.06695024392463,2.47508488294878,2.31034012061215,2.39231742277876,2.99457972421575,3.73226919950145,2.20476675065461,2.34482849699744,1.96347412397489,2.40053792958373,2.85798099512757,1.790772037862,2.11103131238874,2.39506279951758,3.0232553523003,2.16671544496642,1.56559717585423,1.32192809488736,3.59693514238723,1.73552217729654,2.75702324650746,2.53106949272595,2.48284828306847,2.7158933705476,2.99819550315325,3.47118746038699,1.0703893278914,1.33342373372519,1.97819562968165,2.58496250072116,3.08406426478847],[10.4475275409266,10.4320198468125,10.4523853997561,10.7211590148677,11.2458081856384,9.02901108665907,9.22583403553784,9.412739192116421,9.91233549843777,9.316688823436939,9.428025110276749,9.412146396540679,9.620146497339491,10.9260809830959,7.3039633562453,9.247642742898019,9.32424806586749,10.2153509830024,11.2601080099627,10.8557639471917,10.6537676442859,10.6473145095471,10.8149028512154,11.0078962674403,10.9341110109735,9.933439857942711,10.009954550333,10.2267734234482,10.7809497987497,10.646036617491,9.797969484059131,9.8490610025615,10.2090268302306,10.7211590148677,10.8557639471917,9.696480751820189,9.72723977748481,9.49325503512067,10.0461417816447,11.0193335777127,10.534828110721,10.8215188973143,10.8050102485124,11.7900226112658,11.031563370024,9.35887093968562,9.18935511254853,9.033670771393091,9.36430978186441,11.0217850685075,8.283551423174311,8.22211866974429,8.617467465294199,9.414283550503709,8.973180744513019,9.425320822759881,9.308885057411761,9.61402881967061,10.4624204732526,10.7035402591663,10.7662472988462,10.9813603523445,11.5188793273505,11.3227900730377,9.89590846480624,10.3333342781343,10.3711558797312,10.374854737178,10.7157046495212,11.7319596151152,9.92478281821723,9.8490610025615,10.0477374043091,10.6360535917989,9.40224465525408,9.797969484059131,10.0853263897027,10.4360035526972,11.191225812124,10.9299462991725]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Donor<\/th>\n      <th>Dose<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[3,4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Donor","targets":1},{"name":"Dose","targets":2},{"name":"IL1B","targets":3},{"name":"IL6","targets":4},{"name":"IL8","targets":5},{"name":"IL10","targets":6},{"name":"TNFa","targets":7},{"name":"VEGF","targets":8}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <div class="sourceCode" id="cb431"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb431-1"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb431-1" tabindex="-1"></a><span class="fu">datatable</span>(demographic_data)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-3189b2bff132d575565a" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-3189b2bff132d575565a">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16"],["D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16"],[22,35,27,60,43,38,26,36,40,44,46,52,58,65,37,25],["M","F","M","F","F","M","M","F","F","M","M","M","M","M","F","F"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Donor<\/th>\n      <th>Age_yr<\/th>\n      <th>Sex<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":2},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Donor","targets":1},{"name":"Age_yr","targets":2},{"name":"Sex","targets":3}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-79cd6db5f6ff52cb1e57" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-79cd6db5f6ff52cb1e57">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16"],["D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16"],[22,35,27,60,43,38,26,36,40,44,46,52,58,65,37,25],["M","F","M","F","F","M","M","F","F","M","M","M","M","M","F","F"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Donor<\/th>\n      <th>Age_yr<\/th>\n      <th>Sex<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":2},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Donor","targets":1},{"name":"Age_yr","targets":2},{"name":"Sex","targets":3}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 </div>
 </div>
 <div id="overview-of-multi-group-statistical-tests" class="section level2 hasAnchor">
@@ -521,8 +530,8 @@ <h2>Multi-Group Analysis Example<a href="multi-group-and-multi-variable-comparis
 <span id="cb434-20"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb434-20" tabindex="-1"></a></span>
 <span id="cb434-21"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb434-21" tabindex="-1"></a><span class="co"># View results</span></span>
 <span id="cb434-22"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb434-22" tabindex="-1"></a><span class="fu">datatable</span>(dose_friedmanres)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-fe37efcf18abe44feef0" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-fe37efcf18abe44feef0">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6"],["IL1B","IL6","IL8","IL10","TNFa","VEGF"],[0.01399579248765089,0.01558587421705306,4.071975170150775e-10,0.002839556743989798,8.904966446852206e-07,2.338026427265107e-07]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>.y.<\/th>\n      <th>p<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":2},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":".y.","targets":1},{"name":"p","targets":2}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-6dd9719d892636eeaba6" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-6dd9719d892636eeaba6">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6"],["IL1B","IL6","IL8","IL10","TNFa","VEGF"],[0.01399579248765089,0.01558587421705306,4.071975170150775e-10,0.002839556743989798,8.904966446852206e-07,2.338026427265107e-07]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>.y.<\/th>\n      <th>p<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":2},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":".y.","targets":1},{"name":"p","targets":2}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>These results demonstrate that all of our endpoints have significant overall differences across doses (p &lt; 0.05). To determine which pairwise comparisons are significant, we next need to apply a post-hoc test. We will apply a pairwise, paired Wilcoxon test due to our experimental design and data distribution, with the Benjamini-Hochberg (BH) correction for multiple testing:</p>
 <div class="sourceCode" id="cb435"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb435-1"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb435-1" tabindex="-1"></a>dose_wilcox_posthoc_IL1B <span class="ot">&lt;-</span> biomarker_data <span class="sc">%&gt;%</span> </span>
 <span id="cb435-2"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb435-2" tabindex="-1"></a>  <span class="fu">pairwise_wilcox_test</span>(IL1B <span class="sc">~</span> Dose, <span class="at">paired =</span> <span class="cn">TRUE</span>, <span class="at">p.adjust.method =</span> <span class="st">&quot;BH&quot;</span>)</span>
@@ -563,8 +572,8 @@ <h2>Multi-Group Analysis Example<a href="multi-group-and-multi-variable-comparis
 <span id="cb437-19"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb437-19" tabindex="-1"></a></span>
 <span id="cb437-20"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb437-20" tabindex="-1"></a><span class="co"># View results</span></span>
 <span id="cb437-21"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb437-21" tabindex="-1"></a><span class="fu">datatable</span>(dose_wilcox_posthoc)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-d76d627236241d7b8d98" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-d76d627236241d7b8d98">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60"],["IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL6","IL6","IL6","IL6","IL6","IL6","IL6","IL6","IL6","IL6","IL8","IL8","IL8","IL8","IL8","IL8","IL8","IL8","IL8","IL8","IL10","IL10","IL10","IL10","IL10","IL10","IL10","IL10","IL10","IL10","TNFa","TNFa","TNFa","TNFa","TNFa","TNFa","TNFa","TNFa","TNFa","TNFa","VEGF","VEGF","VEGF","VEGF","VEGF","VEGF","VEGF","VEGF","VEGF","VEGF"],["0","0","0","0","0.6","0.6","0.6","1","1","2","0","0","0","0","0.6","0.6","0.6","1","1","2","0","0","0","0","0.6","0.6","0.6","1","1","2","0","0","0","0","0.6","0.6","0.6","1","1","2","0","0","0","0","0.6","0.6","0.6","1","1","2","0","0","0","0","0.6","0.6","0.6","1","1","2"],["0.6","1","2","4","1","2","4","2","4","4","0.6","1","2","4","1","2","4","2","4","4","0.6","1","2","4","1","2","4","2","4","4","0.6","1","2","4","1","2","4","2","4","4","0.6","1","2","4","1","2","4","2","4","4","0.6","1","2","4","1","2","4","2","4","4"],[16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16],[16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16],[17,11,21,18,38,31,25,43,40,62,38,76,76,114,96,83,116,72,118,117,22,53,0,0,78,2,0,0,0,5,133,69,84,65,13,24,11,74,45,41,72,30,14,5,15,3,3,22,15,30,44,17,0,28,16,0,30,2,39,84],[0.006,0.002,0.013,0.008,0.13,0.058,0.025,0.211,0.159,0.782,0.127,0.706,0.706,0.016,0.159,0.464,0.011,0.86,0.008,0.008999999999999999,0.016,0.464,3.05e-05,3.05e-05,0.632,9.16e-05,3.05e-05,3.05e-05,3.05e-05,0.000305,0.000153,0.98,0.433,0.9,0.003,0.021,0.002,0.782,0.41,0.175,0.86,0.094,0.003,0.000305,0.012,0.000153,0.000153,0.016,0.004,0.051,0.231,0.006,3.05e-05,0.039,0.005,3.05e-05,0.051,9.16e-05,0.144,0.433],[0.025,0.017,0.033,0.025,0.186,0.096,0.05,0.234,0.199,0.782,0.254,0.784,0.784,0.039,0.265,0.663,0.037,0.86,0.037,0.037,0.019,0.516,6.1e-05,6.1e-05,0.632,0.000153,6.1e-05,6.1e-05,6.1e-05,0.000436,0.002,0.98,0.619,0.98,0.008999999999999999,0.054,0.008,0.978,0.619,0.35,0.86,0.104,0.008,0.001,0.019,0.0007649999999999999,0.0007649999999999999,0.022,0.008,0.063,0.257,0.013,0.000152,0.064,0.013,0.000152,0.07199999999999999,0.000305,0.18,0.433],["*","*","*","*","ns","ns","*","ns","ns","ns","ns","ns","ns","*","ns","ns","*","ns","*","*","*","ns","****","****","ns","***","****","****","****","***","**","ns","ns","ns","**","ns","**","ns","ns","ns","ns","ns","**","**","*","***","***","*","**","ns","ns","*","***","ns","*","***","ns","***","ns","ns"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>.y.<\/th>\n      <th>group1<\/th>\n      <th>group2<\/th>\n      <th>n1<\/th>\n      <th>n2<\/th>\n      <th>statistic<\/th>\n      <th>p<\/th>\n      <th>p.adj<\/th>\n      <th>p.adj.signif<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":".y.","targets":1},{"name":"group1","targets":2},{"name":"group2","targets":3},{"name":"n1","targets":4},{"name":"n2","targets":5},{"name":"statistic","targets":6},{"name":"p","targets":7},{"name":"p.adj","targets":8},{"name":"p.adj.signif","targets":9}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-88f7c9912b5cbcb55f6d" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-88f7c9912b5cbcb55f6d">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60"],["IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL6","IL6","IL6","IL6","IL6","IL6","IL6","IL6","IL6","IL6","IL8","IL8","IL8","IL8","IL8","IL8","IL8","IL8","IL8","IL8","IL10","IL10","IL10","IL10","IL10","IL10","IL10","IL10","IL10","IL10","TNFa","TNFa","TNFa","TNFa","TNFa","TNFa","TNFa","TNFa","TNFa","TNFa","VEGF","VEGF","VEGF","VEGF","VEGF","VEGF","VEGF","VEGF","VEGF","VEGF"],["0","0","0","0","0.6","0.6","0.6","1","1","2","0","0","0","0","0.6","0.6","0.6","1","1","2","0","0","0","0","0.6","0.6","0.6","1","1","2","0","0","0","0","0.6","0.6","0.6","1","1","2","0","0","0","0","0.6","0.6","0.6","1","1","2","0","0","0","0","0.6","0.6","0.6","1","1","2"],["0.6","1","2","4","1","2","4","2","4","4","0.6","1","2","4","1","2","4","2","4","4","0.6","1","2","4","1","2","4","2","4","4","0.6","1","2","4","1","2","4","2","4","4","0.6","1","2","4","1","2","4","2","4","4","0.6","1","2","4","1","2","4","2","4","4"],[16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16],[16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16],[17,11,21,18,38,31,25,43,40,62,38,76,76,114,96,83,116,72,118,117,22,53,0,0,78,2,0,0,0,5,133,69,84,65,13,24,11,74,45,41,72,30,14,5,15,3,3,22,15,30,44,17,0,28,16,0,30,2,39,84],[0.006,0.002,0.013,0.008,0.13,0.058,0.025,0.211,0.159,0.782,0.127,0.706,0.706,0.016,0.159,0.464,0.011,0.86,0.008,0.008999999999999999,0.016,0.464,3.05e-05,3.05e-05,0.632,9.16e-05,3.05e-05,3.05e-05,3.05e-05,0.000305,0.000153,0.98,0.433,0.9,0.003,0.021,0.002,0.782,0.41,0.175,0.86,0.094,0.003,0.000305,0.012,0.000153,0.000153,0.016,0.004,0.051,0.231,0.006,3.05e-05,0.039,0.005,3.05e-05,0.051,9.16e-05,0.144,0.433],[0.025,0.017,0.033,0.025,0.186,0.096,0.05,0.234,0.199,0.782,0.254,0.784,0.784,0.039,0.265,0.663,0.037,0.86,0.037,0.037,0.019,0.516,6.1e-05,6.1e-05,0.632,0.000153,6.1e-05,6.1e-05,6.1e-05,0.000436,0.002,0.98,0.619,0.98,0.008999999999999999,0.054,0.008,0.978,0.619,0.35,0.86,0.104,0.008,0.001,0.019,0.0007649999999999999,0.0007649999999999999,0.022,0.008,0.063,0.257,0.013,0.000152,0.064,0.013,0.000152,0.07199999999999999,0.000305,0.18,0.433],["*","*","*","*","ns","ns","*","ns","ns","ns","ns","ns","ns","*","ns","ns","*","ns","*","*","*","ns","****","****","ns","***","****","****","****","***","**","ns","ns","ns","**","ns","**","ns","ns","ns","ns","ns","**","**","*","***","***","*","**","ns","ns","*","***","ns","*","***","ns","***","ns","ns"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>.y.<\/th>\n      <th>group1<\/th>\n      <th>group2<\/th>\n      <th>n1<\/th>\n      <th>n2<\/th>\n      <th>statistic<\/th>\n      <th>p<\/th>\n      <th>p.adj<\/th>\n      <th>p.adj.signif<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":".y.","targets":1},{"name":"group1","targets":2},{"name":"group2","targets":3},{"name":"n1","targets":4},{"name":"n2","targets":5},{"name":"statistic","targets":6},{"name":"p","targets":7},{"name":"p.adj","targets":8},{"name":"p.adj.signif","targets":9}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>We now have a dataframe storing all of our pairwise comparison results. However, this is a lot to scroll through, making it hard to interpret. We can generate a publication-quality table by manipulating the table and joining it with the overall test data.</p>
 <div class="sourceCode" id="cb438"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb438-1"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb438-1" tabindex="-1"></a>dose_results_cleaned <span class="ot">&lt;-</span> dose_wilcox_posthoc <span class="sc">%&gt;%</span></span>
 <span id="cb438-2"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb438-2" tabindex="-1"></a>  <span class="fu">unite</span>(comparison, group1, group2, <span class="at">sep =</span> <span class="st">&quot; vs. &quot;</span>) <span class="sc">%&gt;%</span></span>
@@ -576,8 +585,8 @@ <h2>Multi-Group Analysis Example<a href="multi-group-and-multi-variable-comparis
 <span id="cb438-8"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb438-8" tabindex="-1"></a>  <span class="fu">mutate</span>(<span class="fu">across</span>(<span class="st">&#39;Overall&#39;</span><span class="sc">:</span><span class="st">&#39;2 vs. 4&#39;</span>, \(x) <span class="fu">format</span>(x, <span class="at">scientific =</span> <span class="cn">TRUE</span>, <span class="at">digits =</span> <span class="dv">3</span>)))</span>
 <span id="cb438-9"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb438-9" tabindex="-1"></a></span>
 <span id="cb438-10"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb438-10" tabindex="-1"></a><span class="fu">datatable</span>(dose_results_cleaned)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-39a8ff0d38334e0b2c8a" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-39a8ff0d38334e0b2c8a">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6"],["IL1B","IL6","IL8","IL10","TNFa","VEGF"],["1.40e-02","1.56e-02","4.07e-10","2.84e-03","8.90e-07","2.34e-07"],["2.50e-02","2.54e-01","1.90e-02","2.00e-03","8.60e-01","2.57e-01"],["1.70e-02","7.84e-01","5.16e-01","9.80e-01","1.04e-01","1.30e-02"],["3.30e-02","7.84e-01","6.10e-05","6.19e-01","8.00e-03","1.52e-04"],["2.5e-02","3.9e-02","6.1e-05","9.8e-01","1.0e-03","6.4e-02"],["1.86e-01","2.65e-01","6.32e-01","9.00e-03","1.90e-02","1.30e-02"],["9.60e-02","6.63e-01","1.53e-04","5.40e-02","7.65e-04","1.52e-04"],["5.00e-02","3.70e-02","6.10e-05","8.00e-03","7.65e-04","7.20e-02"],["2.34e-01","8.60e-01","6.10e-05","9.78e-01","2.20e-02","3.05e-04"],["1.99e-01","3.70e-02","6.10e-05","6.19e-01","8.00e-03","1.80e-01"],["7.82e-01","3.70e-02","4.36e-04","3.50e-01","6.30e-02","4.33e-01"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Variable<\/th>\n      <th>Overall<\/th>\n      <th>0 vs. 0.6<\/th>\n      <th>0 vs. 1<\/th>\n      <th>0 vs. 2<\/th>\n      <th>0 vs. 4<\/th>\n      <th>0.6 vs. 1<\/th>\n      <th>0.6 vs. 2<\/th>\n      <th>0.6 vs. 4<\/th>\n      <th>1 vs. 2<\/th>\n      <th>1 vs. 4<\/th>\n      <th>2 vs. 4<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Variable","targets":1},{"name":"Overall","targets":2},{"name":"0 vs. 0.6","targets":3},{"name":"0 vs. 1","targets":4},{"name":"0 vs. 2","targets":5},{"name":"0 vs. 4","targets":6},{"name":"0.6 vs. 1","targets":7},{"name":"0.6 vs. 2","targets":8},{"name":"0.6 vs. 4","targets":9},{"name":"1 vs. 2","targets":10},{"name":"1 vs. 4","targets":11},{"name":"2 vs. 4","targets":12}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-a1fc7b73b2fdaea7dc33" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-a1fc7b73b2fdaea7dc33">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6"],["IL1B","IL6","IL8","IL10","TNFa","VEGF"],["1.40e-02","1.56e-02","4.07e-10","2.84e-03","8.90e-07","2.34e-07"],["2.50e-02","2.54e-01","1.90e-02","2.00e-03","8.60e-01","2.57e-01"],["1.70e-02","7.84e-01","5.16e-01","9.80e-01","1.04e-01","1.30e-02"],["3.30e-02","7.84e-01","6.10e-05","6.19e-01","8.00e-03","1.52e-04"],["2.5e-02","3.9e-02","6.1e-05","9.8e-01","1.0e-03","6.4e-02"],["1.86e-01","2.65e-01","6.32e-01","9.00e-03","1.90e-02","1.30e-02"],["9.60e-02","6.63e-01","1.53e-04","5.40e-02","7.65e-04","1.52e-04"],["5.00e-02","3.70e-02","6.10e-05","8.00e-03","7.65e-04","7.20e-02"],["2.34e-01","8.60e-01","6.10e-05","9.78e-01","2.20e-02","3.05e-04"],["1.99e-01","3.70e-02","6.10e-05","6.19e-01","8.00e-03","1.80e-01"],["7.82e-01","3.70e-02","4.36e-04","3.50e-01","6.30e-02","4.33e-01"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Variable<\/th>\n      <th>Overall<\/th>\n      <th>0 vs. 0.6<\/th>\n      <th>0 vs. 1<\/th>\n      <th>0 vs. 2<\/th>\n      <th>0 vs. 4<\/th>\n      <th>0.6 vs. 1<\/th>\n      <th>0.6 vs. 2<\/th>\n      <th>0.6 vs. 4<\/th>\n      <th>1 vs. 2<\/th>\n      <th>1 vs. 4<\/th>\n      <th>2 vs. 4<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Variable","targets":1},{"name":"Overall","targets":2},{"name":"0 vs. 0.6","targets":3},{"name":"0 vs. 1","targets":4},{"name":"0 vs. 2","targets":5},{"name":"0 vs. 4","targets":6},{"name":"0.6 vs. 1","targets":7},{"name":"0.6 vs. 2","targets":8},{"name":"0.6 vs. 4","targets":9},{"name":"1 vs. 2","targets":10},{"name":"1 vs. 4","targets":11},{"name":"2 vs. 4","targets":12}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>To more easily see overall significance patterns, we could also make the same table but with significance stars instead of p-values by keeping the <code>p.adjust.signif</code> column instead of the <code>p.adj</code> column in our post-hoc test results dataframe:</p>
 <div class="sourceCode" id="cb439"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb439-1"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb439-1" tabindex="-1"></a>dose_results_cleaned_2 <span class="ot">&lt;-</span> dose_wilcox_posthoc <span class="sc">%&gt;%</span></span>
 <span id="cb439-2"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb439-2" tabindex="-1"></a>  <span class="fu">unite</span>(comparison, group1, group2, <span class="at">sep =</span> <span class="st">&quot; vs. &quot;</span>) <span class="sc">%&gt;%</span></span>
@@ -589,8 +598,8 @@ <h2>Multi-Group Analysis Example<a href="multi-group-and-multi-variable-comparis
 <span id="cb439-8"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb439-8" tabindex="-1"></a>  <span class="fu">mutate</span>(<span class="fu">across</span>(<span class="st">&#39;Overall&#39;</span><span class="sc">:</span><span class="st">&#39;2 vs. 4&#39;</span>, \(x) <span class="fu">format</span>(x, <span class="at">scientific =</span> <span class="cn">TRUE</span>, <span class="at">digits =</span> <span class="dv">3</span>)))</span>
 <span id="cb439-9"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb439-9" tabindex="-1"></a></span>
 <span id="cb439-10"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb439-10" tabindex="-1"></a><span class="fu">datatable</span>(dose_results_cleaned_2)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-a7dc63436490a80a2955" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-a7dc63436490a80a2955">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6"],["IL1B","IL6","IL8","IL10","TNFa","VEGF"],["1.40e-02","1.56e-02","4.07e-10","2.84e-03","8.90e-07","2.34e-07"],["* ","ns","* ","**","ns","ns"],["* ","ns","ns","ns","ns","* "],["*   ","ns  ","****","ns  ","**  ","*** "],["*   ","*   ","****","ns  ","**  ","ns  "],["ns","ns","ns","**","* ","* "],["ns ","ns ","***","ns ","***","***"],["*   ","*   ","****","**  ","*** ","ns  "],["ns  ","ns  ","****","ns  ","*   ","*** "],["ns  ","*   ","****","ns  ","**  ","ns  "],["ns ","*  ","***","ns ","ns ","ns "]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Variable<\/th>\n      <th>Overall<\/th>\n      <th>0 vs. 0.6<\/th>\n      <th>0 vs. 1<\/th>\n      <th>0 vs. 2<\/th>\n      <th>0 vs. 4<\/th>\n      <th>0.6 vs. 1<\/th>\n      <th>0.6 vs. 2<\/th>\n      <th>0.6 vs. 4<\/th>\n      <th>1 vs. 2<\/th>\n      <th>1 vs. 4<\/th>\n      <th>2 vs. 4<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Variable","targets":1},{"name":"Overall","targets":2},{"name":"0 vs. 0.6","targets":3},{"name":"0 vs. 1","targets":4},{"name":"0 vs. 2","targets":5},{"name":"0 vs. 4","targets":6},{"name":"0.6 vs. 1","targets":7},{"name":"0.6 vs. 2","targets":8},{"name":"0.6 vs. 4","targets":9},{"name":"1 vs. 2","targets":10},{"name":"1 vs. 4","targets":11},{"name":"2 vs. 4","targets":12}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-8a00cb0bb6ba1817b7b6" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-8a00cb0bb6ba1817b7b6">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6"],["IL1B","IL6","IL8","IL10","TNFa","VEGF"],["1.40e-02","1.56e-02","4.07e-10","2.84e-03","8.90e-07","2.34e-07"],["* ","ns","* ","**","ns","ns"],["* ","ns","ns","ns","ns","* "],["*   ","ns  ","****","ns  ","**  ","*** "],["*   ","*   ","****","ns  ","**  ","ns  "],["ns","ns","ns","**","* ","* "],["ns ","ns ","***","ns ","***","***"],["*   ","*   ","****","**  ","*** ","ns  "],["ns  ","ns  ","****","ns  ","*   ","*** "],["ns  ","*   ","****","ns  ","**  ","ns  "],["ns ","*  ","***","ns ","ns ","ns "]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Variable<\/th>\n      <th>Overall<\/th>\n      <th>0 vs. 0.6<\/th>\n      <th>0 vs. 1<\/th>\n      <th>0 vs. 2<\/th>\n      <th>0 vs. 4<\/th>\n      <th>0.6 vs. 1<\/th>\n      <th>0.6 vs. 2<\/th>\n      <th>0.6 vs. 4<\/th>\n      <th>1 vs. 2<\/th>\n      <th>1 vs. 4<\/th>\n      <th>2 vs. 4<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Variable","targets":1},{"name":"Overall","targets":2},{"name":"0 vs. 0.6","targets":3},{"name":"0 vs. 1","targets":4},{"name":"0 vs. 2","targets":5},{"name":"0 vs. 4","targets":6},{"name":"0.6 vs. 1","targets":7},{"name":"0.6 vs. 2","targets":8},{"name":"0.6 vs. 4","targets":9},{"name":"1 vs. 2","targets":10},{"name":"1 vs. 4","targets":11},{"name":"2 vs. 4","targets":12}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <div id="answer-to-environmental-health-question-1-5" class="section level3 hasAnchor">
 <h3>Answer to Environmental Health Question 1<a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#answer-to-environmental-health-question-1-5" class="anchor-section" aria-label="Anchor link to header"></a></h3>
 <div class="question">
@@ -615,15 +624,15 @@ <h3>Single Plot<a href="multi-group-and-multi-variable-comparisons-and-visualiza
 </ul>
 <p>Our results dataframe for IL-1<span class="math inline">\(\beta\)</span> already contains our groups and p-values:</p>
 <div class="sourceCode" id="cb440"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb440-1"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb440-1" tabindex="-1"></a><span class="fu">datatable</span>(dose_wilcox_posthoc_IL1B)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-7d6c91d8b3128295f2e9" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-7d6c91d8b3128295f2e9">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10"],["IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B"],["0","0","0","0","0.6","0.6","0.6","1","1","2"],["0.6","1","2","4","1","2","4","2","4","4"],[16,16,16,16,16,16,16,16,16,16],[16,16,16,16,16,16,16,16,16,16],[17,11,21,18,38,31,25,43,40,62],[0.006,0.002,0.013,0.008,0.13,0.058,0.025,0.211,0.159,0.782],[0.025,0.017,0.033,0.025,0.186,0.096,0.05,0.234,0.199,0.782],["*","*","*","*","ns","ns","*","ns","ns","ns"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>.y.<\/th>\n      <th>group1<\/th>\n      <th>group2<\/th>\n      <th>n1<\/th>\n      <th>n2<\/th>\n      <th>statistic<\/th>\n      <th>p<\/th>\n      <th>p.adj<\/th>\n      <th>p.adj.signif<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":".y.","targets":1},{"name":"group1","targets":2},{"name":"group2","targets":3},{"name":"n1","targets":4},{"name":"n2","targets":5},{"name":"statistic","targets":6},{"name":"p","targets":7},{"name":"p.adj","targets":8},{"name":"p.adj.signif","targets":9}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-e6f579a04f37d03b59d1" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-e6f579a04f37d03b59d1">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10"],["IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B"],["0","0","0","0","0.6","0.6","0.6","1","1","2"],["0.6","1","2","4","1","2","4","2","4","4"],[16,16,16,16,16,16,16,16,16,16],[16,16,16,16,16,16,16,16,16,16],[17,11,21,18,38,31,25,43,40,62],[0.006,0.002,0.013,0.008,0.13,0.058,0.025,0.211,0.159,0.782],[0.025,0.017,0.033,0.025,0.186,0.096,0.05,0.234,0.199,0.782],["*","*","*","*","ns","ns","*","ns","ns","ns"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>.y.<\/th>\n      <th>group1<\/th>\n      <th>group2<\/th>\n      <th>n1<\/th>\n      <th>n2<\/th>\n      <th>statistic<\/th>\n      <th>p<\/th>\n      <th>p.adj<\/th>\n      <th>p.adj.signif<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":".y.","targets":1},{"name":"group1","targets":2},{"name":"group2","targets":3},{"name":"n1","targets":4},{"name":"n2","targets":5},{"name":"statistic","targets":6},{"name":"p","targets":7},{"name":"p.adj","targets":8},{"name":"p.adj.signif","targets":9}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>We can add the position columns using the function <code>add_xy_position()</code>:</p>
 <div class="sourceCode" id="cb441"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb441-1"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb441-1" tabindex="-1"></a>dose_wilcox_posthoc_IL1B <span class="ot">&lt;-</span> dose_wilcox_posthoc_IL1B <span class="sc">%&gt;%</span></span>
 <span id="cb441-2"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb441-2" tabindex="-1"></a>  <span class="fu">add_xy_position</span>(<span class="at">x =</span> <span class="st">&quot;Dose&quot;</span>, <span class="at">step.increase =</span> <span class="dv">2</span>)</span>
 <span id="cb441-3"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb441-3" tabindex="-1"></a></span>
 <span id="cb441-4"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb441-4" tabindex="-1"></a><span class="fu">datatable</span>(dose_wilcox_posthoc_IL1B)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-92c421874717c4c09655" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-92c421874717c4c09655">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10"],["IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B"],["0","0","0","0","0.6","0.6","0.6","1","1","2"],["0.6","1","2","4","1","2","4","2","4","4"],[16,16,16,16,16,16,16,16,16,16],[16,16,16,16,16,16,16,16,16,16],[17,11,21,18,38,31,25,43,40,62],[0.006,0.002,0.013,0.008,0.13,0.058,0.025,0.211,0.159,0.782],[0.025,0.017,0.033,0.025,0.186,0.096,0.05,0.234,0.199,0.782],["*","*","*","*","ns","ns","*","ns","ns","ns"],[4.749999999999999,5.427777777777776,6.105555555555553,6.783333333333331,7.461111111111109,8.138888888888886,8.816666666666663,9.49444444444444,10.17222222222222,10.84999999999999],[["0","0.6"],["0","1"],["0","2"],["0","4"],["0.6","1"],["0.6","2"],["0.6","4"],["1","2"],["1","4"],["2","4"]],[1,1,1,1,2,2,2,3,3,4],[2,3,4,5,3,4,5,4,5,5]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>.y.<\/th>\n      <th>group1<\/th>\n      <th>group2<\/th>\n      <th>n1<\/th>\n      <th>n2<\/th>\n      <th>statistic<\/th>\n      <th>p<\/th>\n      <th>p.adj<\/th>\n      <th>p.adj.signif<\/th>\n      <th>y.position<\/th>\n      <th>groups<\/th>\n      <th>xmin<\/th>\n      <th>xmax<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[4,5,6,7,8,10,12,13]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":".y.","targets":1},{"name":"group1","targets":2},{"name":"group2","targets":3},{"name":"n1","targets":4},{"name":"n2","targets":5},{"name":"statistic","targets":6},{"name":"p","targets":7},{"name":"p.adj","targets":8},{"name":"p.adj.signif","targets":9},{"name":"y.position","targets":10},{"name":"groups","targets":11},{"name":"xmin","targets":12},{"name":"xmax","targets":13}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-c1fb89b44d7155413f8d" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-c1fb89b44d7155413f8d">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10"],["IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B"],["0","0","0","0","0.6","0.6","0.6","1","1","2"],["0.6","1","2","4","1","2","4","2","4","4"],[16,16,16,16,16,16,16,16,16,16],[16,16,16,16,16,16,16,16,16,16],[17,11,21,18,38,31,25,43,40,62],[0.006,0.002,0.013,0.008,0.13,0.058,0.025,0.211,0.159,0.782],[0.025,0.017,0.033,0.025,0.186,0.096,0.05,0.234,0.199,0.782],["*","*","*","*","ns","ns","*","ns","ns","ns"],[4.749999999999999,5.427777777777776,6.105555555555553,6.783333333333331,7.461111111111109,8.138888888888886,8.816666666666663,9.49444444444444,10.17222222222222,10.84999999999999],[["0","0.6"],["0","1"],["0","2"],["0","4"],["0.6","1"],["0.6","2"],["0.6","4"],["1","2"],["1","4"],["2","4"]],[1,1,1,1,2,2,2,3,3,4],[2,3,4,5,3,4,5,4,5,5]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>.y.<\/th>\n      <th>group1<\/th>\n      <th>group2<\/th>\n      <th>n1<\/th>\n      <th>n2<\/th>\n      <th>statistic<\/th>\n      <th>p<\/th>\n      <th>p.adj<\/th>\n      <th>p.adj.signif<\/th>\n      <th>y.position<\/th>\n      <th>groups<\/th>\n      <th>xmin<\/th>\n      <th>xmax<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[4,5,6,7,8,10,12,13]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":".y.","targets":1},{"name":"group1","targets":2},{"name":"group2","targets":3},{"name":"n1","targets":4},{"name":"n2","targets":5},{"name":"statistic","targets":6},{"name":"p","targets":7},{"name":"p.adj","targets":8},{"name":"p.adj.signif","targets":9},{"name":"y.position","targets":10},{"name":"groups","targets":11},{"name":"xmin","targets":12},{"name":"xmax","targets":13}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>Now, we are ready to make a graph of our results. We will use <code>stat_friedman_test()</code> to add our overall p-value and <code>stat_pvalue_manual()</code> to add our pairwise values.</p>
 <div class="sourceCode" id="cb442"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb442-1"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb442-1" tabindex="-1"></a><span class="co"># Set graphing theme</span></span>
 <span id="cb442-2"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb442-2" tabindex="-1"></a><span class="fu">theme_set</span>(<span class="fu">theme_bw</span>())</span>
@@ -715,8 +724,8 @@ <h3>Faceted Plot<a href="multi-group-and-multi-variable-comparisons-and-visualiz
 </ol>
 <p>First, let’s create our annotations dataframe. We will start with the results dataframe from our posthoc testing:</p>
 <div class="sourceCode" id="cb446"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb446-1"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb446-1" tabindex="-1"></a><span class="fu">datatable</span>(dose_wilcox_posthoc)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-effa83a3ed7397929b93" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-effa83a3ed7397929b93">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60"],["IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL6","IL6","IL6","IL6","IL6","IL6","IL6","IL6","IL6","IL6","IL8","IL8","IL8","IL8","IL8","IL8","IL8","IL8","IL8","IL8","IL10","IL10","IL10","IL10","IL10","IL10","IL10","IL10","IL10","IL10","TNFa","TNFa","TNFa","TNFa","TNFa","TNFa","TNFa","TNFa","TNFa","TNFa","VEGF","VEGF","VEGF","VEGF","VEGF","VEGF","VEGF","VEGF","VEGF","VEGF"],["0","0","0","0","0.6","0.6","0.6","1","1","2","0","0","0","0","0.6","0.6","0.6","1","1","2","0","0","0","0","0.6","0.6","0.6","1","1","2","0","0","0","0","0.6","0.6","0.6","1","1","2","0","0","0","0","0.6","0.6","0.6","1","1","2","0","0","0","0","0.6","0.6","0.6","1","1","2"],["0.6","1","2","4","1","2","4","2","4","4","0.6","1","2","4","1","2","4","2","4","4","0.6","1","2","4","1","2","4","2","4","4","0.6","1","2","4","1","2","4","2","4","4","0.6","1","2","4","1","2","4","2","4","4","0.6","1","2","4","1","2","4","2","4","4"],[16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16],[16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16],[17,11,21,18,38,31,25,43,40,62,38,76,76,114,96,83,116,72,118,117,22,53,0,0,78,2,0,0,0,5,133,69,84,65,13,24,11,74,45,41,72,30,14,5,15,3,3,22,15,30,44,17,0,28,16,0,30,2,39,84],[0.006,0.002,0.013,0.008,0.13,0.058,0.025,0.211,0.159,0.782,0.127,0.706,0.706,0.016,0.159,0.464,0.011,0.86,0.008,0.008999999999999999,0.016,0.464,3.05e-05,3.05e-05,0.632,9.16e-05,3.05e-05,3.05e-05,3.05e-05,0.000305,0.000153,0.98,0.433,0.9,0.003,0.021,0.002,0.782,0.41,0.175,0.86,0.094,0.003,0.000305,0.012,0.000153,0.000153,0.016,0.004,0.051,0.231,0.006,3.05e-05,0.039,0.005,3.05e-05,0.051,9.16e-05,0.144,0.433],[0.025,0.017,0.033,0.025,0.186,0.096,0.05,0.234,0.199,0.782,0.254,0.784,0.784,0.039,0.265,0.663,0.037,0.86,0.037,0.037,0.019,0.516,6.1e-05,6.1e-05,0.632,0.000153,6.1e-05,6.1e-05,6.1e-05,0.000436,0.002,0.98,0.619,0.98,0.008999999999999999,0.054,0.008,0.978,0.619,0.35,0.86,0.104,0.008,0.001,0.019,0.0007649999999999999,0.0007649999999999999,0.022,0.008,0.063,0.257,0.013,0.000152,0.064,0.013,0.000152,0.07199999999999999,0.000305,0.18,0.433],["*","*","*","*","ns","ns","*","ns","ns","ns","ns","ns","ns","*","ns","ns","*","ns","*","*","*","ns","****","****","ns","***","****","****","****","***","**","ns","ns","ns","**","ns","**","ns","ns","ns","ns","ns","**","**","*","***","***","*","**","ns","ns","*","***","ns","*","***","ns","***","ns","ns"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>.y.<\/th>\n      <th>group1<\/th>\n      <th>group2<\/th>\n      <th>n1<\/th>\n      <th>n2<\/th>\n      <th>statistic<\/th>\n      <th>p<\/th>\n      <th>p.adj<\/th>\n      <th>p.adj.signif<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":".y.","targets":1},{"name":"group1","targets":2},{"name":"group2","targets":3},{"name":"n1","targets":4},{"name":"n2","targets":5},{"name":"statistic","targets":6},{"name":"p","targets":7},{"name":"p.adj","targets":8},{"name":"p.adj.signif","targets":9}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-df0f55e30b0d718da6c7" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-df0f55e30b0d718da6c7">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60"],["IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL1B","IL6","IL6","IL6","IL6","IL6","IL6","IL6","IL6","IL6","IL6","IL8","IL8","IL8","IL8","IL8","IL8","IL8","IL8","IL8","IL8","IL10","IL10","IL10","IL10","IL10","IL10","IL10","IL10","IL10","IL10","TNFa","TNFa","TNFa","TNFa","TNFa","TNFa","TNFa","TNFa","TNFa","TNFa","VEGF","VEGF","VEGF","VEGF","VEGF","VEGF","VEGF","VEGF","VEGF","VEGF"],["0","0","0","0","0.6","0.6","0.6","1","1","2","0","0","0","0","0.6","0.6","0.6","1","1","2","0","0","0","0","0.6","0.6","0.6","1","1","2","0","0","0","0","0.6","0.6","0.6","1","1","2","0","0","0","0","0.6","0.6","0.6","1","1","2","0","0","0","0","0.6","0.6","0.6","1","1","2"],["0.6","1","2","4","1","2","4","2","4","4","0.6","1","2","4","1","2","4","2","4","4","0.6","1","2","4","1","2","4","2","4","4","0.6","1","2","4","1","2","4","2","4","4","0.6","1","2","4","1","2","4","2","4","4","0.6","1","2","4","1","2","4","2","4","4"],[16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16],[16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16,16],[17,11,21,18,38,31,25,43,40,62,38,76,76,114,96,83,116,72,118,117,22,53,0,0,78,2,0,0,0,5,133,69,84,65,13,24,11,74,45,41,72,30,14,5,15,3,3,22,15,30,44,17,0,28,16,0,30,2,39,84],[0.006,0.002,0.013,0.008,0.13,0.058,0.025,0.211,0.159,0.782,0.127,0.706,0.706,0.016,0.159,0.464,0.011,0.86,0.008,0.008999999999999999,0.016,0.464,3.05e-05,3.05e-05,0.632,9.16e-05,3.05e-05,3.05e-05,3.05e-05,0.000305,0.000153,0.98,0.433,0.9,0.003,0.021,0.002,0.782,0.41,0.175,0.86,0.094,0.003,0.000305,0.012,0.000153,0.000153,0.016,0.004,0.051,0.231,0.006,3.05e-05,0.039,0.005,3.05e-05,0.051,9.16e-05,0.144,0.433],[0.025,0.017,0.033,0.025,0.186,0.096,0.05,0.234,0.199,0.782,0.254,0.784,0.784,0.039,0.265,0.663,0.037,0.86,0.037,0.037,0.019,0.516,6.1e-05,6.1e-05,0.632,0.000153,6.1e-05,6.1e-05,6.1e-05,0.000436,0.002,0.98,0.619,0.98,0.008999999999999999,0.054,0.008,0.978,0.619,0.35,0.86,0.104,0.008,0.001,0.019,0.0007649999999999999,0.0007649999999999999,0.022,0.008,0.063,0.257,0.013,0.000152,0.064,0.013,0.000152,0.07199999999999999,0.000305,0.18,0.433],["*","*","*","*","ns","ns","*","ns","ns","ns","ns","ns","ns","*","ns","ns","*","ns","*","*","*","ns","****","****","ns","***","****","****","****","***","**","ns","ns","ns","**","ns","**","ns","ns","ns","ns","ns","**","**","*","***","***","*","**","ns","ns","*","***","ns","*","***","ns","***","ns","ns"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>.y.<\/th>\n      <th>group1<\/th>\n      <th>group2<\/th>\n      <th>n1<\/th>\n      <th>n2<\/th>\n      <th>statistic<\/th>\n      <th>p<\/th>\n      <th>p.adj<\/th>\n      <th>p.adj.signif<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":".y.","targets":1},{"name":"group1","targets":2},{"name":"group2","targets":3},{"name":"n1","targets":4},{"name":"n2","targets":5},{"name":"statistic","targets":6},{"name":"p","targets":7},{"name":"p.adj","targets":8},{"name":"p.adj.signif","targets":9}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <div class="sourceCode" id="cb447"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb447-1"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb447-1" tabindex="-1"></a>dose_wilcox_posthoc_forgraph <span class="ot">&lt;-</span> dose_wilcox_posthoc <span class="sc">%&gt;%</span></span>
 <span id="cb447-2"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb447-2" tabindex="-1"></a>  </span>
 <span id="cb447-3"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb447-3" tabindex="-1"></a>  <span class="fu">filter</span>(p.adj <span class="sc">&lt;=</span> <span class="fl">0.05</span>) <span class="sc">%&gt;%</span></span>
@@ -728,8 +737,8 @@ <h3>Faceted Plot<a href="multi-group-and-multi-variable-comparisons-and-visualiz
 <span id="cb447-9"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb447-9" tabindex="-1"></a>  <span class="fu">rename</span>(<span class="st">&quot;variable&quot;</span> <span class="ot">=</span> <span class="st">&quot;.y.&quot;</span>, <span class="st">&quot;Dose&quot;</span> <span class="ot">=</span> <span class="st">&quot;group2&quot;</span>)</span>
 <span id="cb447-10"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb447-10" tabindex="-1"></a>  </span>
 <span id="cb447-11"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb447-11" tabindex="-1"></a><span class="fu">datatable</span>(dose_wilcox_posthoc_forgraph)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-e731b225c6ad06c9f790" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-e731b225c6ad06c9f790">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13"],["IL1B","IL1B","IL1B","IL1B","IL6","IL8","IL8","IL8","IL10","TNFa","TNFa","VEGF","VEGF"],["0","0","0","0","0","0","0","0","0","0","0","0","0"],["0.6","1","2","4","4","0.6","2","4","0.6","2","4","1","2"],[16,16,16,16,16,16,16,16,16,16,16,16,16],[16,16,16,16,16,16,16,16,16,16,16,16,16],[17,11,21,18,114,22,0,0,133,14,5,17,0],[0.006,0.002,0.013,0.008,0.016,0.016,3.05e-05,3.05e-05,0.000153,0.003,0.000305,0.006,3.05e-05],[0.025,0.017,0.033,0.025,0.039,0.019,6.1e-05,6.1e-05,0.002,0.008,0.001,0.013,0.000152],["*","*","*","*","*","*","****","****","**","**","**","*","***"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>variable<\/th>\n      <th>group1<\/th>\n      <th>Dose<\/th>\n      <th>n1<\/th>\n      <th>n2<\/th>\n      <th>statistic<\/th>\n      <th>p<\/th>\n      <th>p.adj<\/th>\n      <th>p.adj.signif<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"variable","targets":1},{"name":"group1","targets":2},{"name":"Dose","targets":3},{"name":"n1","targets":4},{"name":"n2","targets":5},{"name":"statistic","targets":6},{"name":"p","targets":7},{"name":"p.adj","targets":8},{"name":"p.adj.signif","targets":9}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-3659cf91fbf2c414ca22" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-3659cf91fbf2c414ca22">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13"],["IL1B","IL1B","IL1B","IL1B","IL6","IL8","IL8","IL8","IL10","TNFa","TNFa","VEGF","VEGF"],["0","0","0","0","0","0","0","0","0","0","0","0","0"],["0.6","1","2","4","4","0.6","2","4","0.6","2","4","1","2"],[16,16,16,16,16,16,16,16,16,16,16,16,16],[16,16,16,16,16,16,16,16,16,16,16,16,16],[17,11,21,18,114,22,0,0,133,14,5,17,0],[0.006,0.002,0.013,0.008,0.016,0.016,3.05e-05,3.05e-05,0.000153,0.003,0.000305,0.006,3.05e-05],[0.025,0.017,0.033,0.025,0.039,0.019,6.1e-05,6.1e-05,0.002,0.008,0.001,0.013,0.000152],["*","*","*","*","*","*","****","****","**","**","**","*","***"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>variable<\/th>\n      <th>group1<\/th>\n      <th>Dose<\/th>\n      <th>n1<\/th>\n      <th>n2<\/th>\n      <th>statistic<\/th>\n      <th>p<\/th>\n      <th>p.adj<\/th>\n      <th>p.adj.signif<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"variable","targets":1},{"name":"group1","targets":2},{"name":"Dose","targets":3},{"name":"n1","targets":4},{"name":"n2","targets":5},{"name":"statistic","targets":6},{"name":"p","targets":7},{"name":"p.adj","targets":8},{"name":"p.adj.signif","targets":9}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>The <code>Dose</code> column will be used to tell <em>ggplot2</em> where to place the annotations on the x axis, but we need to also specify where to add the annotations on the y axis. This will be different for each variable because each variable is on a different scale. We can approach this by computing the maximum value of each variable, then increasing that by 20% to add some space on top of the points.</p>
 <div class="sourceCode" id="cb448"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb448-1"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb448-1" tabindex="-1"></a>sig_labs_y <span class="ot">&lt;-</span> biomarker_data <span class="sc">%&gt;%</span></span>
 <span id="cb448-2"><a href="multi-group-and-multi-variable-comparisons-and-visualizations.html#cb448-2" tabindex="-1"></a>  <span class="fu">summarise</span>(<span class="fu">across</span>(IL1B<span class="sc">:</span>VEGF, \(x) <span class="fu">max</span>(x))) <span class="sc">%&gt;%</span></span>
diff --git a/normality-tests-and-data-transformations.html b/normality-tests-and-data-transformations.html
index f2d48f0..d9318e4 100644
--- a/normality-tests-and-data-transformations.html
+++ b/normality-tests-and-data-transformations.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
diff --git a/overview-of-experimental-design-and-example-data.html b/overview-of-experimental-design-and-example-data.html
index 0ca3708..5e2e59d 100644
--- a/overview-of-experimental-design-and-example-data.html
+++ b/overview-of-experimental-design-and-example-data.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
diff --git a/search_index.json b/search_index.json
index cf3fdfb..7de2fe9 100644
--- a/search_index.json
+++ b/search_index.json
@@ -1 +1 @@
-[["index.html", "TAME 2.0: An Update to the TAME Toolkit for Introductory Data Science, Chemical-Biological Analyses, Machine Learning and Predictive Modeling, and Database Mining for Environmental Health Research Preface", " TAME 2.0: An Update to the TAME Toolkit for Introductory Data Science, Chemical-Biological Analyses, Machine Learning and Predictive Modeling, and Database Mining for Environmental Health Research Rager Lab 2024-10-01 Preface Background Research in exposure science, toxicology, and environmental health is becoming increasingly reliant upon data science and computational methods that can more efficiently extract information from complex datasets. These methods can be used to better identify relationships between exposures to chemicals in the environment and human disease outcomes. Still, there remains a critical gap surrounding the training of researchers on these in silico methods. Objectives We aimed to address this critical gap by developing the inTelligence And Machine lEarning (TAME) Toolkit, promoting trainee-driven data generation, management, and analysis methods to “TAME” data in environmental health studies. This toolkit encompasses training modules, organized as chapters within this Github Bookdown site. TAME site users are welcome to participate in training modules by viewing them online and/or downloading all underlying script, data, and figure input files through the UNC-SRP TAME2 GitHub website. Contributors to TAME 2.0 Julia E. Rager, PhD, MSEE (UNC-Chapel Hill) directed the developed of TAME 2.0. All training resources were developed by combining modules drafted by multiple authors and experts in the field of environmental health and data science. Authors are specifically recognized at the beginning of each module section. Contributors specifically include the following, listed in alphabetical order: Oyemwenosa N. Avenbuan, University of North Carolina at Chapel Hill (UNC-Chapel Hill) Rebecca Boyles, MSPH, Research Triangle Institute (RTI) International Jessie Chappel, PhD, University of North Carolina at Chapel Hill (UNC-Chapel Hill) Chloe K. Chou, PhD Candidate, University of North Carolina at Chapel Hill (UNC-Chapel Hill) Lauren Eaves, PhD, University of North Carolina at Chapel Hill (UNC-Chapel Hill) Rebecca C. Fry, PhD, University of North Carolina at Chapel Hill (UNC-Chapel Hill) Elise Hickman, PhD, University of North Carolina at Chapel Hill (UNC-Chapel Hill) Ilona Jaspers, PhD, University of North Carolina at Chapel Hill (UNC-Chapel Hill) Lauren E. Koval, PhD Candidate, University of North Carolina at Chapel Hill (UNC-Chapel Hill) Paul Kruse, PhD, U.S. Environmental Protection Agency (U.S. EPA) Sarah Miller, PhD Candidate, University of North Carolina at Chapel Hill (UNC-Chapel Hill) David M. Reif, PhD, Division of Translational Toxicology (DTT), National Institutes of Environmental Health Science (NIEHS) Cynthia Rider, PhD, Division of Translational Toxicology (DTT), National Institutes of Environmental Health Science (NIEHS) Kyle Roell, PhD, University of North Carolina at Chapel Hill (UNC-Chapel Hill) Alexis Payton, MS, University of North Carolina at Chapel Hill (UNC-Chapel Hill) Grace Patlewicz, PhD, U.S. Environmental Protection Agency (U.S. EPA) Caroline Ring, PhD, U.S. Environmental Protection Agency (U.S. EPA) Cavin Ward-Caviness, PhD, U.S. Environmental Protection Agency (U.S. EPA) Module Development Overview Training modules were developed to provide applications-driven examples of data organization and analysis methods that can be used to address environmental health questions. Target audiences for these modules include students and professionals in academia, government, and industry that are interested in expanding their skillset. Modules were developed by study coauthors using annotated script formatted for R/RStudio coding language and interface and are now organized into seven chapters. The first group of modules focuses on introductory data science, which included the following topics: data management and sharing practices, file management, and basic data wrangling. The second chapter of modules provides an overview of scripting, using R as the example language. Topics span downloading and programming in R, coding best practices, data manipulation and reshaping, and coding efficiencies. The third chapter demonstrates methods for basic data analysis and visualizations, spanning an introduction to data visualizations, methods to improve data visualizations, normality tests and data transformations, and introduction to statistical tests. The fourth chapter focuses on the translation of wet lab data into dry lab analyses. Topics in this chapter include an introduction to key experimental features and terms; data importing, cleaning and summary statistics from wet lab sources; two-group comparisons and visualizations; multigroup comparisons and visualizations; advanced multigroup comparisons; and advanced data export and sharing. The fifth chapter provides an introduction to machine learning and artifical intelligence (ML/AI). Topics in this chapter include a general introduction of ML/AI and then provides examples of unsupervised machine learning methods, supervised machine learning methods, and interpretation of developed ML models. The sixth chapter provides applications in toxicology and exposure science, spanning descriptive cohort analyses; -omics analyses; mixtures methodologies; toxicokinetic modeling; and chemical read-across.The seventh chapter delves into examples of environmental health database mining, and highlights the helpful databases of Comparative Toxicogenomics Database, Gene Expression Omnibus, Computational Toxicology Dashboard access through APIs, and database integration using geospatial methods. The overall organization of this TAME toolkit is summarized below. Modules are organized into seven chapters that are listed on the left side of this website. Ways to use TAME 2.0 and Underlying Site License Site viewers are encouraged to follow along each module, either starting at the very beginning and reading through the material like a classical book or clicking on specific modules of interest and learning topics of greatest interest to their data science development. Each module is designed to stand-alone. Site license This site is covered by the BY-NC-ND Creative Commons license. This license allows reusers to copy and distribute the material in any medium or format in unadapted form only, for noncommercial purposes only, and only so long as attribution is given to the creator. Figure credit Figures included in TAME 2.0 were either generated in R, created within Biorender.com, and/or pulled from online publications with site licensing and author attribution throughout. Concluding Remarks Together, this valuable resource provides unique opportunities to obtain introductory-level training on current data analysis methods applicable to 21st century exposure science, toxicology, and environmental health. These modules serve as applications-based examples on how to “TAME” data within the environmental health research field, expanding the toolbox for career development and cross-training of scientists in multiple specialties, as well as supporting the next generation of data scientists. Funding This study was supported by the National Institutes of Health (NIH) from the National Institute of Environmental Health Sciences, including the following grant funds and associated programs: P42ES031007: The University of North Carolina (UNC)-Superfund Research Program (SRP) seeks to develop new solutions for reducing exposure to inorganic arsenic and prevent arsenic-induced diabetes through mechanistic and translational research. The UNC-SRP Data Analysis and Management Core (UNC-SRP-DMAC) provides the UNC-SRP with critical expertise in bioinformatics, statistics, data management, and data integration. T32ES007126: The UNC Curriculum in Toxicology and Environmental Medicine (CiTEM) seeks to provide a cutting edge research and mentoring environment to train students and postdoctoral fellows in environmental health and toxicology. Towards this goal, the CiTEM has a T32 Training Program for Pre- and Postdoctoral Training in Toxicology to support the development of future investigators in environmental health and toxicology. Support was additionally provided through the Institute for Environmental Health Solutions (IEHS) at the University of North Carolina (UNC) Gillings School of Global Public Health. The IEHS is aimed at protecting those who are particularly vulnerable to diseases caused by environmental factors, putting solutions directly into the hands of individuals and communities of North Carolina and beyond. "],["fair-data-management-practices.html", "FAIR Data Management Practices Introduction to Training Module Introduction to FAIR Breaking Down FAIR, Letter-by-Letter What Does This Mean for You? Data Repositories for Sharing of Data Recent Shifts in Regulatory Policies for Data Sharing Additional Training Resources on FAIR", " FAIR Data Management Practices This training module was developed by Rebecca Boyles, with contributions from Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module This training module provides a description of FAIR data management practices, and points participants to important resources to help ensure generated data meet current FAIR guidelines. This training module is descriptive content-based (as opposed to coding-based), in order to present information clearly and serve as an important resource alongside the other scripted training activities. Training Module’s Environmental Heatlh Questions This training module was specifically developed to answer the following questions: What is FAIR? When was FAIR first developed? When making data ‘Findable’, who and what should be able to find your data? When saving/formatting your data, which of the following formats is preferred to meet FAIR principles: .pdf, .csv, or a proprietary output file from your lab instrument? How can I find a suitable data repository for my data? Introduction to FAIR Proper data management is of utmost importance while leading data analyses within the field of environmental health science. A method to ensure proper data management is the implementation of Findability, Accessibility, Interoperability, and Reusability (FAIR) practices. A landmark paper that describes FAIR practices in environmental health research is the following: Wilkinson MD, Dumontier M, Aalbersberg IJ, et al. The FAIR Guiding Principles for scientific data management and stewardship. Sci Data. 2016 Mar 15. PMID: 26978244. The FAIR principles describe a framework for data management and stewardship aimed at increasing the value of data by enabling sharing and reuse. These principles were originally developed from discussions during the Jointly Designing a Data FAIRport meeting at the Lorentz Center in Leiden, The Netherlands in 2014, which brought together stakeholders to discuss the creation of an environment for virtual computational science. The resulting principles are technology agnostic, discipline independent, community driven, and internationally adopted. Below is a schematic providing an overview of this guiding principle: Answer to Environmental Health Question 1 &amp; 2 With this background, we can answer Environmental Health Question #1 and #2: What is FAIR and when was it first developed? Answer: FAIR is guiding framework that was recently established to promote best data management practices, to ensure that data are Findable, Accessibility, Interoperable, and Reusable. It was first developed in 2014- which means that these principles are very new and continuing to evolve! Breaking Down FAIR, Letter-by-Letter The aspects of the FAIR principles apply to data and metadata with the aim of making the information available to people and computers as described in the seminal paper by Wilkinson et al., 2016. F (Findable) in FAIR The F in FAIR identifies components of the principles needed to make the meta(data) findable through the application of unique persistent identifiers, thoroughly described, reference the unique identifiers, and that the descriptive information (i.e., metadata) could be searched by both humans and computer systems. F1. (Meta)data are assigned a globally unique and persistent identifier Each dataset is assigned a globally unique and persistent identifier (PID), for example a DOI. These identifiers allow to find, cite and track (meta)data. A DOI looks like: https://doi.org/10.1109/5.771073 Action: Ensure that each dataset is assigned a globally unique and persistent identifier. Certain repositories automatically assign identifiers to datasets as a service. If not, obtain a PID via a PID registration service. F2. Data are described with rich metadata Each dataset is thoroughly (see R1) described: these metadata document how the data was generated, under what term (license) and how it can be (re)used and provide the necessary context for proper interpretation. This information needs to be machine-readable. Action: Fully document each dataset in the metadata, which may include descriptive information about the context, quality and condition, or characteristics of the data. Another researcher in any field, or their computer, should be able to properly understand the nature of your dataset. Be as generous as possible with your metadata (see R1). F3. Metadata clearly and explicitly include the identifier of the data it describes Explanation: The metadata and the dataset they describe are separate files. The association between a metadata file and the dataset is obvious thanks to the mention of the dataset’s PID in the metadata. Action: Make sure that the metadata contains the dataset’s PID. F4. (Meta)data are registered or indexed in a searchable resource Explanation: Metadata are used to build easily searchable indexes of datasets. These resources will allow to search for existing datasets similarly to searching for a book in a library. Action: Provide detailed and complete metadata for each dataset (see F2). Answer to Environmental Health Question 3 With this, we can answer Environmental Health Question #3: When making data ‘Findable’, who and what should be able to find your data? Answer: Both humans and computer systems should be able to find your data. A (Accessible) in FAIR The A components are designed to enable meta(data) be available long-term, accessed by humans and machines using standard communication protocols with clearly described limitations on reuse. A1. (Meta)data are retrievable by their identifier using a standardized communications protocol Explanation: If one knows a dataset’s identifier and the location where it is archived, one can access at least the metadata. Furthermore, the user knows how to proceed to get access to the data. Action: Clearly define who can access the actual data and specify how. It is possible that data will not be downloaded, but rather reused in situ. If so, the metadata must specify the conditions under which this is allowed (sometimes versus the conditions needed to fulfill for external usage/“download”). A1.1 The protocol is open, free, and universally implementable Explanation: Anyone with a computer and an internet connection can access at least the metadata. A1.2 The protocol allows for an authentication and authorization procedure, where necessary Explanation: It often makes sense to request users to create a user account on a repository. This allows to authenticate the owner (or contributor) of each dataset, and to potentially set user specific rights. A2. Metadata are accessible, even when the data are no longer available Explanation: Maintaining all datasets in a readily usable state eternally would require an enormous amount of curation work (adapting to new standards for formats, converting to different format if specifically needed software is discontinued, etc). Keeping the metadata describing each dataset accessible, however, can be done with fewer resources. This allows to build comprehensive data indexes including all current, past, and potentially arising datasets. Action: Provide detailed and complete metadata for each dataset (see R1). I (Interoperable) in FAIR The I components of the principles address needs for data exchange and interpretation by humans and machines which includes the use of controlled vocabularies or ontologies to describe meta(data) and to describe provenance relationships through appropriate data citation. I1. (Meta)data use a formal, accessible, shared, and broadly applicable language Explanation: Interoperability typically means that each computer system has at least knowledge of the other system’s formats in which data is exchanged. If (meta)data are to be searchable and if compatible data sources should be combinable in a (semi)automatic way, computer systems need to be able to decide if the content of datasets are comparable. Action: Provide machine readable data and metadata in an accessible language, using a well-established formalism. Data and metadata are annotated with resolvable vocabularies/ontologies/thesauri that are commonly used in the field (see I2). I2. (Meta)data use vocabularies that follow FAIR principles Explanation: The controlled vocabulary (e.g., MESH) used to describe datasets needs to be documented. This documentation needs to be easily findable and accessible by anyone who uses the dataset. Action: The vocabularies/ontologies/thesauri are themselves findable, accessible, interoperable and thoroughly documented, hence FAIR. Lists of these standards can be found at: NCBO BioPortal, FAIRSharing, OBO Foundry. I3. (Meta)data include qualified references to other (meta)data Explanation: If the dataset builds on another dataset, if additional datasets are needed to complete the data, or if complementary information is stored in a different dataset, this needs to be specified. In particular, the scientific link between the datasets needs to be described. Furthermore, all datasets need to be properly cited (i.e. including their persistent identifiers). Action: Properly cite relevant/associated datasets, by providing their persistent identifiers, in the metadata, and describe the scientific link/relation to your dataset. R (Reusable) in FAIR The R components highlight needs for the meta(data) to be reused and support integration such as sufficient description of the data and data use limitations. R1. Meta(data) are richly described with a plurality of accurate and relevant attributes Explanation: Description of a dataset is required at two different levels: Metadata describing the dataset: what does the dataset contain, how was the data generated, how has it been processed, how can it be reused. Metadata describing the data: any needed information to properly use the data, such as definitions of the variable names Action: Provide complete metadata for each data file. Scope of your data: for what purpose was it generated/collected? Particularities or limitations about the data that other users should be aware of. Date of the dataset generation, lab conditions, who prepared the data, parameter settings, name and version of the software used. Variable names are explained or self-explanatory. Version of the archived and/or reused data is clearly specified and documented. What Does This Mean for You? We advise the following as ‘starting-points’ for participants to start meeting FAIR guidances: Learn how to create a Data Management Plan Keep good documentation (project &amp; data-level) while working Do not use proprietary file formats (.csv is a great go-to formats for your data!) When able, use a domain appropriate metadata standard or ontology Ruthlessly document any steps in a project Most of FAIR can be handled by selecting a good data or software repository Don’t forget to include a license! Answer to Environmental Health Question 4 With these, we can answer Environmental Health Question #4: When saving/formatting your data, which of the following formats is preferred to meet FAIR principles: .pdf, .csv, or a proprietary output file from your lab instrument? Answer: A .csv file is preferred to enhance data sharing. Data Repositories for Sharing of Data When you are organizing your data to deposit online, it is important to identify an appropriate repository to publish your dataset it. A good starting place is a repository registry such as FAIRsharing.org or re3data.org. Journals can also provide helpful resources and starting repository lists, such as Nature and PLOS, which both have published a list of recommended repositories. Funding agencies, including the NIH, can also inform specific repositories. Below are some examples of two main categories of data repositories: 1. Domain Agnostic Data Repositories Domain agnostic repositories allow the deposition of any data type. Some examples include the following: Data in Brief Articles (e.g., Elsevier’s Data in Brief Journal) Dryad Figshare The Dataverse Project Zenodo 2. Domain Specific Data Repositories Domain specific repositories allow the deposition of specific types of data, produced from specific types of technologies or within specific domains. Some examples include the following: Database of Genotypes and Phenotypes Gene Expression Omnibus The Immunology Database and Analysis Portal Metabolomics Workbench (National Metabolomics Data Repository) Microphysiology Systems Database Mouse Genome Informatics Mouse Phenome Database OpenNeuro Protein Data Bank ProteomeXchange Rat Genome Database The Database of Genotypes and Phenotypes Zebrafish Model Organism Database and many, many, many others… Answer to Environmental Health Question 5 With these, we can answer Environmental Health Question #5: How can I find a suitable data repository for my data? Answer: I can search through a data repository registry service or look for recommendations from NIH or other funding agencies. Recent Shifts in Regulatory Policies for Data Sharing The NIH Data Management and Sharing Policy NIH’s data management and sharing (DMS) policy became effective January 2023. This policy specifically lists the expectations that investigators must comply with in order to promote the sharing of scientific data. Information about this recent policy can be found through updated NIH websites. Information about writing an official Data Management and Sharing (DMS) plan for your research can be found through NIH’s Guidance on Writing a Data Management &amp; Sharing Plan. The 2018 Evidence Act The Evidence Act, or Foundations for Evidence-Based Policymaking Act of 2018, was signed into U.S. law on January 14, 2019. The Act requires federal agencies to build the capacity to use evidence and data in their decision-making and policymaking. It also requires agencies to: Develop an evidence-building plan as part of their quadrennial strategic plan &amp; Develop an evaluation plan concurrent with their annual performance plan. The Evidence Act also: Mandates that data be “open by default” Specifies that a comprehensive data inventory should be created for each agency’s open data assets How Does the NIH Data Management and Sharing Policy Intersect with the 2018 Evidence Act? Making your data FAIR, by definition, makes it more shareable and reusable. Many of the requirements in the NIH DMS and the Evidence Act policy overlap with the FAIR principles. The CARE Principles for Indigenous Data Governance While we are experiencing increased requirements for the open sharing of data, it is important to recognize that there are circumstances and populations that should, at the same time, be carefully protected. Examples include human clinical or epidemiological data that may become identifiable upon the sharing of sensitive data. Another example includes the consideration of Indigenous populations. A recent article by Carroll et al. 2021 describes in their abstract: As big data, open data, and open science advance to increase access to complex and large datasets for innovation, discovery, and decision-making, Indigenous Peoples’ rights to control and access their data within these data environments remain limited. Operationalizing the FAIR Principles for scientific data with the CARE Principles for Indigenous Data Governance enhances machine actionability and brings people and purpose to the fore to resolve Indigenous Peoples’ rights to and interests in their data across the data lifecycle. Additional Training Resources on FAIR Many organizations, from specific programs to broad organizations, provide training and resources for scientists in FAIR principles. Some of the notable global organizations organizing and providing training that offer opportunities for community involvement are: Committee on Data for Science and Technology (CODATA) Global Alliance for Genomics &amp; Health GoFAIR Force11 Research Data Alliance Example Workshops discussing FAIR: NAS Implementing FAIR Data for People and Machines: Impacts and Implications (2019). Available at: https://www.nationalacademies.org/our-work/implementing-fair-data-for-people-and-machines-impacts-and-implications NIH Catalyzing Knowledge-driven Discovery in Environmental Health Sciences Through a Harmonized Language, Virtual Workshop (2021). Available at: https://www.niehs.nih.gov/news/events/pastmtg/2021/ehslanguage/index.cfm NIH Trustworthy Data Repositories Workshop (2019). Available at: https://datascience.nih.gov/data-ecosystem/trustworthy-data-repositories-workshop NIH Virtual Workshop on Data Metrics (2020). Available at: https://datascience.nih.gov/data-ecosystem/nih-virtual-workshop-on-data-metrics NIH Workshop on the Role of Generalist Repositories to Enhance Data Discoverability and Reuse: Workshop Summary (2020). Available at: https://datascience.nih.gov/data-ecosystem/nih-data-repository-workshop-summary Example Government Report Documents on FAIR: Collins S, Genova F, Harrower N, Hodson S, Jones S, Laaksonen L, Mietchen D, Petrauskaite R, Wittenburg P. Turning FAIR into reality: Final report and action plan from the European Commission expert group on FAIR data: European Union; 2018. Available at: https://www.vdu.lt/cris/handle/20.500.12259/103794. EU. FAIR Data Advanced Use Cases: From Principles to Practice in the Netherlands. 2018. European Union. Available at: doi:10.5281/zenodo.1250535. NIH. Final NIH Policy for Data Management and Sharing and Supplemental Information. National Institutes of Health. Federal Register, vol. 85, 2020-23674, 30 Oct. 2020, pp. 68890–900. Available at: https://www.federalregister.gov/d/2020-23674. NIH. NIH Strategic Plan for Data Science 2018. National Institutes of Health. Available at: https://datascience.nih.gov/strategicplan. NLM. NLM Strategic Plan 2017 to 2027. U.S. National Library of Medicine, Feb. 2018. Available at: https://www.nlm.nih.gov/about/strategic-plan.html. Example Related Publications on FAIR: Comess S, Akbay A, Vasiliou M, Hines RN, Joppa L, Vasiliou V, Kleinstreuer N. Bringing Big Data to Bear in Environmental Public Health: Challenges and Recommendations. Front Artif Intell. 2020 May;3:31. doi: 10.3389/frai.2020.00031. Epub 2020 May 15. PMID: 33184612; PMCID: PMC7654840. Koers H, Bangert D, Hermans E, van Horik R, de Jong M, Mokrane M. Recommendations for Services in a FAIR Data Ecosystem. Patterns (N Y). 2020 Jul 7;1(5):100058. doi: 10.1016/j.patter.2020.100058. Erratum in: Patterns (N Y). 2020 Sep 11;1(6):100104. PMID: 33205119. Kush RD, Warzel D, Kush MA, Sherman A, Navarro EA, Fitzmartin R, Pétavy F, Galvez J, Becnel LB, Zhou FL, Harmon N, Jauregui B, Jackson T, Hudson L. FAIR data sharing: The roles of common data elements and harmonization. J Biomed Inform. 2020 Jul;107:103421. doi: 10.1016/j.jbi.2020.103421. Epub 2020 May 12. PMID: 32407878. Lin D, Crabtree J, Dillo I, Downs RR, Edmunds R, Giaretta D, De Giusti M, L’Hours H, Hugo W, Jenkyns R, Khodiyar V, Martone ME, Mokrane M, Navale V, Petters J, Sierman B, Sokolova DV, Stockhause M, Westbrook J. The TRUST Principles for digital repositories. Sci Data. 2020 May 14;7(1):144. PMID: 32409645. Thessen AE, Grondin CJ, Kulkarni RD, Brander S, Truong L, Vasilevsky NA, Callahan TJ, Chan LE, Westra B, Willis M, Rothenberg SE, Jarabek AM, Burgoon L, Korrick SA, Haendel MA. Community Approaches for Integrating Environmental Exposures into Human Models of Disease. Environ Health Perspect. 2020 Dec;128(12):125002. PMID: 33369481. Roundtable on Environmental Health Sciences, Research, and Medicine; Board on Population Health and Public Health Practice; Health and Medicine Division; National Academies of Sciences, Engineering, and Medicine. Principles and Obstacles for Sharing Data from Environmental Health Research: Workshop Summary. Washington (DC): National Academies Press (US); 2016 Apr 29. PMID: 27227195. Test Your Knowledge Let’s imagine that you’re a researcher who is planning on gathering a lot of data using the zebrafish model. In order to adequately prepare your studies and steps to ensure data are deposited into proper repositories, you have the idea to check repository information obtained in FAIRsharing.org. What are some example repositories and relevant ontology resources that you could use to organize, deposit, and share your zebrafish data (hint: use the search tool)? "],["data-sharing-through-online-repositories.html", "Data Sharing through Online Repositories An Overview and Example with the Dataverse Repository Introduction to Training Module Data Repositories The Dataverse Project What is a Dataverse? Metadata Creating a Dataverse Creating a Dataset Concluding Remarks", " Data Sharing through Online Repositories An Overview and Example with the Dataverse Repository This training module was developed by Kyle R. Roell, Alexis Payton, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module Submitting data to publicly available repositories is an essential part of ensuring data meet FAIR guidelines, as discussed in detail in the previous training module. There are many benefits to sharing and submitting your researching, such as: Making more use out of data that are generated in your lab More easily sharing and integrating across datasets Ensuring reproducibility in analysis findings and conclusions Improving the tracking and archiving of data sources, and data updates Increasing the awareness and attention surrounding your research as others locate your data through additional online queries Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: How should I structure my data for upload into online repositories? What does the term ‘metadata’ mean and what does it look like? This module will introduce some of the repositories that are commonly used to deposit data, how to set up metadata files, and how to organize example data in preparation for sharing. We will also provide information surrounding best practices for data organization and sharing through these repositories. Additional resources are also provided throughout, as there are many ways to organize, share, and deposit data depending on your data types and structures and overall research goals. Data Repositories There are many publicly available repositories that we should consider when depositing data. Some general repository registries that are helpful to search through include FAIRsharing.org or re3data.org. Journals can also provide helpful resources and starting repository lists, such as Nature and PLOS, which both have published a list of recommended repositories. As detailed in the FAIR training module, there are two main categories of data repositories: 1. Domain Agnostic Data Repositories Domain agnostic repositories allow the deposition of any data type. Some examples include: Data in Brief Articles (e.g., Elsevier’s Data in Brief Journal) Dryad Figshare The Dataverse Project Zenodo 2. Domain Specific Data Repositories Domain specific repositories allow the deposition of specific types of data, produced from specific types of technologies or within specific domains. Some examples include: Database of Genotypes and Phenotypes Gene Expression Omnibus The Immunology Database and Analysis Portal Metabolomics Workbench (National Metabolomics Data Repository) Microphysiology Systems Database Mouse Genome Informatics Mouse Phenome Database OpenNeuro Protein Data Bank ProteomeXchange Rat Genome Database The Database of Genotypes and Phenotypes Zebrafish Model Organism Database and many, many, many others… This training module focuses on providing an example of how to organize and upload data into the Dataverse; though many of the methods described below pertain to other data repositories as well, and also incorporate general data organization and sharing best practices. The Dataverse Project Dataverse, organized through The Dataverse Project, is a popular repository option that allows for upload of most types of material, without any stringent requirements. The Dataverse organization also provides ample resources on how to organize, upload, and share data through Dataverse. These resources include very thorough, readable, and user guides and best practices. Screenshot of the main page of The Dataverse Project An easier way to think about Dataverse is to interpret it similar to a folder system on your computer. A Dataverse is just an online folder that contains files, data, or datasets that are all related to some topic, project, etc. Although Dataverse was started at Harvard and the base Dataverse lives there, there are many versions of Dataverse that are specific to and supported by various institutions. For example, these training modules are being developed primarily by faculty, staff, and students at the University of North Carolina at Chapel Hill. As such, the examples contained in this module will specifically connect with the UNC Dataverse; though many of the methods outlined here are applicable to other Dataverses and additional online repositories, in general. What is a Dataverse? Remember how we pointed out that a Dataverse is similar to a folder system on a computer? Well, here we are going to show you what that actually looks like. But first, something that can be confusing when starting to work with Dataverse is the fact that the term Dataverse is used for both the overarching repository as well as individual subsections (or folders) in which data are stored. For example, the UNC Dataverse is called a Dataverse, but to upload data, you need to upload it to a specific sub-Dataverse. So, what is the difference between the high level UNC Dataverse and smaller, sub-dataverses? Well, nothing, really. The UNC Dataverse is similar to a large folder that says, these are all the projects and research related to or contained within UNC. From there, we want to be more specific about where we store our research, so we are creating more sub-Dataverses (folders) within that higher, overarching UNC Dataverse. As an example, using the UNC Dataverse, here we can see various sub-Dataverses that have been created as repositories for specific projects or types of data. As another example looking within a specific Dataverse, here we can see the Dataverse that hosts datasets and publications for Dr. Julia Rager’s lab, the Ragerlab-Dataverse. Within this Datavere, we can see various datasets produced by her lab. It is worth noting that the datasets may not necessarily be directly related to each other in terms of exact topic, for example, the Ragerlab-Dataverse hosts data pertaining to wildfire smoke exposure as well as chemical exposures and breast cancer. But they are all pertaining to experiments and analyses run within her specific lab. Let’s now start talking more specifically about how to organize data and format files for Dataverse, create your own “Dataverse”, upload datasets, and what this all means! Dataset Structure Before uploading your data to any data repository, it is important to structure your data efficiently and effectively, making it easy for others to navigate, understand, and utilize. While we will cover this in various sections throughout these training modules, here are some basic tips for data structure and organization. Keep all data for one participant or subject within one column (or row) of your dataset Genomic data and other analytical assays tend to have subjects on columns and genes, expression, etc. as the rows Descriptive and demographic data often tend to have subjects or participants as the rows and each descriptor variable (including demographics and any other subject variables) as columns Create succinct, descriptive variable names For example, do not use something like “This Variable Contains Information Regarding Smoking Status”, and instead just using something like, “Smoking_Status” Be aware of using spacing, special characters, and capitalization within variable names Think about transforming data from wide to long format depending on your specific dataset and general conventions Be sure to follow specific guidelines of repository when appropriate TAME 2.0 Module 1.1 FAIR Data Management Practices and TAME 2.0 Module 1.4 Data Wrangling in Excel are also helpful resources to reference when thinking about organizing your data. A general example of an organized, long format dataset in Excel in provided below: Only .csv or .txt files can be uploaded to dataverse; therefore, the metadata and data tabs in an excel file will need to saved and uploaded as two separate .csv or .txt files. Answer to Environmental Health Question 1 With this, we can answer Environmental Health Question 1: How should I structure my data for upload into online repositories? Answer: It is ideal to have data clearly organized and filled, with succinct and descriptive variable names clearly labeled and values filled in. Most commonly, datasets should be saved as separate .csv or .txt files for upload into data repositories. Metadata There are many different definitions of what a metadata file is. Helpful explanations, for example, are provided by the UNC University Libraries: There are many definitions of metadata, but one of the simplest is data about data. More specifically… Metadata (in terms of data management) describe a dataset: how they were collected; when they were collected; what assumptions were made in their methodology; their geographic scope; if there are multiple files, how they relate to one another; the definitions of individual variables and, if applicable, what possible answers were (i.e., to survey questions); the calibration of any equipment used in data collection; the version of software used for analysis; etc. Very often, a dataset that has no metadata is incomprehensible. Metadata ARE data. They are pieces of information that have some meaning in relation to another piece of information. They can be created, managed, stored, and preserved like any other data. Metadata can be applied to anything. A computer file can be described in the same way that a book or piece of art can be described. For example, both can have a title, an author, and a year created. Metadata should be documented for research outputs of any kind. Metadata generally has little value on their own. Metadata adds value to other information, but are usually not valuable in themselves. There are exceptions to this rule, such as text transcription of an audio file. There are three kinds of metadata: Descriptive metadata consist of information about the content and context of your data. Examples: title, creator, subject keywords, and description (abstract) Structural metadata describe the physical structure of compound data. Examples: camera used, aperture, exposure, file format, and relation to other data or files Administrative metadata are information used to manage your data. Examples: when and how they were created, who can access them, software required to use them, and copyright permissions Therefore, after having organized your primary dataset for submission into online repositories, it is equally important to have a metadata file for easy comprehension and utilization of your data for future researchers or anyone downloading your data. While most repositories capture some metadata on the dataset page (e.g., descripton of data, upload date, contact information), there is generally little information about the specific data values and variables. In this section, we review some general guidelines and tips to better annotate your data. First, keep in mind, depending on the specific repository you are using, you may have to follow their metadata standards. But, if uploading to more generalist repository, this may be up to you to define. Generally, a metadata file consists of a set of descriptors for each variable in the data. If you are uploading data that contains many covariates or descriptive variables, it is essential that you provide a metadata file that describes these covariates. Both a description of the variable as well as any specific levels of any categorical or factor type variables. From the dataset presented previously, here we present an example of an associated metadata file: Answer to Environmental Health Question 2 With this, we can answer Environmental Health Question 2: What does the term ‘metadata’ mean and what does it look like? Answer: Metadata refers to the information that describes and explains data. It looks like an additional dataset that provides context with details such as the source, type, owner, and relationships to other datasets. This file can help users understand the relevance of a specific dataset and provide guidance on how to use it. Creating a Dataverse Now, let’s review how to actually create a Dataverse. First, navigate to the parent Dataverse that you would like to use as your primary host website. For example, our group uses the UNC Dataverse. If you do not already have one, create a username and login. Then, from the home Dataverse page, click “Add Data” and select “New Dataverse”. And fill in the information necessary. And that is it. After creating your Dataverse site, you will need to publish it; however, before it is accessible to the public, note that you can actually create a Dataverse within another Dataverse (similar to a folder within a folder on your computer). This makes sense even when you are creating a new Dataverse at the home, UNC Dataverse level, you are still technically creating a new Dataverse within an existing one (the large UNC Dataverse). Here are some tips as you create your Dataverse: Do not recreate a Dataverse that already exists Choose a name that is specific, but general enough that it doesn’t only pertain to one specific dataset You can add more than one contact email, if necessary Creating a Dataset Creating a dataset creates a page for your data containing information about that data, a citation for the data (something valuable and somewhat unique to Dataverse), as well the place from where you data can be directly accessed or downloaded. First, decide the specific Dataverse your data will live and navigate to that specific Dataverse site. Then carry out the following steps to create a dataset: Navigate to the Dataverse page under which your dataset will live Click “Add Data” and then select “New Dataset” Fill in the necessary information Upload your data and metadata file(s) structured as detailed above Now, you have a dataset within your Dataverse. Again, you will have to publish the dataset for someone to have access to it. The easy part of using a more generalist repository like Dataverse, is that you do not have to have a strict data structure adherence. However, this means it is up to you to make sure your data is readable and useable. Concluding Remarks In this training module, we set out to express the importance of uploading data to online repositories, demonstrate what the upload process may look like using a generalist repository (Dataverse), and give some examples and tips on structuring data for upload and creating metadata files. It is important to choose the appropriate repository for your data based on your field of study and specifications of your work. Test Your Knowledge Try creating your own Dataverse repository, format your files to be uploaded to Dataverse, and upload those files to your new repository! "],["file-management-using-github.html", "File Management using Github Introduction to Training Module Creating an Account Creating a Repository Uploading Code Adding Subfolders in a Repository Updating Code Updating Repository Titles and Structure to Support a Manuscript Tracking Code Changes using Github Branches Concluding Remarks", " File Management using Github This training module was developed by Alexis Payton, Lauren E. Koval, Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module Good data practices like file management and code tracking are imperative for data analysis initiatives, especially when working in research teams and/or shared project folders. Often times analyses and manuscripts are edited many times prior to being submitted for a grant or publication. Analysis methods are also shared between members of a research team and to external communities, as further detailed in TAME 2.0 Module 1.1 FAIR Data Management Practices. Therefore, Github has emerged as an effective way to manage, share, and track how code changes over time. Github is an open source or publicly accessible platform designed to facilitate version control and issue tracking of code. It is used by us and many of our colleagues to not only document versions of script written for data analysis and visualization, but to also make our code publicly available for open communication and dissemination of results. This training module serves a launch pad for getting acclimated with Github and includes… Creating an account Uploading code Creating a repository and making it legible for manuscript submission Creating an Account First, users must create their own accounts within github to start uploading/sharing code. To do this, navigate to github.com, click “Sign Up”, and follow the on screen instructions. Creating a Repository A repository, also known as a “repo”, is similar to a project folder that will contain all code pertaining to a specific project (which can be used for specific research programs, grants, or manuscripts, as examples). A repository can be set to public or private. If a repo is initially set to private to keep findings confidential prior to publication, it can always be updated to public once findings are ready for public dissemination. Multiple people can be allowed to work on a project together within a single repository. To access the repositories that are currently available to you through your user account, click the circle in top right-hand corner and click “Your repositories”. To create a new repository, click on the green button that says “New”. Then give your repository a descriptive name. We often edit the repo titles to match the title of specific manuscripts, though specific titling formats are up to the users/team’s preference. For more information, visit Github’s Create a repo documentation. Then click “Add a README file” to initiate the README file, which is important to continually edit to provide analysis-specific background information, and any additional information that would be helpful during and after code is drafted to better facilitate tracking information and project details. We provide further details surrounding specific information that can be included within the README file below. Uploading Code The simplest way to upload code is to first navigate to the repository that you would like to upload your code/associated files to. Note that this could represent a repo that you created or that someone granted you access to. Click “Add file” then click “Upload files”. Drag and drop your file containing your script into github and click “Commit changes”. A more advanced way to upload code is by using the command line, which allows a user to directly interact with the computer or software application. Further documentation can be found here. Adding Subfolders in a Repository To keep the repository organized, it might be necessary to create a new folder (like the folder labeled “1.1. Summary Statistics” in the above screenshot). Files can be grouped into these folders based on the type of analysis. To do so, click on the new file and then click on the pencil icon next to the “Blame” button. Click on the box that contains the title of the file. Write the title of your new folder and then end with a forward slash (/). In the screenshot below, we’re creating a new folder entitled “New Folder”. Click “Commit changes” and your file should now be in a new folder. Updating Code Saving iterations of code can save valuable time later as analyses are constantly being updated and edited. If your code undergoes substantial changes, (e.g., adding/ removing steps or if there’s code that is likely to be beneficial later on, but is no longer relevant to the current analysis), it is helpful to save that version in Github for future reference. To do so, create a subfolder named “Archive” and move the old file into it. If you have multiple versions of a file with the same name, add the current date to prevent the file from being overwritten later on as seen in the screenshot below. Once the old file version has been archived, now upload the most recent version of your code to the main folder. Based on the screenshot above, that would be under “3. ML Visualizations”. Note: If a file is uploaded with the same name it will be overwritten, which can’t be undone! Therefore, put the older file into the archive folder if you’d like it to be saved PRIOR to uploading the new version. Updating Repository Titles and Structure to Support a Manuscript If the code is for a manuscript, it’s helpful to include the table or figure name it pertains to in the manuscript in parentheses. For example, “Baseline Clusters (Figure 3)”. This allows viewers to find find the code for each table or figure faster. Using a README.md file A README.md file is used to describe the overall aims and purpose of the analyses in the repository or a folder within a repository. It is often the first file that someone will look at in a repo/folder, so it is important to include information that would be valuable to an outsider trying to make use of the work. To add a README.md file, click “Add file” and then “Create new file”. Name your file “README.md”. A README.md file uses R markdown syntax. This type of syntax is very helpful as you continue to develop R coding skills, as it provides a mechanism through which your code’s output can be visualized and saved as a rendered file version. There are many helpful resources for R markdown, including some that we find helpful: R Markdown Cheatsheet R Markdown Syntax Overview The final README.md file for the OVERALL repository for manuscript submission should look something like the screenshot below. Always include… The main goal of the project The final manuscript name, year it was published, Pub Med ID (if applicable) Graphical abstract (if needed for publication) Names and brief descriptions of each file Include both the goal of the analysis and the methodology used (ie. Using chi square tests to determine if there are statistically significant differences across demographic groups) If the code was written in the software Jupyter (ie. has the extension .ipynb not .R or .Rmd), NBViewer is a website that can render jupyter notebooks (files). This is helpful, because sometimes the files take too long to render, so link the repository from the NB viewer website. Go to nbviewer.org –&gt; type in the name of the repository –&gt; copy the url and add it to the README.md file The final README.md file for the a subfolder within a repository should look something like the screenshot below. Always include… The name of each file Brief description of each file Include both the goal of the analysis and the methodology used Table or Figure name in the corresponding manuscript (if applicable) Note: That the organization structure for the README.md files are simply recommendations and should be changed based on needs of the project. However, it is important to include information and organize the repository in a way that helps other readers and colleagues navigate it who aren’t familiar with the project. Example Repositories Below are links to repositories that contain code for analyses used in published manuscripts. These are examples of well organized Github repositories. Wildfires and Environmental Justice: Future Wildfire Events Predicted to Disproportionally Impact Socioeconomically Vulnerable Communities in North Carolina Plasma sterols and vitamin D are correlates and predictors of ozone-induced inflammation in the lung: A pilot study Cytokine signature clusters as a tool to compare changes associated with tobacco product use in upper and lower airway samples Tracking Code Changes using Github Branches Github is a useful platform for managing and facilitating code tracking performed by different collaborators through branches. When creating a repository on Github, it automatically creates a default branch entitled “main”. It’s possible to create a new branch which allows a programmer to make changes to files in a repository in isolation from the main branch. This is beneficial, because the same file can be compared across branches, potentially created by different scientists, and merged together to reflect those changes. Note: In order for this to work the file in main branch has to have the same name and the file in the newly created branch. Let’s start by creating a new branch. First, navigate to a repository, select “main” and then “View all branches”. Click “New branch”, give your branch a title, and click “Create new branch”. In the screenshot, you’ll see the new branch entitled “jr-changes”. As a new collaborator interested in comparing and merging code changes to a file, click on the new branch that was just created. Based on the screenshot, that means click “jr-changes”. After uploading the file(s) to this branch, you’ll see a notification that this branch is now a certain number of commits ahead of the main branch. A commit records the number of changes to files in a branch. Based on the screenshot, “jr-changes” is now 2 commits ahead of “main”. Click on “2 commits ahead” and scroll down to compare versions between the “main” and “jr-changes” branches. A pull request will need to be created. A pull request allows other collaborators to see changes made to a file within a branch. These proposed changes can be discussed and amended before merging them into the main branch. For more information, visit Github’s branches, pull requests and comparing branches in pull requests documentation. Go ahead and click on “Create pull request”. Click on “Create pull request” again on the next screen. Select “Merge pull request” and then “Confirm merge”. Concluding Remarks In summary, this training module serves as a basic tutorial for sharing code on Github in a way that is beneficial for scientific research. Concepts discussed include uploading and updating code, making a repository easily readable for manuscript submissions, and tracking code changes across collaborators. We encourage trainees and data scientists to implement code tracking and sharing through Github and to also keep up with current trends in data analysis documentation that continue to evolve over time. Test Your Knowledge Try creating your own Github profile, set up a practice repo with subfolders, and a detailed READ.md file paralleling the suggested formatting and content detailed above for your own data analyses! "],["data-wrangling-in-excel.html", "Data Wrangling in Excel Introduction to Training Module Save a Copy of the Soon-To-Be Organized and Cleaned Dataset as a New File Remove Extraneous White Space Replace Missing Data with “NA” Create a Metadata Tab Abbreviate and Capitalize Categorical Data Alphabetize (Sort) the Data by the Categorical Variable of Interest Create a New Subject Number Column Remove Special Symbols and Dashes Bold all Column Names and Center all Data Create a Subject Identifier Column Separate Subject Demographic Data from Experimental Measurements Convert Data from Wide to Long Format Pivoting Data from a Wide to Long Format Generating Summary-Level Statistics with Pivot Tables Excel vs. R: Which Should You Use? Concluding Remarks", " Data Wrangling in Excel This training module was developed by Alexis Payton, Elise Hickman, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module This module is intended to be a starting guide to cleaning and organizing an example toxicology dataset in Excel. Data wrangling involves cleaning, removing of erroneous data, and restructuring necessary for to preparing wet lab generated data for downstream analyses. These steps will ensure that: Data are amenable to downstream analyses in R, or your preferred programming language Data are clear and easily interpretable by collaborators, reviewers, and readers Click here for more information on data wrangling. In this training tutorial, we’ll make use of an example dataset that needs to be wrangled. The dataset contains concentration values for molecules that were measured using protein-based ELISA technologies. These molecules specifically span 17 sterols and cytokines, selected based upon their important roles in mediating biological responses. These measures were derived from human serum samples. Demographic information also exists for each subject. The following steps detailed in this training module are by no means exhaustive! Further resources are provided at the end. This module provides example steps that are helpful when wrangling your data in Excel. Datasets often come in many different formats from our wet bench colleagues, therefore some steps will likely need to be added, removed, or amended depending on your specific data. Save a Copy of the Soon-To-Be Organized and Cleaned Dataset as a New File Open Microsoft Excel and prior to ANY edits, click “File” –&gt; “Save As” to save a new version of the file that can serve as the cleaned version of the data. This is very important for file tracking purposes, and can help in the instance that the original version needs to be referred back to (e.g., if data are accidentally deleted or modified during downstream steps). The file needs to be named something indicative of the data it contains followed by the current date (e.g., “Allostatic Mediator Data_061622”). The title should be succinct and descriptive. It is okay to use dashes or underscores in the name of the title. Do not include special characters, such as $, #, @, !, %, &amp;, *, (, ), and +. Special characters tend to generate errors on local hard drives when syncing to cloud-based servers, and they are difficult to upload into programming software. Let’s first view what the dataset currently looks like: Helpful Excel Keyboard Shortcuts The following keyboard shortcuts can help you work more efficiently in Excel: Move to the last cell in use on the sheet Control + Fn + Right arrow key (Mac users) Control + End (PC users) Move to the beginning of the sheet Control + Fn + Left arrow key, then same Control + Fn + Up arrow key (Mac users) Control + Home (PC users) Highlight and grab all data Click on the first cell in the upper left hand corner then click and hold Shift + Command + Down arrow key + Right arrow key (Mac users) Shift + Command + Down arrow key + Right arrow key (PC users) Note: This only works if there are no cells with missing information or gaps in the columns/rows used to define the peripheral area. For more available shortcuts on various operating systems click here. Remove Extraneous White Space Before we can begin organizing the data, we need to remove the entirely blank rows of cells. This reduces the file size and allows for the use of the filter function in Excel, as well as other organizing functions, which will be used in the next few steps. This step also makes the data look more tidy and amenable to import for coding purposes. Excel Trick #1: Select all lines that need to be removed and press Control + minus key for Mac and PC users. (Note that there are other ways to do this for larger datasets, but this works fine for this small example.) Excel Trick #2: An easier way to remove blank rows and cells for larger datasets, includes clicking “Find &amp; Select”–&gt; “Special” –&gt; “Blanks” –&gt; click “OK” to select all blank rows and cells. Click “Delete” within the home tab –&gt; “Delete sheet rows”. After removing the blank rows, the file should look like the screenshot below. Replace Missing Data with “NA” There are many ways missing data can be encoded in datasets. This includes values like “blank”, “N/A”, “NA”, or leaving a cell blank. Replacing all missing values with “NA” values is done for 2 reasons: To confirm that the data is indeed missing R reads in “NA” values as missing values To check for missing values, the filter function can be used on each column and only select cells with missing values. You may need to scroll to the bottom of the filter pop up window for numerical data. Enter “NA” into the cell of the filtered column. Double click the bottom right corner of the cell to copy the “NA” down the rest of the column. There was no missing data in this dataset, so this step can be skipped. Create a Metadata Tab Metadata explains what each column represents in the dataset. Metadata is now a required component of data sharing, so it is best to initiate this process prior to data analysis. Ideally, this information is filled in by the scientist(s) who generated the data. Create a new tab (preferably as the first tab) and label it “XXXXX_METADATA” (ie., “Allostatic_METADATA”) Then relabel the original data tab as “XXXX_DATA” (ie., “Allostatic_DATA). Within the metadata tab, create three columns: the first, “Column Identifier”, contains each of the column names found in the data tab; the second, “Code”, contains the individual variable/ abbreviation for each column identifier; the third, “Description” contains additional information and definitions for abbreviations. Abbreviate and Capitalize Categorical Data Categorical data are easier to handle in programming languages when they are capitalized and abbreviated. It also helps reduce typos and potential typing mistakes within your script. For this dataset, the following variables were edited: Group “control” became “NS” for non-smoker “smoker” became “CS” for cigarette smoker Sex “f” became “F” for female “m” became “M” for male Race “AA” became “B” for Black “White” became “W” for White Excel Trick: To change cells that contain the same data simultaneously, navigate to “Edit”, click “Find”, and then “Replace”. Once the categorical data have been abbreviated, add those abbreviations to the metadata and describe what they symbolize. Alphabetize (Sort) the Data by the Categorical Variable of Interest For this dataset, we will sort by the column “Group”. This organizes the data and sets it up for the next step. Highlight all the column headers. Click on the “Sort &amp; Filter” button and click “Filter”. Click on the arrow on cell that contains the column name “Group” and click “Ascending”. Create a New Subject Number Column Analysis-specific subjects are created to give an ordinal subject number to each subject, which allows the scientist to easily identify the number of subjects. In addition, these new ordinal subject numbers will be used to create a subject identifier that combines both a subject’s group and subject number that is helpful for downstream visualization analyses. Relabel the subject number/identifier column as “Original_Subject_Number” and create an ordinal subject number column labeled “Subject_Number”. R reads in spaces between words as periods, therefore it’s common practice to replace spaces with underscores when doing data analysis in R. Avoid using dashes in column names or anywhere else in the dataset. Remove Special Symbols and Dashes Programming languages, in general, do not operate well with special symbols and dashes, particularly when included in column identifiers. For this reason, it is best to remove these while cleaning up your data, prior to importing it into R or your preferred programming software. In this case, this dataset contains dashes and Greek letters within some of the column header identifiers. Here, it is beneficial to remove these dashes (e.g., change IL-10 to IL10) and replace the Greek letters with first letter of the word in English (e.g., change TNF-\\(\\alpha\\) to TNFa). Bold all Column Names and Center all Data These data will likely be shared with collaborators, uploaded onto data deposition websites, and used as supporting information in published manuscripts. For these purposes, it is nice to format data in Excel such that it is visually appealing and easy to digest. For example, here, it is nice to bold column identifiers and center the data, as shown below: Create a Subject Identifier Column The subject identifier column labeled, “Group_Subject_No”, combines the subject number with the variable of interest (ie. Group for this dataset). This is useful for analyses to identify outliers by the subject number and the group. Insert 2 additional columns where the current “Sex” column is. To combine values from two different columns, type “=CONCAT(D1,” _ “,C1)” in the first cell in the first column inserted. Double click the right corner of the cell for the formula to be copied to last row in the dataset. Copy the entire column and paste only the values in the second column by navigating to the drop down arrow next to “Paste” and click “Paste Values”. Label the second column “Group_Subject_No” and delete the first column. Separate Subject Demographic Data from Experimental Measurements This example dataset is very small, so the demographic data (e.g., sex, race, age) was kept within the same file as the experimentally measured molecules. Though in larger datasets (e.g., genome-wide data, exposomic data, etc), it is often beneficial to separate the demographic data into one file that can be labeled according to the following format: “XXX_Subject_Info_061622” (ie. “Allostatic_Subject_Info_061622”). This step was not completed for this current data, since it had a smaller size and the downstream analyses were simple. Convert Data from Wide to Long Format A wide format contains values that DO NOT repeat the subject identifier column. For this dataset, each subject has one row containing all of its data, therefore the subject identifier occurs once in the dataset. Wide Format A long format contains values that DO repeat the subject identifier column. For this dataset, that means a new column was created entitled “Variable” containing all the mediator names and a column entitled “Value” containing all their corresponding values. In the screenshot, an additional column, “Category”, was added to help with the categorization of mediators in R analyses. Long Format The reason a long format is preferred is because it makes visualizations and statistical analyses more efficient in R. In the long format, we were able to add a column entitled “Category” to categorize the mediators into “AL Biomarker” or “Cytokine” allowing us to more easily subset the mediators in R. Read more about wide and long formats here. To convert the data from a wide to long format, follow the steps below: Pivoting Data from a Wide to Long Format To do this, a power query in Excel will be used. Note: If you are working on a Mac, you will need to have at least Excel 2016 installed to follow this tutorial, as Power Query is not avaialble for earlier versions. Add-ins are available for Windows users. See this link for more details. Start by copying all of the data, including the column titles. (Hint: Try using the keyboard shortcut mentioned above.) Click the tab at the top that says “Data”. Then click “Get Data (Power Query)” at the far left. It will ask you to choose a data source. Click “Blank table” in the bottom row. Paste the data into the table. (Hint: Use the shortcut Ctrl + “v”). At this point, your screen should look like the screenshot below. Click “Use first row as headers” and then click “Next” in the bottom right hand corner. Select all the columns with biomarker names. That should be the column “Cortisol” through the end. Click the “Transform” button in the upper left hand corner. Then click “Unpivot columns” in the middle of the pane. The final result should look like the sceenshot below with all the biomarkers now in one column entitled “Attribute” and their corresponding values in another column entitled “Value”. To save this, go back to the “Home” tab and click “Close &amp; load”. You should see something similar to the screenshot below. In the upper right with all the shaded tables (within the “Table” tab), click the arrow to the left of the green table until you see one with no shading. Then click the table with no colors. Click “Convert to Range” within the “Table” tab. This removes the power query capabilities, so that the data is a regular excel sheet. Now the “Category” column can be created to identify the types of biomarkers in the dataset. The allostatic load (AL) biomarkers denoted in the “Category” column include the variables Cortisol, CRP, Fibrinogen, Hba1c, HDL, and Noradrenaline. The rest of the variables were labeled as cytokines. Additionally, we can make this data more closely resemble the final long format screenshot by bolding the headers, centering all the data, etc. We have successfully wrangled our data and the final dataset now looks like this: Generating Summary-Level Statistics with Pivot Tables A PivotTable is a tool in Excel used to summarize numerical data. It’s called a pivot table, because it pivots or changes how the data is displayed to make statistical inferences. This can be useful for generating initial summary-level statistics to guage the distribution of data. To create a PivotTable, start by selecting all of the data. (Hint: Try using the keyboard shortcut mentioned above.) Click “Insert” tab on the upper left-hand side, click “PivotTable”, and click “OK”. The new PivotTable should be available in a new sheet as seen in the screenshot below. A PivotTable will be constructed based on the column headers that can be dragged into the PivotTable fields located on the right-hand side. For example, what if we were interested in determining if there were differences in average expression between non-smokers and cigarette smokers in each category of biomarkers? As seen below, drag the “Group” variable under the “Rows” field and drag the “Value” variable under the “Values” field. Notice that it automatically calculates the sum of the expression values for each group. To change the function to average, click the “i” icon and select “Average”. The output should mirror what’s below with non-smokers having an average expression that’s more than double that of cigarette smokers. Excel vs. R: Which Should You Use? For the most part, it’s better to perform final analyses in R (or another programming language) rather than Excel for the following reasons… R clearly shows the code (instructions), which makes editing, interpretability, and sharing easier. This makes analyses more reproducible and can save time. R has packages that makes more complex analyses possible (i.e., machine learning and heatmaps) that aren’t available in Excel. R can handle larger data sets. R can compute and process data faster. However, Excel is still a software that has many benefits for running analyses including… Excel is user-friendly and most people have experience in navigating the software at a basic level. Excel can be faster for rudimentary statistical analyses and visualizations. Depending on each scientist’s skill-level and the complexity of the analysis, Excel or R could be beneficial. Concluding Remarks In summary, this training module highlights the importance of data wrangling and how to do so in Microsoft Excel for downstream analyses. Concepts discussed include helpful Excel features like power queries and pivot tables and when to use Microsoft Excel vs. R. Additional Resources Data wrangling in Excel can be expedited with knowledge of useful features and functions to format data. Check out the resources below for additional information on Excel tricks. Data Analysis in Excel Excel Spreesheet Hacks Excel for Beginners Test Your Knowledge Try wrangling the “Module1_4_TYKInput.xlsx” to mimic the cleaned versions of the data found in “Module1_4_TYKSolution.xlsx”. This dataset includes sterol and cytokine concentration levels extracted from induced sputum samples collected after ozone exposure. After wrangling, you should end up with a sheet for subject information and a sheet for experimental data. Using the a PivotTable on the cleaned dataset, find the standard deviation of each cytokine variable stratified by the disease status. "],["downloading-and-programming-in-r.html", "Downloading and Programming in R Introduction to Training Module General Introduction and Installation of R and RStudio Introduction to R Packages Scripting Basics Code Troubleshooting Concluding Remarks", " Downloading and Programming in R This training module was developed by Kyle Roell, Elise Hickman, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module In this training module, we will provide a brief introduction of: R R Studio Packages in R Scripting basics Code troubleshooting General Introduction and Installation of R and RStudio What is R? R is a programming language. Computer script (lines of code) can be used to increase data analysis reproducibility, transparency, and methods sharing, and is becoming increasingly incorporated into exposure science, toxicology, and environmental health research. One of the most commonly used coding languages in the field of environmental health science is the R language. Some advantages of using R include the following: Free, open-source programming language that is licensed under the Free Software Foundation’s GNU General Public License Can be run across all major platforms and operating systems, including Unix, Windows, and MacOS Publicly available packages help you carry out analyses efficiently (without you having to code for everything yourself) Large, diverse collection of packages Comprehensive documentation When code is efficiently tracked during development/execution, it promotes reproducible analyses Because of these advantages, R has emerged as an avenue for world-wide collaboration in data science. Other commonly implemented scripting languages in the field of environmental health research include Python and SAS, among others; and these training tutorials focus on R as an important introductory-level example that also houses many relevant packages and example datasets as further described throughout TAME. Downloading and Installing R To download R, first navigate to https://cran.rstudio.com/ and download the .pkg file for your operating system. Install this file according to your computer’s typical program installation steps. What is RStudio? RStudio is an Integrated Development Environment (IDE) for R, which makes it more ‘user friendly’ when developing and using R script. It is a desktop application that can be downloaded for free, online. Downloading and Installing RStudio To download RStudio: Navigate to: https://posit.co/download/rstudio-desktop/ Scroll down and select “Download RStudio” Install according to your computer’s typical program installation steps RStudio Orientation Here is a screenshot demonstrating what the RStudio desktop app looks like: The default RStudio layout has four main panes (numbered above in the blue boxes): Source Editor: allows you to open and edit script files and view data. Console: where you can type code that will execute immediately when you press enter/return. This is also where code from script files will appear when you run the code. Environment: shows you the objects in your environment. Viewer: has a number of useful tabs, including: Files: a file manager that allows you to navigate similar to Finder or File Explorer Plots: where plots you generate by executing code will appear Packages: shows you packages that are loaded (checked) and those that can be loaded (unchecked) Help: where help pages will appear for packages and functions (see below for further instructions on the help option) Under “Tools” → “Global Options,” RStudio panes can be customized to appear in different configurations or with different color themes. A number of other options can also be changed. For example, you can choose to have colors highlighted the color they appear or rainbow colored parentheses that can help you visualize nested code. Introduction to R Packages One of the major benefits to coding in the R language is access to the continually expanding resource of thousands of user-developed packages. Packages represent compilations of code and functions fitted for a specialized focus or purpose. These are often written by R users and submitted to the CRAN, or another host such as BioConductor or Github. Packages aid in improved data analyses and methods sharing. Packages have varying utilities, spanning basic organization and manipulation of data, visualizing data, and more advanced approaches to parse and analyze data, with examples included in all of the proceeding training modules. Examples of some common packages that we’ll be using throughout these training modules include the following: tidyverse: A collection of open source R packages that share an underlying design philosophy, grammar, and data structures of tidy data. For more information on the tidyverse package, see its associated CRAN webpage, primary webpage, and peer-reviewed article released in 2018. ggplot2: A system for creating graphics. Users provide the data and tell R what type of graph to use, how to map variables to aesthetics (elements of the graph), and additional stylistic elements to include in the graph. For more information on the ggplot2 package, see its associated CRAN webpage and R Documentation. More information on these packages, as well as many others, is included throughout TAME training modules. Downloading/Installing R Packages R packages often do not need to be downloaded from a website. Instead, you can install packages and load them through running script in R. Note that you only need to install packages one time, but packages must be loaded each time you start a new R session. # Install the package install.packages(“tidyverse”) # Load the package for use library(tidyverse) Many packages also exist as part of the baseline configuration of an R working environment, and do not require manual loading each time you launch R. These include the following packages: datasets graphics methods stats utils You can learn more about a function by typing one question mark before the name of the function, which will bring up documentation in the Help tab of the Viewer window. Importantly, this documentation includes a description of the different arguments that can be passed to the function and examples for how to use the function. ?install.packages ## starting httpd help server ... done You can learn more about a package by typing two question marks before the name of the package. This will bring up vingettes and help pages associated with that package. ??tidyverse Scripting Basics Data Types Before writing any script, let’s first review different data types in R. Data types are what they imply – the type of data you are handling. It is important to understand data types because functions often require a specific data type as input. R has 5 basic data types: Logical (e.g., TRUE or FALSE) Integer (e.g., 1, 2, 3) Numeric (real or decimal) Character (e.g., ”apple”) Complex (e.g., 1 + 0i) Numeric variables are often stored as “double” values (sometimes shown as &lt; dbl &gt;), or a decimal type with at least two decimal places. Character variables can also be stored as factors, which are data structures that are implemented to store categorical data in a specific order (also known as levels). Data are stored in data structures. There are many different data structures in R. Some packages even implement unique data structures. The most common data structures are: Vectors: also known as an atomic vector, can contain characters, logical values, integers, or numeric values (but all elements must be the same data type). Matrices: a vector with multiple dimensions. Elements must still be all the same data type. Data frames: similar to a matrix but can contain different data types and additional attributes such as row names (and is one of the most common data structures in environmental health research). Tibbles are a stricter type of data frame implemented in the tidyverse package. Lists: a special type of vector that acts as a container – other data structures can be stored within the list, and lists can contain other lists. Lists can contain elements that are different data structures. Figure 1: Created with BioRender.com Writing Script R code is written line by line. It may take just one line or many lines of code for one step to be executed, depending on the number of arguments to the function you are using. R code is executed (run) by selecting the line(s) of code to run and pressing return/enter (or a keyboard shortcut), or by clicking “Run” in the upper right corner of the script. A very simple example of running code is as follows: 3 + 4 ## [1] 7 We can see that when we ran our code, the answer was returned. But what if we want to store that answer? We can assign that number to a variable named x using the assignment operator &lt;-: x &lt;- 3 + 4 Then, if we run a line of code with our variable, we will get that value: x ## [1] 7 The assignment operator can also be used to assign values to any of the data structures discussed above, such as vectors and data frames, as shown here: # Creating a vector of values called my_values my_values &lt;- c(7, 3, 8, 9) # Viewing the vector my_values ## [1] 7 3 8 9 # Creating a data frame of values corresponding to colors my_df &lt;- data.frame(values = my_values, color = c(&quot;Blue&quot;, &quot;Red&quot;, &quot;Yellow&quot;, &quot;Purple&quot;)) # Viewing the data frame my_df ## values color ## 1 7 Blue ## 2 3 Red ## 3 8 Yellow ## 4 9 Purple Comments You may have noticed in the code chunks above that there were # followed by phrases describing the code. R allows for scripts to contain non-code elements, called comments, that will not be run or interpreted. Comments are useful to help make code more interpretable for others or to add reminders of what and why parts of code may have been written. To make a comment, simply use a # followed by the comment. A # only comments out a single line of code. In other words, only that line will be commented and therefore not be run, but lines directly above/below it will still be run: # This is an R comment! For more on comments, see TAME 2.0 Module 2.2 Coding Best Practices. Autofilling RStudio will autofill function names and object names as you type, which can save a lot of time. When you are typing a variable or function name, you can press tab while typing. RStudio will look for variables or functions that match the first few letters you’ve typed. If multiple matches are found, RStudio will provide you with a drop down list to select from, which may be useful when searching through newly installed packages or trying to quickly type variable names in an R script. For example, let’s say we instead named our example data frame something much longer, and we had two data frames with similar names. If we start typing in my_ and pause our typing, all of the objects that start with that name will appear as options in a list. To select which one to autofill, navigate down the list and click return/enter. my_df_with_really_long_name &lt;- data.frame(values = my_values, color = c(&quot;Blue&quot;, &quot;Red&quot;, &quot;Yellow&quot;, &quot;Purple&quot;)) my_df_with_really_long_name_2 &lt;- data.frame(values = my_values, color = c(&quot;Green&quot;, &quot;Teal&quot;, &quot;Magenta&quot;, &quot;Orange&quot;)) Finding and Setting Your Working Directory Another step that is commonly done at the very beginning of your code is setting your working direction. This tells your computer where to look for files that you want to import and where to deposit output files produced during your scripted activities. To view your current working directory, run the following: getwd() To set or change the location of your working directory, run the following: setwd(&quot;/file path to where your input files are&quot;) Note that macOS file paths use / to separate folders, whereas PC file paths use \\. You can easily find the file path to your desired working directory by navigating to “Session”, then “Set Working Directory”, and “Choose Directory”: In the popup box, navigate to the folder you want to set as your working directory and click “Open.” Look in the R console, which will now contain a line of code with setwd() containing your file path. You can copy this line of code to the top of your script for future use. Alternatively, you can navigate to the folder you want in Finder or File Explorer and right click to see the file path. Within your working directory, you can make sub-folders to keep your analyses organized. Here is an example folder hierarchy: How you set up your folder hierarchy is highly dependent on your specific analysis and coding style. However, we recommend that you: Name your script something concise, but descriptive (no acronyms) Consider using dates when appropriate Separate your analysis into logical sections so that script doesn’t get too long or hard to follow Revisit and adapt your organization as the project evolves! Archive old code so you can revisit it A Quick Note About Projects Creating projects allows you to store your progress (open script, global environment) for one project in an R Project File. This facilitates quick transitions between multiple projects. Find detailed information about how to set up projects here. Importing Files After setting the working directory, you can import and export files using various functions based on the type of file being imported or exported. Often, it is easiest to import data into R that are in a comma separated value / comma delimited file (.csv) or tab / text delimited file (.txt). Other datatypes such as SAS data files or large .csv files may require different functions to be more efficiently read in, and some of these file formats will be discussed in future modules. Files can also be imported and exported from Excel using the openxlsx package. Below, we will demonstrate how to read in .csv and .txt files: # Read in the .csv data that&#39;s located in our working directory csv.dataset &lt;- read.csv(&quot;Module2_1_Input/Module2_1_InputData1.csv&quot;) # Read in the .txt data txt.dataset &lt;- read.table(&quot;Module2_1_Input/Module2_1_InputData1.txt&quot;) These datasets now appear as saved dataframes (“csv.dataset” and “txt.dataset”) in our working environment. Viewing Data After data have been loaded into R, or created within R, you will likely want to view what these datasets look like. Datasets can be viewed in their entirety, or datasets can be subsetted to quickly look at part of the data. Here’s some example script to view just the beginnings of a dataframe using the head() function: head(csv.dataset) ## Sample Var1 Var2 Var3 ## 1 sample1 1 2 1 ## 2 sample2 2 4 4 ## 3 sample3 3 6 9 ## 4 sample4 4 8 16 ## 5 sample5 5 10 25 Here, you can see that this automatically brings up a view of the first five rows of the dataframe. Another way to view the first five rows of a dataframe is to run the following: csv.dataset[1:5,] ## Sample Var1 Var2 Var3 ## 1 sample1 1 2 1 ## 2 sample2 2 4 4 ## 3 sample3 3 6 9 ## 4 sample4 4 8 16 ## 5 sample5 5 10 25 This brings us to an important concept - indexing! Brackets are used in R to index. Within the bracket, the first argument represents the row numbers, and the second argument represents the column numbers. A colon between two numbers means to select all of the columns in between the left and right numbers. The above line of code told R to select rows 1 to 5, and, by leaving the column argument blank, all of the columns. Expanding on this, to view the first 5 rows and 2 columns, we can run the following: csv.dataset[1:5, 1:2] ## Sample Var1 ## 1 sample1 1 ## 2 sample2 2 ## 3 sample3 3 ## 4 sample4 4 ## 5 sample5 5 For another example: What if we want to only view the first and third row, and first and fourth column? We can use a vector within the index to do this: csv.dataset[c(1, 3), c(1, 4)] ## Sample Var3 ## 1 sample1 1 ## 3 sample3 9 To view the entire dataset, use the View() function: View(csv.dataset) Another way to view a dataset is to just click on the name of the data in the environment pane. The view window will pop up in the same way that it did with the View() function. Determining Data Structures and Data Types As discussed above, there are a number of different data structures and types that can be used in R. Here, we will demonstrate functions that can be used to identify data structures and types within R objects. The glimpse() function, which is part of the tidyverse package, is helpful because it allows us to see an overview of our column names and the types of data contained within those columns. # Load tidyverse package library(tidyverse) glimpse(csv.dataset) ## Rows: 5 ## Columns: 4 ## $ Sample &lt;chr&gt; &quot;sample1&quot;, &quot;sample2&quot;, &quot;sample3&quot;, &quot;sample4&quot;, &quot;sample5&quot; ## $ Var1 &lt;int&gt; 1, 2, 3, 4, 5 ## $ Var2 &lt;int&gt; 2, 4, 6, 8, 10 ## $ Var3 &lt;int&gt; 1, 4, 9, 16, 25 Here, we see that our Sample column is a character column, while the rest are integers. The class() function is also helpful for understanding objects in our global environment: # What class (data structure) is our object? class(csv.dataset) ## [1] &quot;data.frame&quot; # What class (data type) is a specific column in our data? class(csv.dataset$Sample) ## [1] &quot;character&quot; These functions are particularly helpful when introducing new functions or troubleshooting code because functions often require input data to be a specific structure or data type. Exporting Data Now that we have these datasets saved as dataframes, we can use these as examples to export data files from the R environment back into our local directory. There are many ways to export data in R. Data can be written out into a .csv file, tab delimited .txt file, or RData file, for example. There are also many functions within packages that write out specific datasets generated by that package. To write out to a .csv file: write.csv(csv.dataset, &quot;Module2_1_SameCSVFileNowOut.csv&quot;) To write out a .txt tab delimited file: write.table(txt.dataset, &quot;Module2_1_SameTXTFileNowOut.txt&quot;) R also allows objects to be saved in RData files. These files can be read into R, as well, and will load the object into the current workspace. Entire workspaces are also able to be saved in RData files, such that when you open an RData file, your script and Global Environment will be just as you saved them. Below includes example code to carry out these tasks, and though these files are not provided, they are just example code for future reference. # Read in saved single R data object r.obj = readRDS(&quot;data.rds&quot;) # Write single R object to file saveRDS(object, &quot;single_object.rds&quot;) # Read in multiple saved R objects load(&quot;multiple_data.RData&quot;) # Save multiple R objects save(object1, object2, &quot;multiple_objects.RData&quot;) # Save entire workspace save.image(&quot;entire_workspace.RData&quot;) # Load entire workspace load(&quot;entire_workspace.RData&quot;) Code Troubleshooting Learning how to code is an iterative, exploratory process. The secret to coding is to… Make sure to include “R” and the package and/or function name in your search. Don’t be afraid to try out different solutions until you find one that works for you, but also know when it is time to ask for help. For example, when you have tried solutions available on forums, but they aren’t working for you, or you know a colleague has already spent a significant amount of time developing code for this specific task. Note that when reading question/answer forums, make sure to look at how recent a post is, as packages are updated frequently, and old answers may or may not work. Some common reasons that code doesn’t work and potential solutions to these problems include: Two packages are loaded that have functions with the same name, and the default function is not the one you are intending to run. Solutions: specify the package that you want the function to be called from each time you use it (e.g., dplyr::select()) or re-assign that function at the beginning of your script (e.g., select &lt;- dplyr::select) Your data object is the wrong input type (is a data frame and needs to be a matrix, is character but needs to be numeric) Solution: double check the documentation (?functionname) for the input/variable type needed You accidentally wrote over your data frame or variable with another section of code Solution: re-run your code from the beginning, checking that your input is in the correct format There is a bug in the function/package you are trying to use (this is most common after packages are updated or after you update your version of R) Solution: post an issue on GitHub for that package (or StackOverflow if there is not a GitHub) using a reproducible example There are a number of forums that can be extremely helpful when troubleshooting your code, such as: Stack Overflow: one of the most common forums to post questions related to coding and will often be the first few links in a Google search about any code troubleshooting. It is free to make an account, which allows you to post and answer questions. Cross Validated: a forum focused on statistics, including machine learning, data analysis, data mining, and data visualization, and is best for conceptual questions related to how statistical tests are carried out, when to use specific tests, and how to interpret tests (rather than code execution questions, which are more appropriate to post on Stack Overflow). BioConductor Forum: provides a platform for specific coding and conceptual questions about BioConductor packages. GitHub: can also be used to create posts about specific issues/bugs for functions within that package. Before you post a question, make sure you have thoroughly explored answers to existing similar questions and are able to explain in your question why those haven’t worked for you. You will also need to provide a reproducible example of your error or question, meaning that you provide all information (input data, packages, code) needed such that others can reproduce your exact issues. While demonstrating a reproducible example is beyond the scope of this module, see the below links and packages for help getting started: Detailed step-by-step guides for how to make reproducible examples: How to Reprex by Ariel Muldoon What’s a reproducible example (reprex) and how do I create one? Helpful packages: reprex: part of tidyverse, useful for preparing reproducible code for posting to forums. datapasta: useful for creating code you can copy and paste that creates a new data frame as a subset of your original data. Concluding Remarks Together, this training module provides introductory level information on installing and loading packages in R, scripting basics, importing and exporting data, and code troubleshooting. Additional Resources Coursera Stack Overflow How to Learn R R for Data Science Test Your Knowledge Install R and RStudio on your computer. Launch RStudio and explore installing packages (e.g., tidyverse) and understanding data types using the built-in datasets in R. Make a vector of the letters A-E. Make a data frame of the letters A-E in one column and their corresponding number in the alphabet order in the second column (e.g., A corresponds with 1). "],["coding-best-practices.html", "Coding “Best” Practices Introduction to Training Module Scripting File Types Script Headers and Annotation Coding Style Script Organization Concluding Remarks", " Coding “Best” Practices This training module was developed by Kyle Roell, Alexis Payton, Elise Hickman, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module In this training module, we will be going over coding “best” practices. The reason we put “best” in quotes is because these practices are what we currently consider best or better, though everyone has different coding styles, annotation styles, etc that also change over time. Here, we hope to give you a sense of what we do when coding, why we do it, and why we think it is important. We will also be pointing out other guides to style, annotations, and best practices that we suggest implementing into your own coding. Some of the questions we hope to answer in this module are: What type of scripting file should I use? What should I name my script? What should I put at the top of every script and why is it important? How should I annotate my code? Why are annotations important? How do I implement these coding practices into my own code? Where can I find other resources to help with coding best practices? In the following sections, we will be addressing these questions. Keep in mind that the advice and suggestions in this section are just that: advice and suggestions. So please take them into consideration and integrate them into your own coding style as appropriate. Scripting File Types Two of the most common scripting file types applicable to the R language are .R (normal R files) and .Rmd (R Markdown). Normal R files appear as plain text and can be used for running any normal R code. R Markdown files are used for more intensive documentation of code and allow for a combination of code, non-code text explaining the code, and viewing of code output, tables, and figures that are rendered together into an output file (typically .html, although other formats such as .pdf are also offered). For example, TAME is coded using R Markdown, which allows us to include blocks of non-code text, hyperlinks, annotated code, schematics, and output figures all in one place. We highly encourage the use of R Markdown as the default scripting file type for R-based projects because it produces a polished final document that is easy for others to follow, whereas .R files are more appropriate for short, one-off analyses and writing in-depth functions and packages. However, code executed in normal .R files and R Markdown will produce the same results, and ultimately, which file type to use is personal preference. See below for screenshots that demonstrate some of the stylistic differences between .R, .Rmd, and .Rmd knitted to HTML format: If you are interested in learning more about the basic features of R Markdown and how to use them, see the following resources: RStudio introduction to R Markdown R Markdown Cheat Sheet Bookdown R Markdown guide Including external images in R Markdown with knitr Interactive plots with plotly Interactive data tables with DT Naming the Script File The first thing we need to talk about, which is sometimes overlooked in the discussion of coding practices, is script file naming conventions and high level descriptive headers within a script. It is important to remember to name your code something concise, but descriptive. You want to be able to easily recognize what the script is for and does without a cumbersome, lengthy title. Some tips for naming conventions: Be concise, but descriptive Use dates when appropriate Avoid special characters Use full words if possible, avoiding non-standard acronyms Keep in mind that each script should have a clear purpose within a given project. And, it is sometimes necessary, and often common, to have multiple scripts within one project that all pertain to different parts of the analysis. For example, it may be appropriate to have one script for data cleaning and pre-processing and another script for analyzing data. When scripting an analysis with multiple sub-analyses, some prefer to keep code for each sub-analysis separate (e.g., one file for an ANOVA and one file for a k-means analysis on the same data input), while others prefer to have longer code files with more subsections. Whichever method you choose, we recommend maintaining clear documentation that indicates locations for input and output files for each sub-analysis (e.g., whether global environment objects or output files from a previous script are needed to run the current script). Script Headers and Annotation Script Header Once your script is created and named, it is generally recommended to include a header at the top of the script. The script header can be used for describing: Title of Script - This can be a longer or more readable name than script file name. Author(s) - Who wrote the script? Date - When was the script developed? Description - Provides a more detailed description of the purpose of the script and any notes or special considerations for this particular script. In R, it is common to include multiple #, the comment operator, or a # followed by another special character, to start and end a block of coding annotation or the script header. An example of this in an .R file is shown below: ######################################################################## ######################################################################## ### Script Longer Title ### ### Description of what this script does! ### Also can include special notes or anything else here. ### ### Created by: Kyle Roell and Julia Rager ### Last updated: 01 May 2023 ######################################################################## ######################################################################## This block of comment operators is common in .R but not .Rmd files because .Rmd files have their own specific type of header, known as the YAML, which contains the title, author, date, and formatting outputs for the .Rmd file: We will now review how annotations within the script itself can make a huge difference in understanding the code within. Annotations Before we review coding style considerations, it is important to address code annotating. So, what are annotations and why are they important? Annotations are notes embedded within your code as comments that will not be run. The beauty of annotating your code is that not only others, but future you, will be able to read through and better understand what a particular piece of code does. We suggest annotating your code while you write it and incorporate a lot of description. While not every single line needs an annotation, or a very detailed one, it is helpful to provide comments and annotation as much as you can while maintaining feasibility. General annotation style In general, annotations will be short sentences that describe what your code does or why you are executing that specific code. This can be helpful when you are defining a covariate a specific way, performing a specific analytical technique, or just generally explaining why you are doing what you’re doing. # Performing logistic regression to assess association between xyz and abc # Regression confounders: V1, V2, V3 ... xyz.regression.output = glm(xyz ~ abc + V1 + V2 + V3, family=binomial(), data=example.data) Mid-script headings Another common approach to annotations is to use mid-script type headings to separate out the script into various sections. For example, you might want to create distinct sections for “Loading Packages, Data, and Setup”, “Covariate Definition”, “Correlation Analysis”, “Regression Analysis”, etc. This can help you, and others, reading your script, to navigate the script more easily. It also can be more visually pleasing to see the script split up into multiple sections as opposed to one giant chunk of code interspersed with comments. Similar to above, the following example is specific to .R files. For .Rmd files, sub headers can be created by increasing the number of # before the header. ########################################################################### ########################################################################### ### ### Regression Analyses ### ### You can even add some descriptions or notes here about this section! ### ########################################################################### # Performing logistic regression to assess association between xyz and abc # Regression confounders: V1, V2, V3 ... xyz.regression.output = glm(xyz ~ abc + V1 + V2 + V3, family=binomial(), data=example.data) General tips for annotations: Make comments that are useful and meaningful You don’t need to comment every single line In general, you probably won’t over-comment your script, so more is generally better That being said, don’t write super long paragraphs every few lines Split up your script into various sections using mid-script headings when appropriate Quick, short comments and annotations While it is important to provide descriptive annotations, not every one needs to be a sentence or longer. As stated previously, it is not necessary to comment every single line. Here is an example of very brief commenting: # Loading necessary packages library(ggplot2) # Plotting package In the example above, we can see that these short comments clearly convey what the script does – load the necessary package and indicate what the package is needed for. Short, one line annotations can also be placed after lines to clarify that specific line or within the larger mid-script headings to split up these larger sections of code. Coding Style Coding style is often a contentious topic! There are MANY styles of coding, and no two coders have the same exact style, even if they are following the same reference. Here, we will provide some guides to coding style and go over some of the basic, general tips for making your code readable and efficient. Here is an example showing how you can use spacing to align variable assignment: # Example of using spacing for alignment of variable assignment Longer_variable_name_x = 1 Short_name_y = 2 Note that guides will suggest you use &lt;- as the assignment operator. However, for most situations, &lt;- and = will do the same thing. For spacing around certain symbols and operators: Include a space after if, before parenthesis Include a space on either side of symbols such as &lt; The first (opening) curly brace should not be on its own line, but the second (closing) should # Example of poor style if(Longer_variable_name_x &lt;Short_name_y) {Short_name_y = 0} # Example of better style if (Longer_variable_name_x &lt; Short_name_y) { Short_name_y = 0 } Summary of general tips for coding style: Variable names Make them intuitive, short, but descriptive Use the same convention throughout (ex: separating words with . or _) Data names Use _datatype (e.g., cytokines_df for a dataframe data type) following a short, concise name for your data so that readers know what type of input is being used AND/OR Clearly define important data frames or other input in your code comments (e.g., “cytokines” refers to a data frame containing all of the cytokine data from this experiment) Separate long lines onto two or more lines (typically for loops or functions) Use &lt;- for assignment operator Using spacing appropriately for readability Alignment of lines After certain keywords when appropriate Be consistent throughout Example: if you use Tidyverse conventions, continue to use it throughout your script Try to make your code as readable as possible Common style guides: Google’s Style Guide Tidyverse Style Guide R Blogger’s Best Coding Practices Script Organization Lastly, it is important to note that organizing your script efficiently can help with readability as well. In general, as stated before, the beginning of your script should contain some sort of header lines to describe the script. The basic ordering we suggest for most scripts is: Header section Loading libraries and data Function definitions (if any user defined functions exist) Data and variables manipulation Analyses While following this exact organization isn’t absolute, using this structure or something similar can greatly improve the clarity of your analyses and make them easier for others to follow. Concluding Remarks In this module, we demonstrate basic coding style and best practices. Please reference additional style guides (above) and create your own style from best practices that work for you. Aim to make your code understandable and readable, and consider what a future reader, including yourself, will need to understand why you wrote your code the way you did and how to apply your code to new analyses. Test Your Knowledge Using the input file provided (“Module2_2_TYKInput.R”): Convert the script and annotations into R Markdown format. Improve the organization, comments, and scripting to follow the coding best practices described in this module. List the changes you made at the bottom of the new R Markdown file. Notes on the starting code: This starting code uses dummy data to demonstrate how to make a graph in R that includes bars representing the mean, with standard deviation error bars overlaid. You don’t need to understand every step in the code to be able to improve the existing coding style! You can run each step of the code if needed to understand better what it does. "],["data-manipulation-and-reshaping.html", "Data Manipulation and Reshaping Introduction to Training Module Data Manipulation Using Base R Introduction to Tidyverse Concluding Remarks", " Data Manipulation and Reshaping This training module was developed by Kyle Roell, Alexis Payton, Elise Hickman, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module Data within the fields of exposure science, toxicology, and public health are very rarely prepared and ready for downstream statistical analyses and visualization code. The beginning of almost any scripted analysis includes important formatting steps that make the data easier to read and work with. This can be done in several ways, including: Base R operations and functions, or A collection of packages (and philosophy) known as The Tidyverse. In this training tutorial we will review some of the most common ways you can organize and manipulate data, including: Merging data Filtering and subsetting data Pivoting data wider and longer (also known as casting and melting) These approaches will first be demonstrated using the functions available in base R. Then, the exact same approaches will be demonstrated using the functions and syntax that are part of the Tidyverse package. We will demonstrate these data manipulation and organization methods using an environmentally relevant example data set from a human cohort. This dataset was generated by creating data distributions randomly pulled from our previously published cohorts, resulting in a unique data set for these training purposes. The dataset contains environmental exposure metrics from metal levels obtained using sources of drinking water and human urine samples and associated demographic data. Training Module’s Environmental Health Question This training module was specifically developed to answer the following environmental health question using data manipulation and reshaping approaches: What is the average urinary chromium concentration across different maternal education levels? We’ll use base R and Tidydverse to answer this question, but let’s start with Base R. Workspace Preparation and Data Import Set your working directory In preparation, first let’s set our working directory to the folder path that contains our input files: setwd(&quot;/file path to where your input files are&quot;) Note that macOS file paths use / as folder separators, and PC file paths use \\. Importing example datasets Next, let’s read in our example data sets: demographic_data &lt;- read.csv(&quot;Module2_3_Input/Module2_3_InputData1.csv&quot;) chemical_data &lt;- read.csv(&quot;Module2_3_Input/Module2_3_InputData2.csv&quot;) Viewing example datasets Let’s see what these datasets look like: dim(demographic_data) ## [1] 200 6 dim(chemical_data) ## [1] 200 7 The demographic data set includes 200 rows x 7 columns, while the chemical measurement data set includes 200 rows x 7 columns. We can preview the demographic data frame by using the head() function, which displays all the columns and the first 6 rows of a data frame: head(demographic_data) ## ID BMI MAge MEdu BW GA ## 1 1 27.7 22.99928 3 3180.058 34 ## 2 2 26.8 30.05142 3 3210.823 43 ## 3 3 33.2 28.04660 3 3311.551 40 ## 4 4 30.1 34.81796 3 3266.844 32 ## 5 5 37.4 42.68440 3 3664.088 35 ## 6 6 33.3 24.94960 3 3328.988 40 These demographic data are organized according to subject ID (first column) followed by the following subject information: ID: subject number BMI: body mass index MAge: maternal age in years MEdu: maternal education level; 1 = “less than high school”, 2 = “high school or some college”, 3 = “college or greater” BW: body weight in grams GA: gestational age in weeks We can also preview the chemical dataframe: head(chemical_data) ## ID DWAs DWCd DWCr UAs UCd UCr ## 1 1 6.426464 1.292941 51.67987 10.192695 0.7537104 42.60187 ## 2 2 7.832384 1.798535 50.10409 11.815088 0.9789506 41.30757 ## 3 3 7.516569 1.288461 48.74001 10.079057 0.1903262 36.47716 ## 4 4 5.906656 2.075259 50.92745 8.719123 0.9364825 42.47987 ## 5 5 7.181873 2.762643 55.16882 9.436559 1.4977829 47.78528 ## 6 6 9.723429 3.054057 51.14812 11.589403 1.6645837 38.26386 These chemical data are organized according to subject ID (first column), followed by measures of: DWAs: drinking water arsenic levels in µg/L DWCd: drinking water cadmium levels in µg/L DWCr: drinking water chromium levels in µg/L UAs: urinary arsenic levels in µg/L UCd: urinary cadmium levels in µg/L UCr: urinary chromium levels in µg/L Data Manipulation Using Base R Merging Data Using Base R Syntax Merging datasets represents the joining together of two or more datasets, using a common identifier (generally some sort of ID) to connect the rows. This is useful if you have multiple datasets describing different aspects of the study, different variables, or different measures across the same samples. Samples could correspond to the same study participants, animals, cell culture samples, environmental media samples, etc, depending on the study design. In the current example, we will be joining human demographic data and environmental metals exposure data collected from drinking water and human urine samples. Let’s start by merging the example demographic data with the chemical measurement data using the base R function merge(). To learn more about this function, you can type ?merge, which brings up helpful information in the R console. To merge these datasets with the merge function, use the following code. The by = argument specifies the column used to match the rows of data. full.data &lt;- merge(demographic_data, chemical_data, by = &quot;ID&quot;) dim(full.data) ## [1] 200 12 This merged dataframe contains 200 rows x 12 columns. Viewing this merged dataframe, we can see that the merge() function retained the first column in each original dataframe (ID), though did not replicate it since it was used as the identifier for merging. All other columns include their original data, just merged together by the IDs in the first column. head(full.data) ## ID BMI MAge MEdu BW GA DWAs DWCd DWCr UAs ## 1 1 27.7 22.99928 3 3180.058 34 6.426464 1.292941 51.67987 10.192695 ## 2 2 26.8 30.05142 3 3210.823 43 7.832384 1.798535 50.10409 11.815088 ## 3 3 33.2 28.04660 3 3311.551 40 7.516569 1.288461 48.74001 10.079057 ## 4 4 30.1 34.81796 3 3266.844 32 5.906656 2.075259 50.92745 8.719123 ## 5 5 37.4 42.68440 3 3664.088 35 7.181873 2.762643 55.16882 9.436559 ## 6 6 33.3 24.94960 3 3328.988 40 9.723429 3.054057 51.14812 11.589403 ## UCd UCr ## 1 0.7537104 42.60187 ## 2 0.9789506 41.30757 ## 3 0.1903262 36.47716 ## 4 0.9364825 42.47987 ## 5 1.4977829 47.78528 ## 6 1.6645837 38.26386 These datasets were actually quite easy to merge, since they had the same exact column identifier and number of rows. You can edit your script to include more specifics in instances when these may differ across datasets that you would like to merge. This option allows you to edit the name of the column that is used in each dataframe. Here, these are still the same “ID”, but you can see that adding the by.x and by.y arguments allows you to specify instances when different column names are used in the two datasets. full.data &lt;- merge(demographic_data, chemical_data, by.x = &quot;ID&quot;, by.y = &quot;ID&quot;) # Viewing data head(full.data) ## ID BMI MAge MEdu BW GA DWAs DWCd DWCr UAs ## 1 1 27.7 22.99928 3 3180.058 34 6.426464 1.292941 51.67987 10.192695 ## 2 2 26.8 30.05142 3 3210.823 43 7.832384 1.798535 50.10409 11.815088 ## 3 3 33.2 28.04660 3 3311.551 40 7.516569 1.288461 48.74001 10.079057 ## 4 4 30.1 34.81796 3 3266.844 32 5.906656 2.075259 50.92745 8.719123 ## 5 5 37.4 42.68440 3 3664.088 35 7.181873 2.762643 55.16882 9.436559 ## 6 6 33.3 24.94960 3 3328.988 40 9.723429 3.054057 51.14812 11.589403 ## UCd UCr ## 1 0.7537104 42.60187 ## 2 0.9789506 41.30757 ## 3 0.1903262 36.47716 ## 4 0.9364825 42.47987 ## 5 1.4977829 47.78528 ## 6 1.6645837 38.26386 Note that after merging datasets, it is always helpful to check that the merging was done properly before proceeding with your data analysis. Helpful checks could include viewing the merged dataset, checking the numbers of rows and columns to make sure chunks of data are not missing, and searching for values (or strings) that exist in one dataset but not the other, among other mechanisms of QA/QC. Filtering and Subsetting Data Using Base R Syntax Filtering and subsetting data are useful tools when you need to focus on specific parts of your dataset for downstream analyses. These could represent, for example, specific samples or participants that meet certain criteria that you are interested in evaluating. It is also useful for removing unneeded variables or samples from dataframes as you are working through your script. Note that in the examples that follow, we will create new dataframes that are distinguished from our original dataframe by adding sequential numbers to the end of the dataframe name (e.g., subset.data1, subset.data2, subset.data3). This style of dataframe naming is useful for the simple examples we are demonstrating, but in a full scripted analysis, we encourage the use of more descriptive dataframe names. For example, if you are subsetting your data to include only the first 100 rows, you could name that dataframe “data.first100.” For this example, let’s first define a vector of columns that we want to keep in our analysis, then subset the data by keeping only the columns specified in our vector: # Defining a vector of columns to keep in the analysis subset.columns &lt;- c(&quot;BMI&quot;, &quot;MAge&quot;, &quot;MEdu&quot;) # Subsetting the data by selecting the columns represented in the defined &#39;subset.columns&#39; vector subset.data1 &lt;- full.data[,subset.columns] # Viewing the top of this subsetted dataframe head(subset.data1) ## BMI MAge MEdu ## 1 27.7 22.99928 3 ## 2 26.8 30.05142 3 ## 3 33.2 28.04660 3 ## 4 30.1 34.81796 3 ## 5 37.4 42.68440 3 ## 6 33.3 24.94960 3 We can also easily subset data based on row numbers. For example, to keep only the first 100 rows: subset.data2 &lt;- full.data[1:100,] # Viewing the dimensions of this new dataframe dim(subset.data2) ## [1] 100 12 To remove the first 100 rows, we use the same code as above, but include a - sign before our vector to indicate that these rows should be removed: subset.data3 &lt;- full.data[-c(1:100),] # Viewing the dimensions of this new dataframe dim(subset.data3) ## [1] 100 12 Conditional statements are also written to filter and subset data. A conditional statement is written to execute one block of code if the statement is true and a different block of code if the statement is false. A conditional statement requires a Boolean or true/false statement that will be either TRUE or FALSE. A couple of the more commonly used functions used to create conditional statements include… if(){} or an if statement means “execute R code when the condition is met”. if(){} else{} or an if/else statement means “execute R code when condition 1 is met, if not execute R code for condition 2”. ifelse() is a function that executes the same logic as an if/else statement. The first argument specifies a condition to be met. If that condition is met, R code in the second argument is executed, and if that condition is not met, R code in the third argument is executed. There are six comparison operators that are used to created these Boolean values: == means “equals”. != means “not equal”. &lt; means “less than”. &gt; means “greater than”. &lt;= means “less than or equal to”. &gt;= mean “greater than or equal to”. There are also three logical operators that are used to create these Boolean values: &amp; means “and”. | means “or”. ! means “not”. We can filter data based on conditions using the subset() function. For example, the following code filters for subjects whose BMI is greater than 25 and who have a college education: subset.data4 &lt;- subset(full.data, BMI &gt; 25 &amp; MEdu == 3) Additionally, we can subset and select specific columns we would like to keep, using the select argument within the subset() function: # Filtering for subjects whose BMI is less than 22 or greater than 27 # Also selecting the BMI, maternal age, and maternal education columns subset.data5 &lt;- subset(full.data, BMI &lt; 22 | BMI &gt; 27, select = subset.columns) For more information on the subset() function, see its associated documentation. Melting and Casting Data using Base R Syntax Melting and casting refers to the conversion of data to “long” or “wide” form as discussed previously in TAME 2.0 Module 1.4 Data Wrangling in Excel. You will often see data within the environmental health field in wide format, though long format is necessary for some procedures, such as plotting with ggplot2 and performing certain analyses. Here, we’ll illustrate some example script to melt and cast data using the reshape2 package. Let’s first install and load the reshape2 package: if (!requireNamespace(&quot;reshape2&quot;)) install.packages(&quot;reshape2&quot;); library(reshape2) ## ## Attaching package: &#39;reshape2&#39; ## The following object is masked from &#39;package:tidyr&#39;: ## ## smiths Using the fully merged dataframe, let’s remind ourselves what these data look like in the current dataframe format: head(full.data) ## ID BMI MAge MEdu BW GA DWAs DWCd DWCr UAs ## 1 1 27.7 22.99928 3 3180.058 34 6.426464 1.292941 51.67987 10.192695 ## 2 2 26.8 30.05142 3 3210.823 43 7.832384 1.798535 50.10409 11.815088 ## 3 3 33.2 28.04660 3 3311.551 40 7.516569 1.288461 48.74001 10.079057 ## 4 4 30.1 34.81796 3 3266.844 32 5.906656 2.075259 50.92745 8.719123 ## 5 5 37.4 42.68440 3 3664.088 35 7.181873 2.762643 55.16882 9.436559 ## 6 6 33.3 24.94960 3 3328.988 40 9.723429 3.054057 51.14812 11.589403 ## UCd UCr ## 1 0.7537104 42.60187 ## 2 0.9789506 41.30757 ## 3 0.1903262 36.47716 ## 4 0.9364825 42.47987 ## 5 1.4977829 47.78528 ## 6 1.6645837 38.26386 These data are represented by single subject identifiers listed as unique IDs per row, with associated environmental measures and demographic data organized across the columns. Thus, this dataframe is currently in wide (also known as casted) format. Let’s convert this dataframe to long (also known as melted) format. Here, will will specify that we want a row for each unique sample ID + variable measure pair by using id = \"ID\": full.melted &lt;- melt(full.data, id = &quot;ID&quot;) # Viewing this new dataframe head(full.melted) ## ID variable value ## 1 1 BMI 27.7 ## 2 2 BMI 26.8 ## 3 3 BMI 33.2 ## 4 4 BMI 30.1 ## 5 5 BMI 37.4 ## 6 6 BMI 33.3 You can see here that each measure that was originally contained as a unique column has been reoriented, such that the original column header is now listed throughout the second column labeled variable. Then, the third column contains the value of this variable. Let’s see an example view of the middle of this new dataframe: full.melted[1100:1110,1:3] ## ID variable value ## 1100 100 DWAs 7.928885 ## 1101 101 DWAs 8.677403 ## 1102 102 DWAs 8.115183 ## 1103 103 DWAs 7.134189 ## 1104 104 DWAs 8.816142 ## 1105 105 DWAs 7.487227 ## 1106 106 DWAs 7.541973 ## 1107 107 DWAs 6.313516 ## 1108 108 DWAs 6.654474 ## 1109 109 DWAs 7.564429 ## 1110 110 DWAs 7.357122 Here, we can see a different variable (DWAs) now being listed. This continues throughout the entire dataframe, which has the following dimensions: dim(full.melted) ## [1] 2200 3 Let’s now re-cast this dataframe back into wide format using the dcast() function. Here, we are telling the dcast() function to give us a sample (ID) for every variable in the column labeled variable. The column names from the variable column and corresponding values from the value column are then used to fill in the dataset: full.cast &lt;- dcast(full.melted, ID ~ variable) head(full.cast) ## ID BMI MAge MEdu BW GA DWAs DWCd DWCr UAs ## 1 1 27.7 22.99928 3 3180.058 34 6.426464 1.292941 51.67987 10.192695 ## 2 2 26.8 30.05142 3 3210.823 43 7.832384 1.798535 50.10409 11.815088 ## 3 3 33.2 28.04660 3 3311.551 40 7.516569 1.288461 48.74001 10.079057 ## 4 4 30.1 34.81796 3 3266.844 32 5.906656 2.075259 50.92745 8.719123 ## 5 5 37.4 42.68440 3 3664.088 35 7.181873 2.762643 55.16882 9.436559 ## 6 6 33.3 24.94960 3 3328.988 40 9.723429 3.054057 51.14812 11.589403 ## UCd UCr ## 1 0.7537104 42.60187 ## 2 0.9789506 41.30757 ## 3 0.1903262 36.47716 ## 4 0.9364825 42.47987 ## 5 1.4977829 47.78528 ## 6 1.6645837 38.26386 Here, we can see that this dataframe is back in its original casted (or wide) format. Now that we’re familiar with some base R functions to reshape our data, let’s answer our original question: What is the average urinary chromium concentration for each maternal education level? Although it is not necessary to calculate the average, we could first subset our data frame to only include the two columns we are interested in (MEdu and UCr): subset.data6 &lt;- full.data[,c(&quot;MEdu&quot;, &quot;UCr&quot;)] head(subset.data6) ## MEdu UCr ## 1 3 42.60187 ## 2 3 41.30757 ## 3 3 36.47716 ## 4 3 42.47987 ## 5 3 47.78528 ## 6 3 38.26386 Next, we will make a new data frame for each maternal education level: # Creating new data frames based on maternal education category data.matedu.1 &lt;- subset(subset.data6, MEdu == 1) data.matedu.2 &lt;- subset(subset.data6, MEdu == 2) data.matedu.3 &lt;- subset(subset.data6, MEdu == 3) # Previewing the first data frame to make sure our function is working as specified head(data.matedu.1) ## MEdu UCr ## 14 1 38.59349 ## 18 1 47.77878 ## 37 1 35.33980 ## 63 1 34.72255 ## 66 1 34.13982 ## 76 1 31.38145 Last, we can calculate the average urinary chromium concentration using each of our data frames: mean(data.matedu.1$UCr) ## [1] 39.88055 mean(data.matedu.2$UCr) ## [1] 40.61807 mean(data.matedu.3$UCr) ## [1] 40.41556 With this, we can answer our Environmental Health Question: What is the average urinary chromium concentration across different maternal education levels? Answer: The average urinary Chromium concentrations are 39.9 µg/L for participants with less than high school education, 40.6 µg/L for participants with high school or some college education, and 40.4 µg/L for participants with college education or greater. Introduction to Tidyverse Tidyverse is a collection of packages that are commonly used to more efficiently organize and manipulate datasets in R. This collection of packages has its own specific type of syntax and formatting that differ slightly from base R functions. There are eight core tidyverse packages: For data visualization and exploration: ggplot2 For data wrangling and transformation: dplyr tidyr stringr forcats For data import and management: tibble readr For functional programming: purr Here, we will carry out all the of the same data organization exercises demonstrated above using packages that are part of The Tidyverse, specifically using functions that are part of the dplyr and tidyr packages. Downloading and Loading the Tidyverse Package If you don’t have tidyverse already installed, you will need to install it using: if(!require(tidyverse)) install.packages(&quot;tidyverse&quot;) And then load the tidyverse package using: library(tidyverse) Note that by loading the tidyverse package, you are also loading all of the packages included within The Tidyverse and do not need to separately load these packages. Merging Data Using Tidyverse Syntax To merge the same example dataframes using tidyverse, you can run the following script: full.data.tidy &lt;- inner_join(demographic_data, chemical_data, by = &quot;ID&quot;) head(full.data.tidy) ## ID BMI MAge MEdu BW GA DWAs DWCd DWCr UAs ## 1 1 27.7 22.99928 3 3180.058 34 6.426464 1.292941 51.67987 10.192695 ## 2 2 26.8 30.05142 3 3210.823 43 7.832384 1.798535 50.10409 11.815088 ## 3 3 33.2 28.04660 3 3311.551 40 7.516569 1.288461 48.74001 10.079057 ## 4 4 30.1 34.81796 3 3266.844 32 5.906656 2.075259 50.92745 8.719123 ## 5 5 37.4 42.68440 3 3664.088 35 7.181873 2.762643 55.16882 9.436559 ## 6 6 33.3 24.94960 3 3328.988 40 9.723429 3.054057 51.14812 11.589403 ## UCd UCr ## 1 0.7537104 42.60187 ## 2 0.9789506 41.30757 ## 3 0.1903262 36.47716 ## 4 0.9364825 42.47987 ## 5 1.4977829 47.78528 ## 6 1.6645837 38.26386 Note that you can still merge dataframes that have different ID column names with the argument by = c(\"ID.x\", \"ID.y\"). tidyverse also has other join, functions, shown in the graphic below (source): inner_join keeps only rows that have matching ID variables in both datasets full_join keeps the rows in both datasets left_join matches rows based on the ID variables in the first dataset (and omits any rows from the second dataset that do not have matching ID variables in the first dataset) right_join matches rows based on ID variables in the second dataset (and omits any rows from the first dataset that do not have matching ID variables in the second dataset) anti_join(x,y) keeps the rows that are unique to the first dataset anti_join(y,x) keeps the rows that are unique to the second dataset The Pipe Operator One of the most important elements of Tidyverse syntax is use of the pipe operator (%&gt;%). The pipe operator can be used to chain multiple functions together. It takes the object (typically a dataframe) to the left of the pipe operator and passes it to the function to the right of the pipe operator. Multiple pipes can be used in chain to execute multiple data cleaning steps without the need for intermediate dataframes. The pipe operator can be used to pass data to functions within all of the Tidyverse universe packages, not just the functions demonstrated here. Below, we can see the same code executed above, but this time with the pipe operator. The demographic_data dataframe is passed to inner_join() as the first argument to that function, with the following arguments remaining the same. full.data.tidy2 &lt;- demographic_data %&gt;% inner_join(chemical_data, by = &quot;ID&quot;) head(full.data.tidy2) ## ID BMI MAge MEdu BW GA DWAs DWCd DWCr UAs ## 1 1 27.7 22.99928 3 3180.058 34 6.426464 1.292941 51.67987 10.192695 ## 2 2 26.8 30.05142 3 3210.823 43 7.832384 1.798535 50.10409 11.815088 ## 3 3 33.2 28.04660 3 3311.551 40 7.516569 1.288461 48.74001 10.079057 ## 4 4 30.1 34.81796 3 3266.844 32 5.906656 2.075259 50.92745 8.719123 ## 5 5 37.4 42.68440 3 3664.088 35 7.181873 2.762643 55.16882 9.436559 ## 6 6 33.3 24.94960 3 3328.988 40 9.723429 3.054057 51.14812 11.589403 ## UCd UCr ## 1 0.7537104 42.60187 ## 2 0.9789506 41.30757 ## 3 0.1903262 36.47716 ## 4 0.9364825 42.47987 ## 5 1.4977829 47.78528 ## 6 1.6645837 38.26386 Because the pipe operator is often used in a chain, it is best practice is to start a new line after each pipe operator, with the new lines of code indented. This makes code with multiple piped steps easier to follow. However, if just one function is being executed, the pipe operator can be used on the same line as the input and function or omitted altogether (as shown in the previous two code chunks). Here is an example of placing the function to the right of the pipe operator on a new line, with placeholder functions shown as additional steps: full.data.tidy3 &lt;- demographic_data %&gt;% inner_join(chemical_data, by = &quot;ID&quot;) %&gt;% additional_function_1() %&gt;% additional_function_2() Filtering and Subsetting Data Using Tidyverse Syntax Column-wise functions The select() function is used to subset columns in Tidyverse. Here, we can use our previously defined vector subset.columns in the select() function to keep only the columns in our subset.columns vector. The all_of() function tells the select() to keep all of the columns that match elements of the subset.columns vector. subset.tidy1 &lt;- full.data.tidy %&gt;% select(all_of(subset.columns)) head(subset.tidy1) ## BMI MAge MEdu ## 1 27.7 22.99928 3 ## 2 26.8 30.05142 3 ## 3 33.2 28.04660 3 ## 4 30.1 34.81796 3 ## 5 37.4 42.68440 3 ## 6 33.3 24.94960 3 There are many different ways that select() can be used. See below for some examples using dummy variable names: # Select specific ranges in the dataframe data &lt;- data %&gt;% select(start_column_1:end_column_1) data &lt;- data %&gt;% select(c(start_column_1:end_column_1, start_column_2:end_column_2)) # Select columns that match the elements in a character vector an an additional range of columns data &lt;- data %&gt;% select(c(all_of(character_vector), start_column_1:end_column_1)) To select columns that have names that contain specific strings, you can use functions such as starts_with(), ends_with(), and contains(). These functions allow you to ignore the case of the strings with ignore.case = TRUE. These arguments can be combined with specific column names and other selection ranges. data &lt;- data %&gt;% select(starts_with(&quot;starting_string&quot;)) data &lt;- data %&gt;% select(other_column_to_keep, starts_with(&quot;starting_string&quot;)) To remove columns using tidyverse, you can use similar code, but include a - sign before the argument defining the columns. # Removing columns subset.tidy2 &lt;- full.data.tidy %&gt;% select(-all_of(subset.columns)) # Viewing this new dataframe head(subset.tidy2) ## ID BW GA DWAs DWCd DWCr UAs UCd UCr ## 1 1 3180.058 34 6.426464 1.292941 51.67987 10.192695 0.7537104 42.60187 ## 2 2 3210.823 43 7.832384 1.798535 50.10409 11.815088 0.9789506 41.30757 ## 3 3 3311.551 40 7.516569 1.288461 48.74001 10.079057 0.1903262 36.47716 ## 4 4 3266.844 32 5.906656 2.075259 50.92745 8.719123 0.9364825 42.47987 ## 5 5 3664.088 35 7.181873 2.762643 55.16882 9.436559 1.4977829 47.78528 ## 6 6 3328.988 40 9.723429 3.054057 51.14812 11.589403 1.6645837 38.26386 Row-wise functions The slice() function can be used to keep or remove a certain number of rows based on their position within the dataframe. For example, we can retain only the first 100 rows using the following code: subset.tidy3 &lt;- full.data.tidy %&gt;% slice(1:100) dim(subset.tidy3) ## [1] 100 12 Or, we can remove the first 100 rows: subset.tidy4 &lt;- full.data.tidy %&gt;% slice(-c(1:100)) dim(subset.tidy4) ## [1] 100 12 The related functions slice_min() and slice_max() can be used to select rows with the smallest or largest values of a variable. The filter() function can be used to keep or remove specific rows based on conditional statements. For example, we can keep only rows where BMI is greater than 25 and age is greater than 31: subset.tidy5 &lt;- full.data.tidy %&gt;% filter(BMI &gt; 25 &amp; MAge &gt; 31) dim(subset.tidy5) ## [1] 49 12 Combining column and row-wise functions Now, we can see how Tidyverse makes it easy to chain together multiple data manipulation steps. Here, we first filter rows based on values for BMI and age, then we select our columns of interest: subset.tidy6 &lt;- full.data.tidy %&gt;% filter(BMI &gt; 25 &amp; MAge &gt; 31) %&gt;% select(BMI, MAge, MEdu) head(subset.tidy6) ## BMI MAge MEdu ## 1 30.1 34.81796 3 ## 2 37.4 42.68440 3 ## 3 36.9 33.58589 3 ## 4 33.7 33.82961 3 ## 5 25.7 37.08028 3 ## 6 28.4 47.85761 3 Melting and Casting Data Using Tidyverse Syntax To melt and cast data in Tidyverse, you can use the pivot functions (i.e., pivot_longer() or pivot_wider()). The first argument in the pivot_longer() function specifies which columns should be pivoted. This can be specified with either positive or negative selection - i.e., naming columns to pivot with a vector or range or naming columns not to pivot with a - sign. Here, we are telling the function to pivot all of the columns except the ID column, which we need to keep to be able to trace back which values came from which subject. The names_to = argument allows you to set what you want to name the column that stores the variable names (the column names in wide format). The values_to = argument allows you to set what you want to name the column that stores the values. We almost always call these columns “var” and “value”, respectively, but you can name them anything that makes sense for your dataset. full.pivotlong &lt;- full.data.tidy %&gt;% pivot_longer(-ID, names_to = &quot;var&quot;, values_to = &quot;value&quot;) head(full.pivotlong, 15) ## # A tibble: 15 × 3 ## ID var value ## &lt;int&gt; &lt;chr&gt; &lt;dbl&gt; ## 1 1 BMI 27.7 ## 2 1 MAge 23.0 ## 3 1 MEdu 3 ## 4 1 BW 3180. ## 5 1 GA 34 ## 6 1 DWAs 6.43 ## 7 1 DWCd 1.29 ## 8 1 DWCr 51.7 ## 9 1 UAs 10.2 ## 10 1 UCd 0.754 ## 11 1 UCr 42.6 ## 12 2 BMI 26.8 ## 13 2 MAge 30.1 ## 14 2 MEdu 3 ## 15 2 BW 3211. To pivot our data back to wide format, we can use pivot_wider(), which will pull the column names from the column specified in the names_from = argument and the corresponding values from the column specified in the values_from = argument. full.pivotwide &lt;- full.pivotlong %&gt;% pivot_wider(names_from = &quot;var&quot;, values_from = &quot;value&quot;) head(full.pivotwide) ## # A tibble: 6 × 12 ## ID BMI MAge MEdu BW GA DWAs DWCd DWCr UAs UCd UCr ## &lt;int&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; ## 1 1 27.7 23.0 3 3180. 34 6.43 1.29 51.7 10.2 0.754 42.6 ## 2 2 26.8 30.1 3 3211. 43 7.83 1.80 50.1 11.8 0.979 41.3 ## 3 3 33.2 28.0 3 3312. 40 7.52 1.29 48.7 10.1 0.190 36.5 ## 4 4 30.1 34.8 3 3267. 32 5.91 2.08 50.9 8.72 0.936 42.5 ## 5 5 37.4 42.7 3 3664. 35 7.18 2.76 55.2 9.44 1.50 47.8 ## 6 6 33.3 24.9 3 3329. 40 9.72 3.05 51.1 11.6 1.66 38.3 Now that we’re familiar with some tidyverse functions to reshape our data, let’s answer our original question: What is the average urinary Chromium concentration for each maternal education level? We can use the group_by() function to group our dataset by education class, then the summarize function to calculate the mean of our variable of interest within each class. Note how much shorter and more efficient this code is than the code we used to calculate the same values using base R! full.data %&gt;% group_by(MEdu) %&gt;% summarize(Avg_UCr = mean(UCr)) ## # A tibble: 3 × 2 ## MEdu Avg_UCr ## &lt;int&gt; &lt;dbl&gt; ## 1 1 39.9 ## 2 2 40.6 ## 3 3 40.4 For more detailed and advanced examples of pivoting in Tidyverse, see the Tidyverse Pivoting Vignette. Concluding Remarks This training module provides an introductory level overview of data organization and manipulation basics in base R and Tidyverse, including merging, filtering, subsetting, melting, and casting, and demonstrates these methods with an environmentally relevant dataset. These methods are used regularly in scripted analyses and are important preparation steps for almost all downstream analyses and visualizations. Test Your Knowledge What subjects, arranged from highest to lowest drinking water cadmium levels, had babies at at least 35 weeks and had urinary cadmium levels of at least 1.5 µg/L? Hint: Try using the arrange() function from the tidyverse package. "],["improving-coding-efficiencies.html", "Improving Coding Efficiencies Introduction to Training Module Loops Functions List operations Concluding Remarks Additional Resources", " Improving Coding Efficiencies This training module was developed by Elise Hickman, Alexis Payton, Kyle Roell, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module In this module, we’ll explore how to improve coding efficiency. Coding efficiency involves performing a task in as few lines as possible and can… Shorten code by eliminating redundancies Reduce the number of typos Help other coders understand script better Specific approaches that we will discuss in this module include loops, functions, and list operations, which can all be used to make code more succinct. A loop is employed when we want to perform a repetitive task, while a function contains a block of code organized together to perform one specific task. List operations, in which the same function is applied to a list of dataframes, can also be used to code more efficiently. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: Are there statistically significant differences in drinking water arsenic, cadmium, and chromium between normal weight (BMI &lt; 25) and overweight (BMI \\(\\geq\\) 25) subjects? Are there statistically significant differences in drinking water arsenic, cadmium, and chromium between underweight (BMI &lt; 18.5) and non-underweight (BMI \\(\\geq\\) 18.5) subjects? Are there statistically significant difference in drinking water arsenic, cadmium, and chromium between non-obese (BMI &lt; 29.9) and obese (BMI \\(\\geq\\) 29.9) subjects? We will demonstrate how this analysis can be approached using for loops, functions, or list operations. We will introduce the syntax and structure of each approach first, followed by application of the approach to our data. First, let’s prepare the workspace and familiarize ourselves with the dataset we are going to use. Data Import and Workspace Preparation Installing required packages If you already have these packages installed, you can skip this step, or you can run the below code which checks installation status for you. We will be using the tidyverse package for data manipulation steps and the rstatix package for statistical tests, as it provides pipe friendly adaptations of the base R statistical tests and returns results in a dataframe rather than a list format, making results easier to access. This brings up an important aspect of coding efficiency - sometimes, there is already a package that has been designed with functions to help you execute your desired analysis in an efficient way, so you don’t need to write custom functions yourself! So, don’t forget to explore packages relevant to your analysis before spending a lot of time developing custom solutions (although, sometimes this is necessary). if (!requireNamespace(&quot;tidyverse&quot;)) install.packages(&quot;tidyverse&quot;) if (!requireNamespace(&quot;rstatix&quot;)) install.packages(&quot;rstatix&quot;) Loading required packages library(tidyverse) library(rstatix) Setting your working directory setwd(&quot;/file path to where your input files are&quot;) Importing example dataset The first example dataset contains subject demographic data, and the second dataset contains corresponding chemical data. Familiarize yourself with these data used previously in TAME 2.0 Module 2.3 Data Manipulation and Reshaping. # Load the demographic data demographic_data &lt;- read.csv(&quot;Module2_4_Input/Module2_4_InputData1.csv&quot;) # View the top of the demographic dataset head(demographic_data) ## ID BMI MAge MEdu BW GA ## 1 1 27.7 22.99928 3 3180.058 34 ## 2 2 26.8 30.05142 3 3210.823 43 ## 3 3 33.2 28.04660 3 3311.551 40 ## 4 4 30.1 34.81796 3 3266.844 32 ## 5 5 37.4 42.68440 3 3664.088 35 ## 6 6 33.3 24.94960 3 3328.988 40 # Load the chemical data chemical_data &lt;- read.csv(&quot;Module2_4_Input/Module2_4_InputData2.csv&quot;) # View the top of the chemical dataset head(chemical_data) ## ID DWAs DWCd DWCr UAs UCd UCr ## 1 1 6.426464 1.292941 51.67987 10.192695 0.7537104 42.60187 ## 2 2 7.832384 1.798535 50.10409 11.815088 0.9789506 41.30757 ## 3 3 7.516569 1.288461 48.74001 10.079057 0.1903262 36.47716 ## 4 4 5.906656 2.075259 50.92745 8.719123 0.9364825 42.47987 ## 5 5 7.181873 2.762643 55.16882 9.436559 1.4977829 47.78528 ## 6 6 9.723429 3.054057 51.14812 11.589403 1.6645837 38.26386 Preparing the example dataset For ease of analysis, we will merge these two datasets before proceeding. # Merging data full_data &lt;- inner_join(demographic_data, chemical_data, by = &quot;ID&quot;) # Previewing new data head(full_data) ## ID BMI MAge MEdu BW GA DWAs DWCd DWCr UAs ## 1 1 27.7 22.99928 3 3180.058 34 6.426464 1.292941 51.67987 10.192695 ## 2 2 26.8 30.05142 3 3210.823 43 7.832384 1.798535 50.10409 11.815088 ## 3 3 33.2 28.04660 3 3311.551 40 7.516569 1.288461 48.74001 10.079057 ## 4 4 30.1 34.81796 3 3266.844 32 5.906656 2.075259 50.92745 8.719123 ## 5 5 37.4 42.68440 3 3664.088 35 7.181873 2.762643 55.16882 9.436559 ## 6 6 33.3 24.94960 3 3328.988 40 9.723429 3.054057 51.14812 11.589403 ## UCd UCr ## 1 0.7537104 42.60187 ## 2 0.9789506 41.30757 ## 3 0.1903262 36.47716 ## 4 0.9364825 42.47987 ## 5 1.4977829 47.78528 ## 6 1.6645837 38.26386 Continuous demographic variables, like BMI, are often dichotomized (or converted to a categorical variable with two categories representing higher vs. lower values) to increase statistical power in analyses. This is particularly important for clinical data that tend to have smaller sample sizes. In our initial dataframe, BMI is a continuous or numeric variable; however, our questions require us to dichotomize BMI. We can use the following code, which relies on if/else logic (see TAME 2.0 Module 2.3 Data Manipulation and Reshaping for more information) to generate a new column representing our dichotomized BMI variable for our first environmental health question. # Adding dichotomized BMI column full_data &lt;- full_data %&gt;% mutate(Dichotomized_BMI = ifelse(BMI &lt; 25, &quot;Normal&quot;, &quot;Overweight&quot;)) # Previewing new data head(full_data) ## ID BMI MAge MEdu BW GA DWAs DWCd DWCr UAs ## 1 1 27.7 22.99928 3 3180.058 34 6.426464 1.292941 51.67987 10.192695 ## 2 2 26.8 30.05142 3 3210.823 43 7.832384 1.798535 50.10409 11.815088 ## 3 3 33.2 28.04660 3 3311.551 40 7.516569 1.288461 48.74001 10.079057 ## 4 4 30.1 34.81796 3 3266.844 32 5.906656 2.075259 50.92745 8.719123 ## 5 5 37.4 42.68440 3 3664.088 35 7.181873 2.762643 55.16882 9.436559 ## 6 6 33.3 24.94960 3 3328.988 40 9.723429 3.054057 51.14812 11.589403 ## UCd UCr Dichotomized_BMI ## 1 0.7537104 42.60187 Overweight ## 2 0.9789506 41.30757 Overweight ## 3 0.1903262 36.47716 Overweight ## 4 0.9364825 42.47987 Overweight ## 5 1.4977829 47.78528 Overweight ## 6 1.6645837 38.26386 Overweight We can see that we now have created a new column entitled Dichotomized_BMI that we can use to perform a statistical test to assess if there are differences between drinking water metals between normal and overweight subjects. Loops We will start with loops. There are three main types of loops in R: for, while, and repeat. We will focus on for loops in this module, but for more in-depth information on loops, including the additional types of loops, see here. Before applying loops to our data, let’s discuss how for loops work. The basic structure of a for loop is shown here: # Basic structure of a for loop for (i in 1:4){ print(i) } ## [1] 1 ## [1] 2 ## [1] 3 ## [1] 4 for loops always start with for followed by a statement in parentheses. The argument in the parentheses tells R how to iterate (or repeat) through the code in the curly brackets. Here, we are telling R to iterate through the code in curly brackets 4 times. Each time we told R to print the value of our iterator, or i, which has a value of 1, 2, 3, and then 4. Loops can also iterate through columns in a dataset. For example, we can use a for loop to print the ages of each subject: # Creating a smaller dataframe for our loop example full_data_subset &lt;- full_data[1:6, ] # Finding the total number of rows or subjects in the dataset number_of_rows &lt;- length(full_data_subset$MAge) # Creating a for loop to iterate from 1 to the last row for (i in 1:number_of_rows){ # Printing each subject age # Need to put `[i]` to index the correct value corresponding to the row we are evaluating print(full_data_subset$MAge[i]) } ## [1] 22.99928 ## [1] 30.05142 ## [1] 28.0466 ## [1] 34.81796 ## [1] 42.6844 ## [1] 24.9496 Now that we know how a for loop works, how can we apply this approach to determine whether there are statistically significant differences in drinking water arsenic, cadmium, and chromium between normal weight (BMI &lt; 25) and overweight (BMI \\(\\geq\\) 25) subjects. Because our data are normally distributed and there are two groups that we are comparing, we will use a t-test applied to each metal measured in drinking water. Testing for assumptions is outside the scope of this module, but see TAME 2.0 Module 3.3 Normality Tests and Data Transformation for more information on this topic. Running a t-test in R is very simple, which we can demonstrate by running a t-test on the drinking water arsenic data: # Running t-test and storing results in t_test_res t_test_res &lt;- full_data %&gt;% t_test(DWAs ~ Dichotomized_BMI) # Viewing results t_test_res ## # A tibble: 1 × 8 ## .y. group1 group2 n1 n2 statistic df p ## * &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;int&gt; &lt;int&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; ## 1 DWAs Normal Overweight 96 104 -0.728 192. 0.468 We can see that our p-value is 0.468. Because this is greater than 0.05, we cannot reject the null hypothesis that normal weight and overweight subjects are exposed to the same drinking water arsenic concentrations. Although this was a very simple line of code to run, what if we have many columns we want to run the same t-test on? We can use a for loop to iterate through these columns. Let’s break down the steps of our for loop before executing the code. First, we will define the variables (columns) we want to run our t-test on. This is different from our approach above, because in those code chunks, we were using numbers to indicate the number of iterations through the loop. Here, we are naming the specific variables instead, and R will iterate though each of these variables. Note that we could omit this step and instead use the numeric column index of our variables of interest [7:9]. However, naming the specific columns makes this approach more robust because if additional data are added to or removed from our dataframe, the numeric column index of our variables could change. Which approach you choose really depends on the purpose of your loop! Second, we will create an empty dataframe where we will store the results generated by our for loop. Third, we will actually run our for loop. This will tell R: for each variable in our vars_of_interest vector, run a t-test with that variable (and store the results in a temporary dataframe called “res”), then add those results to our final results dataframe. A row will be added to the results dataframe each time R iterates through a new variable, resulting in a dataframe that stores the results of all of our t-tests. # Defining variables (columns) we want to run a t-test on vars_of_interest &lt;- c(&quot;DWAs&quot;, &quot;DWCd&quot;, &quot;DWCr&quot;) # Creating an empty dataframe to store results t_test_res_DW &lt;- data.frame() # Running for loop for (i in vars_of_interest) { # Storing the results of each iteration of the loop in a temporary results dataframe res &lt;- full_data %&gt;% # Writing the formula needed for each iteration of the loop t_test(as.formula(paste(i, &quot;~ Dichotomized_BMI&quot;, sep = &quot;&quot;))) # Adding a row to the results dataframe each time the loop is iterated t_test_res_DW &lt;- bind_rows(t_test_res_DW, res) } # Viewing our results t_test_res_DW ## .y. group1 group2 n1 n2 statistic df p ## 1 DWAs Normal Overweight 96 104 -0.7279621 192.3363 0.468 ## 2 DWCd Normal Overweight 96 104 -0.5894360 196.1147 0.556 ## 3 DWCr Normal Overweight 96 104 0.1102933 197.9870 0.912 With this, we can answer Environmental Health Question #1: Are there statistically significant differences in drinking water arsenic, cadmium, and chromium between normal weight (BMI &lt; 25) and overweight (BMI \\(\\geq\\) 25) subjects? Answer: No, there are not any statistically significant differences in drinking water metals between normal weight and overweight subjects. Formulas and Pasting Note the use of the code as.formula(paste0(i, \"~ Dichotomized_BMI\")). Let’s take a quick detour to discuss the use of the as.formula() and paste() functions, as these are important functions often used in loops and user-defined functions. Many statistical test functions and regression functions require one argument to be a formula, which is typically formatted as y ~ x, where y is the dependent variable of interest and x is an independent variable. For some functions, additional variables can be included on the right side of the formula to represent covariates (additional variables of interest). The function as.formula() returns the argument in parentheses in formula format so that it can be correctly passed to other functions. We can demonstrate that here by assigning a dummy variable j the character string var1: # Assigning variable j &lt;- &quot;var1&quot; # Demonstrating output of as.formula() as.formula(paste(j, &quot; ~ Dichotomized_BMI&quot;, sep = &quot;&quot;)) ## var1 ~ Dichotomized_BMI We can use the paste() function to combine strings of characters. The paste function takes each argument (as many arguments as is needed) and pastes them together into one character string, with the separator between arguments set by the sep = argument. When our y variable is changing with each iteration of our for loop, we can use the paste() function to write our formula correctly by telling the function to paste the variable i, followed by the rest of our formula, which stays the same for each iteration of the loop. Let’s examine the output of just the paste() part of our code: paste(j, &quot; ~ Dichotomized_BMI&quot;, sep = &quot;&quot;) ## [1] &quot;var1 ~ Dichotomized_BMI&quot; The paste() function is very flexible and can be useful in many other settings when you need to create one character string from arguments from different sources! Notice that the output looks different from the output of as.formula(). There is a returned index ([1]), and there are quotes around the character string. The last function we will highlight here is the noquote() function, which can be helpful if you’d like a string without quotes: noquote(paste(j, &quot; ~ Dichotomized_BMI&quot;, sep = &quot;&quot;)) ## [1] var1 ~ Dichotomized_BMI However, this still returns an indexed number, so there are times when it will not allow code to execute properly (for example, when we need a formula format). Next, we will learn about functions and apply them to our dataset to answer our additional environmental health questions. Functions Functions are useful when you want to execute a block of code organized together to perform one specific task, and you want to be able to change parameters for that task easily rather than having to copy and paste code over and over that largely stays the same but might have small modifications in certain arguments. The basic structure of a function is as follows: function_name &lt;- function(parameter_1, parameter_2...){ # Function body (where the code goes) insert_code_here # What the function returns return() } A function requires you to name it as we did with function_name. In parentheses, the function requires you to specify the arguments or parameters. Parameters (i.e., parameter_1) act as placeholders in the body of the function. This allows us to change the values of the parameters each time a function is called, while the majority of the code remains the same. Lastly, we have a return() statement, which specifies what object (i.e., vector, dataframe, etc.) we want to retrieve from a function. Although a function can display the last expression from the function body in the absence of a return() statement, it’s a good habit to include it as the last expression. It is important to note that, although functions can take many input parameters and execute large code chunks, they can only return one item, whether that is a value, vector, dataframe, plot, code output, or list. When writing your own functions, it is important to describe the purpose of the function, its input, its parameters, and its output so that others can understand what your functions does and how to use it. This can be defined either in text above a code chunk if you are using R Markdown or as comments within the code itself. We’ll start with a simple function. Let’s say we want to convert temperatures from Fahrenheit to Celsius. We can write a function that takes the temperature in Fahrenheit and converts it to Celsius. Note that we have given our parameters descriptive names (fahrenheit_temperature, celsius_temperature), which makes our code more readable than if we assigned them dummy names such as x and y. # Function to convert temperatures in Fahrenheit to Celsius ## Parameters: temperature in Fahrenheit (input) ## Output: temperature in Celsius fahrenheit_to_celsius &lt;- function(fahrenheit_temperature){ celsius_temperature &lt;- (fahrenheit_temperature - 32) * (5/9) return(celsius_temperature) } Notice that the above code block was run, but there isn’t an output. Rather, running the code assigns the function code to that function. When you run code defining a function, that function will appear in your Global Environment under the “Functions” section. We can see the output of the function by providing an input value. Let’s start by converting 41 degrees Fahrenheit to Celsius: # Calling the function # Here, 41 is the `fahrenheit_temperature` in the function fahrenheit_to_celsius(41) ## [1] 5 41 degrees Fahrenheit is equivalent to 5 degrees Celsius. We can also have the function convert a vector of values. # Defining vector of temperatures vector_of_temperatures &lt;- c(81,74,23,65) # Calling the function fahrenheit_to_celsius(vector_of_temperatures) ## [1] 27.22222 23.33333 -5.00000 18.33333 Before getting back to answer our environmental health related questions, let’s look at one more example of a function. This time we’ll create a function that can calculate the circumference of a circle based on its radius in inches. Here you can also see a different style of commenting to describe the function’s purpose, inputs, and outputs. circle_circumference &lt;- function(radius){ # Calculating a circle&#39;s circumference based on the radius inches # :parameters: radius # :output: circumference and radius # Calculating diameter first diameter &lt;- 2 * radius # Calculating circumference circumference &lt;- pi * diameter return(circumference) } # Calling function circle_circumference(3) ## [1] 18.84956 So, if a circle had a radius of 3 inches, its circumference would be ~19 inches. What if we were interested in seeing the diameter to double check our code? diameter ## Error in eval(expr, envir, enclos): object &#39;diameter&#39; not found R throws an error, because the variable diameter was created inside the function and the function only returned the circumference variable. This is actually one of the ways that functions can improve coding efficiency - by not needing to store intermediate variables that aren’t of interest to the main goal of the code or analysis. However, there are two ways we can still see the diameter variable: Put print statements in the body of the function (print(diameter)). Have the function return a different variable or list of variables (c(circumference, diameter)). See the below section on List Operation for more on this topic. We can now move on to using a more complicated function to answer all three of our environmental health questions without repeating our earlier code three times. The main difference between each of our first three environmental health questions is the BMI cutoff used to dichotomize the BMI variable, so we can use that as one of the parameters for our function. We can also use arguments in our function to name our groups. We can adapt our previous for loop code into a function that will take different BMI cutoffs and return statistical results by including parameters to define the parts of the analysis that will change with each unique question. For example: Changing the BMI cutoff from a number (in our previous code) to our parameter name that specifies the cutoff Changing the group names for assigning category (in our previous code) to our parameter names # Function to dichotomize BMI into different categories and return results of t-test on drinking water metals between dichotomized groups ## Parameters: ### input_data: dataframe containing BMI and drinking water metals levels ### bmi_cutoff: numeric value specifying the cut point for dichotomizing BMI ### lower_group_name: name for the group of subjects with BMIs lower than the cutoff ### upper_group_name: name for the group of subjects with BMIs higher than the cutoff ### variables: vector of variable names that statistical test should be run on ## Output: dataframe with statistical results for each variable in the variables vector bmi_DW_ttest &lt;- function(input_data, bmi_cutoff, lower_group_name, upper_group_name, variables){ # Creating dichotomized variable dichotomized_data &lt;- input_data %&gt;% mutate(Dichotomized_BMI = ifelse(BMI &lt; bmi_cutoff, lower_group_name, upper_group_name)) # Creating an empty dataframe to store results t_test_res_DW &lt;- data.frame() # Running for loop for (i in variables) { # Storing the results of each iteration of the loop in a temporary results dataframe res &lt;- dichotomized_data %&gt;% # Writing the formula needed for each iteration of the loop t_test(as.formula(paste(i, &quot;~ Dichotomized_BMI&quot;, sep = &quot;&quot;))) # Adding a row to the results dataframe each time the loop is iterated t_test_res_DW &lt;- bind_rows(t_test_res_DW, res) } # Return results return(t_test_res_DW) } For the first example of using the function, we have included the name of each argument for clarity, but this isn’t necessary if you pass in the arguments in the order they were defined when writing the function. # Defining variables (columns) we want to run a t-test on vars_of_interest &lt;- c(&quot;DWAs&quot;, &quot;DWCd&quot;, &quot;DWCr&quot;) # Apply function for normal vs. overweight (bmi_cutoff = 25) bmi_DW_ttest(input_data = full_data, bmi_cutoff = 25, lower_group_name = &quot;Normal&quot;, upper_group_name = &quot;Overweight&quot;, variables = vars_of_interest) ## .y. group1 group2 n1 n2 statistic df p ## 1 DWAs Normal Overweight 96 104 -0.7279621 192.3363 0.468 ## 2 DWCd Normal Overweight 96 104 -0.5894360 196.1147 0.556 ## 3 DWCr Normal Overweight 96 104 0.1102933 197.9870 0.912 Here, we can see the same results as above in the Loops section. We can next apply the function to answer our additional environmental health questions: # Apply function for underweight vs. non-underweight (bmi_cutoff = 18.5) bmi_DW_ttest(full_data, 18.5, &quot;Underweight&quot;, &quot;Non-Underweight&quot;, vars_of_interest) ## .y. group1 group2 n1 n2 statistic df p ## 1 DWAs Non-Underweight Underweight 166 34 -0.86947835 53.57143 0.388 ## 2 DWCd Non-Underweight Underweight 166 34 -0.97359810 55.45450 0.334 ## 3 DWCr Non-Underweight Underweight 166 34 0.04305105 56.08814 0.966 # Apply function for non-obese vs. obese (bmi_cutoff = 29.9) bmi_DW_ttest(full_data, 29.9, &quot;Non-Obese&quot;, &quot;Obese&quot;, vars_of_interest) ## .y. group1 group2 n1 n2 statistic df p ## 1 DWAs Non-Obese Obese 144 56 -1.9312097 86.80253 0.0567 ## 2 DWCd Non-Obese Obese 144 56 0.3431076 94.52209 0.7320 ## 3 DWCr Non-Obese Obese 144 56 -0.6878311 89.61818 0.4930 With this, we can answer Environmental Health Questions #2 &amp; #3: Are there statistically significant differences in drinking water arsenic, cadmium, and chromium between underweight (BMI &lt; 18.5) and non-underweight (BMI \\(\\geq\\) 18.5) subjects or between non-obese (BMI &lt; 29.9) and obese (BMI \\(\\geq\\) 29.9) subjects? Answer: No, there are not any statistically significant differences in drinking water metals between underweight and non-underweight subjects or between non-obese and obese subjects. Here, we were able to answer all three of our environmental health questions within relatively few lines of code by using a function to efficiently assess different variations on our analysis. In the last section of this module, we will demonstrate how to use list operations to improve coding efficiency. List operations Lists are a data type in R that can store other data types (including lists, to make nested lists). This allows you to store multiple dataframes in one object and apply the same functions to each dataframe in the list. Lists can also be helpful for storing the results of a function if you would like to be able to access multiple outputs. For example, if we return to our example of a function that calculates the circumference of a circle, we can store both the diameter and circumference as list objects. The function will then return a list containing both of these values when called. # Adding list element to our function circle_circumference_4 &lt;- function(radius){ # Calculating a circle&#39;s circumference and diameter based on the radius in inches # :parameters: radius # :output: list that contains diameter [1] and circumference [2] # Calculating diameter first diameter &lt;- 2 * radius # Calculating circumference circumference &lt;- pi * diameter # Storing results in a named list results &lt;- list(&quot;diameter&quot; = diameter, &quot;circumference&quot; = circumference) # Return results results } # Calling function circle_circumference_4(10) ## $diameter ## [1] 20 ## ## $circumference ## [1] 62.83185 We can also call the results individually using the following code: # Storing results of function circle_10 &lt;- circle_circumference_4(10) # Viewing only diameter ## Method 1 circle_10$diameter ## [1] 20 ## Method 2 circle_10[1] ## $diameter ## [1] 20 # Viewing only circumference ## Method 1 circle_10$circumference ## [1] 62.83185 ## Method 2 circle_10[2] ## $circumference ## [1] 62.83185 In the context of our dataset, we can use list operations to clean up and combine our results from all three BMI stratification approaches. This is often necessary to prepare data to share with collaborators or for supplementary tables in a manuscript. Let’s revisit our code for producing our statistical results, this time assigning our results to a dataframe rather than viewing them. # Defining variables (columns) we want to run a t-test on vars_of_interest &lt;- c(&quot;DWAs&quot;, &quot;DWCd&quot;, &quot;DWCr&quot;) # Normal vs. overweight (bmi_cutoff = 25) norm_vs_overweight &lt;- bmi_DW_ttest(input_data = full_data, bmi_cutoff = 25, lower_group_name = &quot;Normal&quot;, upper_group_name = &quot;Overweight&quot;, variables = vars_of_interest) # Underweight vs. non-underweight (bmi_cutoff = 18.5) under_vs_nonunderweight &lt;- bmi_DW_ttest(full_data, 18.5, &quot;Underweight&quot;, &quot;Non-Underweight&quot;, vars_of_interest) # Non-obese vs. obese (bmi_cutoff = 29.9) nonobese_vs_obese &lt;- bmi_DW_ttest(full_data, 29.9, &quot;Non-Obese&quot;, &quot;Obese&quot;, vars_of_interest) # Viewing one results dataframe as an example norm_vs_overweight ## .y. group1 group2 n1 n2 statistic df p ## 1 DWAs Normal Overweight 96 104 -0.7279621 192.3363 0.468 ## 2 DWCd Normal Overweight 96 104 -0.5894360 196.1147 0.556 ## 3 DWCr Normal Overweight 96 104 0.1102933 197.9870 0.912 For publication purposes, let’s say we want to make the following formatting changes: Keep only the comparison of interest (for example Normal vs. Overweight) and the associated p-value, removing columns that are not as useful for interpreting or sharing the results Rename the .y. column so that its contents are clearer Collapse all of our data into one final dataframe We can first write a function to execute these cleaning steps: # Function to clean results dataframes ## Parameters: ### input_data: dataframe containing results of t-test ## Output: cleaned dataframe data_cleaning &lt;- function(input_data) { data &lt;- input_data %&gt;% # Rename .y. column rename(&quot;Variable&quot; = &quot;.y.&quot;) %&gt;% # Merge group1 and group2 unite(Comparison, group1, group2, sep = &quot; vs. &quot;) %&gt;% # Keep only columns of interest select(c(Variable, Comparison, p)) return(data) } Then, we can make a list of our dataframes to clean and apply: # Making list of dataframes t_test_res_list &lt;- list(norm_vs_overweight, under_vs_nonunderweight, nonobese_vs_obese) # Viewing list of dataframes head(t_test_res_list) ## [[1]] ## .y. group1 group2 n1 n2 statistic df p ## 1 DWAs Normal Overweight 96 104 -0.7279621 192.3363 0.468 ## 2 DWCd Normal Overweight 96 104 -0.5894360 196.1147 0.556 ## 3 DWCr Normal Overweight 96 104 0.1102933 197.9870 0.912 ## ## [[2]] ## .y. group1 group2 n1 n2 statistic df p ## 1 DWAs Non-Underweight Underweight 166 34 -0.86947835 53.57143 0.388 ## 2 DWCd Non-Underweight Underweight 166 34 -0.97359810 55.45450 0.334 ## 3 DWCr Non-Underweight Underweight 166 34 0.04305105 56.08814 0.966 ## ## [[3]] ## .y. group1 group2 n1 n2 statistic df p ## 1 DWAs Non-Obese Obese 144 56 -1.9312097 86.80253 0.0567 ## 2 DWCd Non-Obese Obese 144 56 0.3431076 94.52209 0.7320 ## 3 DWCr Non-Obese Obese 144 56 -0.6878311 89.61818 0.4930 And we can apply the cleaning function to each of the dataframes using the lapply() function, which takes a list as the first argument and the function to apply to each list element as the second argument: # Applying cleaning function t_test_res_list_cleaned &lt;- lapply(t_test_res_list, data_cleaning) # Vieweing cleaned dataframes head(t_test_res_list_cleaned) ## [[1]] ## Variable Comparison p ## 1 DWAs Normal vs. Overweight 0.468 ## 2 DWCd Normal vs. Overweight 0.556 ## 3 DWCr Normal vs. Overweight 0.912 ## ## [[2]] ## Variable Comparison p ## 1 DWAs Non-Underweight vs. Underweight 0.388 ## 2 DWCd Non-Underweight vs. Underweight 0.334 ## 3 DWCr Non-Underweight vs. Underweight 0.966 ## ## [[3]] ## Variable Comparison p ## 1 DWAs Non-Obese vs. Obese 0.0567 ## 2 DWCd Non-Obese vs. Obese 0.7320 ## 3 DWCr Non-Obese vs. Obese 0.4930 Last, we can collapse our list down into one dataframe using the do.call() and rbind.data.frame() functions, which together, take the elements of the list and collapse them into a dataframe by binding the rows together: t_test_res_cleaned &lt;- do.call(rbind.data.frame, t_test_res_list_cleaned) # Viewing final dataframe t_test_res_cleaned ## Variable Comparison p ## 1 DWAs Normal vs. Overweight 0.4680 ## 2 DWCd Normal vs. Overweight 0.5560 ## 3 DWCr Normal vs. Overweight 0.9120 ## 4 DWAs Non-Underweight vs. Underweight 0.3880 ## 5 DWCd Non-Underweight vs. Underweight 0.3340 ## 6 DWCr Non-Underweight vs. Underweight 0.9660 ## 7 DWAs Non-Obese vs. Obese 0.0567 ## 8 DWCd Non-Obese vs. Obese 0.7320 ## 9 DWCr Non-Obese vs. Obese 0.4930 The above example is just that - an example to demonstrate the mechanics of using list operations. However, there are actually a couple of even more efficient ways to execute the above cleaning steps: Build cleaning steps into the analysis function if you know you will not need to access the raw results dataframe. Bind all three dataframes together, then execute the cleaning steps. We will demonstrate #2 below: # Start by binding the rows of each of the results dataframes t_test_res_cleaned_2 &lt;- bind_rows(norm_vs_overweight, under_vs_nonunderweight, nonobese_vs_obese) %&gt;% # Rename .y. column rename(&quot;Variable&quot; = &quot;.y.&quot;) %&gt;% # Merge group1 and group2 unite(Comparison, group1, group2, sep = &quot; vs. &quot;) %&gt;% # Keep only columns of interest select(c(Variable, Comparison, p)) # Viewing results t_test_res_cleaned_2 ## Variable Comparison p ## 1 DWAs Normal vs. Overweight 0.4680 ## 2 DWCd Normal vs. Overweight 0.5560 ## 3 DWCr Normal vs. Overweight 0.9120 ## 4 DWAs Non-Underweight vs. Underweight 0.3880 ## 5 DWCd Non-Underweight vs. Underweight 0.3340 ## 6 DWCr Non-Underweight vs. Underweight 0.9660 ## 7 DWAs Non-Obese vs. Obese 0.0567 ## 8 DWCd Non-Obese vs. Obese 0.7320 ## 9 DWCr Non-Obese vs. Obese 0.4930 As you can see, this dataframe is the same as the one we produced using list operations. It was produced using fewer lines of code and without the need for a user-defined function! For our purposes, this was a more efficient approach. However, we felt it was important to demonstrate the mechanics of list operations because there may be times where you do need to keep dataframes separate during specific analyses. Concluding Remarks This module provided an introduction to loops, functions, and list operations and demonstrated how to use them to efficiently analyze an environmentally relevant dataset. When and how you implement these approaches depends on your coding style and the goals of your analysis. Although here we were focused on statistical tests and data cleaning, these flexible approaches can be used in a variety of data analysis steps. We encourage you to implement loops, functions, and list operations in your analyses when you find the need to iterate through statistical tests, visualizations, data cleaning, or other common workflow elements! Additional Resources Intro2r Loops Intro2r Functions in R Hadley Wickham Advanced R - Functionals Test Your Knowledge Use the same input data we used in this module to answer the following questions and produce a cleaned, publication-ready data table of results. Note that these data are normally distributed, so you can use a t-test. Are there statistically significant differences in urine metal concentrations (ie. arsenic levels, cadmium levels, etc.) between younger (MAge &lt; 40) and older (MAge \\(\\geq\\) 40) mothers? Are there statistically significant differences in urine metal concentrations (ie. arsenic levels, cadmium levels, etc.) between between normal weight (BMI &lt; 25) and overweight (BMI \\(\\geq\\) 25) subjects? "],["data-visualizations.html", "Data Visualizations Introduction to Data Visualizations Introduction to Training Module Density Plot Visualization Boxplot Visualization Correlation Visualizations Heatmap Visualization Concluding Remarks", " Data Visualizations This training module was developed by Alexis Payton, Kyle Roell, Lauren E. Koval, Elise Hickman, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Data Visualizations Selecting an approach to visualize data is an important consideration when presenting scientific research, given that figures have the capability to summarize large amounts of data efficiently and effectively. (At least that’s the goal!) This module will focus on basic data visualizations that we view to be most commonly used, both in and outside of the field of environmental health research, many of which you have likely seen before. This module is not meant to be an exhaustive representation of all figure types, rather it serves as an introduction to some types of figures and how to approach choosing the one that most optimally displays your data and primary findings. When selecting a data visualization approach, here are some helpful questions you should first ask yourself: What message am I trying to convey with this figure? How does this figure highlight major findings from the paper? Who is the audience? What type of data am I working with? A Guide To Getting Data Visualization Right is a great resource for determining which figure is best suited for various types of data. More complex methodology-specific charts are presented in succeeding TAME modules. These include visualizations for: Two Group Comparisons (e.g.,boxplots and logistic regression) in Module 3.4 Introduction to Statistical Tests and Module 4.4 Two Group Comparisons and Visualizations Multi-Group Comparisons (e.g.,boxplots) in Module 3.4 Introduction to Statistical Tests and Module 4.5 Mult-Group Comparisons and Visualizations Supervised Machine Learning (e.g.,decision boundary plots, variable importance plots) in Module 5.3 Supervised ML Model Interpretation Unsupervised Machine Learning Principal Component Analysis (PCA) plots and heatmaps in Module 5.4 Unsupervised Machine Learning I: K-Means Clustering &amp; PCA Dendrograms, clustering visualizations, heatmaps, and variable contribution plots in Module 5.5 Unsupervised Machine Learning II: Additional Clustering Applications -Omics Expression (e.g.,MA plots and volcano plots) in Module 6.2 -Omics and Systems Biology: Transcriptomic Applications Mixtures Methods Forest Plots in Module 6.3 Mixtures I: Overview and Quantile G-Computation Application Trace Plots in Module 6.4 Mixtures II: BKMR Application Sufficient Similarity (e.g.,heatmaps, clustering) in Module 6.5 Mixtures III: Sufficient Similarity Toxicokinetic Modeling (e.g.,line graph, dose response) in Module 6.6 Toxicokinetic Modeling Introduction to Training Module Visualizing data is an important step in any data analysis, including those carried out in environmental health research. Often, visualizations allow scientists to better understand trends and patterns within a particular dataset under evaluation. Even after statistical analysis of a dataset, it is important to then communicate these findings to a wide variety of target audiences. Visualizations are a vital part of communicating complex data and results to target audiences. In this module, we highlight some figures that can be used to visualize larger, more high-dimensional datasets using figures that are more simple (but still relevant!) than methods presented later on in TAME. This training module specifically reviews the formatting of data in preparation of generating visualizations, scaling datasets, and then guides users through the generation of the following example data visualizations: Density plots Boxplots Correlation plots Heatmaps These visualization approaches are demonstrated using a large environmental chemistry dataset. This example dataset was generated through chemical speciation analysis of smoke samples collected during lab-based simulations of wildfire events. Specifically, different biomass materials (eucalyptus, peat, pine, pine needles, and red oak) were burned under two combustion conditions of flaming and smoldering, resulting in the generation of 12 different smoke samples. These data have been previously published in the following environmental health research studies, with data made publicly available: Rager JE, Clark J, Eaves LA, Avula V, Niehoff NM, Kim YH, Jaspers I, Gilmour MI. Mixtures modeling identifies chemical inducers versus repressors of toxicity associated with wildfire smoke. Sci Total Environ. 2021 Jun 25;775:145759. doi: 10.1016/j.scitotenv.2021.145759. Epub 2021 Feb 10. PMID: 33611182. Kim YH, Warren SH, Krantz QT, King C, Jaskot R, Preston WT, George BJ, Hays MD, Landis MS, Higuchi M, DeMarini DM, Gilmour MI. Mutagenicity and Lung Toxicity of Smoldering vs. Flaming Emissions from Various Biomass Fuels: Implications for Health Effects from Wildland Fires. Environ Health Perspect. 2018 Jan 24;126(1):017011. doi: 10.1289/EHP2200. PMID: 29373863. GGplot2 ggplot2 is a powerful package used to create graphics in R. It was designed based on the philosophy that every figure can be built using a dataset, a coordinate system, and a geom that specifies the type of plot. As a result, it is fairly straightforward to create highly customizable figures and is typically preferred over using base R to generate graphics. We’ll generate all of the figures in this module using ggplot2. For additional resources on ggplot2 see ggplot2 Posit Documentation and Data Visualization with ggplot2. Script Preparations Cleaning the global environment rm(list=ls()) Installing required R packages If you already have these packages installed, you can skip this step, or you can run the below code which checks installation status for you if (!requireNamespace(&quot;GGally&quot;)) install.packages(&quot;GGally&quot;); if (!requireNamespace(&quot;corrplot&quot;)) install.packages(&quot;corrplot&quot;); if (!requireNamespace(&quot;pheatmap&quot;)) install.packages(&quot;pheatmap&quot;); Loading R packages required for this session library(tidyverse) library(GGally) library(corrplot) library(reshape2) library(pheatmap) Set your working directory setwd(&quot;/filepath to where your input files are&quot;) Importing example dataset Then let’s read in our example dataset. As mentioned in the introduction, this example dataset represents chemical measurements across 12 different biomass burn scenarios representing potential wildfire events. Let’s upload and view these data: # Load the data smoke_data &lt;- read.csv(&quot;Module3_1_Input/Module3_1_InputData.csv&quot;) # View the top of the dataset head(smoke_data) ## Chemical.Category Chemical CASRN Eucalyptus_Smoldering ## 1 n-Alkanes 2-Methylnonadecane 1560-86-7 0.06 ## 2 n-Alkanes 3-Methylnonadecane 6418-45-7 0.04 ## 3 n-Alkanes Docosane 629-97-0 0.21 ## 4 n-Alkanes Dodecylcyclohexane 1795-17-1 0.04 ## 5 n-Alkanes Eicosane 112-95-8 0.11 ## 6 n-Alkanes Heneicosane 629-94-7 0.13 ## Eucalyptus_Flaming Peat_Smoldering Peat_Flaming Pine_Smoldering Pine_Flaming ## 1 0.06 1.36 0.06 0.06 0.06 ## 2 0.04 1.13 0.90 0.47 0.04 ## 3 0.25 9.46 0.57 0.16 0.48 ## 4 0.04 0.25 0.04 0.04 0.04 ## 5 0.25 7.55 0.54 0.17 0.29 ## 6 0.28 6.77 0.34 0.13 0.42 ## Pine_Needles_Smoldering Pine_Needles_Flaming Red_Oak_Smoldering ## 1 0.06 0.06 0.06 ## 2 0.04 0.72 0.04 ## 3 0.32 0.18 0.16 ## 4 0.12 0.04 0.04 ## 5 0.28 0.16 0.15 ## 6 0.30 0.13 0.13 ## Red_Oak_Flaming Units ## 1 0.13 ng_per_uL ## 2 0.77 ng_per_uL ## 3 0.36 ng_per_uL ## 4 0.04 ng_per_uL ## 5 0.38 ng_per_uL ## 6 0.69 ng_per_uL Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: How do the distributions of the chemical concentration data differ based on each biomass burn scenario? Are there correlations between biomass burn conditions based on the chemical concentration data? Under which biomass burn conditions are concentrations of certain chemical categories the highest? We can create a density plot to answer the first question. Similar to a histogram, density plots are an effective way to show overall distributions of data and can be useful to compare across various test conditions or other stratifications of the data. In this example of a density plot, we’ll visualize the distributions of chemical concentration data on the x axis. A density plot automatically displays where values are concentrated on the y axis. Additionally, we’ll want to have multiple density plots within the same figure for each biomass burn condition. Before the data can be visualized, it needs to be converted from a wide to long format. This is because we need to have variable or column names entitled Chemical_Concentration and Biomass_Burn_Condition that can be placed into ggplot(). For review on converting between long and wide formats and using other tidyverse tools, see TAME 2.0 Module 2.3 Data Manipulation &amp; Reshaping. longer_smoke_data = pivot_longer(smoke_data, cols = 4:13, names_to = &quot;Biomass_Burn_Condition&quot;, values_to = &quot;Chemical_Concentration&quot;) head(longer_smoke_data) ## # A tibble: 6 × 6 ## Chemical.Category Chemical CASRN Units Biomass_Burn_Condition ## &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; ## 1 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus_Smoldering ## 2 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus_Flaming ## 3 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat_Smoldering ## 4 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat_Flaming ## 5 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine_Smoldering ## 6 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine_Flaming ## # ℹ 1 more variable: Chemical_Concentration &lt;dbl&gt; Scaling dataframes for downstream data visualizations A data preparation method that is commonly used to convert values into those that can be used to better illustrate overall data trends is data scaling. Scaling can be achieved through data transformations or normalization procedures, depending on the specific dataset and goal of analysis/visualization. Scaling is often carried out using data vectors or columns of a dataframe. For this example, we will normalize the chemical concentration dataset using a basic scaling and centering procedure using the base R function, scale(). This algorithm results in the normalization of a dataset using the mean value and standard deviation. This scaling step will convert chemical concentration values in our dataset into normalized values across samples, such that each chemical’s concentration distributions are more easily comparable between the different biomass burn conditions. For more information on the scale() function, see its associated RDocumentation and helpful tutorial on Implementing the scale() function in R. scaled_longer_smoke_data = longer_smoke_data %&gt;% # scaling within each chemical group_by(Chemical) %&gt;% mutate(Scaled_Chemical_Concentration = scale(Chemical_Concentration)) %&gt;% ungroup() head(scaled_longer_smoke_data) # see the new scaled values now in the last column (column 7) ## # A tibble: 6 × 7 ## Chemical.Category Chemical CASRN Units Biomass_Burn_Condition ## &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; ## 1 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus_Smoldering ## 2 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus_Flaming ## 3 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat_Smoldering ## 4 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat_Flaming ## 5 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine_Smoldering ## 6 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine_Flaming ## # ℹ 2 more variables: Chemical_Concentration &lt;dbl&gt;, ## # Scaled_Chemical_Concentration &lt;dbl[,1]&gt; We can see that in the Scaled_Chemical_Concentration column, each chemical is scaled based on a normal distribution centered around 0, with values now less than or greater than zero. Now that we have our dataset formatted, let’s plot it. Density Plot Visualization The following code can be used to generate a density plot: ggplot(scaled_longer_smoke_data, aes(x = Scaled_Chemical_Concentration, color = Biomass_Burn_Condition)) + geom_density() Answer to Environmental Health Question 1, Method I With this method, we can answer Environmental Health Question #1: How do the distributions of the chemical concentration data differ based on each biomass burn scenario? Answer: In general, there are a high number of chemicals that were measured at relatively lower abundances across all smoke samples (hence, the peak in occurrence density occurring towards the left, before 0). The three conditions of smoldering peat, flaming peat, and flaming pine contained the most chemicals at the highest relative concentrations (hence, these lines are the top three lines towards the right). Boxplot Visualization A boxplot can also be used to answer our first environmental health question: How do the distributions of the chemical concentration data differ based on each biomass burn scenario?. A boxplot also displays a data’s distribution, but it incorporates a visualization of a five number summary (i.e., minimum, first quartile, median, third quartile, and maximum). Any outliers are displayed as dots. For this example, let’s have Scaled_Chemical_Concentration on the x axis and Biomass_Burn_Condition on the y axis. The scaled_longer_smoke_data dataframe is the format we need, so we’ll use that for plotting. ggplot(scaled_longer_smoke_data, aes(x = Scaled_Chemical_Concentration, y = Biomass_Burn_Condition, color = Biomass_Burn_Condition)) + geom_boxplot() Answer to Environmental Health Question 1, Method II With this alternative method, we can answer, in a different way, Environmental Health Question #1: How do the distributions of the chemical concentration data differ based on each biomass burn scenario? Answer, Method II: The median chemical concentration is fairly low (less than 0) for all biomass burn conditions. Overall, there isn’t much variation in chemical concentrations with the exception of smoldering peat, flaming peat, and flaming eucalyptus. Correlation Visualizations Let’s turn our attention to the second environmental health question: Are there correlations between biomass burn conditions based on the chemical concentration data? We’ll use two different correlation visualizations to answer this question using the GGally package. GGally is a package that serves as an extension of ggplot2, the baseline R plotting system based on the grammar of graphics. GGally is very useful for creating plots that compare groups or features within a dataset, among many other utilities. Here we will demonstrate the ggpairs() function within GGally using the scaled chemistry dataset. This function will produce an image that shows correlation values between biomass burn sample pairs and also illustrates the overall distributions of values in the samples. For more information on GGally, see its associated RDocumentation and example helpful tutorial. GGally requires a wide dataframe with ids (i.e.,Chemical) as the rows and the variables that will be compared to each other (i.e.,Biomass_Burn_Condition) as the columns. Let’s create that dataframe. # first selecting the chemical, biomass burn condition, and # the scaled chemical concentration columns wide_scaled_data = scaled_longer_smoke_data %&gt;% pivot_wider(id_cols = Chemical, names_from = &quot;Biomass_Burn_Condition&quot;, values_from = &quot;Scaled_Chemical_Concentration&quot;) %&gt;% # converting the chemical names to row names column_to_rownames(var = &quot;Chemical&quot;) head(wide_scaled_data) ## Eucalyptus_Smoldering Eucalyptus_Flaming Peat_Smoldering ## 2-Methylnonadecane -0.3347765 -0.3347765 2.841935 ## 3-Methylnonadecane -0.8794448 -0.8794448 1.649829 ## Docosane -0.3465132 -0.3327216 2.842787 ## Dodecylcyclohexane -0.4240624 -0.4240624 2.646734 ## Eicosane -0.3802202 -0.3195928 2.841691 ## Heneicosane -0.3895328 -0.3166775 2.835527 ## Peat_Flaming Pine_Smoldering Pine_Flaming ## 2-Methylnonadecane -0.3347765 -0.3347765 -0.3347765 ## 3-Methylnonadecane 1.1161291 0.1183422 -0.8794448 ## Docosane -0.2223890 -0.3637526 -0.2534201 ## Dodecylcyclohexane -0.4240624 -0.4240624 -0.4240624 ## Eicosane -0.1940076 -0.3542370 -0.3022707 ## Heneicosane -0.2875354 -0.3895328 -0.2486793 ## Pine_Needles_Smoldering Pine_Needles_Flaming ## 2-Methylnonadecane -0.3347765 -0.3347765 ## 3-Methylnonadecane -0.8794448 0.6984509 ## Docosane -0.3085863 -0.3568568 ## Dodecylcyclohexane 0.7457649 -0.4240624 ## Eicosane -0.3066012 -0.3585675 ## Heneicosane -0.3069635 -0.3895328 ## Red_Oak_Smoldering Red_Oak_Flaming ## 2-Methylnonadecane -0.3347765 -0.1637228 ## 3-Methylnonadecane -0.8794448 0.8144726 ## Docosane -0.3637526 -0.2947948 ## Dodecylcyclohexane -0.4240624 -0.4240624 ## Eicosane -0.3628981 -0.2632960 ## Heneicosane -0.3895328 -0.1175398 By default, ggpairs() displays Pearson’s correlations. To show Spearman’s correlations takes more nuance, but can be done using the code that has been commented out below. # ggpairs with Pearson&#39;s correlations wide_scaled_data = data.frame(as.matrix(wide_scaled_data)) ggpairs(wide_scaled_data) # ggpairs with Spearman&#39;s correlations # pearson_correlations = cor(wide_scaled_data, method = &quot;spearman&quot;) # ggpairs(wide_scaled_data, upper = list(continuous = wrap(ggally_cor, method = &quot;spearman&quot;))) Many of these biomass burn conditions have significant correlations denoted by the asterisks. ’*’: p value &lt; 0.1 ’**’: p value &lt; 0.05 ’***’: p value &lt; 0.01 The upper right portion displays the correlation values, where a value less than 0 indicates negative correlation and a value greater than 0 signifies positive correlation. The diagonal shows the density plots for each variable. The lower left portion visualizes the values of the two variables compared using a scatterplot. Answer to Environmental Health Question 2 With this, we can answer Environmental Health Question #2: Are there correlations between biomass burn conditions based on the chemical concentration data? Answer: There is low correlation between many of the variables (-0.5 &lt; correlation value &lt; 0.5). Eucalyptus flaming and pine flaming are significantly positively correlated along with peat flaming and pine needles flaming (correlation value ~0.7 and p value &lt; 0.001). We can visualize correlations another way using the other function from GGally, ggcorr(), which visualizes each correlation as a square. Note that this function calculates Pearson’s correlations by default. However, this can be changed using the method parameter shown in the code commented out below. # Pearson&#39;s correlations ggcorr(wide_scaled_data) # Spearman&#39;s correlations # ggcorr(wide_scaled_data, method = &quot;spearman&quot;) We’ll visualize correlations between each of the groups using one more figure using the corrplot() function from the corrplot package. # Need to supply corrplot with a correlation matrix, here, using the &#39;cor&#39; function corrplot(cor(wide_scaled_data)) Each of these correlation figures displays the same information, but the one you choose to use is a matter of personal preference. Click on the following resources for additional information on ggpairs() and corrplot(). Heatmap Visualization Last, we’ll turn our attention to answering the final environmental health question: Under which biomass burn conditions are concentrations of certain chemical categories the highest? This can be addressed with the help of a heatmap. Heatmaps are a highly effective method of viewing an entire dataset at once. Heatmaps can appear similar to correlation plots, but typically illustrate other values (e.g., concentrations, expression levels, presence/absence, etc) besides correlation values. They are used to draw patterns between two variables of highest interest (that comprise the x and y axis, though additional bars can be added to display other layers of information). In this instance, we’ll use a heatmap to determine whether there are patterns apparent between chemical categories and biomass burn condition on chemical concentrations. For this example, we can plot Biomass_Burn_Condition and Chemical.Category on the axes and fill in the values with Scaled_Chemical_Concentration. When generating heatmaps, scaled values are often used to better distinguish patterns between groups/samples. In this example, we also plan to display the median scaled concentration value within the heatmap as an additional layer of helpful information to aid in interpretation. To do so, we’ll need to take the median chemical concentration for each biomass burn condition within each chemical category. However, since we want ggplot() to visualize the median scaled values with the color of the tiles this step was already necessary. # We&#39;ll find the median value and add that data to the dataframe as an additional column heatmap_df = scaled_longer_smoke_data %&gt;% group_by(Biomass_Burn_Condition, Chemical.Category) %&gt;% mutate(Median_Scaled_Concentration = median(Scaled_Chemical_Concentration)) head(heatmap_df) ## # A tibble: 6 × 8 ## # Groups: Biomass_Burn_Condition, Chemical.Category [6] ## Chemical.Category Chemical CASRN Units Biomass_Burn_Condition ## &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; ## 1 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus_Smoldering ## 2 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus_Flaming ## 3 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat_Smoldering ## 4 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat_Flaming ## 5 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine_Smoldering ## 6 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine_Flaming ## # ℹ 3 more variables: Chemical_Concentration &lt;dbl&gt;, ## # Scaled_Chemical_Concentration &lt;dbl[,1]&gt;, Median_Scaled_Concentration &lt;dbl&gt; Now we can plot the data and add the Median_Scaled_Concentration to the figure using geom_text(). Note that specifying the original Scaled_Chemical_Concentration in the fill parameter will NOT give you the same heatmap as specifying the median values in ggplot(). ggplot(data = heatmap_df, aes(x = Chemical.Category, y = Biomass_Burn_Condition, fill = Median_Scaled_Concentration)) + geom_tile() + # function used to specify a heatmap for ggplot geom_text(aes(label = round(Median_Scaled_Concentration, 2))) # adding concentration values as text, rounding to two values after the decimal Answer to Environmental Health Question 3 With this, we can answer Environmental Health Question #3: Under which biomass burn conditions are concentrations of certain chemical categories the highest? Answer: Peat flaming has the highest concentrations of inorganics and ions. Eucalyptus smoldering has the highest concentrations of levoglucosans. Pine smoldering has the highest concentrations of methoxyphenols. Peat smoldering has the highest concentrations of n-alkanes. Pine needles smoldering has highest concentrations of PAHs. This same heatmap can be achieved another way using the pheatmap() function from the pheatmap package. Using this function requires us to use a wide dataset, which we need to create. It will contain Chemical.Category, Biomass_Burn_Condition and Scaled_Chemical_Concentration. heatmap_df2 = scaled_longer_smoke_data %&gt;% group_by(Biomass_Burn_Condition, Chemical.Category) %&gt;% # using the summarize function instead of mutate function as was done previously since we only need the median values now summarize(Median_Scaled_Concentration = median(Scaled_Chemical_Concentration)) %&gt;% # transforming the data to a wide format pivot_wider(id_cols = Biomass_Burn_Condition, names_from = &quot;Chemical.Category&quot;, values_from = &quot;Median_Scaled_Concentration&quot;) %&gt;% # converting the chemical names to row names column_to_rownames(var = &quot;Biomass_Burn_Condition&quot;) head(heatmap_df2) ## Inorganics Ions Levoglucosan Methoxyphenols ## Eucalyptus_Flaming 0.05405359 0.05273246 0.4208870 -0.44781893 ## Eucalyptus_Smoldering -0.68595076 -0.80160192 1.7772753 -0.06449444 ## Peat_Flaming 2.24332901 1.77515899 -0.9383328 -0.51488738 ## Peat_Smoldering -0.51860591 -0.36146158 -0.8041211 0.05720971 ## Pine_Flaming -0.02063532 -0.05999543 -0.1992054 -0.50269422 ## Pine_Needles_Flaming 0.36405527 0.82229035 -0.8570130 -0.46331332 ## PAHs n-Alkanes ## Eucalyptus_Flaming 1.2885776 -0.3790357 ## Eucalyptus_Smoldering -0.4724635 -0.3465132 ## Peat_Flaming -0.5369746 -0.3093608 ## Peat_Smoldering -0.3162278 2.8238921 ## Pine_Flaming 1.7825403 -0.3347765 ## Pine_Needles_Flaming -0.4179505 -0.3850613 Now let’s generate the same heatmap this time using the pheatmap() function: pheatmap(heatmap_df2, # removing the clustering option from both rows and columns cluster_rows = FALSE, cluster_cols = FALSE, # adding the values for each cell, making those values black, and changing the font size display_numbers = TRUE, number_color = &quot;black&quot;, fontsize = 12) Notice that the pheatmap() function does not include axes or legend titles as with ggplot(), however those can be added to the figure after exporting from R in MS Powerpoint or Adobe. Additional parameters, including cluster_rows, for the pheatmap() function are discussed further in TAME 2.0 Module 5.4 Unsupervised Machine Learning. For basic heatmaps like the ones shown here, ggplot() or pheatmap() can both be used however, both have their pros and cons. For example, ggplot() figures tend to be more customizable and easily combined with other figures, while pheatmap() has additional parameters built into the function that can make plotting certain features advantageous like clustering. Concluding Remarks In conclusion, this training module provided example code to create highly customizable data visualizations using ggplot2 pertinent to environmental health research. Test Your Knowledge Replicate the figure below! The heatmap still visualizes the median chemical concentrations, but this time we’re separating the burn conditions, allowing us to determine if the concentrations of chemicals released are contingent upon the burn condition. For additional figures available and to view aspects of figures that can be changed in GGplot2, check out this GGPlot2 Cheat Sheet. You might need it to make this figure! Hint 1: Use the separate() function from tidyverse to split Biomass_Burn_Condition into Biomass and Burn_Condition. Hint 2: Use the function facet_wrap() within ggplot() to separate the heatmaps by Burn_Condition. "],["improving-data-visualizations.html", "Improving Data Visualizations Introduction to Data Visulization Conventions Introduction to Training Module Creating an Improved Boxplot Visualization Creating an Improved Heatmap Visualization Creating Multi-Plot Figures Concluding Remarks", " Improving Data Visualizations This training module was developed by Alexis Payton, Elise Hickman, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Data Visulization Conventions Data visualizations are used to convey key takeaways from a research study’s findings in a clear, succinct manner to highlight data trends, patterns, and/or relationships. In environmental health research, this is of particular importance for high-dimensional datasets that can typically be parsed using multiple methods, potentially resulting in many different approaches to visualize data. As a consequence, researchers are often faced with an overwhelming amount of options when deciding which visualization scheme(s) most optimally translate their results for effective dissemination. Effective data visualization approaches are vital to a researcher’s success for many reasons. For instance, manuscript readers or peer reviewers often scroll through a study’s text and focus on the quality and novelty of study figures before deciding whether to read/review the paper. Therefore, the importance of data visualizations cannot be understated in any research field. As a high-level introduction, it is important that we first communicate some traits that we think are imperative towards ensuring a successful data visualization approach as described in more detail below. Keys to successful data visualizations: Consider your audience, data type, and research question prior to selecting a figure to visualize your data For example, if more computationally complex methods are used in a manuscript that is intended for a journal with an audience that doesn’t have that same level of expertise, consider spending time focusing on how those results are presented in an approachable way for that audience. For a review of how to choose a rudimentary chart based on the data type, check out How to Choose the Right Data Visualization. Some of these basic charts will be presented in this module, while more complex analysis-specific visualizations, especially ones developed for high-dimensional data will be presented in later modules. Take the legibility of the figure into account This includes avoiding abbreviations when possible. (If they can’t be avoided explain them in the caption.) All titles should be capitalized, including titles for the legend(s) and axes. Underscores and periods between words should be replaced with spaces. Consider the legibility of the figure if printed in black and white. (However, that’s not as important these days.) Lastly, feel free to describe your plot in further detail in the caption to aid the reader in understanding the results presented. Minimize text Main titles aren’t necessary for single paneled figures (like the examples below), because in a publication the title of the figure is right underneath each figure. It’s good practice to remove this kind of extraneous text, which can make the figure seem more cluttered. Titles can be helpful in multi-panel figures, especially if there are multiple panels with the same figure type that present slightly different results. For example, in the Test Your Knowledge section, you’ll need to create two heatmaps, but one displays data under smoldering conditions and the other displays data under flaming conditions. In general, try to reduce the amount of extraneous text in a plot to keep a reader focused on the most important elements and takeaways in the plot. Use the minimal number of figures you need to support your narrative It is important to include an optimal number of figures within manuscripts and scientific reports. Too many figures might overwhelm the overall narrative, while too few might not provide enough substance to support your main findings. It can be helpful to also consider placing some figures in supplemental material to aid in the overall flow of your scientific writing. Select an appropriate color palette Packages have been developed to offer color palettes including MetBrewer and RColorBrewer. In addition, ggsci is a package that offers a collection of color palettes used in various scientific journals. For more information, check out MetBrewer, see its associated RDocumentation and example tutorial. For more information on RColorBrewer, see its associated RDocumentation and example tutorial. For more information on ggsci, see its associated RDocumentation. In general, it’s better to avoid bright and flashy colors that can be difficult to read. It’s advisable to use colors for manuscript figures that are color-blind friendly. Check out these Stack overflow answers about color blind-safe color palettes and packages. Popular packages for generating colorblind-friendly palettes include viridis and rcartocolor. Use color strategically Color can be used to visualize a variable. There are three ways to categorize color schemes - sequential, diverging, and qualitative. Below, definitions are provided for each along with example figures that we’ve previously published that illustrate each color scheme. In addition, figure titles and captions are also provided for context. Note that some of these figures have been simplified from what was originally published to show more streamlined examples for TAME. Sequential: intended for ordered categorical data (i.e., disease severity, likert scale, quintiles). The choropleth map below is from Winker, Payton et. al. Figure 2: Figure 1. Geospatial distribution of the risk of future wildfire events across North Carolina. Census tracts in North Carolina were binned into quintiles based on Wildfire Hazard Potential (WHP) with 1 (pale orange) having the lowest risk and 5 (dark red) having the highest risk. Figure regenerated here in alignment with its published Creative Commons Attribution 4.0 International License Diverging: intended to emphasize continuous data at extremes of the data range (typically using darker colors) and mid-range values (typically using lighter colors). This color scheme is ideal for charts like heatmaps. The heatmap below is from Payton, Perryman et. al. Figure 3: Figure 6. Individual cytokine expression levels across all subjects. Cytokine concentrations were derived from nasal lavage fluid samples. On the x axis, subjects were ordered first according to tobacco use status, starting with non-smokers then cigarette smokers and e-cigarette users. Within tobacco use groups, subjects are ordered from lowest to highest average cytokine concentration from left to right. Within each cluster shown on the y axis, cytokines are ordered from lowest to highest average cytokine concentration from bottom to top. Figure regenerated here in alignment with its published Creative Commons Attribution 4.0 International License Qualitative: intended for nominal categorical data to visualize clear differences between groups (i.e., soil types and exposure groups). The dendrogram below is from Koval et. al. Figure 4: Figure 2. Translating chemical use inventory data to inform human exposure patterning. Groups A-I illustrate the identified clusters of exposure source categories. Figure regenerated here in alignment with its published Creative Commons Attribution 4.0 International License Consider ordering axes to reveal patterns relevant to the research questions Ordering the axes can reveal potential patterns that may not be clear in the visualization otherwise. In the cytokine expression heatmap above, there are not clear differences in cytokine expression across the tobacco use groups. However, e-cigarette users seem to have slightly more muted responses compared to non-smokers and cigarette smokers in clusters B and C, which was corroborated in subsequent statistical analyses. It is also evident that Cluster A had the lowest cytokine concentrations, followed by Cluster B, and then Cluster C with the greatest concentrations. What makes these figures so compelling is how the aspects introduced above were thoughtfully incorporated. In the next section, we’ll put those principles into practice using data that were described and referenced previously in TAME 2.0 Module 3.1 Data Visualizations. Introduction to Training Module In this module, ggplot2, R’s data visualization package will be used to walk through ways to improve data visualizations. We’ll recreate two figures (i.e., the boxplot and heatmap) constructed previously in TAME 2.0 Module 3.1 Data Visualizations and improve them so they are publication-ready. Additionally, we’ll write figure titles and captions to contextualize the results presented for each visualization. When writing figure titles and captions, it is helpful to address the research question or overall concept that the figure seeks to capture rather than getting into the weeds of specific methods the plot is based on. This is especially important when visualizing more complex methods that your audience might not have as much knowledge on. Script Preparations Cleaning the global environment rm(list=ls()) Installing required R packages If you already have these packages installed, you can skip this step, or you can run the below code which checks installation status for you if (!requireNamespace(&quot;MetBrewer&quot;)) install.packages(&quot;MetBrewer&quot;); if (!requireNamespace(&quot;RColorBrewer&quot;)) install.packages(&quot;RColorBrewer&quot;); if (!requireNamespace(&quot;pheatmap&quot;)) install.packages(&quot;pheatmap&quot;); if (!requireNamespace(&quot;cowplot&quot;)) install.packages(&quot;cowplot&quot;); Loading required R packages library(tidyverse) library(MetBrewer) library(RColorBrewer) library(pheatmap) library(cowplot) Set your working directory setwd(&quot;/filepath to where your input files are&quot;) Importing example dataset Let’s now read in our example dataset. As mentioned in the introduction, this example dataset represents chemical measurements across 12 different biomass burn scenarios, representing chemicals emitted during potential wildfire events. Let’s upload and view these data: # Load the data smoke_data &lt;- read.csv(&quot;Module3_2_Input/Module3_2_InputData.csv&quot;) # View the top of the dataset head(smoke_data) ## Chemical.Category Chemical CASRN Eucalyptus_Smoldering ## 1 n-Alkanes 2-Methylnonadecane 1560-86-7 0.06 ## 2 n-Alkanes 3-Methylnonadecane 6418-45-7 0.04 ## 3 n-Alkanes Docosane 629-97-0 0.21 ## 4 n-Alkanes Dodecylcyclohexane 1795-17-1 0.04 ## 5 n-Alkanes Eicosane 112-95-8 0.11 ## 6 n-Alkanes Heneicosane 629-94-7 0.13 ## Eucalyptus_Flaming Peat_Smoldering Peat_Flaming Pine_Smoldering Pine_Flaming ## 1 0.06 1.36 0.06 0.06 0.06 ## 2 0.04 1.13 0.90 0.47 0.04 ## 3 0.25 9.46 0.57 0.16 0.48 ## 4 0.04 0.25 0.04 0.04 0.04 ## 5 0.25 7.55 0.54 0.17 0.29 ## 6 0.28 6.77 0.34 0.13 0.42 ## Pine_Needles_Smoldering Pine_Needles_Flaming Red_Oak_Smoldering ## 1 0.06 0.06 0.06 ## 2 0.04 0.72 0.04 ## 3 0.32 0.18 0.16 ## 4 0.12 0.04 0.04 ## 5 0.28 0.16 0.15 ## 6 0.30 0.13 0.13 ## Red_Oak_Flaming Units ## 1 0.13 ng_per_uL ## 2 0.77 ng_per_uL ## 3 0.36 ng_per_uL ## 4 0.04 ng_per_uL ## 5 0.38 ng_per_uL ## 6 0.69 ng_per_uL Now that we’ve been able to view the dataset, let’s come up with questions that can be answered with our boxplot and heatmap figure. This will inform how we format the dataframe for visualization. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: Boxplot: How do the distributions of the chemical concentration data differ based on each biomass burn scenario? Heatmap: Which classes of chemicals show the highest concentrations across the evaluated biomass burn conditions? How can these figures be combined into a single plot that can be then be exported from R? Formatting dataframes for downstream visualization code First, format the dataframe by changing it from a wide to long format and normalizing the chemical concentration data. For more details on this data reshaping visit TAME 2.0 Module 2.3 Data Manipulation &amp; Reshaping. scaled_longer_smoke_data = pivot_longer(smoke_data, cols = 4:13, names_to = &quot;Biomass_Burn_Condition&quot;, values_to = &quot;Chemical_Concentration&quot;) %&gt;% # scaling within each chemical group_by(Chemical) %&gt;% mutate(Scaled_Chemical_Concentration = scale(Chemical_Concentration)) %&gt;% ungroup() head(scaled_longer_smoke_data) ## # A tibble: 6 × 7 ## Chemical.Category Chemical CASRN Units Biomass_Burn_Condition ## &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; ## 1 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus_Smoldering ## 2 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus_Flaming ## 3 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat_Smoldering ## 4 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat_Flaming ## 5 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine_Smoldering ## 6 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine_Flaming ## # ℹ 2 more variables: Chemical_Concentration &lt;dbl&gt;, ## # Scaled_Chemical_Concentration &lt;dbl[,1]&gt; Creating an Improved Boxplot Visualization As we did in the previous module, a boxplot will be constructed to answer the first environmental heath question: How do the distributions of the chemical concentration data differ based on each biomass burn scenario?. Let’s remind ourselves of the original figure from the previous module. Based on the figure above, peat smoldering has the highest median scaled chemical concentration. However, this was difficult to determine given that the burn conditions aren’t labeled on the x axis and a sequential color palette was used, making it difficult to identify the correct boxplot with its burn condition in the legend. If you look closely, the colors in the legend are in a reverse order of the colors assigned to the boxplots. Let’s identify some elements of this graph that can be modified to make it easier to answer our research question. There are four main aspects we can adjust on this figure: 1. The legibility of the text in the legend and axes. Creating spaces between the text or exchanging the underscores for spaces improves the legibility of the figure. 2. The order of the boxplots. Ordering the biomass burn conditions from highest to lowest based on their median scaled chemical concentration allows the reader to easily determine the biomass burn condition that had the greatest or least chemical concentrations relative to each other. In R, this can be done by putting the Biomass_Burn_Condition variable into a factor. 3. Use of color. Variables can be visualized using color, text, size, etc. In this figure, it is redundant to have the biomass burn condition encoded in the legend and the color. Instead this variable can be put on the y axis and the legend will be removed to be more concise. The shades of the colors will also be changed, but to keep each burn condition distinct from each other, colors will be chosen that are distinct from one another. Therefore, we will choose a qualitative color scheme. 4. Show all data points when possible. Many journals now require that authors report every single value when making data visualizations, particularly for small n studies using bar graphs and boxplots to show results. Instead of just displaying the mean/median and surrounding data range, it is advised to show how every replicate landed in the study range when possible. Note that this requirement is not feasible for studies with larger sample sizes though should be considered for smaller in vitro and animal model studies. Let’s start with addressing #1: Legibility of Axis Text. The legend title and axis titles can easily be changed with ggplot(), so that will be done later. To remove the underscore from the Biomass_Burn_Condition column, we can use the function gsub(), which will replace all of the underscores with spaces, resulting in a cleaner-looking graph. # First adding spaces between the biomass burn conditions scaled_longer_smoke_data = scaled_longer_smoke_data %&gt;% mutate(Biomass_Burn_Condition = gsub(&quot;_&quot;, &quot; &quot;, Biomass_Burn_Condition)) # Viewing dataframe head(scaled_longer_smoke_data) ## # A tibble: 6 × 7 ## Chemical.Category Chemical CASRN Units Biomass_Burn_Condition ## &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; ## 1 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus Smoldering ## 2 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus Flaming ## 3 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat Smoldering ## 4 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat Flaming ## 5 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine Smoldering ## 6 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine Flaming ## # ℹ 2 more variables: Chemical_Concentration &lt;dbl&gt;, ## # Scaled_Chemical_Concentration &lt;dbl[,1]&gt; #2. Reordering the boxplots based on the median scaled chemical concentration. After calculating the median scaled chemical concentration for each biomass burn condition, the new dataframe will be arranged from lowest to highest median scaled concentration from the top of the dataframe to the bottom. This order will be saved in a vector, median_biomass_order. Although the biomass burn conditions are saved from lowest to highest concentration, ggplot() will plot them in reverse order with the highest concentration at the top and the lowest at the bottom of the y axis. Axis reordering can also be accomplished using reorder within the ggplot() function as described here and here. median_biomass = scaled_longer_smoke_data %&gt;% group_by(Biomass_Burn_Condition) %&gt;% summarize(Median_Concentration = median(Scaled_Chemical_Concentration)) %&gt;% # arranges dataframe from lowest to highest from top to bottom arrange(Median_Concentration) head(median_biomass) ## # A tibble: 6 × 2 ## Biomass_Burn_Condition Median_Concentration ## &lt;chr&gt; &lt;dbl&gt; ## 1 Red Oak Smoldering -0.459 ## 2 Eucalyptus Smoldering -0.451 ## 3 Pine Smoldering -0.424 ## 4 Pine Needles Smoldering -0.370 ## 5 Pine Needles Flaming -0.350 ## 6 Red Oak Flaming -0.337 # Saving that order median_biomass_order = median_biomass$Biomass_Burn_Condition # Putting into factor to organize the burn conditions scaled_longer_smoke_data$Biomass_Burn_Condition = factor(scaled_longer_smoke_data$Biomass_Burn_Condition, levels = median_biomass_order) # Final dataframe to be used for plotting head(scaled_longer_smoke_data) ## # A tibble: 6 × 7 ## Chemical.Category Chemical CASRN Units Biomass_Burn_Condition ## &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;fct&gt; ## 1 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus Smoldering ## 2 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus Flaming ## 3 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat Smoldering ## 4 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat Flaming ## 5 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine Smoldering ## 6 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine Flaming ## # ℹ 2 more variables: Chemical_Concentration &lt;dbl&gt;, ## # Scaled_Chemical_Concentration &lt;dbl[,1]&gt; Now that the dataframe has been finalized, we can plot the new boxplot. The final revision, #3: Making Use of Color, will be addressed with ggplot(). However, a palette can be chosen from the MetBrewer package. # Choosing the &quot;Jurarez&quot; palette from the `MetBrewer` package # `n = 12`, since there are 12 biomass burn conditions juarez_colors = met.brewer(name = &quot;Juarez&quot;, n = 12)[1:12] #4. Show all data points when possible will also be addressed with ggplot() by simply using geom_point(). FigureX1 = ggplot(scaled_longer_smoke_data, aes(x = Scaled_Chemical_Concentration, y = Biomass_Burn_Condition, color = Biomass_Burn_Condition)) + geom_boxplot() + # jittering the points, so they&#39;re not all on top of each other and adding transparency geom_point(position = position_jitter(h = 0.1), alpha = 0.7) + theme_light() + # changing the theme theme(axis.text = element_text(size = 9), # changing size of axis labels axis.title = element_text(face = &quot;bold&quot;, size = rel(1.5)), # changes axis titles legend.position = &quot;none&quot;) + # removes legend xlab(&#39;Scaled Chemical Concentration (pg/uL)&#39;) + ylab(&#39;Biomass Burn Condition&#39;) + # changing axis labels scale_color_manual(values = c(juarez_colors)) # changing the colors FigureX1 An appropriate title for this figure could be: “Figure X. Chemical concentration distributions of biomass burn conditions. The boxplots are based on the scaled chemical concentration values, which used the raw chemical concentrations values scaled within each chemical. The individual dots represent the concentrations of each chemical. The biomass burn conditions on the y axis are ordered from greatest (top) to least (bottom) based on median scaled chemical concentration.” Answer to Environmental Health Question 1 With this, we can answer Environmental Health Question #1: Which biomass burn condition has the highest total chemical concentration? Answer: Smoldering peat has the highest median chemical concentration, however the median concentrations are comparable across all biomass burn conditions. All the flaming conditions have the highest median chemical concentrations and more overall variation than their respective smoldering conditions with the exception of smoldering peat. You may notice that the scaled chemical concentration was put on the x axis and burn condition was put on the y axis and not vice versa. When names are longer in length, they are more legible if placed on the y axis. Other aspects of the figure were changed in the latest version, but those are minor compared to changing the order of the boxplots, revamping the text, and changing the usage of color. For example, the background was changed from gray to white. Figure backgrounds are generally white, because the figure is easier to read if the paper is printed in black and white. A plot’s background can easily be changed to white in R using theme_light(), theme_minimal(), or theme_bw(). Posit provides a very helpful GGplot2 cheat sheet for changing a figure’s parameters. Creating an Improved Heatmap Visualization We’ll use a heatmap to answer the second environmental health question: Which classes of chemicals show the highest concentrations across the evaluated biomass burn conditions? Let’s view the original heatmap from the previous module and find aspects of it that can be improved. # Changing the biomass condition variable back to a character from a factor scaled_longer_smoke_data$Biomass_Burn_Condition = as.character(scaled_longer_smoke_data$Biomass_Burn_Condition) # Calculating the median value within each biomass burn condition and category scaled_longer_smoke_data = scaled_longer_smoke_data %&gt;% group_by(Biomass_Burn_Condition, Chemical.Category) %&gt;% mutate(Median_Scaled_Concentration = median(Scaled_Chemical_Concentration)) # Plotting ggplot(data = scaled_longer_smoke_data, aes(x = Chemical.Category, y = Biomass_Burn_Condition, fill = Median_Scaled_Concentration)) + geom_tile() + geom_text(aes(label = round(Median_Scaled_Concentration, 2))) # adding concentration values as text, rounding to two values after the decimal From the figure above, it’s clear that certain biomass burn conditions are associated with higher chemical concentrations for some of the chemical categories. For example, peat flaming exposure was associated with higher levels of inorganics and ions, while pine smoldering exposure was associated with higher levels of methoxyphenols. Although these are important findings, it is still difficult to determine if there are greater similarities in chemical profiles based on the biomass or the incineration temperature. Therefore, let’s identify some elements of this chart that can be modified to make it easier to answer our research question. There are three main aspects we can adjust on this figure: 1. The legibility of the text in the legend and axes. Similar to what we did previously, we’ll replace underscores and periods with spaces in the axis labels and titles. 2. The order of the axis labels. Ordering the biomass burn condition and chemical category from highest to lowest based on their median scaled chemical concentration allows the reader to easily determine the biomass burn condition that had the greatest or least total chemical concentrations relative to each other. From the previous boxplot figure, biomass burn condition is already in this order, however we need to order the chemical category by putting the variable into a factor. 3. Use of color. Notice that in the boxplot we used a qualitative palette, which is best for creating visual differences between different classes or groups. In this heatmap, we’ll choose a diverging color palette that uses two or more contrasting colors. A diverging color palette is able to highlight mid range with a lighter color and values at either extreme with a darker color or vice versa. #1: Legibility of Text can be addressed in ggplot() and so can #2: Reordering the heatmap. Biomass_Burn_Condition has already been reordered and put into a factor, but we need to do the same with Chemical.Category. Similar to before, median scaled chemical concentration for each chemical category will be calculated. However, this time the new dataframe will be arranged from highest to lowest median scaled concentration from the top of the dataframe to the bottom. ggplot() will plot them in the SAME order with the highest concentration on the left side and the lowest on the right side of the figure. # Order the chemical category by the median scaled chemical concentration median_chemical = scaled_longer_smoke_data %&gt;% group_by(Chemical.Category) %&gt;% summarize(Median_Concentration = median(Scaled_Chemical_Concentration)) %&gt;% arrange(-Median_Concentration) head(median_chemical) ## # A tibble: 6 × 2 ## Chemical.Category Median_Concentration ## &lt;chr&gt; &lt;dbl&gt; ## 1 Inorganics -0.265 ## 2 n-Alkanes -0.335 ## 3 Ions -0.359 ## 4 Levoglucosan -0.417 ## 5 Methoxyphenols -0.434 ## 6 PAHs -0.459 # Saving that order median_chemical_order = median_chemical$Chemical.Category # Putting into factor to organize the chemical categories scaled_longer_smoke_data$Chemical.Category = factor(scaled_longer_smoke_data$Chemical.Category, levels = median_chemical_order) # Putting burn conditons back into a factor to organize them scaled_longer_smoke_data$Biomass_Burn_Condition = factor(scaled_longer_smoke_data$Biomass_Burn_Condition, levels = median_biomass_order) # Viewing the dataframe to be plotted head(scaled_longer_smoke_data) ## # A tibble: 6 × 8 ## # Groups: Biomass_Burn_Condition, Chemical.Category [6] ## Chemical.Category Chemical CASRN Units Biomass_Burn_Condition ## &lt;fct&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;fct&gt; ## 1 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus Smoldering ## 2 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus Flaming ## 3 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat Smoldering ## 4 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat Flaming ## 5 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine Smoldering ## 6 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine Flaming ## # ℹ 3 more variables: Chemical_Concentration &lt;dbl&gt;, ## # Scaled_Chemical_Concentration &lt;dbl[,1]&gt;, Median_Scaled_Concentration &lt;dbl&gt; Now that the dataframe has been finalized, we can plot the new boxplot. The final revision, #3: Making Use of Color, will be addressed with ggplot(). Here a palette is chosen from the RColorBrewer package. # Only needed to choose 2 colors for &#39;low&#39; and &#39;high&#39; in the heatmap # `n = 8` in the code to generate more colors that can be chosen from rcolorbrewer_colors = brewer.pal(n = 8, name = &#39;Accent&#39;) FigureX2 = ggplot(data = scaled_longer_smoke_data, aes(x = Chemical.Category, y = Biomass_Burn_Condition, fill = Median_Scaled_Concentration)) + geom_tile(color = &#39;white&#39;) + # adds white space between the tiles geom_text(aes(label = round(Median_Scaled_Concentration, 2))) + # adding concentration values as text theme_minimal() + # changing the theme theme(axis.text = element_text(size = 9), # changing size of axis labels axis.title = element_text(face = &quot;bold&quot;, size = rel(1.5)), # changes axis titles legend.title = element_text(face = &#39;bold&#39;, size = 10), # changes legend title legend.text = element_text(size = 9)) + # changes legend text labs(x = &#39;Chemical Category&#39;, y = &#39;Biomass Burn Condition&#39;, fill = &quot;Scaled Chemical\\nConcentration (pg/mL)&quot;) + # changing axis labels scale_fill_gradient(low = rcolorbrewer_colors[5], high = rcolorbrewer_colors[6]) # changing the colors FigureX2 An appropriate title for this figure could be: “Figure X. Chemical category concentrations across biomass burn conditions. Scaled chemical concentration values are based on the raw chemical concentration values scaled within each chemical. Chemical category on the x axis is ordered from highest to lowest median concentration from left to right. Biomass burn condition on the y axis is ordered from the highest to lowest median concentration from top to bottom. The values in each tile represent the median scaled chemical concentration.” Answer to Environmental Health Question 2 With this, we can answer Environmental Health Question #2: Which classes of chemicals show the highest concentrations across the evaluated biomass burn conditions? Answer: Ordering the axes from highest to lowest concentration didn’t help organize the data as much as we would’ve liked given some of the variance of chemical concentrations across the chemical categories. Nevertheless, it’s still clear that peat flaming produces the highest concentration of inorganics and ions, peat smoldering with n-Alkanes, eucalyptus smoldering with Levoglucosan, pine smoldering with methoxyphenols, and pine flaming with PAHs. In addition, flaming conditions seem to have higher levels of inorganics and ions while smoldering conditions seem to have higher levels of levoglucosan and PAHs. It would be helpful if there was a way to group these chemical profiles based on similarity and that’s where the pheatmap() function can be helpful when it can be difficult to spot those patterns using visual inspection alone. Just for fun, let’s briefly visualize a hierarchical clustering heatmap, which will be used to group both the biomass burn conditions and chemical categories based on their chemical concentrations. In this module, we’ll focus only on the pheatmap() visualization, but more information on hierarchical clustering can be found in Module 5.5 Unsupervised Machine Learning II: Additional Clustering Applications. As we showed in the previous module, this function requires a wide dataframe which we’ll need to create. It will contain Chemical.Category, Biomass_Burn_Condition and Scaled_Chemical_Concentration. heatmap_df2 = scaled_longer_smoke_data %&gt;% group_by(Biomass_Burn_Condition, Chemical.Category) %&gt;% # using the summarize function instead of mutate function as was done previously since we only need the median values now summarize(Median_Scaled_Concentration = median(Scaled_Chemical_Concentration)) %&gt;% # transforming the data to a wide format pivot_wider(id_cols = Biomass_Burn_Condition, names_from = &quot;Chemical.Category&quot;, values_from = &quot;Median_Scaled_Concentration&quot;) %&gt;% # converting the chemical names to row names column_to_rownames(var = &quot;Biomass_Burn_Condition&quot;) head(heatmap_df2) ## Inorganics n-Alkanes Ions Levoglucosan ## Red Oak Smoldering -0.6236516 -0.3895328 -0.8011155 1.1042911 ## Eucalyptus Smoldering -0.6859508 -0.3465132 -0.8016019 1.7772753 ## Pine Smoldering -0.8012749 -0.3637526 -0.8016019 0.9961855 ## Pine Needles Smoldering -0.8436790 -0.3076779 -0.4611222 -0.6357051 ## Pine Needles Flaming 0.3640553 -0.3850613 0.8222903 -0.8570130 ## Red Oak Flaming 0.4858796 -0.3016948 0.6793662 -0.8642615 ## Methoxyphenols PAHs ## Red Oak Smoldering 0.04674277 -0.5179641 ## Eucalyptus Smoldering -0.06449444 -0.4724635 ## Pine Smoldering 2.74452541 -0.4580922 ## Pine Needles Smoldering -0.22835290 -0.3162278 ## Pine Needles Flaming -0.46331332 -0.4179505 ## Red Oak Flaming -0.51488738 -0.5318127 Now let’s generate the same heatmap this time using the pheatmap() function: # creating a color palette blue_pink_palette = colorRampPalette(c(rcolorbrewer_colors[5], rcolorbrewer_colors[6])) pheatmap(heatmap_df2, # changing the color scheme color = blue_pink_palette(40), # hierarchical clustering of the biomass burn conditions cluster_rows = TRUE, # creating white space between the two largest clusters cutree_row = 2, # adding the values for each cell and making those values black display_numbers = TRUE, number_color = &quot;black&quot;, # changing the font size and the angle of the column names fontsize = 12, angle_col = 45) By using incorporating the dendrogram into the visualization, it’s easier to see that the chemical profiles have greater similarities within incineration temperatures rather than biomasses (with the exception of pine needles smoldering). Creating Multi-Plot Figures We can combine figures using the plot_grid() function from the cowplot package. For additional information on the plot_grid() function and parameters that can be changed see Arranging Plots in a Grid. Other packages that have figure combining capabilities include the patchwork package and the grid_arrange() function from the gridExtra package. Figures can also be combined after they’re exported from R using other applications like MS powerpoint and Adobe pdf. FigureX = plot_grid(FigureX1, FigureX2, # Adding labels, changing size their size and position labels = &quot;AUTO&quot;, label_size = 15, label_x = 0.04, rel_widths = c(1, 1.5)) FigureX An appropriate title for this figure could be: “Figure X. Chemical concentration distributions across biomass burn conditions. (A) The boxplots are based on the scaled chemical concentration values, which used the raw chemical concentrations values scaled within each chemical. The individual dots represent the concentrations of each chemical. The biomass burn conditions on the y axis are ordered from greatest (top) to least (bottom) based on median scaled chemical concentration. (B) The heatmap visualizes concentrations across chemical categories. Chemical category on the x axis is ordered from highest to lowest median concentration from left to right. Biomass burn condition on the y axis is ordered from the highest to lowest median concentration from top to bottom. The values in each tile represent the median scaled chemical concentration. By putting these two figures side by side, it’s now easier to compare the distributions of each biomass burn condition in figure A alongside the median chemical category concentrations in figure B that are responsible for the variation seen on the left. Concluding Remarks In conclusion, this training module provided information and example code for improving, streamlining, and making ggplot2 figures publication ready. Keep in mind that concepts and ideas presented in this module can be subjective and might need to be amended given the situation, dataset, and visualization. Additional Resources Beginner’s Guide to Data Visualizations and Improving Data Visualizations in R Generating Colors for Visualizations Additional Hands on Training Brewer, Cynthia A. 1994. Color use guidelines for mapping and visualization. Chapter 7 (pp. 123-147) in Visualization in Modern Cartography Hattab, G., Rhyne, T.-M., &amp; Heider, D. (2020). Ten simple rules to colorize biological data visualization. PLOS Computational Biology, 16(10), e1008259. PMID: 33057327 Lastly, for researchers who are newer to R programming, ggpubr is a package specifically designed to create publication-ready graphs similar to ggplot2 with more concise syntax. This package is particularly useful for statistically relevant visualizations, which are further explored in later modules including, TAME 2.0 Module 3.4 Introduction to Statistical Tests, TAME 2.0 Module 4.4 Two Group Comparisons and Visualizations, and TAME 2.0 Module 4.5 Multigroup Comparisons and Visualizations. Test Your Knowledge Replicate the figure below! The heatmap is the same as the “Test Your Knowledge” figure from TAME 2.0 Module 3.1 Data Visualizations. This time we’ll focus on making the figure look more publication ready by cleaning up the titles, cleaning up the labels, and changing the colors. The heatmap still visualizes the median chemical concentrations, but this time we’re separating the burn conditions, allowing us to determine if the concentrations of chemicals released are contingent upon the burn condition. Hint: To view additional aspects of figures that can be changed in ggplot2 check out this GGPlot2 Cheat Sheet. It might come in handy! "],["normality-tests-and-data-transformations.html", "Normality Tests and Data Transformations Introduction to Training Module What is a Normal Distribution? Qualitative Assessment of Normality Quantitative Normality Assessment Data Transformation Additional Considerations Regarding Normality Concluding Remarks", " Normality Tests and Data Transformations This training module was developed by Elise Hickman, Alexis Payton, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module When selecting the appropriate statistical tests to evaluate potential trends in your data, selection often relies upon whether or not underlying data are normally distributed. Many statistical tests and methods that are commonly implemented in exposure science, toxicology, and environmental health research rely on assumptions of normality. Applying a statistical test intended for data with a specific distribution when your data do not fit within that distribution can generate unreliable results, with the potential for false positive and false negative findings. Thus, one of the most common statistical tests to perform at the beginning of an analysis is a test for normality. In this training module, we will: Review the normal distribution and why it is important Demonstrate how to test whether your variable distributions are normal… Qualitatively, with histograms and Q-Q plots Quantitatively, with the Shapiro-Wilk test Discuss data transformation approaches Demonstrate log2 data transformation for non-normal data Discuss additional considerations related to normality We will demonstrate normality assessment using example data derived from a study in which chemical exposure profiles were collected across study participants through silicone wristbands. This exposure monitoring technique has been described through previous publications, including the following examples: O’Connell SG, Kincl LD, Anderson KA. Silicone wristbands as personal passive samplers. Environ Sci Technol. 2014 Mar 18;48(6):3327-35. doi: 10.1021/es405022f. Epub 2014 Feb 26. PMID: 24548134; PMCID: PMC3962070. Kile ML, Scott RP, O’Connell SG, Lipscomb S, MacDonald M, McClelland M, Anderson KA. Using silicone wristbands to evaluate preschool children’s exposure to flame retardants. Environ Res. 2016 May;147:365-72. doi: 10.1016/j.envres.2016.02.034. Epub 2016 Mar 3. PMID: 26945619; PMCID: PMC4821754. Hammel SC, Hoffman K, Phillips AL, Levasseur JL, Lorenzo AM, Webster TF, Stapleton HM. Comparing the Use of Silicone Wristbands, Hand Wipes, And Dust to Evaluate Children’s Exposure to Flame Retardants and Plasticizers. Environ Sci Technol. 2020 Apr 7;54(7):4484-4494. doi: 10.1021/acs.est.9b07909. Epub 2020 Mar 11. PMID: 32122123; PMCID: PMC7430043. Levasseur JL, Hammel SC, Hoffman K, Phillips AL, Zhang S, Ye X, Calafat AM, Webster TF, Stapleton HM. Young children’s exposure to phenols in the home: Associations between house dust, hand wipes, silicone wristbands, and urinary biomarkers. Environ Int. 2021 Feb;147:106317. doi: 10.1016/j.envint.2020.106317. Epub 2020 Dec 17. PMID: 33341585; PMCID: PMC7856225. In the current example dataset, chemical exposure profiles were obtained from the analysis of silicone wristbands worn by 97 participants for one week. Chemical concentrations on the wristbands were measured with gas chromatography mass spectrometry. The subset of chemical data used in this training module are all phthalates, a group of chemicals used primarily in plastic products to increase flexibility and durability. Script Preparations Cleaning the global environment rm(list=ls()) Installing required R packages If you already have these packages installed, you can skip this step, or you can run the below code which checks installation status for you if (!requireNamespace(&quot;openxlsx&quot;)) install.packages(&quot;openxlsx&quot;); if (!requireNamespace(&quot;tidyverse&quot;)) install.packages(&quot;tidyverse&quot;); if (!requireNamespace(&quot;ggpubr&quot;)) install.packages(&quot;ggpubr&quot;); Loading R packages required for this session library(openxlsx) # for importing data library(tidyverse) # for manipulating and plotting data library(ggpubr) # for making Q-Q plots with ggplot Set your working directory setwd(&quot;/filepath to where your input files are&quot;) Importing example dataset # Import data wrist_data &lt;- read.xlsx(&quot;Module3_3_Input/Module3_3_InputData.xlsx&quot;) # Viewing the data head(wrist_data) ## S_ID Age DEP DBP BBP DEHA DEHP DEHT ## 1 1 24.76986 335.58857 574.5443 40.67286 755.8157 10621.7029 30420.68 ## 2 2 25.39452 56.38286 1075.7114 243.48857 2716.7314 3036.5757 23991.82 ## 3 3 34.55068 515.65429 121.1657 205.86857 3286.5886 3056.2743 46188.06 ## 4 4 23.83562 1009.00714 373.4957 66.97571 3966.5371 729.7971 17900.74 ## 5 5 39.29315 33.74143 104.0629 77.17286 1654.3317 2599.7129 13597.44 ## 6 6 36.15616 168.79714 503.8300 61.98429 398.6314 1492.6143 29875.76 ## DINP TOTM ## 1 26534.290 1447.86000 ## 2 10073.704 39.46143 ## 3 1842.949 112.67714 ## 4 78779.567 92.31000 ## 5 3682.956 161.84571 ## 6 23845.493 182.56429 Our example dataset contains subject IDs (S_ID), subject ages, and measurements of 8 different phthalates from silicone wristbands: DEP: Diethyl phthalate DBP : Dibutyl phthalate BBP : Butyl benzyl phthalate DEHA : Di(2-ethylhexyl) adipate DEHP : Di(2-ethylhexyl) phthalate DEHT: Di(2-ethylhexyl) terephthalate DINP : Diisononyl phthalate TOTM : Trioctyltrimellitate The units for the chemical data are nanogram of chemical per gram of silicone wristband (ng/g) per day the participant wore the wristband. One of the primary questions in this study was whether there were significant differences in chemical exposure between subjects with different levels of social stress or between subjects with differing demographic characteristics. However, before we can analyze the data for significant differences between groups, we first need to assess whether our numeric variables are normally distributed. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: Are these data normally distributed? How does the distribution of data influence the statistical tests performed on the data? Before answering these questions, let’s define normality and how to test for it in R. What is a Normal Distribution? A normal distribution is a distribution of data in which values are distributed roughly symmetrically out from the mean such that 68.3% of values fall within one standard deviation of the mean, 95.4% of values fall within 2 standard deviations of the mean, and 99.7% of values fall within three standard deviations of the mean. Figure 5: Figure Credit: D Wells, CC BY-SA 4.0 https://creativecommons.org/licenses/by-sa/4.0, via Wikimedia Commons Common parametric statistical tests, such as t-tests, one-way ANOVAs, and Pearson correlations, rely on the assumption that data fall within the normal distribution for calculation of z-scores and p-values. Non-parametric tests, such as the Wilcoxon Rank Sum test, Kruskal-Wallis test, and Spearman Rank correlation, do not rely on assumptions about data distribution. Some of the aforementioned between-group comparisons were introduced in TAME 2.0 Module 3.4 Introduction to Statistical Tests. They, along with non-parametric tests, are explored further in later modules including TAME 2.0 Module 4.4 Two-Group Comparisons &amp; Visualizations and TAME 2.0 Module 4.5 Multi-group Comparisons &amp; Visualizations. Qualitative Assessment of Normality We can begin by assessing the normality of our data through plots. For example, plotting data using histograms, densities, or Q-Q plots can graphically help inform if a variable’s values appear to be normally distributed or not. We will start with visualizing our data distributions with histograms. Histograms Let’s start with visualizing the distribution of the participant’s ages using the hist() function that is part of base R. hist(wrist_data$Age) We can edit some of the parameters to improve this basic histogram visualization. For example, we can decrease the size of each bin using the breaks parameter: hist(wrist_data$Age, breaks = 10) The hist() function is useful for plotting single distributions, but what if we have many variables that need normality assessment? We can leverage ggplot2’s powerful and flexible graphics functions such as geom_histogram() and facet_wrap() to inspect histograms of all of our variables in one figure panel. For more information on data manipulation in general, see TAME 2.0 Module 2.3 Data Manipulation &amp; Reshaping and for more on ggplot2 including the use of facet_wrap(), see TAME 2.0 Module 3.2 Improving Data Visualizations. First, we’ll pivot our data to longer to prepare for plotting. Then, we’ll make our plot. We can use the theme_set() function to set a default graphing theme for the rest of the script. A graphing theme represents a set of default formatting parameters (mostly colors) that ggplot will use to make your graphs. theme_bw() is a basic theme that includes a white background for the plot and dark grey axis text and minor axis lines. The theme that you use is a matter of personal preference. For more on the different themes available through ggplot2, see here. # Pivot data longer to prepare for plotting wrist_data_long &lt;- wrist_data %&gt;% pivot_longer(!S_ID, names_to = &quot;variable&quot;, values_to = &quot;value&quot;) # Set theme for graphing theme_set(theme_bw()) # Make figure panel of histograms ggplot(wrist_data_long, aes(value)) + geom_histogram(fill = &quot;gray40&quot;, color = &quot;black&quot;, binwidth = function(x) {(max(x) - min(x))/25}) + facet_wrap(~ variable, scales = &quot;free&quot;) + labs(y = &quot;# of Observations&quot;, x = &quot;Value&quot;) From these histograms, we can see that our chemical variables do not appear to be normally distributed. Q-Q Plots Q-Q (quantile-quantile) plots are another way to visually assess normality. Similar to the histogram above, we can create a single Q-Q plot for the age variable using base R functions. Normal Q-Q plots (Q-Q plots where the theoretical quantiles are based on a normal distribution) have theoretical quantiles on the x-axis and sample quantiles, representing the distribution of the variable of interest from the dataset, on the y-axis. If the variable of interest is normally distributed, the points on the graph will fall along the reference line. # Plot points qqnorm(wrist_data$Age) # Add a reference line for theoretically normally distributed data qqline(wrist_data$Age) Small variations from the reference line, as seen above, are to be expected for the most extreme values. Overall, we can see that the age data are relatively normally distributed, as the points fall along the reference line. To make a figure panel with Q-Q plots for all of our variables of interest, we can use the ggqqplot() function within the ggpubr package. This function generates Q-Q plots and has arguments that are similar to ggplot2. ggqqplot(wrist_data_long, x = &quot;value&quot;, facet.by = &quot;variable&quot;, ggtheme = theme_bw(), scales = &quot;free&quot;) With this figure panel, we can see that the chemical data have very noticeable deviations from the reference, suggesting non-normal distributions. To answer our first environmental health question, age is the only variable that appears to be normally distributed in our dataset. This is based on our histograms and Q-Q plots with data centered in the middle and spreading with a distribution on both the lower and upper sides that follow typical normal data distributions. However, chemical concentrations appear to be non-normally distributed. Next, we will implement a quantitative approach to assessing normality, based on a statistical test for normality. Quantitative Normality Assessment Single Variable Normality Assessment We will use the Shapiro-Wilk test to quantitatively assess whether our data distribution is normal, again looking at the age data. This test can be carried out simply using the shapiro.test() function from the base R stats package. When using this test and interpreting its results, it is important to remember that the null hypothesis is that the sample distribution is normal, and a significant p-value means the distribution is non-normal. shapiro.test(wrist_data$Age) ## ## Shapiro-Wilk normality test ## ## data: wrist_data$Age ## W = 0.9917, p-value = 0.8143 This test resulted in a p-value of 0.8143, so we cannot reject the null hypothesis (that data are normally distributed). This means that we can assume that age is normally distributed, which is consistent with our visualizations above. Multiple Variable Normality Assessment With a large dataset containing many variables of interest (e.g., our example data with multiple chemicals), it is more efficient to test each column for normality and then store those results in a dataframe. We can use the base R function apply() to apply the Shapiro Wilk test over all of the numeric columns of our dataframe. This function generates a list of results, with a list element for each variable tested. There are also other ways that you could iterate through each of your columns, such as a for loop or a function as discussed in TAME 2.0 Module 2.4 Improving Coding Efficiencies. # Apply Shapiro Wilk test shapiro_res &lt;- apply(wrist_data %&gt;% select(-S_ID), 2, shapiro.test) # View first three list elements glimpse(shapiro_res[1:3]) ## List of 3 ## $ Age:List of 4 ## ..$ statistic: Named num 0.992 ## .. ..- attr(*, &quot;names&quot;)= chr &quot;W&quot; ## ..$ p.value : num 0.814 ## ..$ method : chr &quot;Shapiro-Wilk normality test&quot; ## ..$ data.name: chr &quot;newX[, i]&quot; ## ..- attr(*, &quot;class&quot;)= chr &quot;htest&quot; ## $ DEP:List of 4 ## ..$ statistic: Named num 0.225 ## .. ..- attr(*, &quot;names&quot;)= chr &quot;W&quot; ## ..$ p.value : num 2.74e-20 ## ..$ method : chr &quot;Shapiro-Wilk normality test&quot; ## ..$ data.name: chr &quot;newX[, i]&quot; ## ..- attr(*, &quot;class&quot;)= chr &quot;htest&quot; ## $ DBP:List of 4 ## ..$ statistic: Named num 0.658 ## .. ..- attr(*, &quot;names&quot;)= chr &quot;W&quot; ## ..$ p.value : num 1.08e-13 ## ..$ method : chr &quot;Shapiro-Wilk normality test&quot; ## ..$ data.name: chr &quot;newX[, i]&quot; ## ..- attr(*, &quot;class&quot;)= chr &quot;htest&quot; We can then convert those list results into a dataframe. Each variable is now in a row, with columns describing outputs of the statistical test. # Create results dataframe shapiro_res &lt;- do.call(rbind.data.frame, shapiro_res) # View results dataframe shapiro_res ## statistic p.value method data.name ## Age 0.9917029 8.143367e-01 Shapiro-Wilk normality test newX[, i] ## DEP 0.2248611 2.736536e-20 Shapiro-Wilk normality test newX[, i] ## DBP 0.6584967 1.076529e-13 Shapiro-Wilk normality test newX[, i] ## BBP 0.2367689 3.757059e-20 Shapiro-Wilk normality test newX[, i] ## DEHA 0.6646692 1.454576e-13 Shapiro-Wilk normality test newX[, i] ## DEHP 0.6163531 1.519572e-14 Shapiro-Wilk normality test newX[, i] ## DEHT 0.8072684 6.315917e-10 Shapiro-Wilk normality test newX[, i] ## DINP 0.5741864 2.486638e-15 Shapiro-Wilk normality test newX[, i] ## TOTM 0.3397424 6.901903e-19 Shapiro-Wilk normality test newX[, i] Finally, we can clean up our results dataframe and add a column that will quickly tell us whether our variables are normally or non-normally distributed based on the Shapiro-Wilk normality test results. # Clean dataframe shapiro_res &lt;- shapiro_res %&gt;% # Add normality conclusion mutate(normal = ifelse(p.value &lt; 0.05, F, T)) %&gt;% # Remove columns that do not contain informative data select(c(p.value, normal)) # View cleaned up dataframe shapiro_res ## p.value normal ## Age 8.143367e-01 TRUE ## DEP 2.736536e-20 FALSE ## DBP 1.076529e-13 FALSE ## BBP 3.757059e-20 FALSE ## DEHA 1.454576e-13 FALSE ## DEHP 1.519572e-14 FALSE ## DEHT 6.315917e-10 FALSE ## DINP 2.486638e-15 FALSE ## TOTM 6.901903e-19 FALSE The results from the Shapiro-Wilk test demonstrate that age data are normally distributed, while the chemical concentration data are non-normally distributed. These results support the conclusions we made based on our qualitative assessment above with histograms and Q-Q plots. Answer to Environmental Health Question 1 With this, we can now answer Environmental Health Question #1: Are these data normally distributed? Answer: Age is normally distributed, while chemical concentrates are non-normally distributed. Answer to Environmental Health Question 2 We can also answer Environmental Health Question #2: How does the distribution of data influence the statistical tests performed on the data? Answer: Parametric statistical tests should be used when analyzing the age data, and non-parametric tests should be used when analyzing the chemical concentration data Data Transformation There are a number of approaches that can be used to change the range and/or distribution of values within each variable. Typically, the purpose for applying these changes is to reduce bias in a dataset, remove known sources of variation, or prepare data for specific downstream analyses. The following are general definitions for common terms used when discussing these changes: Transformation refers to any process used to change data into other, related values. Normalization and standardization are types of data transformation. Transformation can also refer to performing the same mathematical operation on every value in your dataframe. For example, taking the log2 or log10 of every value is referred to as log transformation. Normalization is the process of transforming variables so that they are on a similar scale and therefore are comparable. This can be important when variables in a dataset contain a mixture of data types that are represented by vastly different numeric magnitudes or when there are known sources of variability across samples. Normalization methods are highly dependent on the type of input data. One example of normalization is min-max scaling, which results in a range for each variable of 0 to 1. Although normalization in computational methodologies typically refers to min-max scaling or other similar methods where the variable’s range is bounded by specific values, wet-bench approaches also employ normalization - for example, using a reference gene for RT-qPCR assays or dividing a total protein amount for each sample by the volume of each sample to obtain a concentration. Standardization, also known as Z-score normalization, is a specific type of normalization that involves subtracting each value from the mean of that variable and dividing by that variable’s standard deviation. The standardized values for each variable will have a mean of 0 and a standard deviation of 1. The scale() function in R performs standardization by default when the data are centered (argument center = TRUE is included within the scale function). Transformation of example data When data are non-normally distributed, such as with the chemical concentrations in our example dataset, it may be desirable to transform the data so that the distribution becomes closer to a normal distribution, particularly if there are only parametric tests available to test your hypothesis. A common transformation used in environmental health research is log2 transformation, in which data are transformed by taking the log2 of each value in the dataframe. Let’s log2 transform our chemical data and examine the resulting histograms and Q-Q plots to qualitatively assess whether data appear more normal following transformation. We will apply a pseudo-log2 transformation, where we will add 1 to each value before log2 transforming so that all resulting values are positive and any zeroes in the dataframe do not return -Inf. # Apply psuedo log2 (pslog2) transformation to chemical data wrist_data_pslog2 &lt;- wrist_data %&gt;% mutate(across(DEP:TOTM, ~ log2(.x + 1))) # Pivot data longer wrist_data_pslog2_long &lt;- wrist_data_pslog2 %&gt;% pivot_longer(!S_ID, names_to = &quot;variable&quot;, values_to = &quot;value&quot;) # Make figure panel of histograms ggplot(wrist_data_pslog2_long, aes(value)) + geom_histogram(fill = &quot;gray40&quot;, color = &quot;black&quot;, binwidth = function(x) {(max(x) - min(x))/25}) + facet_wrap(~ variable, scales = &quot;free&quot;) + labs(y = &quot;# of Observations&quot;, x = &quot;Value&quot;) # Make a figure panel of Q-Q plots ggqqplot(wrist_data_pslog2_long, x = &quot;value&quot;, facet.by = &quot;variable&quot;, ggtheme = theme_bw(), scales = &quot;free&quot;) Both the histograms and the Q-Q plots demonstrate that our log2 transformed data are more normally distributed than the raw data graphed above. Let’s apply the Shapiro-Wilk test to our log2 transformed data to determine if the chemical distributions are normally distributed. # Apply Shapiro Wilk test shapiro_res_pslog2 &lt;- apply(wrist_data_pslog2 %&gt;% select(-S_ID), 2, shapiro.test) # Create results dataframe shapiro_res_pslog2 &lt;- do.call(rbind.data.frame, shapiro_res_pslog2) # Clean dataframe shapiro_res_pslog2 &lt;- shapiro_res_pslog2 %&gt;% ## Add normality conclusion mutate(normal = ifelse(p.value &lt; 0.05, F, T)) %&gt;% ## Remove columns that do not contain informative data select(c(p.value, normal)) # View cleaned up dataframe shapiro_res_pslog2 ## p.value normal ## Age 0.814336705 TRUE ## DEP 0.001335217 FALSE ## DBP 0.368954224 TRUE ## BBP 0.052805523 TRUE ## DEHA 0.979072298 TRUE ## DEHP 0.304963678 TRUE ## DEHT 0.770066136 TRUE ## DINP 0.883662530 TRUE ## TOTM 0.004399442 FALSE The results from the Shapiro-Wilk test demonstrate that the the log2 chemical concentration data are more normally distributed than the raw data. Overall, the p-values, even for the chemicals that are still non-normally distributed, are much higher, and only 2 out of the 8 chemicals are non-normally distributed by the Shapiro-Wilk test. We can also calculate average p-values across all variables for our raw and log2 transformed data to further demonstrate this point. # Calculate the mean Shapiro-Wilk p-value for the raw chemical data mean(shapiro_res$p.value) ## [1] 0.09048186 # Calculate the mean Shapiro-Wilk p-value for the pslog2 transformed chemical data mean(shapiro_res_pslog2$p.value) ## [1] 0.4643995 Therefore, the log2 chemical data would be most appropriate to use if researchers are wanting to perform parametric statistical testing (and particularly if there is not a non-parametric statistical test for a given experimental design). It is important to note that if you proceed to statistical testing using log2 or other transformed data, graphs you make of significant results should use the transformed values on the y-axis, and findings should be interpreted in the context of the transformed values. Additional Considerations Regarding Normality The following sections detail additional considerations regarding normality. Similar to other advice in TAME, appropriate methods for handling normality assessment and normal versus non-normal data can be dependent on your field, lab, endpoints of interest, and downstream analyses. We encourage you to take those elements of your study into account, alongside the guidance provided here, when assessing normality. Regardless of the specific steps you take, be sure to report normality assessment steps and the data transformation or statistical test decisions you make based on them in your final report or manuscript. Determining which data should go through normality testing: Values for all samples (rows) that will be going into statistical testing should be tested for normality. If you are only going to be statistically testing a subset of your data, perform the normality test on that subset. Another way to think of this is that data points that are on the same graph together and/or that have been used as input for a statistical test should be tested for normality together. Analyzing datasets with a mixture of normally and non-normally distributed variables: There are a couple of different routes you can pursue if you have a mixture of normally and non-normally distributed variables in your dataframe: Perform parametric statistical tests on the normally distributed variables and non-parametric tests on the non-normally distributed variable. Perform the statistical test across all variables that fits with the majority of the variable distributions in your dataset. Our preference is to perform one test across all variables of the same data type/endpoint (e.g., all chemical concentrations, all cytokine concentrations). Aim to choose an approach that fits best rather than perfectly. Improving efficiency for normality assessment: If you find yourself frequently performing the same normality assessment workflow, consider writing a function that will execute each normality testing step (making a histogram, making a Q-Q plot, determining Shapiro-Wilk normality variable by variable, and determining the average Shapiro-Wilk p-value across all variables) and store the results in a list for easy inspection. Concluding Remarks In conclusion, this training module serves as an introduction to and step by step tutorial for normality assessment and data transformations. Approaches described in this training module include visualizations to qualitatively assess normality, statistical tests to quantitatively assess normality, data transformation, and other distribution considerations relating to normality. These methods are an important step in data characterization and exploration prior to downstream analyses and statistical testing, and they can be applied to nearly all studies carried out in environmental health research. Additional Resources Descriptive Statistics and Normality Tests for Statistical Data STHDA Normality Test in R Normalization vs. Standardization Test Your Knowledge Use the input file provided (“Module3_3_TYKInput.xlsx”), which represents a similar dataset to the one used in the module, to answer the following questions: Are any variables normally distributed in the raw data? Does psuedo log2 transforming the values make the distributions overall more or less normally distributed? What are the average Shapiro-Wilk p-values for the raw and psuedo log2 transformed data? "],["intoduction-to-statistical-tests.html", "Intoduction to Statistical Tests Introduction to Training Module Assessing Normality &amp; Homogeneity of Variance Two-Group Visualizations and Statistical Comparisons using the T-Test Three-Group Visualizations and Statistical Comparisons using an ANOVA Regression Modeling and Visualization: Linear and Logistic Regressions Statistical Evaluations of Categorical Data using the Chi-Squared Test and Fisher’s Exact Test Concluding Remarks", " Intoduction to Statistical Tests This training module was developed by Alexis Payton, Kyle Roell, Elise Hickman, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module This training module provides a brief introduction to some of the most commonly implemented statistics and associated visualizations used in exposure science, toxicology, and environmental health studies. This module first uploads an example dataset that is similar to the data used in TAME 2.0 Module 2.3 Data Manipulation &amp; Reshaping], though it includes some expanded subject information data to allow for more example statistical tests. Then, methods to evaluate data normality are presented, including visualization-based and statistical-based approaches. Basic statistical tests discussed in this module include: T test Analysis of Variance (ANOVA) with a Tukey’s Post-Hoc test Regression Modeling (Linear and Logistic) Chi-squared test Fisher’s exact test These statistical tests are very simple, with more extensive examples and associated descriptions of statistical models in the proceeding applications-based training modules in: TAME 2.0 Module 4.4 Two-Group Comparisons &amp; Visualizations TAME 2.0 Module 4.5 Multi-Group Comparisons &amp; Visualizations TAME 2.0 Module 4.6 Advanced Multi-Group Comparisons &amp; Visualizations Script Preparations Cleaning the global environment rm(list=ls()) Installing required R packages If you already have these packages installed, you can skip this step, or you can run the below code which checks installation status for you if (!requireNamespace(&quot;tidyverse&quot;)) install.packages(&quot;tidyverse&quot;); if (!requireNamespace(&quot;car&quot;)) install.packages(&quot;car&quot;); if (!requireNamespace(&quot;ggpubr&quot;)) install.packages(&quot;ggpubr&quot;); if(!requireNamespace(&quot;effects&quot;)) install.packages(&quot;effects&quot;); Loading R packages required for this session library(tidyverse) # all tidyverse packages, including dplyr and ggplot2 library(car) # package for statistical tests library(ggpubr) # ggplot2 based plots library(effects) # for linear modeling Set your working directory setwd(&quot;/filepath to where your input files are&quot;) Importing example datasets Let’s read in our example dataset. Note that these data are similar to those used previously, except that demographic and chemical measurement data were previously merged, and a few additional columns of subject information/demographics were added to serve as more thorough examples of data for use in this training module. # Loading data full.data &lt;- read.csv(&quot;Module3_4_Input/Module3_4_InputData.csv&quot;) Let’s view the top of the first 9 columns of data in this dataframe: full.data[1:10,1:9] ## ID BMI BMIcat MAge MEdu BW GA Smoker Smoker3 ## 1 1 27.7 Overweight 22.99928 College_Degree 3180.058 34 NS Never ## 2 2 26.8 Overweight 30.05142 College_Degree 3210.823 43 S Never ## 3 3 33.2 Overweight 28.04660 College_Degree 3311.551 40 NS Never ## 4 4 30.1 Overweight 34.81796 College_Degree 3266.844 32 S Never ## 5 5 37.4 Overweight 42.68440 College_Degree 3664.088 35 NS Never ## 6 6 33.3 Overweight 24.94960 College_Degree 3328.988 40 NS Never ## 7 7 24.8 Overweight 29.54798 College_Degree 3061.949 30 NS Never ## 8 8 16.9 Underweight 24.94954 College_Degree 3332.539 38 NS Current ## 9 9 36.9 Overweight 33.58589 College_Degree 3260.482 39 NS Never ## 10 10 21.7 Normal 39.29018 College_Degree 3141.723 35 NS Current These represent the subject information/demographic data, which include the following columns: ID: subject number BMI: body mass index BMIcat: BMI &lt;= 18.5 binned as “Underweight”, 18.5 &lt; BMI &lt;= 24.5 binned as “Normal”, BMI &gt; 24.5 binned as “Overweight” MAge: maternal age in years MEdu: maternal education level; “No_HS_Degree” = “less than high school”, “No_College_Degree” = “high school or some college”, “College_Degree” = “college or greater” BW: body weight in grams GA: gestational age in weeks Smoker: “NS” = non-smoker, “S” = smoker Smoker3: “Never”, “Former”, or “Current” smoking status Let’s now view the remaining columns (columns 10-15) in this dataframe: full.data[1:10,10:15] ## DWAs DWCd DWCr UAs UCd UCr ## 1 6.426464 1.292941 51.67987 10.192695 0.7537104 42.60187 ## 2 7.832384 1.798535 50.10409 11.815088 0.9789506 41.30757 ## 3 7.516569 1.288461 48.74001 10.079057 0.1903262 36.47716 ## 4 5.906656 2.075259 50.92745 8.719123 0.9364825 42.47987 ## 5 7.181873 2.762643 55.16882 9.436559 1.4977829 47.78528 ## 6 9.723429 3.054057 51.14812 11.589403 1.6645837 38.26386 ## 7 6.268547 1.218410 52.08578 8.887948 0.6347667 39.45535 ## 8 6.718448 1.414975 54.96740 9.304968 0.6658849 45.09987 ## 9 9.074928 2.727755 55.72826 10.818153 1.6585757 42.58577 ## 10 5.771691 2.410993 47.06552 8.747217 1.7354305 34.80661 These columns represent the environmental exposure measures, including: DWAs: drinking water arsenic levels in µg/L DWCd: drinking water cadmium levels in µg/L DWCr: drinking water chromium levels in µg/L UAs: urinary arsenic levels in µg/L UCd: urinary cadmium levels in µg/L UCr: urinary chromium levels in µg/L Now that the script is prepared and the data are uploaded, we can start by asking some initial questions about the data that can be answered by running some basic statistical tests and visualizations. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: Are there statistically significant differences in BMI between non-smokers and smokers? Are there statistically significant differences in BMI between current, former, and never smokers? Is there a relationship between maternal BMI and birth weight? Are maternal age and gestational age considered potential covariates in the relationship between maternal BMI and birth weight? Are there statistically significant differences in gestational age based on whether a subject is a non-smoker or a smoker? Is there a relationship between smoking status and BMI? Assessing Normality &amp; Homogeneity of Variance Statistical test selection often relies upon whether or not the underlying data are normally distributed and that variance across the groups is the same (homogeneity of variances). Many statistical tests and methods that are commonly implemented in exposure science, toxicology, and environmental health research rely on assumptions of normality. Thus, one of the most common statistical tests to perform at the beginning of an analysis is a test for normality. As discussed in the previous module, there are a few ways to evaluate the normality of a dataset: First, you can visually gauge whether a dataset appears to be normally distributed through plots. For example, plotting data using histograms, densities, or Q-Q plots can graphically help inform if a variable’s values appear to be normally distributed or not. Second, you can evaluate normality using statistical tests, such as the Kolmogorov-Smirnov (K-S) test and Shapiro-Wilk test. When using these tests and interpreting their results, it is important to remember that the null hypothesis is that the sample distribution is normal, and a significant p-value means the distribution is non-normal. Let’s start with the first approach based on data visualizations. In this module, we’ll primarily be generating figures using the ggubr package which is specifically designed to generate ggplot2-based figures using more streamlined coding syntax. In addition, this package has statistical parameters for plotting that are useful for basic statistical analysis, especially for people with introductory experience to plotting in R. For further documentation on ggubr, click here. Let’s begin with a histogram to view the distribution of BMI data using the gghistogram() function from the ggubr package: gghistogram(data = full.data, x = &quot;BMI&quot;, bins = 20) Let’s also view the Q–Q (quantile-quantile) plot using the ggqqplot() function also from the ggubr package: ggqqplot(full.data$BMI, ylab = &quot;BMI&quot;) From these visualizations, the BMI variable appears to be normally distributed, with data centered in the middle and spreading with a distribution on both the lower and upper sides that follow typical normal data distributions. Let’s now implement the second approach based on statistical tests for normality. Here, let’s use the Shapiro-Wilk test as an example, again looking at the BMI data. shapiro.test(full.data$BMI) ## ## Shapiro-Wilk normality test ## ## data: full.data$BMI ## W = 0.99232, p-value = 0.3773 This test resulted in a p-value of 0.3773, so we cannot reject the null hypothesis (that the BMI data are normally distributed). These findings support the assumption that these data are normally distributed. Next, we’ll assess homogeneity of variance using the Levene’s test. This will be done using the leveneTest()function from the car package: # First convert the smoker variable to a factor full.data$Smoker = factor(full.data$Smoker, levels = c(&quot;NS&quot;, &quot;S&quot;)) leveneTest(BMI ~ Smoker, data = full.data) ## Levene&#39;s Test for Homogeneity of Variance (center = median) ## Df F value Pr(&gt;F) ## group 1 0.2631 0.6086 ## 198 The p value, (Pr&gt;F), is 0.6086 indicating that variance in BMI across the smoking groups is the same. Therefore, the assumptions of a t-test, including normality and homogeneity of variance, have been met. Two-Group Visualizations and Statistical Comparisons using the T-Test T-tests are commonly used to test for a significant difference between the means of two groups in normally distributed data. In this example, we will be answering Environmental Health Question 1: Are there statistically significant differences in BMI between non-smokers and smokers? We will specifically implement a two sample t-test (or independent samples t-test). Let’s first visualize the BMI data across these two groups using boxplots: ggboxplot(data = full.data, x = &quot;Smoker&quot;, y = &quot;BMI&quot;) From this plot, it looks like non-smokers (labeled “NS”) may have significantly higher BMI than smokers (labeled “S”), though we need statistical evaluation of these data to more thoroughly evaluate this potential data trend. It is easy to perform a t-test on these data using the t.test() function from the base R stats package: t.test(data = full.data, BMI ~ Smoker) ## ## Welch Two Sample t-test ## ## data: BMI by Smoker ## t = 2.5372, df = 80.362, p-value = 0.01311 ## alternative hypothesis: true difference in means between group NS and group S is not equal to 0 ## 95 percent confidence interval: ## 0.583061 4.823447 ## sample estimates: ## mean in group NS mean in group S ## 26.11176 23.40851 Answer to Environmental Health Question 1 With this, we can answer Environmental Health Question #1: Are there statistically significant differences in BMI between non-smokers and smokers? Answer: From this statistical output, we can see that the overall mean BMI in non-smokers (group “NS”) is 26.1, and the overall mean BMI in smokers (group “S”) is 23.4. We can also see that the resulting p-value comparison between the means of these two groups is, indeed, significant (p-value = 0.013), meaning that the means between these groups are significantly different (i.e., are not equal). It’s also helpful to save these results into a variable within the R global environment, which then allows us to access specific output values and extract them more easily for our records. For example, we can run the following to specifically extract the resulting p-value from this test: ttest.res &lt;- t.test(data = full.data, BMI ~ Smoker) # making a list in the R global environment with the statistical results signif(ttest.res$p.value, 2) # pulling the p-value and using the `signif` function to round to 2 significant figures ## [1] 0.013 Three-Group Visualizations and Statistical Comparisons using an ANOVA Analysis of Variance (ANOVA) is a statistical method that can be used to compare means across three or more groups in normally distributed data. To demonstrate an ANOVA test on this dataset, let’s answer Environmental Health Question 2: Are there statistically significant differences in BMI between current, former, and never smokers? To do this we’ll use the Smoker3 variable from our dataset. Let’s again start by viewing these data distributions using a boxplot: ggboxplot(data = full.data, x = &quot;Smoker3&quot;, y = &quot;BMI&quot;) From this cursory review of the data, it looks like the current smokers likely demonstrate significantly different BMI measures than the former and never smokers, though we need statistical tests to verify this potential trend. We also require statistical tests to evaluate potential differences (or lack of differences) between former and never smokers. Let’s now run the ANOVA to compare BMI between smoking groups, using the aov() function to fit an ANOVA model: smoker_anova = aov(data = full.data, BMI ~ Smoker3) smoker_anova ## Call: ## aov(formula = BMI ~ Smoker3, data = full.data) ## ## Terms: ## Smoker3 Residuals ## Sum of Squares 2046.713 6817.786 ## Deg. of Freedom 2 197 ## ## Residual standard error: 5.882861 ## Estimated effects may be unbalanced We need to extract the typical ANOVA results table using either the summary() or anova() function on the resulting fitted object: anova(smoker_anova) ## Analysis of Variance Table ## ## Response: BMI ## Df Sum Sq Mean Sq F value Pr(&gt;F) ## Smoker3 2 2046.7 1023.36 29.57 5.888e-12 *** ## Residuals 197 6817.8 34.61 ## --- ## Signif. codes: 0 &#39;***&#39; 0.001 &#39;**&#39; 0.01 &#39;*&#39; 0.05 &#39;.&#39; 0.1 &#39; &#39; 1 This table outputs a lot of information, including the F value referring to the resulting F-statistic, Pr(&gt;F) referring to the p-value of the F-statistic, and other values that are described in detail through other available resources including this helpful video through PennState’s statistics online resources. Answer to Environmental Health Question 2 With this, we can answer Environmental Health Question #2: Are there statistically significant differences in BMI between current, former, never smokers? Answer: From this ANOVA output table, we can conclude that the group means across all three groups are not equal given that the p value, written as Pr(&gt;F) is significant (p value = 5.88 x 10-12). However, it doesn’t tell us which groups differ from each other and that’s where post hoc tests like Tukey’s are useful. Let’s run a Tukey’s post hoc test using the TukeyHSD() function in base R to determine which of the current, former, and never smokers have significant differences in BMI: smoker_tukey = TukeyHSD(smoker_anova) smoker_tukey ## Tukey multiple comparisons of means ## 95% family-wise confidence level ## ## Fit: aov(formula = BMI ~ Smoker3, data = full.data) ## ## $Smoker3 ## diff lwr upr p adj ## Former-Current 7.4436765 4.203006 10.684347 0.0000005 ## Never-Current 8.1167857 5.595470 10.638102 0.0000000 ## Never-Former 0.6731092 -2.011764 3.357982 0.8245448 Although the above Tukey object contains a column p adj, those are the raw unadjusted p values. It is common practice to adjust p values from multiple comparisons to prevent the reporting of false positives or reporting of a significant difference that doesn’t actually exist (Feise, 2002). There are a couple of different methods that are used to adjust p values including the Bonferroni and the Benjamini &amp; Hochberg approaches. For this example, we’ll use the p.adjust() function to obtain the Benjamini &amp; Hochberg adjusted p values. Check out the associated RDocumentation to discover other methods that can be used to adjust p values using the p.adjust() function: # First converting the Tukey object into a dataframe smoker_tukey_df = data.frame(smoker_tukey$Smoker3) %&gt;% # renaming the `p adj` to `P Value` for clarity rename(`P Value` = p.adj) # Adding a column with the adjusted p values smoker_tukey_df$`P Adj` = p.adjust(smoker_tukey_df$`P Value`, method = &quot;fdr&quot;) smoker_tukey_df ## diff lwr upr P Value P Adj ## Former-Current 7.4436765 4.203006 10.684347 5.064863e-07 7.597295e-07 ## Never-Current 8.1167857 5.595470 10.638102 3.432921e-12 1.029876e-11 ## Never-Former 0.6731092 -2.011764 3.357982 8.245448e-01 8.245448e-01 Answer to Environmental Health Question 2 We can use this additional information to further answer Environmental Health Question #2: Are there statistically significant differences in BMI between current, former, and never smokers? Answer: Current smokers have significantly lower BMIs than people who have never smoked and people who have formerly smoked. This is made evident by the 95% confidence intervals (lwr and upr) that don’t cross 0 and the p values that are less than 0.05 even after adjusting. Regression Modeling and Visualization: Linear and Logistic Regressions Regression modeling aims to find a relationship between a dependent variable (or outcome, response, y) and an independent variable (or predictor, explanatory variable, x). There are many forms of regression analysis, but here we will focus on two: linear regression and logistic regression. In brief, linear regression is generally used when you have a continuous dependent variable and there is assumed to be some sort of linear relationship between the dependent and independent variables. Conversely, logistic regression is often used when the dependent variable is dichotomous. Let’s first run through an example linear regression model to answer Environmental Health Question 3: Is there a relationship between maternal BMI and birth weight? Linear Regression We will first visualize the data and a run simple correlation analysis to evaluate whether these data are generally correlated. Then, we will run a linear regression to evaluate the relationship between these variables in more detail. Plotting the variables against one another and adding a linear regression line using the function ggscatter() from the ggubr package: ggscatter(full.data, x = &quot;BMI&quot;, y = &quot;BW&quot;, # Adding a linear line with 95% condfidence intervals as the shaded region add = &quot;reg.line&quot;, conf.int = TRUE, # Customize reg. line add.params = list(color = &quot;blue&quot;, fill = &quot;lightgray&quot;), # Adding Pearson&#39;s correlation coefficient cor.coef = TRUE, cor.method = &quot;pearson&quot;, cor.coeff.args = list(label.sep = &quot;\\n&quot;)) We can also run a basic correlation analysis between these two variables using the cor.test() function. This function uses the Pearson’s correlation test as default, which we can implement here due to the previously discussed assumption of normality for this dataset. Note that other tests are needed in instances when data are not normally distributed (e.g., Spearman Rank). This function is used here to extract the Pearson’s correlation coefficient and p-value (which also appear above in the upper left corner of the graph): cor.res &lt;- cor.test(full.data$BW, full.data$BMI) signif(cor.res$estimate, 2) ## cor ## 0.25 signif(cor.res$p.value, 2) ## [1] 0.00039 Together, it looks like there may be an association between BW and BMI, based on these correlation results, demonstrating a significant p-value of 0.0004. To test this further, let’s run a linear regression analysis using the lm() function, using BMI (X) as the independent variable and BW as the dependent variable (Y): crude_lm &lt;- lm(data = full.data, BW ~ BMI) summary(crude_lm) # viewing the results summary ## ## Call: ## lm(formula = BW ~ BMI, data = full.data) ## ## Residuals: ## Min 1Q Median 3Q Max ## -598.39 -116.72 8.11 136.54 490.11 ## ## Coefficients: ## Estimate Std. Error t value Pr(&gt;|t|) ## (Intercept) 3069.201 52.576 58.38 &lt; 2e-16 *** ## BMI 7.208 1.997 3.61 0.000388 *** ## --- ## Signif. codes: 0 &#39;***&#39; 0.001 &#39;**&#39; 0.01 &#39;*&#39; 0.05 &#39;.&#39; 0.1 &#39; &#39; 1 ## ## Residual standard error: 188 on 198 degrees of freedom ## Multiple R-squared: 0.06176, Adjusted R-squared: 0.05702 ## F-statistic: 13.03 on 1 and 198 DF, p-value: 0.0003876 Answer to Environmental Health Question 3 With this, we can answer Environmental Health Question #3: Is there a relationship between maternal BMI and birth weight? Answer: Not only is there a slight positive correlation between maternal BMI and BW as indicated by ~0.25 correlation coefficient, this linear relationship is significant due to the p-value being ~0.0004. Additionally, we can derive confidence intervals for the BMI estimate using: confint(crude_lm)[&quot;BMI&quot;,] ## 2.5 % 97.5 % ## 3.270873 11.145740 Notice that the r-squared (R2) value in regression output is the squared value of the previously calculated correlation coefficient (R). signif(sqrt(summary(crude_lm)$r.squared), 2) ## [1] 0.25 In epidemiological studies, the potential influence of confounders is considered by including important covariates within the final regression model. Let’s go ahead and investigate Environmental Health Question 4: Are maternal age and gestational age considered potential covariates in the relationship between maternal BMI and birth weight? We can do that by adding those variables to the linear model. adjusted_lm = lm(data = full.data, BW ~ BMI + MAge + GA) summary(adjusted_lm) ## ## Call: ## lm(formula = BW ~ BMI + MAge + GA, data = full.data) ## ## Residuals: ## Min 1Q Median 3Q Max ## -454.04 -111.24 5.79 116.46 488.41 ## ## Coefficients: ## Estimate Std. Error t value Pr(&gt;|t|) ## (Intercept) 2247.995 117.946 19.060 &lt; 2e-16 *** ## BMI 6.237 1.774 3.515 0.000547 *** ## MAge 4.269 1.887 2.263 0.024752 * ## GA 19.612 2.656 7.385 4.28e-12 *** ## --- ## Signif. codes: 0 &#39;***&#39; 0.001 &#39;**&#39; 0.01 &#39;*&#39; 0.05 &#39;.&#39; 0.1 &#39; &#39; 1 ## ## Residual standard error: 165.5 on 196 degrees of freedom ## Multiple R-squared: 0.2805, Adjusted R-squared: 0.2695 ## F-statistic: 25.47 on 3 and 196 DF, p-value: 5.884e-14 Let’s further visualize these regression modeling results by adding a regression line to the original scatterplot. Before doing so, we’ll use the effect() function from the effects package to make estimated predictions of birth weight values for the crude and adjusted linear models. The crude model only has BMI as the dependent variable, while the adjusted model includes BMI, maternal age, and gestational age as dependent variables. This function creates a table that contains 5 columns: fitted values for BMI (BMI), predictor values (fit), standard errors of the predictions (se), lower confidence limits (lower), and upper confidence limits (upper). An additional column, Model, was added to specify whether the values correspond to the crude or adjusted model. For additional information on visualizing adjusted linear models, see Plotting Adjusted Associations in R. crude_lm_predtable = data.frame(effect(term = &quot;BMI&quot;, mod = crude_lm), Model = &quot;Crude&quot;) adjusted_lm_predtable = data.frame(effect(term = &quot;BMI&quot;, mod = adjusted_lm), Model = &quot;Adjusted&quot;) # Viewing one of the tables crude_lm_predtable ## BMI fit se lower upper Model ## 1 10 3141.284 33.63898 3074.948 3207.621 Crude ## 2 19 3206.159 18.54497 3169.588 3242.730 Crude ## 3 28 3271.034 14.21563 3243.000 3299.067 Crude ## 4 36 3328.700 24.86346 3279.669 3377.732 Crude ## 5 45 3393.575 41.18575 3312.356 3474.794 Crude Now we can plot each linear model and their corresponding 95% confidence intervals (CI). It’s easier to visualize this using ggplot2 instead of ggubr so that’s what we’ll use: options(repr.plot.width=9, repr.plot.height=6) # changing dimensions of the entire figure ggplot(full.data, aes(x = BMI, y = BW)) + geom_point() + # Crude line geom_line(data = crude_lm_predtable, mapping = aes(x = BMI, y = fit, color = Model)) + # Adjusted line geom_line(data = adjusted_lm_predtable, mapping = aes(x = BMI, y = fit, color = Model)) + # Crude 95% CI geom_ribbon(data = crude_lm_predtable, mapping = aes(x = BMI, y = fit, ymin = lower, ymax = upper, fill = Model), alpha = 0.25) + # Adjusted 95% CI geom_ribbon(data = adjusted_lm_predtable, mapping = aes(x = BMI, y = fit, ymin = lower, ymax = upper, fill = Model), alpha = 0.25) Answer to Environmental Health Question 4 With this, we can answer Environmental Health Question #4: Are maternal age and gestational age considered potential covariates in the relationship between maternal BMI and birth weight? Answer: BMI is still significantly associated with BW and the included covariates are also shown to be significantly related to birth weight in this model. However, the addition of gestational age and maternal age did not have much of an impact on modifying the relationship between BMI and birth weight. Logistic Regression To carry out a logistic regression, we need to evaluate one continuous variable (here, we select gestational age, using the GA variable) and one dichotomous variable (here, we select smoking status, using the Smoker variable) to evaluate Environmental Health Question 5: Are there statistically significant differences in gestational age based on whether a subject is a non-smoker or a smoker? Because smoking status is a dichotomous variable, we will use logistic regression to look at this relationship. Let’s first visualize these data using a stacked bar plot for the dichotomous smoker dataset: ggboxplot(data = full.data, x = &quot;Smoker&quot;, y = &quot;GA&quot;) With this visualization, it’s difficult to tell whether or not there are significant differences in maternal education based on smoking status. Let’s now run the statistical analysis, using logistic regression modeling: # Before running the model, &quot;Smoker&quot;, needs to be binarized to 0&#39;s or 1&#39;s for the glm function glm_data = full.data %&gt;% mutate(Smoker = ifelse(Smoker == &quot;NS&quot;, 0,1)) # Use GLM (generalized linear model) and specify the family as binomial # This tells GLM to run a logistic regression log.res = glm(Smoker ~ GA, family = &quot;binomial&quot;, data = glm_data) summary(log.res) # viewing the results ## ## Call: ## glm(formula = Smoker ~ GA, family = &quot;binomial&quot;, data = glm_data) ## ## Coefficients: ## Estimate Std. Error z value Pr(&gt;|z|) ## (Intercept) 1.26669 1.37042 0.924 0.3553 ## GA -0.06764 0.03796 -1.782 0.0747 . ## --- ## Signif. codes: 0 &#39;***&#39; 0.001 &#39;**&#39; 0.01 &#39;*&#39; 0.05 &#39;.&#39; 0.1 &#39; &#39; 1 ## ## (Dispersion parameter for binomial family taken to be 1) ## ## Null deviance: 218.10 on 199 degrees of freedom ## Residual deviance: 214.89 on 198 degrees of freedom ## AIC: 218.89 ## ## Number of Fisher Scoring iterations: 4 Similar to the regression modeling analysis, we can also derive confidence intervals: confint(log.res)[&quot;GA&quot;,] ## Waiting for profiling to be done... ## 2.5 % 97.5 % ## -0.14301103 0.00640128 Answer to Environmental Health Question 5 With this, we can answer Environmental Health Question #5: Are there statistically significant differences in maternal education level based on whether they are a non-smoker or a smoker? Answer: Collectively, these results show a non-significant p-value relating gestational age to smoking status. The confidence intervals also overlap across zero. Therefore, these data do not demonstrate a significant association between gestational age and smoking status. Statistical Evaluations of Categorical Data using the Chi-Squared Test and Fisher’s Exact Test Chi-squared test and Fisher’s exact tests are used primarily when evaluating data distributions between two categorical variables. The difference between a Chi-squared test and the Fisher’s exact test surrounds the specific procedure being run. The Chi-squared test is an approximation and is run with larger sample sizes to determine whether there is a statistically significant difference between the expected vs. observed frequencies in one or more categories of a contingency table. The Fisher’s exact test is similar, though is an exact measure that can be run on any sample size, including smaller sample sizes. The number of samples or subjects (n) considered to be sufficiently large enough is subjective, contingent upon the research question being asked, and the experimental design. However, smaller sample sizes can be more permissible if the sample is normally distributed, but generally speaking having n &gt; 30 is a common convention in statistics (Alexander, 2022). For this example, we are interested in evaluating the potential relationship between two categorical variables: smoking status (using the Smoker variable) and categorical BMI group (using the BMIcat variable) to address Environmental Health Question 6: Is there a relationship between smoking status and BMI? To run these categorical statistical tests, let’s first create and view a 2-way contingency table describing the frequencies of observations across the categorical BMI and smoking groups: ContingencyTable &lt;- with(full.data, table(BMIcat, Smoker)) ContingencyTable ## Smoker ## BMIcat NS S ## Normal 43 14 ## Overweight 87 22 ## Underweight 23 11 Now let’s run the Chi-squared test on this table: chisq.test(ContingencyTable) ## ## Pearson&#39;s Chi-squared test ## ## data: ContingencyTable ## X-squared = 2.1849, df = 2, p-value = 0.3354 Note that we can also run the Chi-squared test using the following code, without having to generate the contingency table: chisq.test(full.data$BMIcat, full.data$Smoker) ## ## Pearson&#39;s Chi-squared test ## ## data: full.data$BMIcat and full.data$Smoker ## X-squared = 2.1849, df = 2, p-value = 0.3354 Or: with(full.data, chisq.test(BMIcat, Smoker)) ## ## Pearson&#39;s Chi-squared test ## ## data: BMIcat and Smoker ## X-squared = 2.1849, df = 2, p-value = 0.3354 Answer to Environmental Health Question 6 Note that these all produce the same results. With this, we can answer Environmental Health Question #6: Is there a relationship between smoking status and BMI? Answer: This results in a p-value = 0.34, demonstrating that there is no significant relationship between BMI categories and smoking status. We can also run a Fisher’s Exact Test when considering sample sizes. We won’t run this here due to computing time, but here is some example code for your records: #With small sample sizes, can use Fisher&#39;s Exact Test #fisher.test(full.data$BMI, full.data$Smoker) Concluding Remarks In conclusion, this training module serves as a high-level introduction to basic statistics and visualization methods. Statistical approaches described in this training module include tests for normality, t-test, analysis of variance, regression modeling, chi-squared test, and Fisher’s exact test. Visualization approaches include boxplots, histograms, scatterplots, and regression lines. These methods serve as an important foundation for nearly all studies carried out in environmental health research. Test Your Knowledge If we’re interested in investigating if there are significant differences in birth weight based on maternal education level, which statistical test should you use? Is that relationship considered to be statistically significant and how can we visualize the distributions of these groups? "],["overview-of-experimental-design-and-example-data.html", "Overview of Experimental Design and Example Data Introduction to Training Module Replicates Orientation to Example Data for Chapter 4 Concluding Remarks", " Overview of Experimental Design and Example Data This training module was developed by Elise Hickman, Sarah Miller, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module Converting wet lab experimentation data into dry lab analyses facilitates reproducibility and transparency in data analysis. This is helpful for consistency across members of the same research group, review of analyses by collaborators or reviewers, and implementation of similar future analyses. In comparison with analysis workflows that use subscription- or license-based applications, such as Prism or SAS, analysis workflows that leverage open-source programming languages such as R also increase accessibility of analyses. Additionally, scripted analyses minimize the risk for copy-paste error, which can occur when cleaning experimental data, transferring it to an analysis application, and exporting and formatting analysis results. Some of the barriers in converting wet lab experimentation into dry lab analyses include data cleaning, selection and implementation of appropriate statistical tests, and reporting results. This chapter will provide introductory material guiding wet-bench scientists in R analyses, bridging the gap between commonly available R tutorials (which, while helpful, may not provide sufficient level of detail or relevant examples) and intensive data science workflows (which may be too detailed). In this module, we will provide an overview of key experimental design features and terms that will be used throughout this chapter, and we will provide a detailed overview of the example data. In the subsequent modules, we will dive into analyzing the example data. Replicates One of the most important components of selecting an appropriate analysis is first understanding how data should be compared between samples, which often means addressing experimental replicates. There are two main types of replicates that are used in environmental health research: biological replicates and technical replicates. Biological Replicates Biological replicates are the preferred unit of statistical comparison because they represent biologically distinct samples, demonstrating biological variation in the system. What is considered to be a biological replicate can depend on what model system is being used. For example, in studies with human clinical samples or cells from different human donors, the different humans are considered the biological replicates. In studies using animals as model organisms, individual animals are typically considered biological replicates, although this can vary depending on the experimental design. In studies that use cell lines, which are derived from one human or animal and are modified to continuously grow in culture, a biological replicate could be either cells from different passages (different thawed aliquots) grown in completely separate flasks, all experimented with on the same day, or repeating an experiment on the same set of cells (one thawed aliquot) but on separate experimental days, so the cells have grown/replicated between experiments. The final “N” that you report should reflect your biological replicates, or independent experiments. What constitutes an independent experiment or biological replicate is highly field-, lab-, organism-, and endpoint-dependent, so make sure to discuss this within your research group in the experiment planning phase and again before your analysis begins. No matter what you choose, ensure that when you report your results, you are transparent about what your biological replicates are. For example, the below diagram (adapted from BitesizeBio) illustrates different ways of defining replicates in experiments with cell lines: Figure 6: Created with BioRender.com N = 3 cells could be considered technical replicates if the endpoint of interest is very low throughput, such as single cell imaging or analyses. N = 3 cell culture wells is a more common approach to technical replicates and is typically used when one sample is collected from each well, such as in the case of media or cell lysate collection. Note that each well within the Week 1 biological replicate would be considered a technical replicate for Week 1’s experiment. Similarly, each well within the Week 2 biological replicate would be considered a technical replicate for Week 2’s experiment. For more on technical replicates, see the next section. Although N = 3 cell lines is a less common approach to biological replicates, some argue for this approach because each cell line is typically derived from one biological source. In this scenario, each of the cell lines would be unique but would represent the same cell type or lineage (e.g., for respiratory epithelium, A549, 16HBE, and BEAS-2B cell lines). Also note that to perform statistical analyses, an N of at least 3 biological replicates is needed, and an even higher N may be needed for a sufficiently powered study. Although power calculations are outside the scope of this module, we encourage you to use power calculation resources, such as G*Power to assist in selecting an appropriate N for your study. Technical Replicates Technical replicates are repeated measurements on the same sample or biological source, demonstrating the variation underlying protocols, equipment, and sample handling. In environmental health research, there can be technical replicates separately related to either the experimental design or the downstream analyses. Technical replicates related to experimental design refer to the chemical exposure for cell-based (in vitro) experiments, where there may be multiple wells of cells from the same passage or human/mouse exposed to the same treatment. Technical replicates related to downstream analyses refer to the endpoints that are measured after chemical exposure in each sample. To illustrate this, consider an experiment where cells from four unique human donors (D1-D4) are grown in cell culture plates, and then three wells of cells from each donor are exposed to a chemical treatment (Tx) or a vehicle control (Ctrl). The plate layout might look something like this, with technical replicates related to experimental design, i.e. chemical exposure, in the same color: Figure 7: Created with BioRender.com For this experiment, we have four biological replicates (the four donors) and three technical exposure replicates per dose (because three wells from each donor were exposed to each condition). The technical replicates here capture potential unintended variation between wells in cell growth and chemical exposure. Following the exposure of the cells to a chemical of interest, the media is collected from each well and assayed using a plate reader assay for concentrations of a marker of inflammation. For each sample collected (from each well), there are three technical replicates used to measure the concentration of the inflammatory marker. The purpose of these technical replicates is to capture potential unintended well-to-well variation in the plate reader assay. The plate layout might look something like this, with the letter and number in each well of the plate layout representing the well in the exposure plate layout that the media sample being assayed came from: Figure 8: Created with BioRender.com Technical replicates should typically be averaged before performing any statistical analysis. For the experiment described above, we would: Average the technical replicates for the plate reader assay to obtain one value per original cell culture well for inflammatory marker concentration. Then, average the technical replicates for the chemical exposure to obtain one value per biological replicate (donor). This would result in a dataset with eight values (four control and four treatment) for statistical analysis. Number and inclusion of technical replicates The above example is just one approach to experimental design. As mentioned above in the biological replicates section, selection of appropriate biological and technical replicates can vary greatly depending on your model organism, experimental design, assay, and standards in the field. For example, there may be cases where well-to-well variation for certain assays is minimal compared with variation between biological replicates, or when including technical replicates for each donor is experimentally or financially unfeasible, resulting in a lack of technical replicates. Matched Experimental Design Matching (also known as paired or repeated measures) in an experimental design is also a very important concept when selecting the appropriate statistical analysis. In experiments with matched design, multiple measurements are collected from the same biological replicate. This typically provides increased statistical power because changes are observed within each biological replicate relative to its starting point. In environmental health research, this can include study designs such as: Samples were collected from the same individuals, animals, or cell culture wells pre- and post-exposure. Cells from the same biological replicate were exposed to different doses of a chemical. The experimental design described above represents a matched design because cells from the same donor are exposed to both the treatment and the vehicle control. Orientation to Example Data for Chapter 4 In this chapter, we will be using an example dataset derived from an in vitro, or cell culture, experiment. Before diving into analysis of these data in the subsequent modules, we will provide an overview of where these data came from and preview what the input data frames look like. Experimental Design In this experiment, primary human bronchial epithelial cells (HBECs) from sixteen different donors were exposed to the gas acrolein, which is emitted from the combustion of fossil fuels, tobacco, wood, and plastic. Inhalation exposure to acrolein is associated with airway inhalation, and this study aimed to understand how exposure to acrolein changes secretion of markers of inflammation. Prior to experimentation, the HBECs were grown on a permeable membrane support for 24 days with air on one side and liquid media on the other side, allowing them to differentiate into a form that is very similar to what is found in the human body. The cells were then exposed for 2 hours to 0 (filtered air), 0.6, 1, 2, or 4 ppm acrolein, with two technical replicate wells from each donor per dose. Twenty-four hours later, the media was collected, and concentrations of inflammatory markers were measured using an enzyme-linked immunosorbent assay (ELISA). Figure 9: Created with BioRender.com Note that this is a matched experimental design because cells from every donor were exposed to every concentration of acrolein, rather than cells from different donors being exposed to each of the different doses. Starting Data Next, let’s familiarize ourselves with the data that resulted from this experiment. There are two input data files, one that contains cytokine concentration data and one that contains demographic information about the donors: The cytokine data contains information about the cytokine measurements for each of the six proteins measured in the basolateral media for each sample (units = pg/mL), which can be identified by the donor, dose, and replicate columns. The demographic data contains information about the age and sex of each donor. In the subsequent modules, we’ll be using these data to assess whether exposure to acrolein significantly changes secretion of inflammatory markers and whether donor characteristics, such as sex and age, modify these responses. Concluding Remarks This module reviewed important components of experimental design, such as replicates and matching, which are critical for data pre-processing and selecting appropriate statistical tests. Test Your Knowledge Read the following experimental design descriptions. For each description, determine the number of biological replicates (per group), the number of technical replicates, and whether the experimental design is matched. One hundred participants are recruited to a study aiming to determine whether people who use e-cigarettes have different concentrations of inflammatory markers in their airways. Fifty participants are non e-cigarette users and 50 participants are e-cigarette users. After the airway samples are collected, each sample is analyzed with an ELISA, with three measurements taken per sample. Twenty mice are used in a study aiming to understand the effects of particulate matter on cardiovascular health. The mice are randomized such that half of the mice are exposed to filtered air and half are exposed to particulate matter. During the exposures, the mice are continuously monitored for endpoints such as heart rate and heart function. One month later, the mice that were exposed to particulate matter are exposed to filtered air, and the mice that were exposed to filtered air are exposed to particulate matter, with the same cardiovascular endpoints collected. "],["data-import-processing-and-summary-statistics.html", "Data Import, Processing, and Summary Statistics Introduction to Training Module Data Import Handling Missing Values Averaging Replicates Descriptive Statistics Normality Assessment and Data Transformation Concluding Remarks", " Data Import, Processing, and Summary Statistics This training module was developed by Elise Hickman, Alexis Payton, Sarah Miller, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module The first steps in any scripted analysis of wet-bench data include importing the data, cleaning the data to prepare for analyses, and conducting preliminary data exploration steps, such as addressing missing values, calculating summary statistics, and assessing normality. Although less exciting than diving right into the statistical analysis, these steps are crucial in guiding downstream analyses and ensuring accurate results. In this module, we will discuss each of these steps and work through them using an example dataset (introduced in TAME 2.0 Module 4.1 Overview of Experimental Design and Example Data of inflammatory markers secreted by airway epithelial cells after exposure to different concentrations of acrolein. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: What is the mean concentration of each inflammatory biomarker by acrolein concentration? Are our data normally distributed? Data Import First, we need to import our data. Data can be imported into R from many different file formats, including .csv (as demonstrated in previous chapters), .txt, .xlsx, and .pdf. Often, data are formatted in Excel prior to import, and the openxlsx package provides helpful functions that allow the user to import data from Excel, create workbooks for storing results generated in R, and export data from R to Excel workbooks. Below, we will use the read.xlsx() function to import our data directly from Excel. Other useful packages include pdftools (PDF import), tm (text mining of PDFs), and plater (plate reader formatted data import). Workspace Preparation and Data Import Set working directory In preparation, first let’s set our working directory to the folder path that contains our input files: setwd(&quot;/filepath to where your input files are&quot;) Load required packages And load required packages: library(openxlsx) # for importing Excel files library(DT) # for easier viewing of data tables library(tidyverse) # for data cleaning and graphing library(imputeLCMD) # for data imputation with QRILC library(table1) # for summary table library(vtable) # for summary table library(ggpubr) # for making Q-Q plots with ggplot Import example datasets Next, let’s read in our example datasets: biomarker_data &lt;- read.xlsx(&quot;Module4_2_Input/Module4_2_InputData1.xlsx&quot;) demographic_data &lt;- read.xlsx(&quot;Module4_2_Input/Module4_2_InputData2.xlsx&quot;) View example datasets First, let’s preview our example data. Using the datatable() function from the DT package allows us to interactively scroll through our biomarker data. datatable(biomarker_data) We can see that our biomarker data are arranged with samples in rows and sample information and biomarker measurements in the columns. datatable(demographic_data) Our demographic data provide information about the donors that our cells came from, matching to the Donor column in our biomarker data. Handling Missing Values Next, we will investigate whether we have missing values and which variables and donors have missing values. # Calculate the total number of NAs per variable biomarker_data %&gt;% summarise(across(IL1B:VEGF, ~sum(is.na(.)))) ## IL1B IL6 IL8 IL10 TNFa VEGF ## 1 0 0 6 5 0 0 # Calculate the number of missing values per subject biomarker_data %&gt;% group_by(Donor) %&gt;% summarise(across(IL1B:VEGF, ~sum(is.na(.)))) ## # A tibble: 16 × 7 ## Donor IL1B IL6 IL8 IL10 TNFa VEGF ## &lt;chr&gt; &lt;int&gt; &lt;int&gt; &lt;int&gt; &lt;int&gt; &lt;int&gt; &lt;int&gt; ## 1 D1 0 0 0 0 0 0 ## 2 D10 0 0 0 2 0 0 ## 3 D11 0 0 0 0 0 0 ## 4 D12 0 0 2 0 0 0 ## 5 D13 0 0 0 3 0 0 ## 6 D14 0 0 0 0 0 0 ## 7 D15 0 0 0 0 0 0 ## 8 D16 0 0 0 0 0 0 ## 9 D2 0 0 2 0 0 0 ## 10 D3 0 0 0 0 0 0 ## 11 D4 0 0 0 0 0 0 ## 12 D5 0 0 0 0 0 0 ## 13 D6 0 0 0 0 0 0 ## 14 D7 0 0 0 0 0 0 ## 15 D8 0 0 0 0 0 0 ## 16 D9 0 0 2 0 0 0 Here, we can see that we do have a few missing values. What should we do with these values? Missing Values and Data Imputation Missing values Before deciding what to do about our missing values, it’s important to understand why they are missing. There are a few different types of missing values that could be present in a dataset: Missing completely at random (MCAR): has nothing to do with the experimental unit being studied (e.g., a sample is damaged or lost in the lab) Missing at random (MAR): there may be a systematic difference between missing and measured values, but they can be explained by observed differences in the data or experimental unit Missing not at random (MNAR): data are missing due to factors that are not observed/measured (e.g., measurement for a specific endpoint is below the limit of detection (LOD) of an assay) We know from the researchers who generated this dataset that the values are missing because these specific proteins were below the limit of detection for the assay for certain samples; therefore, our data are missing not at random. This can help us with our choice of imputation method, described below. Imputation Imputation is the assignment of a value to a missing data point by inferring that value from other properties of the dataset or externally defined limits. Whether or not you should impute your data is not a one-size-fits-all approach and may vary depending on your field, experimental design, the type of data, and the type of missing values in your dataset. Two questions you can ask yourself when deciding whether or not to impute data are: Is imputation needed for downstream analyses? Some analyses are not permissive to including NAs or 0s; others are. Will imputing values bias my analyses unnecessarily? If so, consider analyzing subsets of the data that are complete separately. There are many different imputation methods (too many to cover them all in this module); here, we will introduce a few that we use most often. We encourage you to explore these in more depth and to understand typical imputation workflows for your lab, data type, and/or discipline. For variables where imputed values are expected to be generally bound by the existing range of data (e.g., MCAR): missForest For variables with samples below the limit of detection for the assay, such as for mass spectrometry or ELISAs (e.g., MNAR) Replace non-detects with the limit of detection divided by the square root of 2 Quantile Regression Imputation of Left-Censored Data (QRILC) GSimp (can also be used to impute values above a specific threshold) If you do impute missing values, make sure to include both your raw and imputed data, along with detailed information about the imputation method, within your manuscript, supplemental information, and/or GitHub. You can even present summary statistics for both raw and imputed data for additional transparency. Imputation of Our Data Before imputing our data, it is a good idea to implement a background filter that checks to see if a certain percentage of values for each variable are missing. For variables with a very high percentage of missing values, imputation can be unreliable because there is not enough information for the imputation algorithm to reference. The threshold for what this percentage should be can vary by study design and the extent to which your data are subset into groups that may have differing biomarker profiles; however, a common threshold we frequently use is to remove variables with missing data for 25% or more of samples. We can use the following code to calculate the percentage values missing for each endpoint: biomarker_data %&gt;% summarise(across(IL1B:VEGF, ~sum(is.na(.))/nrow(biomarker_data)*100)) ## IL1B IL6 IL8 IL10 TNFa VEGF ## 1 0 0 3.75 3.125 0 0 Here, we can see that only about 3-4% of values are missing for our variables with missing data, so we will proceed to imputation with our dataset as-is. We will impute values using QRILC, which pulls from the left side of the data distribution (the lower values) to impute missing values. We will write a function that will apply QRILC imputation to our dataframe. This function takes a data frame with missing values as input and returns a dataframe with QRILC imputed values in place of NAs as output. QRILC_imputation = function(df){ # Normalize data before applying QRILC per QRILC documentation ## Select only numeric columns, log2 transform, and convert to a matrix ### 4 comes from there being 3 metadata columns before the numeric data starts QRILC_prep = df[,4:dim(df)[2]] %&gt;% mutate_all(., function(x) log2(x + 1)) %&gt;% as.matrix() imputed_QRILC_object = impute.QRILC(QRILC_prep, tune.sigma = 0.1) QRILC_log2_df = data.frame(imputed_QRILC_object[1]) # Converting back the original scale QRILC_df = QRILC_log2_df %&gt;% mutate_all(., function(x) 2^x - 1) # Adding back in metadata columns QRILC_df = cbind(Donor = df$Donor, Dose = df$Dose, Replicate = df$Replicate, QRILC_df) return(QRILC_df) } Now we can apply the QRILC function to our dataframe. We use the function set.seed() to ensure that the QRILC function generates the same numbers each time we run the script. For more on setting seeds, see here. # Set random seed to ensure reproducibility in results set.seed(1104) # Apply function biomarker_data_imp &lt;- QRILC_imputation(biomarker_data) Averaging Replicates The last step we need to take before our data are ready for analysis is averaging the two technical replicates for each donor and dose. We will do this by creating an ID column that represents the donor and dose together and using that column to group and average the data. This results in a dataframe where our rows contain data representing each biological replicate exposed to each of the five concentrations of acrolein. biomarker_data_imp_avg &lt;- biomarker_data_imp %&gt;% # Create an ID column that represents the donor and dose unite(Donor_Dose, Donor, Dose, sep = &quot;_&quot;) %&gt;% # Average replicates with each unique Donor_Dose group_by(Donor_Dose) %&gt;% summarize(across(IL1B:VEGF, mean)) %&gt;% # Round results to the same number of significant figures as the original data mutate(across(IL1B:VEGF, \\(x) round(x, 2))) %&gt;% # Separate back out the Donor_Dose column separate(Donor_Dose, into = c(&quot;Donor&quot;, &quot;Dose&quot;), sep = &quot;_&quot;) # View new data frame datatable(biomarker_data_imp_avg) Descriptive Statistics Generating descriptive statistics (e.g., mean, median, mode, range, standard deviation) can be helpful for understanding the general distribution of your data and for reporting results either in the main body of a manuscript/report (for small datasets) or in the supplementary material (for larger datasets). There are a number of different approaches that can be used to calculate summary statistics, including functions that are part of base R and that are part of packages. Here, we will demonstrate a few different ways to efficiently calculate descriptive statistics across our dataset. Method #1 - Tidyverse and Basic Functions The mean, or average of data points, is one of the most commonly reported summary statistics and is often reported as mean ± standard deviation to demonstrate the spread in the data. Here, we will make a table of mean ± standard deviation for each of our biomarkers across each of the dose groups using tidyverse functions. # Calculate means biomarker_group_means &lt;- biomarker_data_imp_avg %&gt;% group_by(Dose) %&gt;% summarise(across(IL1B:VEGF, \\(x) mean(x))) # View data datatable(biomarker_group_means) You’ll notice that there are a lot of decimal places in our calculated means, while in our original data, there are only two decimal places. We can add a step to round the data to our above code chunk to produce cleaner results. # Calculate means biomarker_group_means &lt;- biomarker_data_imp_avg %&gt;% group_by(Dose) %&gt;% summarise(across(IL1B:VEGF, \\(x) mean(x))) %&gt;% mutate(across(IL1B:VEGF, \\(x) round(x, 2))) # View data datatable(biomarker_group_means) With this, we can answer Environmental Health Question 1: What is the mean concentration of each inflammatory biomarker by acrolein concentration? Answer: With the above table, we can see the mean concentrations for each of our inflammatory biomarkers by acrolein dose. IL-8 overall has the highest concentrations, followed by VEGF and IL-6. For IL1B, IL8, TNFa, and VEGF, it appears that the concentration of the biomarker goes up with increasing dose. We can use very similar code to calculate our standard deviations: # Calculate means biomarker_group_sds &lt;- biomarker_data_imp_avg %&gt;% group_by(Dose) %&gt;% summarise(across(IL1B:VEGF, \\(x) sd(x))) %&gt;% mutate(across(IL1B:VEGF, \\(x) round(x, 1))) # View data datatable(biomarker_group_sds) Now we’ve calculated both the means and standard deviations! However, these are typically presented as mean ± standard deviation. We can merge these dataframes by executing the following steps: Pivot each dataframe to a long format, with each row containing the value for one biomarker at one dose. Create a variable that represents each unique row (combination of Dose and variable). Join the data frames by row. Unite the two columns with mean and standard deviation, with ± in between them. Pivot the dataframe wider so that the dataframe resembles what we started with for the means and standard deviations. First, we’ll pivot each dataframe to a long format and create a variable that represents each unique row. # Pivot dataframes longer and create variable column for each row biomarker_group_means_long &lt;- pivot_longer(biomarker_group_means, !Dose, names_to = &quot;variable&quot;, values_to = &quot;mean&quot;) %&gt;% unite(Dose_variable, Dose, variable, remove = FALSE) biomarker_group_sds_long &lt;- pivot_longer(biomarker_group_means, !Dose, names_to = &quot;variable&quot;, values_to = &quot;sd&quot;) %&gt;% unite(Dose_variable, Dose, variable, remove = FALSE) # Preview what dataframe looks like datatable(biomarker_group_means_long) Next, we will join the mean and standard deviation datasets. Notice that we are only joining the Dose_variable and sd columns from the standard deviation data frame to prevent duplicate columns (Dose, variable) from being included. # Merge the dataframes by row biomarker_group_summstats &lt;- left_join(biomarker_group_means_long, biomarker_group_sds_long %&gt;% select(c(Dose_variable, sd)), by = &quot;Dose_variable&quot;) # Preview the new dataframe datatable(biomarker_group_summstats) Then, we can unite the mean and standard deviation columns and add the ± symbol between them by storing that character as a variable and pasting that variable in our paste() function. # Store plus/minus character plusminus &lt;-&quot;\\u00b1&quot; Encoding(plusminus)&lt;-&quot;UTF-8&quot; # Create new column with mean +/- standard deviation biomarker_group_summstats &lt;- biomarker_group_summstats %&gt;% mutate(mean_sd = paste(mean, plusminus, sd, sep = &quot; &quot;)) # Preview the new dataframe datatable(biomarker_group_summstats) Last, we can pivot the dataframe wider to revert it to its original layout, which is more digestible to read. # Pivot dataframe wider biomarker_group_summstats &lt;- biomarker_group_summstats %&gt;% # Remove columns we don&#39;t need any more select(-c(Dose_variable, mean, sd)) %&gt;% # Pivot wider pivot_wider(id_cols = Dose, names_from = &quot;variable&quot;, values_from = &quot;mean_sd&quot;) # View final dataframe datatable(biomarker_group_summstats) These data are now in a publication-ready format that can be exported to a .txt, .csv., or .xlsx file for easy sharing. Method #2 - Applying a List of Functions Calculating our mean and standard deviation separately using tidyverse wasn’t too difficult, but what if we want to calculate other descriptive statistics, such as minimum, median, and maximum? We could use the above approach, but we would need to make a separate dataframe for each and then merge them all together. Instead, we can use the map_dfr() function from the purrr package, which is also part of tidyverse. This function takes a list of functions you want to apply to your data and applies these functions over specified columns in the data. Let’s see how it works: # Define summary functions summary_functs &lt;- lst(min, median, mean, max, sd) # Apply functions to data, grouping by dose # .id = &quot;statistic&quot; tells the function to create a column describing which statistic that row is reporting biomarker_descriptive_stats_all &lt;- map_dfr(summary_functs, ~ summarize(biomarker_data_imp_avg %&gt;% group_by(Dose), across(IL1B:VEGF, .x)), .id = &quot;statistic&quot;) # View data datatable(biomarker_descriptive_stats_all) Depending on your final goal, descriptive statistics data can then be extracted from this dataframe and cleaned up or reformatted as needed to create a publication-ready table! Other Methods There are also packages that have been developed for specifically making summary tables, such as table1 and vtable. These packages can create summary tables in HTML format, which appear nicely in R Markdown and can be copied and pasted into Word. Here, we will briefly demonstrate how these packages work, and we encourage you to explore more using the package vignettes! Table1 The table1 package makes summary tables using the function table1(), which takes the columns that you want in the rows of the table on the left side of the first argument, followed by | and then the grouping variable. The output table can be customized in a number of ways, including what summary statistics are output and whether or not statistical comparisons are run between groups (see package vignette for more details). # Get names of all of the columns to include in the table paste(names(biomarker_data_imp_avg %&gt;% select(IL1B:VEGF)), collapse=&quot; + &quot;) ## [1] &quot;IL1B + IL6 + IL8 + IL10 + TNFa + VEGF&quot; # Make the table table1(~ IL1B + IL6 + IL8 + IL10 + TNFa + VEGF | Dose, data = biomarker_data_imp_avg) Vtable The vtable package includes the function st(), which can also be used to make HTML tables (and other output formats; see out argument). For example: # HTML output st(biomarker_data_imp_avg, group = &#39;Dose&#39;) Table 1: Summary Statistics Dose 0 0.6 1 2 4 Variable N Mean SD N Mean SD N Mean SD N Mean SD N Mean SD IL1B 16 9.8 1.9 16 10 1.9 16 11 1.5 16 11 1.6 16 12 2 IL6 16 597 658 16 670 810 16 565 652 16 536 474 16 288 297 IL8 16 21987 7570 16 24703 10768 16 22607 10648 16 41687 17939 16 94439 33535 IL10 16 1.4 1.3 16 0.44 0.56 16 1.2 1.2 16 0.88 0.43 16 1.2 0.78 TNFa 16 3.2 1.8 16 3.1 1.7 16 4.1 2.4 16 5.2 2 16 6.7 2.7 VEGF 16 971 417 16 1023 486 16 1196 621 16 1754 749 16 1617 836 # Dataframe output st(biomarker_data_imp_avg, group = &#39;Dose&#39;, out = &#39;return&#39;) ## Variable N Mean SD N Mean SD N Mean SD N Mean SD N Mean ## 1 Dose 0 0.6 1 2 4 ## 2 IL1B 16 9.8 1.9 16 10 1.9 16 11 1.5 16 11 1.6 16 12 ## 3 IL6 16 597 658 16 670 810 16 565 652 16 536 474 16 288 ## 4 IL8 16 21987 7570 16 24703 10768 16 22607 10648 16 41687 17939 16 94439 ## 5 IL10 16 1.4 1.3 16 0.44 0.56 16 1.2 1.2 16 0.88 0.43 16 1.2 ## 6 TNFa 16 3.2 1.8 16 3.1 1.7 16 4.1 2.4 16 5.2 2 16 6.7 ## 7 VEGF 16 971 417 16 1023 486 16 1196 621 16 1754 749 16 1617 ## SD ## 1 ## 2 2 ## 3 297 ## 4 33535 ## 5 0.78 ## 6 2.7 ## 7 836 Similar to table1, see the package vignette for detailed information about how to customize tables using this package. Normality Assessment and Data Transformation The last step we will take before beginning to test our data for statistical differences between groups (in the next module) is to understand our data’s distribution through normality assessment. This will inform which statistical tests we will perform on our data. For more detail on normality testing, including detailed explanations of each type of normality assessment and explanations of the code underlying the following graphs and tables, see TAME Module 3.3, Normality Tests and Data Transformations. We’ll start by looking at histograms of our data for qualitative normality assessment: # Set theme theme_set(theme_bw()) # Pivot data longer to prepare for plotting biomarker_data_imp_avg_long &lt;- biomarker_data_imp_avg %&gt;% pivot_longer(-c(Donor, Dose), names_to = &quot;variable&quot;, values_to = &quot;value&quot;) # Make figure panel of histograms ggplot(biomarker_data_imp_avg_long, aes(value)) + geom_histogram(fill = &quot;gray40&quot;, color = &quot;black&quot;, binwidth = function(x) {(max(x) - min(x))/25}) + facet_wrap(~ variable, scales = &quot;free&quot;, nrow = 2) + labs(y = &quot;# of Observations&quot;, x = &quot;Value&quot;) From these histograms, we can see that IL-1B appears to be normally distributed, while the other endpoints do not appear to be normally distributed. We can also use Q-Q plots to assess normality qualitatively: ggqqplot(biomarker_data_imp_avg_long, x = &quot;value&quot;, facet.by = &quot;variable&quot;, ggtheme = theme_bw(), scales = &quot;free&quot;) With this figure panel, we can see that most of the variables have very noticeable deviations from the reference, suggesting non-normal distributions. To assess normality quantitatively, we can use the Shapiro-Wilk test. Note that the null hypothesis is that the sample distribution is normal, and a significant p-value means the distribution is non-normal. # Apply Shapiro Wilk test to dataframe shapiro_res &lt;- apply(biomarker_data_imp_avg %&gt;% select(IL1B:VEGF), 2, shapiro.test) # Create results data frame shapiro_res &lt;- do.call(rbind.data.frame, shapiro_res) # Clean data frame shapiro_res &lt;- shapiro_res %&gt;% ## Add normality conclusion mutate(normal = ifelse(p.value &lt; 0.05, F, T)) %&gt;% ## Remove columns that do not contain informative data select(c(p.value, normal)) # View cleaned up data frame datatable(shapiro_res) With this, we can answer Environmental Health Question 2: Are our data normally distributed? Answer: The results from the Shapiro-Wilk test demonstrate that the IL1B data are normally distributed, while the other variables are non-normally distributed. These results support the conclusions we made based on our qualitative assessment above with histograms and Q-Q plots. Log2 Transforming and Re-Assessing Normality Log2 transformation is a common transformation used in environmental health research and can move data closer to a normal distribution. For more on data transformation, see TAME Module 3.3, Normality Tests and Data Transformations. We will pseudo-log2 transform our data, which adds a 1 to each value before log2 transformation and ensures that resulting values are positive real numbers. Let’s see if the log2 data are more normally distributed than the raw data. # Apply log2 transformation to data biomarker_data_imp_avg_log2 &lt;- biomarker_data_imp_avg %&gt;% mutate(across(IL1B:VEGF, ~ log2(.x + 1))) Make histogram panel: # Pivot data longer and make figure panel of histograms biomarker_data_imp_avg_log2_long &lt;- biomarker_data_imp_avg_log2 %&gt;% pivot_longer(-c(Donor, Dose), names_to = &quot;variable&quot;, values_to = &quot;value&quot;) # Make histogram panel ggplot(biomarker_data_imp_avg_log2_long, aes(value)) + geom_histogram(fill = &quot;gray40&quot;, color = &quot;black&quot;, binwidth = function(x) {(max(x) - min(x))/25}) + facet_wrap(~ variable, scales = &quot;free&quot;) + labs(y = &quot;# of Observations&quot;, x = &quot;Value&quot;) Make Q-Q plot panel: ggqqplot(biomarker_data_imp_avg_log2_long, x = &quot;value&quot;, facet.by = &quot;variable&quot;, ggtheme = theme_bw(), scales = &quot;free&quot;) Run Shapiro-Wilk test: # Apply Shapiro Wilk test shapiro_res_log2 &lt;- apply(biomarker_data_imp_avg_log2 %&gt;% select(IL1B:VEGF), 2, shapiro.test) # Create results data frame shapiro_res_log2 &lt;- do.call(rbind.data.frame, shapiro_res_log2) # Clean data frame shapiro_res_log2 &lt;- shapiro_res_log2 %&gt;% ## Add normality conclusion mutate(normal = ifelse(p.value &lt; 0.05, F, T)) %&gt;% ## Remove columns that do not contain informative data select(c(p.value, normal)) # View cleaned up data frame shapiro_res_log2 ## p.value normal ## IL1B 0.250821677 TRUE ## IL6 0.001445717 FALSE ## IL8 0.017386740 FALSE ## IL10 0.018935719 FALSE ## TNFa 0.194774045 TRUE ## VEGF 0.047231367 FALSE The histograms and Q-Q plots demonstrate that the log2 data are more normally distributed than the raw data. The results from the Shapiro-Wilk test also demonstrate that the the log2 data are more normally distributed as a whole than the raw data. Overall, the p-values, even for the variables that are still non-normally distributed, are much higher. So, should we proceed with the raw data or the log2 data? This depends on what analyses we plan to do. In general, it is best to keep the data in as close to its raw format as possible, so if all of our analyses are available with a non-parametric test, we could use our raw data. However, some statistical tests do not have a non-parametric equivalent, in which case it would likely be best to use the log2-transformed data. For subsequent modules, we will proceed with the log2 data for consistency; however, choices regarding normality assessment can vary, so be sure to discuss these choices within your research group before proceeding with your analysis. For more on decisions regarding normality, see TAME Module 3.3 Normality Tests and Data Transformations. For more on parametric vs. non-parametric tests, see TAME Module 4.4 Two Group Comparisons and Visualizations and TAME Module 4.5 Multi-Group Comparisons and Visualizations. Concluding Remarks Taken together, this module demonstrates important data processing steps necessary before proceeding with between-group statistical testing, including data import, handling missing values, averaging replicates, generating descriptive statistics tables, and assessing normality. Careful consideration and description of these steps in the methods section of a manuscript or report increases reproducibility of analyses and helps to improve the accuracy and statistical validity of subsequent statistical results. Test Your Knowledge Functional endpoints from these cultures were also measured. These endpoints were: 1) Membrane Permeability (MemPerm), 2) Trans-Epithelial Electrical Resistance (TEER), 3) Ciliary Beat Frequency (CBF), and 4) Expression of Mucin (MUC5AC). Work through the same processes demonstrated in this module using the provided data (“Module4_2_TYKInput.xlsx”) to answer the following questions: How many technical replicates are there for each dose? Are there any missing values? What are the average values for each endpoint by dose? Are the raw data normally distributed? "],["data-import-from-pdf-sources.html", "Data Import from PDF Sources Introduction to Training Module Importing Data from Many Single PDFs with the Same Formatting Importing Data Stored in PDF Tables Concluding Remarks", " Data Import from PDF Sources This training module was developed by Elise Hickman, Alexis Payton, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module Most tutorials for R rely on importing .csv, .xlsx, or .txt files, but there are numerous other file formats that can store data, and these file formats can be more difficult to import into R. PDFs can be particularly difficult to interface with in R because they are not formatted with defined rows/columns/cells as is done in Excel or .csv/.txt formatting. In this module, we will demonstrate how to import data from from PDFs into R and format it such that it is amenable for downstream analyses or export as a table. Familiarity with tidyverse, for loops, and functions will make this module much more approachable, so be sure to review TAME 2.0 Modules 2.3 Data Manipulation and Reshaping and 2.4 Improving Coding Efficiencies if you need a refresher. Overview of Example Data To demonstrate import of data from PDFs, we will be leveraging two example datasets, described in more detail in their respective sections later on in the module. PDFs generated by Nanoparticle Tracking Analysis (NTA), a technique used to quantify the size and distribution of particles (such as extracellular vesicles) in a sample. We will be extracting data from an experiment in which epithelial cells were exposed to four different environmental chemicals or a vehicle control, and secreted particles were isolated and characterized using NTA. A PDF containing information about variables collected as part of a study whose samples are part of NIH’s BioLINCC Repository. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: Which chemical(s) increase and decrease the concentration of particles secreted by epithelial cells? How many variables total are available to us to request from the study whose data are store in the repository, and what are these variables? Importing Data from Many Single PDFs with the Same Formatting Getting Familiar with the Example Dataset The following example is based on extracting data from PDFs generated by Nanoparticle Tracking Analysis (NTA), a technique used to quantify the size and distribution of particles in a sample. Each PDF file is associated with one sample, and each PDF contains multiple values that we want to extract. Although this is a very specific type of data, keep in mind that this general approach can be applied to any data stored in PDF format - you will just need to make modifications based on the layout of your PDF file! For this example, we will be extracting data from 5 PDFs that are identically formatted but contain information unique to each sample. The samples represent particles isolated from epithelial cell media following an experiment where cells were exposed to four different environmental chemicals (labeled “A”, “B”, “C”, and “D”) or a vehicle control (labeled “Ctrl”). Here is what a full view of one of the PDFs looks like, with values we want to extract highlighted in yellow: Our goal is to extract these values and end up with a dataframe that looks like this, with each sample in a row and each variable in a column: If your files are not already named in a way that reflects unique sample information, such as the date of the experiment or sample ID, update your file names to contain this information before proceeding with the script. Here are the names for the example PDF files: Workspace Preparation and Data Import Installing and loading required R packages If you already have these packages installed, you can skip this step, or you can run the below code which checks installation status for you. We will be using the pdftools and tm packages to extract text from the PDF. And instead of using head() to preview dataframes, we will be using the function datatable() from the DT package. This function produces interactive tables and generates better formatting for viewing dataframes that have long character strings (like the ones we will be viewing in this section). if (!requireNamespace(&quot;pdftools&quot;)) install.packages(&quot;pdftools&quot;) if (!requireNamespace(&quot;tm&quot;)) install.packages(&quot;tm&quot;) if (!requireNamespace(&quot;DT&quot;)) install.packages(&quot;DT&quot;) if (!requireNamespace(&quot;janitor&quot;)) install.packages(&quot;janitor&quot;) Next, load the packages. library(tidyverse) library(pdftools) library(tm) library(DT) library(janitor) Initial data import from PDF files The following code stores the file names of all of the files in your directory that end in .pdf. To ensure that only PDFs of interest are imported, consider making a subfolder within your directory containing only the PDF extraction script file and the PDFs you want to extract data from. pdf_list &lt;- list.files(path = &quot;./Module4_3_Input&quot;, pattern = &quot;488.pdf$&quot;) We can see that each of our file names are now contained in the list. head(pdf_list) ## [1] &quot;20230214_0002_Expt1_A_size_488.pdf&quot; ## [2] &quot;20230214_0006_Expt1_Ctrl_size_488.pdf&quot; ## [3] &quot;20230214_0014_Expt1_C_size_488.pdf&quot; ## [4] &quot;20230214_0023_Expt1_D_size_488.pdf&quot; ## [5] &quot;20230214_0024_Expt1_B_size_488.pdf&quot; Next, we need to make a dataframe to store the extracted data. The PDF Identifier column will store the file name, and the Text column will store extracted text from the PDF. pdf_raw &lt;- data.frame(&quot;PDF Identifier&quot; = c(), &quot;Text&quot; = c()) The following code uses a for loop to loop through each file (as stored in the pdf_list vector) and extract the text from the PDF. Sometimes this code generates duplicates, so we will also remove the duplicates with distinct(). for (i in 1:length(pdf_list)){ # Iterating through each pdf file and separating each line of text document_text = pdf_text(paste(&quot;./Module4_3_Input/&quot;, pdf_list[i], sep = &quot;&quot;)) %&gt;% strsplit(&quot;\\n&quot;) # Saving the name of each PDF file and its text document = data.frame(&quot;PDF Identifier&quot; = gsub(x = pdf_list[i], pattern = &quot;.pdf&quot;, replacement = &quot;&quot;), &quot;Text&quot; = document_text, stringsAsFactors = FALSE) colnames(document) &lt;- c(&quot;PDF Identifier&quot;, &quot;Text&quot;) # Appending the new text data to the dataframe pdf_raw &lt;- rbind(pdf_raw, document) } pdf_raw &lt;- pdf_raw %&gt;% distinct() The new dataframe contains the data from all of the PDFs, with the PDF Identifier column containing the name of the input PDF file that corresponds to the text in the column next to it. datatable(pdf_raw) Extracting Variables of Interest Specific variables of interest can be extracted from the pdf_raw dataframe by filtering the dataframe for rows that contain a specific character string. This character string could be the variable of interest (if that word or set of words is unique and only occurs in that one place in the document) or a character string that occurs in the same line of the PDF as your variable of interest. Examples of both of these approaches are shown below. It is important to note that there can be different numbers of spaces in each row and after each semicolon, which will change the sep argument for each variable. For example, there are a different number of spaces after the semicolon for “Dilution Factor” than there are for “Concentration” (see above PDF screen shot for reference). We will work through an example for the first variable of interest, dilution factor, in detail. First, we can see what the dataframe looks like when we just filter rows based on keeping only rows that contain the string “Dilution Factor” in the text column using the grepl() function. dilution_factor_df &lt;- pdf_raw %&gt;% filter(grepl(&quot;Dilution Factor&quot;, Text)) datatable(dilution_factor_df) The value we are trying to extract is at the end of a long character string. We will want to use the tidyverse function separate() to isolate those values, but we need to know what part of the character string will separate the dilution factor values from the rest of the text. To determine this, we can call just one of the data cells and copy the semicolon and following spaces for use in the separate() function. # Return the value in the first row and second column. dilution_factor_df[1,2] ## [1] &quot; Temperature: 24.64 °C sensed Dilution Factor: 200&quot; Building on top of the previous code, we can now separate the dilution factor value from the rest of the text in the string. The separate() function takes an input data column and separates it into two or more columns based on the character passed to the separation argument. Here, everything before the separation string is discarded by setting the first new column to NA. Everything after the separation string will be stored in a new column called Dilution Factor, The starting Text column is removed by default. dilution_factor_df &lt;- pdf_raw %&gt;% filter(grepl(&quot;Dilution Factor&quot;, Text)) %&gt;% separate(Text, into = c(NA, &quot;Dilution Factor&quot;), sep = &quot;: &quot;) datatable(dilution_factor_df) For the “Original Concentration” variable, we filter rows by the string “pH” because the word concentration is found in multiple locations in the document. concentration_df = pdf_raw %&gt;% filter(grepl(&quot;pH&quot;, Text)) %&gt;% separate(Text, c(NA, &quot;Concentration&quot;), sep = &quot;: &quot;) datatable(concentration_df) With the dilution factor variable, there were no additional characters after the value of interest, but here, “Particles / mL” remains and needs to be removed so that the data can be used in downstream analyses. We can add an additional cleaning step to remove “Particles / mL” from the data and add the units to the column title. sep = \" P\" refers to the space before and first letter of the string to be removed. concentration_df = pdf_raw %&gt;% filter(grepl(&quot;pH&quot;, Text)) %&gt;% separate(Text, c(NA, &quot;Concentration&quot;), sep = &quot;: &quot;) %&gt;% separate(Concentration, c(&quot;Concentration (Particles/ mL)&quot;, NA), sep = &quot; P&quot;) datatable(concentration_df) Next, we want to extract size distribution data from the lower table. Note that the space in the first separate() function comes from the space between the “Number” and “Concentration” column in the string, and the space in the second separate() function comes from the space between the variable name and the number of interest. We can also convert values to numeric since they are currently stored as characters. size_distribution_df = pdf_raw %&gt;% filter(grepl(&quot;X10&quot;, Text)| grepl(&quot;X50 &quot;, Text)| grepl(&quot;X90&quot;, Text) | grepl(&quot;Mean&quot;, Text)| grepl(&quot;StdDev&quot;, Text)) %&gt;% separate(Text, c(&quot;Text&quot;, NA), sep = &quot; &quot;) %&gt;% separate(Text, c(&quot;Text&quot;, &quot;Size&quot;), sep = &quot; &quot;) %&gt;% mutate(Size = as.numeric(Size)) %&gt;% pivot_wider(names_from = Text, values_from = Size) datatable(size_distribution_df) Creating the final dataframe Now that we have created dataframes for all of the variables that we are interested in, we can join them together into one final dataframe. # Make list of all dataframes to include all_variables &lt;- list(dilution_factor_df, concentration_df, size_distribution_df) # Combine dataframes using reduce function. Sometimes, duplicate rows are generated by full_join. full_df = all_variables %&gt;% reduce(full_join, by = &quot;PDF Identifier&quot;) %&gt;% distinct() # View new dataframe datatable(full_df) For easier downstream analysis, the last step is to separate the PDF Identifier column into an informative sample ID that matches up with other experimental data. final_df &lt;- full_df %&gt;% separate(&#39;PDF Identifier&#39;, # Split sample identifier column into new columns, retaining the original column into = c(&quot;Date&quot;, &quot;FileNumber&quot;, &quot;Experiment Number&quot;, &quot;Sample_ID&quot;, &quot;Size&quot;, &quot;Wavelength&quot;), sep = &quot;_&quot;, remove = FALSE) %&gt;% select(-c(FileNumber, Size)) %&gt;% # Remove uninformative columns mutate(across(&#39;Dilution Factor&#39;:&#39;StdDev&#39;, as.numeric)) # Change variables to numeric where appropriate datatable(final_df) Let’s make a graph to help us answer Environmental Health Question 1. theme_set(theme_bw()) data_for_graphing &lt;- final_df %&gt;% clean_names() data_for_graphing$sample_id &lt;- factor(data_for_graphing$sample_id, levels = c(&quot;Ctrl&quot;, &quot;A&quot;, &quot;B&quot;, &quot;C&quot;, &quot;D&quot;)) ggplot(data_for_graphing, aes(x = sample_id, y = concentration_particles_m_l)) + geom_bar(stat = &quot;identity&quot;, fill = &quot;gray70&quot;, color = &quot;black&quot;) + ylab(&quot;Particle Concentration (Particles/mL)&quot;) + xlab(&quot;Exposure&quot;) With this, we can answer Environmental Health Question #1: Which chemical(s) increase and decrease the concentration of particles secreted by epithelial cells? Answer: Chemicals B and C appear to increase the concentration of secreted particles. However, additional replicates of this experiment are needed to assess statistical significance. Importing Data Stored in PDF Tables The above workflow is useful if you just want to extract a few specific values from PDFs, but isn’t as useful if data are already in a table format in a PDF. The tabulapdf package provides helpful functions for extracting dataframes from tables in PDF format. Getting Familiar with the Example Dataset The following example is based on extracting dataframes from a long PDF containing many individual data tables. This particular PDF came from the NIH’s BioLINCC Repository and details variables that researchers can request from the repository. Variables are part of larger datasets that contain many variables, with each dataset in a separate table. All of the tables are stored in one PDF file, and some of the tables are longer than one page (this will become relevant later on!). Similar to the first PDF workflow, remember that this is a specific example intended to demonstrate how to work through extracting data from PDFs. Modifications will need to be made for differently formatted PDFs. Here is what the first three pages of our 75-page starting PDF look like: If we zoom in a bit more on the first page, we can see that the dataset name is defined in bold above each table. This formatting is consistent throughout the PDF. The zoomed in view also allows us to see the columns and their contents more clearly. Some are more informative than others. The columns we are most interested in are listed below along with a description to guide you through the contents. Num: The number assigned to each variable in the dataset. This numbering restarts with 1 for each table. Variable: The variable name. Type: The type (or class) of the variable, either numeric or character. Label: A description of the variable and values associated with the variable. After extracting the data, we want to end up with a dataframe that contains all of the variables, their corresponding columns, and a column that indicates which dataset the variable is associated with: Workspace Preparation and Data Import Installing and loading required R packages Similar to previous sections, we need to install and load a few packages before proceeding. The tabulapdf package needs to be installed in a specific way as shown below and can sometimes be difficult to install on Macs. If errors are produced, follow the troubleshooting tips outlined in this Stack Overflow solution. # To install all of the packages except for tabulapdf if (!requireNamespace(&quot;stringr&quot;)) install.packages(&quot;stringr&quot;) if (!requireNamespace(&quot;pdftools&quot;)) install.packages(&quot;pdftools&quot;) if (!requireNamespace(&quot;rJava&quot;)) install.packages(&quot;rJava&quot;) # To install tabulapdf if (!require(&quot;remotes&quot;)) { install.packages(&quot;remotes&quot;) } library(remotes) remotes::install_github(c(&quot;ropensci/tabulizerjars&quot;, &quot;ropensci/tabulapdf&quot;), force=TRUE, INSTALL_opts = &quot;--no-multiarch&quot;) ## ## * checking for file &#39;C:\\Users\\Jessie PC\\AppData\\Local\\Temp\\Rtmp6pdi2m\\remotes57483a856b72\\ropensci-tabulizerjars-d1924e0/DESCRIPTION&#39; ... OK ## * preparing &#39;tabulizerjars&#39;: ## * checking DESCRIPTION meta-information ... OK ## * checking for LF line-endings in source and make files and shell scripts ## * checking for empty or unneeded directories ## * building &#39;tabulizerjars_1.0.1.tar.gz&#39; ## glue (1.7.0 -&gt; 1.8.0) [CRAN] ## ## There is a binary version available but the source version is later: ## binary source needs_compilation ## glue 1.7.0 1.8.0 TRUE ## ## Binaries will be installed ## package &#39;glue&#39; successfully unpacked and MD5 sums checked ## Warning: cannot remove prior installation of package &#39;glue&#39; ## Warning in file.copy(savedcopy, lib, recursive = TRUE): problem copying ## C:\\Users\\Jessie ## PC\\AppData\\Local\\R\\win-library\\4.4\\00LOCK\\glue\\libs\\x64\\glue.dll to ## C:\\Users\\Jessie PC\\AppData\\Local\\R\\win-library\\4.4\\glue\\libs\\x64\\glue.dll: ## Permission denied ## Warning: restored &#39;glue&#39; ## ## The downloaded binary packages are in ## C:\\Users\\Jessie PC\\AppData\\Local\\Temp\\Rtmp6pdi2m\\downloaded_packages ## * checking for file &#39;C:\\Users\\Jessie PC\\AppData\\Local\\Temp\\Rtmp6pdi2m\\remotes5748719d13b\\ropensci-tabulapdf-7325a8d/DESCRIPTION&#39; ... OK ## * preparing &#39;tabulapdf&#39;: ## * checking DESCRIPTION meta-information ... OK ## * checking for LF line-endings in source and make files and shell scripts ## * checking for empty or unneeded directories ## Removed empty directory &#39;tabulapdf/docs&#39; ## * building &#39;tabulapdf_1.0.5-4.tar.gz&#39; Load packages: library(tabulapdf) library(tidyverse) library(janitor) library(pdftools) library(stringr) Initial data import from PDF file The extract_tables() function automatically extracts tables from PDFs and stores them as tibbles (a specific tidyverse data structure similar to a dataframe) within a list. One table is extracted per page, even if the table spans multiple pages. This line of code can take a few seconds to run depending on the length of your PDF. tables &lt;- extract_tables(&quot;Module4_3_Input/Module4_3_InputData4.pdf&quot;, output = &quot;tibble&quot;) Glimpsing the first three elements in the tables list, we can see that each list element is a dataframe containing the columns from the PDF tables. glimpse(tables[1:3]) ## List of 3 ## $ : spc_tbl_ [30 × 7] (S3: spec_tbl_df/tbl_df/tbl/data.frame) ## ..$ Num : num [1:30] 1 NA NA NA 2 NA NA 3 NA NA ... ## ..$ Variable: chr [1:30] &quot;AAQ_1000&quot; NA NA NA ... ## ..$ Type : chr [1:30] &quot;Num&quot; NA NA NA ... ## ..$ Len : num [1:30] 8 NA NA NA 8 NA NA 8 NA NA ... ## ..$ Format : chr [1:30] &quot;2.&quot; NA NA NA ... ## ..$ Informat: chr [1:30] &quot;2.&quot; NA NA NA ... ## ..$ Label : chr [1:30] &quot;In the past 3 days, how much of the time did your asthma keep you from&quot; &quot;doing your usual activities at work, school, or at home? 0=None of the&quot; &quot;time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All&quot; &quot;of the time&quot; ... ## ..- attr(*, &quot;spec&quot;)= ## .. .. cols( ## .. .. Num = col_double(), ## .. .. Variable = col_character(), ## .. .. Type = col_character(), ## .. .. Len = col_double(), ## .. .. Format = col_character(), ## .. .. Informat = col_character(), ## .. .. Label = col_character() ## .. .. ) ## ..- attr(*, &quot;problems&quot;)=&lt;externalptr&gt; ## $ : spc_tbl_ [46 × 7] (S3: spec_tbl_df/tbl_df/tbl/data.frame) ## ..$ Num : num [1:46] 1 2 3 4 5 6 NA NA 7 NA ... ## ..$ Variable: chr [1:46] &quot;ABP_1000&quot; &quot;ABP_1010&quot; &quot;ABP_1020&quot; &quot;ABP_1030&quot; ... ## ..$ Type : chr [1:46] &quot;Num&quot; &quot;Num&quot; &quot;Num&quot; &quot;Num&quot; ... ## ..$ Len : num [1:46] 8 8 8 8 8 8 NA NA 8 NA ... ## ..$ Format : num [1:46] 2 2 2 2 2 2 NA NA 2 NA ... ## ..$ Informat: num [1:46] 2 2 2 2 2 2 NA NA 2 NA ... ## ..$ Label : chr [1:46] &quot;Are you currently retired? 1=Yes,0=No&quot; &quot;Are you retired because of asthma? 1=Yes,0=No&quot; &quot;Are you currently unemployed? 1=Yes,0=No&quot; &quot;Are you unemployed because of asthma? 1=Yes,0=No&quot; ... ## ..- attr(*, &quot;spec&quot;)= ## .. .. cols( ## .. .. Num = col_double(), ## .. .. Variable = col_character(), ## .. .. Type = col_character(), ## .. .. Len = col_double(), ## .. .. Format = col_double(), ## .. .. Informat = col_double(), ## .. .. Label = col_character() ## .. .. ) ## ..- attr(*, &quot;problems&quot;)=&lt;externalptr&gt; ## $ : spc_tbl_ [21 × 7] (S3: spec_tbl_df/tbl_df/tbl/data.frame) ## ..$ Num : num [1:21] 20 NA NA 21 NA NA 22 NA NA 23 ... ## ..$ Variable: chr [1:21] &quot;ABP_1190&quot; NA NA &quot;ABP_1200&quot; ... ## ..$ Type : chr [1:21] &quot;Num&quot; NA NA &quot;Num&quot; ... ## ..$ Len : num [1:21] 8 NA NA 8 NA NA 8 NA NA 8 ... ## ..$ Format : chr [1:21] &quot;2.&quot; NA NA &quot;2.&quot; ... ## ..$ Informat: chr [1:21] &quot;2.&quot; NA NA &quot;2.&quot; ... ## ..$ Label : chr [1:21] &quot;How much bother is the worry that you will have an asthma attack when&quot; &quot;visiting a new place? 0=No bother at all, 1=Minor irritation, 2=Slight&quot; &quot;bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery&quot; &quot;How much bother is the worry that you will catch a cold? 0=No bother at&quot; ... ## ..- attr(*, &quot;spec&quot;)= ## .. .. cols( ## .. .. Num = col_double(), ## .. .. Variable = col_character(), ## .. .. Type = col_character(), ## .. .. Len = col_double(), ## .. .. Format = col_character(), ## .. .. Informat = col_character(), ## .. .. Label = col_character() ## .. .. ) ## ..- attr(*, &quot;problems&quot;)=&lt;externalptr&gt; Exploring further, here is how each dataframe is formatted: datatable(tables[[1]]) Notice that, although the dataframe format mirrors the PDF table format, the label column is stored across multiple rows with NAs in the other columns of that row because the text was across multiple lines. In our final dataframe, we will want the entire block of text in one cell. We can also remove the “Len”, “Format”, and “Informat” columns because they are not informative and they are not found in every table. Next, we will walk through how to clean up this table using a series of steps in tidyverse. Cleaning dataframes First, we will select the columns we are interested in and use the fill() function to change the NAs in the “Num” column so that each line of text in the “Label” column has the correct “Num” value in the same row. cleaned_table1 &lt;- data.frame(tables[[1]]) %&gt;% # Extract the first table in the list # Select only the columns of interest select(c(Num, Variable, Type, Label)) %&gt;% # Change the &quot;Num&quot; column to numeric, which is required for the fill function mutate(Num = as.numeric(Num)) %&gt;% # Fill in the NAs in the &quot;Num&quot; column down the column fill(Num, .direction = &quot;down&quot;) datatable(cleaned_table1) We still need to move all of the Label text for each variable into one cell in one row instead of across multiple rows. For this, we can use the unlist() function. Here is a demonstration of how the unlist() function works using just the first variable: cleaned_table1_var1 &lt;- cleaned_table1 %&gt;% # Filter dataframe to just contain rows associated with the first variable filter(Num == 1) %&gt;% # Paste all character strings in the Label column with a space in between them into a new column called &quot;new_label&quot; mutate(new_label = paste(unlist(Label), collapse = &quot; &quot;)) datatable(cleaned_table1_var1) We now have all of the text we want in one cell, but we have duplicate rows that we don’t need. We can get rid of these rows by assigning blank values “NA” and then omitting rows that contain NAs. cleaned_table1_var1 &lt;- cleaned_table1_var1 %&gt;% mutate(across(Variable, na_if, &quot;&quot;)) %&gt;% na.omit() datatable(cleaned_table1_var1) We need to apply this code to the whole dataframe and not just one variable, so we can add group_by(Num) to our cleaning workflow, followed by the code we just applied to our filtered dataframe. cleaned_table1 &lt;- data.frame(tables[[1]]) %&gt;% # Extract the first table in the list # Select only the columns of interest select(c(Num, Variable, Type, Label)) %&gt;% # Change the &quot;Num&quot; column to numeric, which is required for the fill function mutate(Num = as.numeric(Num)) %&gt;% # Fill in the NAs in the &quot;Num&quot; column down the column fill(Num, .direction = &quot;down&quot;) %&gt;% # Group by variable number group_by(Num) %&gt;% # Unlist the text replace the text in the &quot;Label&quot; column with the unlisted text mutate(Label = paste(unlist(Label), collapse =&quot; &quot;)) %&gt;% # Make blanks in the &quot;Variable&quot; column into NAs mutate(across(Variable, na_if, &quot;&quot;)) %&gt;% # Remove rows with NAs na.omit() datatable(cleaned_table1) Ultimately, we need to clean up each dataframe in the list the same way, and we need all of the dataframes to be in one dataframe, instead of in a list. There are a couple of different ways to do this. Both rely on the code shown above for cleaning up each dataframe. Option #1 uses a for loop, while Option #2 uses application of a function on the list of dataframes. Both result in the same ending dataframe! Option #1 # Create a dataframe for storing variables variables &lt;- data.frame() # Make a for loop to format each dataframe and add it to the variables for (i in 1:length(tables)) { table &lt;- data.frame(tables[[i]]) %&gt;% select(c(Num, Variable, Type, Label)) %&gt;% mutate(Num = as.numeric(Num)) %&gt;% fill(Num, .direction = &quot;down&quot;) %&gt;% group_by(Num) %&gt;% mutate(Label = paste(unlist(Label), collapse =&quot; &quot;)) %&gt;% mutate(across(Variable, na_if, &quot;&quot;)) %&gt;% na.omit() variables &lt;- bind_rows(variables, table) } # View resulting dataframe datatable(variables) Option #2 # Write a function that applies all of the cleaning steps to an dataframe (output = cleaned dataframe) clean_tables &lt;- function(data) { data &lt;- data %&gt;% select(c(Num, Variable, Type, Label)) %&gt;% mutate(Num = as.numeric(Num)) %&gt;% fill(Num, .direction = &quot;down&quot;) %&gt;% group_by(Num) %&gt;% mutate(Label = paste(unlist(Label), collapse =&quot; &quot;)) %&gt;% mutate(across(Variable, na_if, &quot;&quot;)) %&gt;% na.omit() return(data) } # Apply the function over each table in the list of tables tables_clean &lt;- lapply(X = tables, FUN = clean_tables) # Unlist the dataframes and combine them into one dataframe tables_clean_unlisted &lt;- do.call(rbind, tables_clean) # View resulting dataframe datatable(tables_clean_unlisted) Adding Dataset Names We now have a dataframe with all of the information from the PDFs contained in one long table. However, now we need to add back in the label on top of each table. We can’t do this with the tabulapdf package because the name isn’t stored in the table. But we can use the pdftools package for this! First, we will read in the pdf using the PDF tools package. This results in a vector containing a long character string for each page of the PDF. Notice a few features of these character strings: Each line is separated by \\n Elements [1] and [2] of the vector contain the text “dataset Name:”, while element [3] does not because the third page was a continuation of the table from the second page and therefore did not have a table title. table_names &lt;- pdf_text(&quot;Module4_3_Input/Module4_3_InputData4.pdf&quot;) head(table_names[1:3]) ## [1] &quot; 10:13 Tuesday, February 11, 2020 1\\n\\n\\nData Set Name: aaaq.sas7bdat\\n\\nNum Variable Type Len Format Informat Label\\n 1 AAQ_1000 Num 8 2. 2. In the past 3 days, how much of the time did your asthma keep you from\\n doing your usual activities at work, school, or at home? 0=None of the\\n time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All\\n of the time\\n 2 AAQ_1010 Num 8 2. 2. During the past 3 days, how often have you had asthma symptoms? 0=Not\\n at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or\\n more times per day\\n 3 AAQ_1020 Num 8 2. 2. During the past 3 days, how often have you used your rescue inhaler or\\n nebulizer medication ? 0=Not at all, 1=Once per day, 2=2-3 times per day,\\n 3=4-5 times per day, 4=6 or more times per day\\n 4 AAQ_1030 Num 8 2. 2. During the past 3 days, how many total times did your asthma symptoms\\n wake you up from sleep? 0=Not at all, 1=1 time in the last 3 days, 2=2-3\\n times in the last 3 days, 3=4-5 times in the last 3 days, 4=&gt;=6 times in the\\n last 3 days\\n 5 AAQ_1040 Num 8 2. 2. How would you rate the amount of impairment you have experienced due\\n to your asthma in the past 3 days? 0=No impairment, 1=Mild impairment,\\n 2=Moderate impairment, 3=Severe impairment, 4=Very severe impairment\\n 6 AAQ_1050 Num 8 2. 2. How stressed or frightened were you by your asthma symptoms in the past\\n 3 days? 0=Not at all, 1=Mildly, 2=Moderately, 3=Severely, 4=Very\\n severely\\n 7 AAQ_1060 Num 8 2. 2. Why do you think your asthma was worse in the past 3 days compared to\\n what is normal for you? 0=I have not been worse over the past 3 days. My\\n asthma symptoms have been usual., 1=Common cold, 2=Allergies,\\n 3=Pollution or chemical irritant, 4=Too little asth\\n 8 VNUM_C Char 3 $3. $3. Visit Number (character)\\n 9 VNUM Num 8 Visit Number (numeric)\\n 10 VDATE Num 8 Number of days from Visit 1 to this visit\\n 11 RAND_ID Char 6 Randomized Master ID\\n 12 ENROLL_TYPE Char 15 Enrollment Type (Screen Fail, Randomized, Healthy Control)\\n 13 ENROLL_ORDER Num 8 Enrollment Order Number\\n&quot; ## [2] &quot; 10:13 Tuesday, February 11, 2020 2\\n\\n\\n\\nData Set Name: abp.sas7bdat\\n\\nNum Variable Type Len Format Informat Label\\n 1 ABP_1000 Num 8 2. 2. Are you currently retired? 1=Yes,0=No\\n 2 ABP_1010 Num 8 2. 2. Are you retired because of asthma? 1=Yes,0=No\\n 3 ABP_1020 Num 8 2. 2. Are you currently unemployed? 1=Yes,0=No\\n 4 ABP_1030 Num 8 2. 2. Are you unemployed because of asthma? 1=Yes,0=No\\n 5 ABP_1040 Num 8 2. 2. Do you get paid to do work? 1=Yes,0=No\\n 6 ABP_1050 Num 8 2. 2. How much does your asthma bother you at your paid work? 0=No bother\\n at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of\\n bother, 5=Makes my life a misery\\n 7 ABP_1060 Num 8 2. 2. Overall, how much does your asthma bother you when you do jobs\\n around the house? 0=No bother at all, 1=Minor irritation, 2=Slight bother,\\n 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery,\\n 0=None of these really apply to me\\n 8 ABP_1070 Num 8 2. 2. Overall, how much does your asthma bother your social life? 0=No bother\\n at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of\\n bother, 5=Makes my life a misery\\n 9 ABP_1080 Num 8 2. 2. Overall, how much does your asthma bother your personal life? 0=No\\n bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother,\\n 4=A lot of bother, 5=Makes my life a misery, 0=None of these really\\n apply to me\\n 10 ABP_1090 Num 8 2. 2. Are you involved in leisure activities, such as: walking for pleasure,\\n sports, exercise, travelling, taking vacations? 1=Yes,0=No\\n 11 ABP_1100 Num 8 2. 2. When involved in leisure activities, how much does your asthma bother\\n you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate\\n bother, 4=A lot of bother, 5=Makes my life a misery\\n 12 ABP_1110 Num 8 2. 2. Would you say that you can&#39;t do some of these sorts of things because of\\n asthma? 1=Yes,0=No\\n 13 ABP_1120 Num 8 2. 2. How much does your asthma bother you when you sleep? 0=No bother at\\n all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of\\n bother, 5=Makes my life a misery\\n 14 ABP_1130 Num 8 2. 2. How much does the cost of your asthma medicines bother you? 0=No\\n bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother,\\n 4=A lot of bother, 5=Makes my life a misery\\n 15 ABP_1140 Num 8 2. 2. Do you get free prescriptions? 1=Yes,0=No\\n 16 ABP_1150 Num 8 2. 2. How much does the inconvenience or embarrassment of taking your\\n asthma medicines bother you? 0=No bother at all, 1=Minor irritation,\\n 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life\\n a misery\\n 17 ABP_1160 Num 8 2. 2. How much do coughs and colds bother you? 0=No bother at all, 1=Minor\\n irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother,\\n 5=Makes my life a misery, 0=None of these really apply to me\\n 18 ABP_1170 Num 8 2. 2. Feeling upset is also a bother. Does your asthma make you feel anxious,\\n depressed, tired, or helpless? 1=Yes,0=No\\n 19 ABP_1180 Num 8 2. 2. Feeling upset is also a bother. Does your asthma make you feel anxious,\\n depressed, tired, or helpless? 0=No bother at all, 1=Minor irritation,\\n 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life\\n a misery\\n&quot; ## [3] &quot; 10:13 Tuesday, February 11, 2020 3\\n\\n\\nNum Variable Type Len Format Informat Label\\n 20 ABP_1190 Num 8 2. 2. How much bother is the worry that you will have an asthma attack when\\n visiting a new place? 0=No bother at all, 1=Minor irritation, 2=Slight\\n bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery\\n 21 ABP_1200 Num 8 2. 2. How much bother is the worry that you will catch a cold? 0=No bother at\\n all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of\\n bother, 5=Makes my life a misery\\n 22 ABP_1210 Num 8 2. 2. How much bother is the worry that you will let others down? 0=No bother\\n at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of\\n bother, 5=Makes my life a misery\\n 23 ABP_1220 Num 8 2. 2. How much bother is the worry that your health may get worse in the\\n future? 0=No bother at all, 1=Minor irritation, 2=Slight bother,\\n 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery\\n 24 ABP_1230 Num 8 2. 2. How much bother is the worry that you won&#39;t be able to cope with an\\n asthma attack? 0=No bother at all, 1=Minor irritation, 2=Slight bother,\\n 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery\\n 25 VNUM_C Char 3 $3. $3. Visit Number (character)\\n 26 VNUM Num 8 Visit Number (numeric)\\n 27 VDATE Num 8 Number of days from Visit 1 to this visit\\n 28 RAND_ID Char 6 Randomized Master ID\\n 29 ENROLL_TYPE Char 15 Enrollment Type (Screen Fail, Randomized, Healthy Control)\\n 30 ENROLL_ORDER Num 8 Enrollment Order Number\\n&quot; Similar to the table cleaning section, we will work through an example of extracting the text of interest from one of these character vectors, then apply the same code to all of the character vectors. First, we will select just the first element in the vector and make it into a dataframe. # Create dataframe dataset_name_df_var1 &lt;- data.frame(strsplit(table_names[1], &quot;\\n&quot;)) # Clean column name colnames(dataset_name_df_var1) &lt;- c(&quot;Text&quot;) # View dataframe datatable(dataset_name_df_var1) Next, we will extract the dataset name using the same approach used in extracting values from the nanoparticle tracking example above and assign the name to a variable. We filter by the string “Data Set Name” because this is the start of the text string in the row where our dataset name is stored and is the same across all of our datasets. # Create dataframe dataset_name_df_var1 &lt;- dataset_name_df_var1 %&gt;% filter(grepl(&quot;Data Set Name&quot;, dataset_name_df_var1$Text)) %&gt;% separate(Text, into = c(NA, &quot;dataset&quot;), sep = &quot;Data Set Name: &quot;) # Assign variable dataset_name_var1 &lt;- dataset_name_df_var1[1,1] # View variable name dataset_name_var1 ## [1] &quot;aaaq.sas7bdat&quot; Now that we have the dataset name stored as a variable, we can create a dataframe that will correspond to the rows in our variables dataframe. The challenge is that each dataset contains a different number of variables! We can determine how many rows each dataset contains by returning to our variables dataframe and calculating the number of rows associated with each dataset. The following code splits the variables dataframe into a list of dataframes by each occurrence of 1 in the “Num” column (when the numbering restarts for a new dataset). # Calculate the number of rows associated with each dataset for reference dataset_list &lt;- split(variables, cumsum(variables$Num == 1)) glimpse(dataset_list[1:3]) ## List of 3 ## $ 1:&#39;data.frame&#39;: 13 obs. of 4 variables: ## ..$ Num : num [1:13] 1 2 3 4 5 6 7 8 9 10 ... ## ..$ Variable: chr [1:13] &quot;AAQ_1000&quot; &quot;AAQ_1010&quot; &quot;AAQ_1020&quot; &quot;AAQ_1030&quot; ... ## ..$ Type : chr [1:13] &quot;Num&quot; &quot;Num&quot; &quot;Num&quot; &quot;Num&quot; ... ## ..$ Label : chr [1:13] &quot;In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, sch&quot;| __truncated__ &quot;During the past 3 days, how often have you had asthma symptoms? 0=Not at all, 1=Once per day, 2=2-3 times per d&quot;| __truncated__ &quot;During the past 3 days, how often have you used your rescue inhaler or nebulizer medication ? 0=Not at all, 1=O&quot;| __truncated__ &quot;During the past 3 days, how many total times did your asthma symptoms wake you up from sleep? 0=Not at all, 1=1&quot;| __truncated__ ... ## $ 2:&#39;data.frame&#39;: 30 obs. of 4 variables: ## ..$ Num : num [1:30] 1 2 3 4 5 6 7 8 9 10 ... ## ..$ Variable: chr [1:30] &quot;ABP_1000&quot; &quot;ABP_1010&quot; &quot;ABP_1020&quot; &quot;ABP_1030&quot; ... ## ..$ Type : chr [1:30] &quot;Num&quot; &quot;Num&quot; &quot;Num&quot; &quot;Num&quot; ... ## ..$ Label : chr [1:30] &quot;Are you currently retired? 1=Yes,0=No&quot; &quot;Are you retired because of asthma? 1=Yes,0=No&quot; &quot;Are you currently unemployed? 1=Yes,0=No&quot; &quot;Are you unemployed because of asthma? 1=Yes,0=No&quot; ... ## $ 3:&#39;data.frame&#39;: 11 obs. of 4 variables: ## ..$ Num : num [1:11] 1 2 3 4 5 6 7 8 9 10 ... ## ..$ Variable: chr [1:11] &quot;ACT_1&quot; &quot;ACT_2&quot; &quot;ACT_3&quot; &quot;ACT_4&quot; ... ## ..$ Type : chr [1:11] &quot;Num&quot; &quot;Num&quot; &quot;Num&quot; &quot;Num&quot; ... ## ..$ Label : chr [1:11] &quot;In the past 4 weeks, how much of the time did your asthma keep you from getting as much done at work, school or&quot;| __truncated__ &quot;During the past 4 weeks, how often have your had shortness of breath? 1=All of the time, 2=Most of the time, 3=&quot;| __truncated__ &quot;During the past 4 weeks, how often did your asthma symptoms wake you up at night or earlier than usual in the m&quot;| __truncated__ &quot;During the past 4 weeks, how often have you used your rescue inhaler or nebulizer medication? 1=All of the time&quot;| __truncated__ ... The number of rows in each list is the number of variables in that dataset. We can use this value in creating our dataframe of dataset names. # Store the number of rows in a variable n_rows = nrow(data.frame(dataset_list[1])) # Repeat the dataset name for the number of variables there are dataset_name_var1 = data.frame(&quot;dataset_name&quot; = rep(dataset_name_var1, times = n_rows)) # View data farme datatable(dataset_name_var1) We now have a dataframe that can be joined with our variables dataframe for the first table. We can apply this approach to each table in our original PDF using a for loop. # Make dataframe to store dataset names dataset_names &lt;- data.frame() # Create list of datasets dataset_list &lt;- split(variables, cumsum(variables$Num == 1)) # Remove elements from the table_names vector that do not contain the string &quot;Data Set Name&quot; table_names_filtered &lt;- stringr::str_subset(table_names, &#39;Data Set Name&#39;) # Populate dataset_names dataframe for (i in 1:length(table_names_filtered)) { # Get dataset name dataset_name_df &lt;- data.frame(strsplit(table_names_filtered[i], &quot;\\n&quot;)) base::colnames(dataset_name_df) &lt;- c(&quot;Text&quot;) dataset_name_df &lt;- dataset_name_df %&gt;% filter(grepl(&quot;Data Set Name&quot;, dataset_name_df$Text)) %&gt;% separate(Text, into = c(NA, &quot;dataset&quot;), sep = &quot;Data Set Name: &quot;) dataset_name &lt;- dataset_name_df[1,1] # Determine number of variables in that dataset data_set &lt;- data.frame(dataset_list[i]) n_rows = nrow(data_set) # Repeat the dataset name for the number of variables there are dataset_name = data.frame(&quot;Data Set Name&quot; = rep(dataset_name, times = n_rows)) # Bind to dataframe dataset_names &lt;- bind_rows(dataset_names, dataset_name) } # Rename column colnames(dataset_names) &lt;- c(&quot;Data Set Name&quot;) # View datatable(dataset_names) Combining Dataset Names and Variable Information Last, we will merge together the dataframe containing dataset names and variable information. # Merge together final_variable_df &lt;- cbind(dataset_names, variables) %&gt;% rename(&quot;Variable Description&quot; = &quot;Label&quot;, &quot;Variable Number Within Dataset&quot; = &quot;Num&quot;) %&gt;% clean_names() datatable(final_variable_df) We can also determine how many total variables we have, all of which are accessible via the table we just generated. # Total number of variables nrow(final_variable_df) ## [1] 1190 # Total number of variables With this, we can answer Environmental Health Question #2: How many variables total are available to us to request from the study whose data are stored in the repository, and what are these variables? Answer: There are 1190 variable available to us. We can browse through the variables, including the sub-table they were from, the type of variable they are, and how they were derived using the table we generated. Concluding Remarks This training module provides example case studies demonstrating how to import PDF data into R and clean it so that it is more useful and accessible for analyses. The approaches demonstrated in this module, though specific to our specific example data, can be adapted to many different types of PDF data. Test Your Knowledge Using the same input files that we used in part 1, “Importing Data from Many Single PDFs with the Same Formatting”, found in the Module4_3_TYKInput folder, extract the remaining variables of interest (Original Concentration and Positions Removed) from the PDFs and summarize them in one dataframe. "],["two-group-comparisons-and-visualizations.html", "Two-Group Comparisons and Visualizations Introduction to Training Module Overview of Two Group Statistical Tests Statistical vs Biological Significance Unpaired Test Example Paired Test Example Visualizing Results Concluding Remarks", " Two-Group Comparisons and Visualizations This training module was developed by Elise Hickman, Alexis Payton, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module Two group statistical comparisons, in which we want to know whether the means between two different groups are significantly different, are some of the most common statistical tests in environmental health research and even biomedical research as a field. In this training module, we will demonstrate how to run two group statistical comparisons and how to present publication-quality figures and tables of these results. We will continue to use the same example dataset as used in this chapter’s previous modules, which represents concentrations of inflammatory biomarkers secreted by airway epithelial cells after exposure to different concentrations of acrolein. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: Are there significant differences in inflammatory biomarker concentrations between cells from male and female donors at baseline? Are there significant differences in inflammatory biomarker concentrations between cells exposed to 0 and 4 ppm acrolein? Workspace Preparation and Data Import Here, we will import the processed data that we generated at the end of TAME 2.0 Module 4.2 Data Import, Processing, and Summary Statistics. These data, along with the associated demographic data, were introduced in TAME 2.0 Module 4.1 Overview of Experimental Design and Example Data. These data represent log2 concentrations of inflammatory biomarkers secreted by airway epithelial cells after exposure to four different concentrations of acrolein (plus filtered air as a control). We will also load packages that will be needed for the analysis, including previously introduced packages such as openxlsx, tidyverse, DT, and ggpubr, and additional packages relevant to statistical analysis and graphing that will be discussed in greater detail below. # Load packages library(openxlsx) library(tidyverse) library(DT) library(rstatix) library(ggpubr) # Import data biomarker_data &lt;- read.xlsx(&quot;Module4_4_Input/Module4_4_InputData1.xlsx&quot;) demographic_data &lt;- read.xlsx(&quot;Module4_4_Input/Module4_4_InputData2.xlsx&quot;) # View data datatable(biomarker_data) datatable(demographic_data) Overview of Two Group Statistical Tests Before applying statistical tests to our data, let’s first review common two-group statistical tests, their underlying assumptions, and variations on these tests. Common Tests The two most common two-group statistical tests are the… t-test (also known as the student’s t-test) and the Wilcoxon test (also known as the Wilcox test, Wilcoxon test, or Mann Whitney test) Both of these tests are testing the null hypothesis that the means of the two populations (groups) are the same; the alternative hypothesis is that they are not the same. A significant p-value means that we can reject the null hypothesis that the means of the two groups are the same. Whether or not a p-value meets criteria for significance is experiment-specific, though commonly implemented p-value filters for significance include p&lt;0.05 and p&lt;0.01. P-values can also be called alpha values, and they indicate the probability of a type I error, or false positive, where the null hypothesis is rejected despite it actually being true. On the other hand, a type II error, or false negative, occurs when the null hypothesis is not rejected when it actually should have been. Assumptions The main difference between these two tests is in the assumption about the underlying distribution of the data. T-tests assume that the data are pulled from a normal distribution, while Wilcoxon tests do not assume that the data are pulled from a normal distribution. Therefore, it is most appropriate to use a t-test when data are, in general, normally distributed and a Wilcoxon test when data are not normally distributed. Additional assumptions underlying t-tests and Wilcoxon test are: The dependent variable is continuous or ordinal (discrete, ordered values). The data is collected from a representative, random sample. T-tests also assume that: The standard deviations of the two groups are approximately equal (also called homogeneity of variance). When to Use a Parametric vs Non-Parametric Test? Deciding whether to use a parametric or non-parametric test isn’t a one size fits all approach, and the decision should be made holistically for each dataset. Typically, parametric tests should be used when the data are normally distributed, continuous, random sampled, without extreme outliers, and representative of independent samples or participants. A non-parametric test can be used when the sample size (n) is small, outliers are present in the dataset, and/or the data are not normally distributed. This decision matters more when dealing with smaller sample sizes (n&lt;10) as smaller sample sizes are more prone to being skewed, and parametric tests are more sensitive to outliers. Therefore, when dealing with a smaller n, it might be best to perform a data transformation as discussed in TAME 2.0 Module 3.3 Normality Testing &amp; Data Transformations and then perform a parametric test if more parametric assumptions are able to be met, or to use non-parametric tests. For larger sample sizes (n&gt;50), outliers can potentially be removed and the dataset can be retested for assumptions. Lastly, what’s considered “small” or “large” in regards to sample size can be subjective and should be taken into consideration within the context of the experiment. Variations Unequal Variance: When the assumption of homogeneity of variance is not met, a Welch’s t-test is generally preferred over a student’s t-test. This can be implemented easily by setting var.equal = FALSE as an argument to the function executing the t-test (e.g., t.test(), t_test()). For more on testing homogeneity of variance in R, see here. Paired vs Unpaired: Variations on the t-test and Wilcoxon test are used when the experimental design is paired (also called repeated measures or matching). This occurs when there are different treatments, exposures, or time points collected from the same biological/experimental unit. For example, cells from the same donor or passage number exposed to different concentrations of a chemical represents a paired design. Matched/paired experiments have increased power to detect significant differences because samples can be compared back to their own controls. One vs Two-Sided: A one-sided test evaluates the hypothesis that the mean of the treatment group significantly differs in a specific direction from the control. A two-sided test evaluates the hypothesis that the mean of the treatment group significantly differs from the control but does not specify a direction for that change. A two-sided test is the preferred approach and the default in R because, typically, either direction of change is possible and represents an informative finding. However, one-sided tests may be appropriate if an effect can only possibly occur in one direction. This can be implemented by setting alternative = \"one.sided\" within the statistical testing function. Which test should I choose? We provide the following flowchart to help guide your choice of statistical test to compare two groups: Figure 10: Created with BioRender.com Statistical vs Biological Significance Another important topic to discuss before proceeding to statistical testing is the true meaning of statistical significance. Statistical significance simply means that it is unlikely that the patterns being observed are due to random chance. However, just because an effect is statistically significant does not mean that it is biologically significant (i.e., has notable biological consequences). Often, there also needs to be a sufficient magnitude of effect (also called effect size) for the effects on a system to be meaningful. Although a p-value &lt; 0.05 is often considered the threshold for significance, this is just a standard threshold set to a generally “acceptable” amount of error (5%). What about a p-value of 0.058 with a very large biological effect? Accounting for effect size is also why filters such as log2 fold change are often applied alongside p-value filters in -omics based analysis. In discussions of effect size, the population size is also a consideration - a small percentage increase in a very large population can represent tens of thousands of individuals (or more). Another consideration is that we frequently do not know what magnitude of biological effect should be considered “significant.” These discussions can get complicated very quickly, and here we do not propose to have a solution to these thought experiments; rather, we recommend considering both statistical and biological significance when interpreting data. And, as stated in other sections of TAME, transparent reporting of statistical results will aid the audience in interpreting the data through their preferred perspectives. Unpaired Test Example We will start by performing a statistical test to determine whether there are significant differences in biomarker concentrations between male and female donors at baseline (0 ppm exposure). Previously we determined that the majority of our data was non-normally distributed (see TAME 2.0 Module 4.2 Data Import, Processing, and Summary Statistics), so we’ll skip testing for that assumption in this module. Based on those results, we will use the Wilcoxon test to determine if there are significant differences between groups. The Wilcoxon test does not assume homogeneity of variance, so we do not need to test for that prior to applying the test. This is an unpaired analysis because samples collected from the cells derived from male and female donor cells are different sets of cells (i.e., independent from each other). Thus, the specific statistical test applied will be the Wilcoxon Rank Sum test. First, we will filter our dataframe to only data representing the control (0 ppm) exposure: biomarker_data_malevsfemale &lt;- biomarker_data %&gt;% filter(Dose == &quot;0&quot;) Next, we need to add the demographic data to our dataframe: biomarker_data_malevsfemale &lt;- biomarker_data_malevsfemale %&gt;% left_join(demographic_data %&gt;% select(Donor, Sex), by = &quot;Donor&quot;) Here is what our data look like now: datatable(biomarker_data_malevsfemale) We can demonstrate the basic anatomy of the Wilcoxon test function wilcox.test() by running the function on just one variable. wilcox.test(IL1B ~ Sex, data = biomarker_data_malevsfemale) ## ## Wilcoxon rank sum exact test ## ## data: IL1B by Sex ## W = 29, p-value = 0.8371 ## alternative hypothesis: true location shift is not equal to 0 The p-value of 0.8371 indicates that males and females do not have significantly different concentrations of IL-1\\(\\beta\\). The wilcox.test() function is part of the pre-loaded package stats. The package rstatix provides identical statistical tests to stats but in a pipe-friendly (tidyverse-friendly) format, and these functions output results as dataframes rather than the text displayed above. biomarker_data_malevsfemale %&gt;% wilcox_test(IL1B ~ Sex) ## # A tibble: 1 × 7 ## .y. group1 group2 n1 n2 statistic p ## * &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;int&gt; &lt;int&gt; &lt;dbl&gt; &lt;dbl&gt; ## 1 IL1B F M 7 9 29 0.837 Here, we can see the exact same results as with the wilcox.test() function. For the rest of this module, we’ll proceed with using the rstatix version of statistical testing functions. Although it is simple to run the Wilcoxon test with the code above, it’s impractical for a large number of endpoints and doesn’t store the results in an organized way. Instead, we can run the Wilcoxon test over every variable of interest using a for loop. There are also other ways you could approach this, such as a function applied over a list. This for loop runs the Wilcoxon test on each endpoint, stores the results in a dataframe, and then binds together the results dataframes for each variable of interest. Note that you could easily change wilcox_test() to t_test() and add additional arguments to modify the way the statistical test is run. # Create a vector with the names of the variables you want to run the test on endpoints &lt;- colnames(biomarker_data_malevsfemale %&gt;% select(IL1B:VEGF)) # Create dataframe to store results sex_wilcoxres &lt;- data.frame() # Run for loop for (i in 1:length(endpoints)) { # Assign a name to the endpoint variable. endpoint &lt;- endpoints[i] # Run wilcox test and store in results dataframe. res_df &lt;- biomarker_data_malevsfemale %&gt;% wilcox_test(as.formula(paste0(endpoint, &quot;~ Sex&quot;, sep = &quot;&quot;))) # Bind results from this test with other tests in this loop sex_wilcoxres &lt;- rbind(sex_wilcoxres, res_df) } # View results sex_wilcoxres ## # A tibble: 6 × 7 ## .y. group1 group2 n1 n2 statistic p ## &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;int&gt; &lt;int&gt; &lt;dbl&gt; &lt;dbl&gt; ## 1 IL1B F M 7 9 29 0.837 ## 2 IL6 F M 7 9 28.5 0.791 ## 3 IL8 F M 7 9 34 0.832 ## 4 IL10 F M 7 9 43 0.252 ## 5 TNFa F M 7 9 39 0.458 ## 6 VEGF F M 7 9 30 0.916 With this, we can answer Environmental Health Question #1: Are there significant differences in inflammatory biomarker concentrations between cells from male and female donors at baseline? Answer: There are not any significant differences in concentrations of any of our biomarkers between male and female donors at baseline. Adjusting for Multiple Hypothesis Testing Above, we compared concentrations between males and females for six different endpoints or variables. Each time we run a comparison (with a p-value threshold of &lt; 0.05), we are accepting that there is a 5% chance that a significant result will actually be due to random chance and that we are rejecting the null hypothesis when it is actually true (type I error). Since we are testing six different hypotheses simultaneously, what is the probability then of observing at least one significant result due just to chance? \\[\\mathbb{P}({\\rm At Least One Significant Result}) = 1 - \\mathbb{P}({\\rm NoSignificantResults}) = 1 - (1 - 0.05)^{6} = 0.26\\] Here, we can see that we have a 26% chance of observing at least one significant result, even if all the tests are actually not significant. This chance increases as our number of endpoints increases; therefore, adjusting for multiple hypothesis testing becomes even more important with larger datasets. Many methods exist for adjusting for multiple hypothesis testing, with some of the most popular including Bonferroni, False Discovery Rate (FDR), and Benjamini-Hochberg (BH). However, opinions about when and how to adjust for multiple hypothesis testing can vary and also depend on the question you are trying to answer. For example, when there are a low number of variables (e.g., &lt; 10), it’s often not necessary to adjust for multiple hypothesis testing, and when there are many variables (e.g., 100s to 1000s), it is necessary, but what about for an intermediate number of comparisons? Whether or not to apply multiple hypothesis test correction also depends on whether each endpoint is of interest on its own or whether the analysis seeks to make general statements about all of the endpoints together and on whether reducing type I or type II error is most important in the analysis. For this analysis, we will not adjust for multiple hypothesis testing due to our relatively low number of variables. For more on multiple hypothesis testing, check out the following publications: Mohieddin J; Naser AP. “Why, When and How to Adjust Your P Values?”. Cell Journal (Yakhteh), 20, 4, 2018, 604-607. doi: 10.22074/cellj.2019.5992 PUBMID: 30124010 Feise, R.J. Do multiple outcome measures require p-value adjustment?. BMC Med Res Methodol 2, 8 (2002). https://doi.org/10.1186/1471-2288-2-8 PUBMID: 12069695 Paired Test Example To demonstrate an example of a paired two-group test, we can also determine whether exposure to 4 ppm acrolein significantly changes biomarker concentrations. This is now a paired design because each donor’s cells were exposed to both 0 and 4 ppm acrolein. To prepare the data, we will filter the dataframe to only include 0 and 4 ppm: biomarker_data_0vs4 &lt;- biomarker_data %&gt;% filter(Dose == &quot;0&quot; | Dose == &quot;4&quot;) Let’s view the dataframe. Note how the measurements for each donor are next to each other - this an important element of the default handling of the paired analysis in R. The dataframe should have the donors in the same order for the 0 and 4 ppm data. datatable(biomarker_data_0vs4) We can now run the same type of loop that we ran before, changing the independent variable in the formula to ~ Dose and adding paired = TRUE to the wilcox_test() function. # Create a vector with the names of the variables you want to run the test on endpoints &lt;- colnames(biomarker_data_0vs4 %&gt;% select(IL1B:VEGF)) # Create dataframe to store results dose_wilcoxres &lt;- data.frame() # Run for loop for (i in 1:length(endpoints)) { # Assign a name to the endpoint variable. endpoint &lt;- endpoints[i] # Run wilcox test and store in results dataframe. res_df &lt;- biomarker_data_0vs4 %&gt;% wilcox_test(as.formula(paste0(endpoint, &quot;~ Dose&quot;, sep = &quot;&quot;)), paired = TRUE) # Bind results from this test with other tests in this loop dose_wilcoxres &lt;- rbind(dose_wilcoxres, res_df) } # View results dose_wilcoxres ## # A tibble: 6 × 7 ## .y. group1 group2 n1 n2 statistic p ## &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;int&gt; &lt;int&gt; &lt;dbl&gt; &lt;dbl&gt; ## 1 IL1B 0 4 16 16 18 0.00763 ## 2 IL6 0 4 16 16 114 0.0155 ## 3 IL8 0 4 16 16 0 0.0000305 ## 4 IL10 0 4 16 16 65 0.9 ## 5 TNFa 0 4 16 16 5 0.000305 ## 6 VEGF 0 4 16 16 28 0.0386 Although this dataframe contains useful information about our statistical test, such as the groups being compared, the sample size (n) of each group, and the test statistic, what we really want (and what would likely be shared in supplemental material), is a more simplified version of these results in table format and more detailed information (n, specific statistical test, groups being compared) in the table legend. We can clean up the results using the following code to make clearer column names and ensure that the p-values are formatted consistently. dose_wilcoxres &lt;- dose_wilcoxres %&gt;% select(c(.y., p)) %&gt;% mutate(p = format(p, digits = 3, scientific = TRUE)) %&gt;% rename(&quot;Variable&quot; = &quot;.y.&quot;, &quot;P-Value&quot; = &quot;p&quot;) datatable(dose_wilcoxres) With this, we can answer Environmental Health Question #2: Are there significant differences in inflammatory biomarker concentrations between cells exposed to 0 and 4 ppm acrolein? Answer: Yes, there are significant differences in IL-1\\(\\beta\\), IL-6, IL-8, TNF-\\(\\alpha\\), and VEGF concentrations between cells exposed to 0 and 4 ppm acrolein. Visualizing Results Now, let’s visualize our results using ggplot2. For an introduction to ggplot2 visualizations, see TAME 2.0 Modules 3.1 Data Visualizations and 3.2 Improving Data Visualizations, as well as the extensive online documentation available for ggplot2. Single Plots We will start by making a very basic box and whisker plot of the IL-1\\(\\beta\\) data with individual data points overlaid. It is best practice to show all data points, allowing the reader to view the whole spread of the data, which can be obscured by plots such as bar plots with mean and standard error. # Setting theme for plot theme_set(theme_bw()) # Making plot ggplot(biomarker_data_0vs4, aes(x = Dose, y = IL1B)) + geom_boxplot() + geom_jitter(position = position_jitter(0.15)) We could add statistical markings to denote significance to this graph manually in PowerPoint or Adobe Illustrator, but there are actually R packages that act as extensions to ggplot2 and will do this for you! Two of our favorites are ggpubr and ggsignif. Here is an example using ggpubr: ggplot(biomarker_data_0vs4, aes(x = Dose, y = IL1B)) + geom_boxplot() + geom_jitter(position = position_jitter(0.15)) + # Adding a p value from a paired Wilcoxon test stat_compare_means(method = &quot;wilcox.test&quot;, paired = TRUE) We can further clean up our figure by modifying elements of the plot’s theme, including the font sizes, axis range, colors, and the way that the statistical results are presented. Perfecting figures can be time consuming but ultimately worth it, because clear figures aid greatly in presenting a coherent story that is understandable to readers/listeners. ggplot(biomarker_data_0vs4, aes(x = Dose, y = IL1B)) + # outlier.shape = NA removes outliers geom_boxplot(aes(fill = Dose), outlier.shape = NA) + # Changing box plot colors scale_fill_manual(values = c(&quot;#BFBFBF&quot;, &quot;#EE2B2B&quot;)) + geom_jitter(size = 3, position = position_jitter(0.15)) + # Adding a p value from a paired Wilcoxon test stat_compare_means(method = &quot;wilcox.test&quot;, paired = TRUE, # Changing the value to asterisks and moving to the middle of the plot label = &quot;p.signif&quot;, label.x = 1.5, label.y = 4.5, size = 12) + ylim(2.5, 5) + # Changing y axis label labs(y = &quot;Log2(IL-1\\u03B2 (pg/mL))&quot;) + # Removing legend theme(legend.position = &quot;none&quot;, axis.title = element_text(color = &quot;black&quot;, size = 15), axis.title.x = element_text(vjust = -0.75), axis.title.y = element_text(vjust = 2), axis.text = element_text(color = &quot;black&quot;, size = 12)) Multiple plots Making one plot was relatively straightforward, but to graph all of our endpoints, we would either need to repeat that code chunk for each individual biomarker or write a function to create similar plots given a specific biomarker as input. Then, we would need to stitch together the individual plots in external software or using a package such as patchwork (which is a great package if you need to combine individual figures from different sources or different size ratios!). While these are workable solutions and would get us to the same place, ggplot2 actually contains a function - facet_wrap() - that can be used to graph multiple endpoints from the same groups in one figure panel, which takes care of a lot of the work for us! To prepare our data for facet plotting, first we will pivot it longer: biomarker_data_0vs4_long &lt;- biomarker_data_0vs4 %&gt;% pivot_longer(-c(Donor, Dose), names_to = &quot;variable&quot;, values_to = &quot;value&quot;) datatable(biomarker_data_0vs4_long) Then, we can use similar code to what we used to make our single graph, with a few modifications to plot multiple panels simultaneously and adjust the style of the plot. Although it is beyond the scope of this module to explain the mechanics of each line of code, here are a few specific things to note about the code below that may be helpful when constructing similar plots: To create the plot with all six endpoints instead of just one, we: Changed input dataframe from wide to long format Changed y = from one specific endpoint to value Added the facet_wrap() argument ~ variable tells the function to make an individual plot for each variable nrow = 2 tells the function to put the plots into two rows scales = \"free_y\" tells the function to allow each individual graph to have a unique y-scale that best shows all of the data on that graph labeller feeds the edited (more stylistically correct) names for each panel to the function To ensure that the statistical results appear cleanly, within stat_compare_means(), we: Added hide.ns = TRUE so that only significant results are shown Added label.x.npc = \"center\" and hjust = 0.5 to ensure that asterisks are centered on the plot and that the text is center justified To add padding along the y axis, allowing space for significance asterisks, we added scale_y_continuous(expand = expansion(mult = c(0.1, 0.4))) # Create clean labels for the graph titles new_labels &lt;- c(&quot;IL10&quot; = &quot;IL-10&quot;, &quot;IL1B&quot; = &quot;IL-1\\u03B2 &quot;, &quot;IL6&quot; = &quot;IL-6&quot;, &quot;IL8&quot; = &quot;IL-8&quot;, &quot;TNFa&quot; = &quot;TNF-\\u03b1&quot;, &quot;VEGF&quot; = &quot;VEGF&quot;) # Make graph ggplot(biomarker_data_0vs4_long, aes(x = Dose, y = value)) + # outlier.shape = NA removes outliers geom_boxplot(aes(fill = Dose), outlier.shape = NA) + # Changing box plot colors scale_fill_manual(values = c(&quot;#BFBFBF&quot;, &quot;#EE2B2B&quot;)) + geom_jitter(size = 1.5, position = position_jitter(0.15)) + # Adding a p value from a paired Wilcoxon test stat_compare_means(method = &quot;wilcox.test&quot;, paired = TRUE, # Changing the value to asterisks and moving to the middle of the plot label = &quot;p.signif&quot;, size = 10, hide.ns = TRUE, label.x.npc = &quot;center&quot;, hjust = 0.5) + # Adding padding y axis scale_y_continuous(expand = expansion(mult = c(0.1, 0.4))) + # Changing y axis label ylab(expression(Log[2]*&quot;(Concentration (pg/ml))&quot;)) + # Faceting by each biomarker facet_wrap(~ variable, nrow = 2, scales = &quot;free_y&quot;, labeller = labeller(variable = new_labels)) + # Removing legend theme(legend.position = &quot;none&quot;, axis.title = element_text(color = &quot;black&quot;, size = 12), axis.title.x = element_text(vjust = -0.75), axis.title.y = element_text(vjust = 2), axis.text = element_text(color = &quot;black&quot;, size = 10), strip.text = element_text(size = 12, face = &quot;bold&quot;)) An appropriate title for this figure could be: “Figure X. Exposure to 4 ppm acrolein increases inflammatory biomarker secretion in primary human bronchial epithelial cells. Groups were compared using the Wilcoxon signed rank test. * p &lt; 0.05, ** p &lt; 0.01, *** p &lt; 0.001, **** p &lt; 0.0001, n = 16 per group (paired).” Concluding Remarks In this module, we introduced two-group statistical tests, which are some of the most common statistical tests applied in biomedical research. We applied these tests to our example dataset and demonstrated how to produce publication-quality tables and figures of our results. Implementing a workflow such as this enables efficient analysis of wet-bench generated data and customization of output figures and tables suited to your personal preferences. Test Your Knowledge Functional endpoints from these cultures were also measured. These endpoints were: 1) Membrane Permeability (MemPerm), 2) Trans-Epithelial Electrical Resistance (TEER), 3) Ciliary Beat Frequency (CBF), and 4) Expression of Mucin (MUC5AC). These data were already processed and tested for normality (see Test Your Knowledge for TAME 2.0 Module 4.2 Data Import, Processing, and Summary Statistics), with results indicating that two of the endpoints are normally distributed and two non-normally distributed. Due to the relatively low n of this dataset, we therefore recommend using non-parametric statistical tests. Use the same processes demonstrated in this module and the provided data (“Module4_4_TYKInput1.xlsx” (functional data) and “Module4_4_TYKInput2.xlsx” (demographic data)), run analyses and make publication-quality figures and tables to answer the following questions to determine: Are there significant differences in functional endpoints between cells from male and female donors at baseline? Are there significant differences in functional endpoints between cells exposed to 0 and 4 ppm acrolein? Go ahead and use non-parametric tests for these analyses. "],["multi-group-and-multi-variable-comparisons-and-visualizations.html", "Multi-Group and Multi-Variable Comparisons and Visualizations Introduction to Training Module Overview of Multi-Group Statistical Tests Multi-Group Analysis Example Visualization of Multi-Group Statistical Results Concluding Remarks", " Multi-Group and Multi-Variable Comparisons and Visualizations This training module was developed by Elise Hickman, Alexis Payton, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module In the previous module, we covered how to apply two-group statistical testing, one of the most basic types of statistical tests. In this module, we will build on the concepts introduced previously to apply statistical testing to datasets with more than two groups, which are also very common in environmental health research. We will review common multi-group overall effects tests and post-hoc tests, and we will demonstrate how to apply these tests and how to graph the results using the same example dataset as in previous modules in this chapter, which represents concentrations of inflammatory biomarkers secreted by airway epithelial cells after exposure to different concentrations of acrolein. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: Are there significant differences in inflammatory biomarker concentrations between different doses of acrolein? Do TNF-\\(\\alpha\\) concentrations significantly increase with increasing dose of acrolein? Workspace Preparation and Data Import Here, we will import the processed data that we generated at the end of TAME 2.0 Module 4.2, introduced in TAME 2.0 Module 4.1 Overview of Experimental Design and Example Data and the associated demographic data. These data represent log2 concentrations of inflammatory biomarkers secreted by airway epithelial cells after exposure to four different concentrations of acrolein (plus filtered air as a control). We will also load packages that will be needed for the analysis, including previously introduced packages such as openxlsx, tidyverse, DT, ggpubr, and rstatix. # Load packages library(openxlsx) library(tidyverse) library(DT) library(rstatix) library(ggpubr) # Import data biomarker_data &lt;- read.xlsx(&quot;Module4_5_Input/Module4_5_InputData1.xlsx&quot;) demographic_data &lt;- read.xlsx(&quot;Module4_5_Input/Module4_5_InputData2.xlsx&quot;) # View data datatable(biomarker_data) datatable(demographic_data) Overview of Multi-Group Statistical Tests Before applying statistical tests to our data, let’s first review the mechanics of multi-group statistical tests, including overall effects tests and post-hoc tests. Figure 11: Created with BioRender.com Overall Effects Tests The first step for multi-group statistical testing is to run an overall effects test. The null hypothesis for the overall effects test is that there are no differences among group means. A significant p-value rejects the null hypothesis that the groups are drawn from populations with the same mean and indicates that at least one group differs significantly from the overall mean. Similar to two-group statistical testing, choice of the specific overall statistical test to run depends on whether the data are normally or non-normally distributed and whether the experimental design is paired: Figure 12: Created with BioRender.com Importantly, overall effects tests return one p-value regardless of the number of groups being compared. To determine which pairwise comparisons are significant, post-hoc testing is needed. Post-Hoc Testing If significance is obtained with an overall effects test, we can use post-hoc testing to determine which specific pairs of groups are significantly different from each other. Just as with two group statistical tests and overall effects multi-group statistical tests, choosing the appropriate post-hoc test depends on the data’s normality and whether the experimental design is paired: Figure 13: Created with BioRender.com Note that the above diagram represents commonly selected post-hoc tests; others may also be appropriate depending on your specific experimental design. As with other aspects of the analysis, be sure to report which post-hoc test(s) you performed! Correcting for Multiple Hypothesis Testing Correcting for multiple hypothesis testing is important for both the overall effects test (if you are running it over many endpoints) and post-hoc tests; however, it is particularly important for post-hoc tests. This is because even an analysis of a relatively small number of experimental groups results in quite a few pairwise comparisons. Comparing each of our five dose groups to each other in our example data, there are 10 separate statistical tests being performed! Therefore, it is generally advisable to adjust pairwise post-hoc testing p-values. The Tukey’s HSD function within rstatix does this automatically, while pairwise t-tests, pairwise Wilcoxon tests, and Dunn’s test do not. P-value adjustment can be added to their respective rstatix functions using the p.adjust.method = argument. When applying a post-hoc test, you may choose to compare every group to every other group, or you may only be interested in significant differences between specific groups (e.g., treatment groups vs. a control). This choice will be governed by your hypothesis. Statistical testing functions will typically default to comparing all groups to each other, but the comparisons can be defined using the comparisons = argument if you want to restrict the test to specific comparisons. It is important to decide at the beginning of your analysis which comparisons are relevant to your hypothesis because the number of pairwise tests performed in the post-hoc analysis will influence how much the resulting p-values will be adjusted for multiple hypothesis testing. Which test should I choose? Use the following flowchart to help guide your choice of statistical test to compare multiple groups: Figure 14: Created with BioRender.com Multi-Group Analysis Example To determine whether there are significant differences across all of our doses, the Friedman test is the most appropriate due to our matched experimental design and non-normally distributed data. The friedman_test() function is part of the rstatix package. This package also has many other helpful functions for statistical tests that are pipe/tidyverse friendly. To demonstrate how this test works, we will first perform the test on one variable: biomarker_data %&gt;% friedman_test(IL1B ~ Dose | Donor) ## # A tibble: 1 × 6 ## .y. n statistic df p method ## * &lt;chr&gt; &lt;int&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;chr&gt; ## 1 IL1B 16 12.5 4 0.0140 Friedman test A p-value of 0.01 indicates that we can reject the null hypothesis that all of our data are drawn from groups that have equivalent means. Now, we can run a for loop similar to our two-group comparisons in TAME 2.0 Module 4.4, Two Group Comparisons and Visualizations to determine the overall p-value for each endpoint: # Create a vector with the names of the variables you want to run the test on endpoints &lt;- colnames(biomarker_data %&gt;% select(IL1B:VEGF)) # Create data frame to store results dose_friedmanres &lt;- data.frame() # Run for loop for (i in 1:length(endpoints)) { # Assign a name to the endpoint variable. endpoint &lt;- endpoints[i] # Run wilcox test and store in results data frame. res &lt;- biomarker_data %&gt;% friedman_test(as.formula(paste0(endpoint, &quot;~ Dose | Donor&quot;, sep = &quot;&quot;))) %&gt;% select(c(.y., p)) dose_friedmanres &lt;- rbind(dose_friedmanres, res) } # View results datatable(dose_friedmanres) These results demonstrate that all of our endpoints have significant overall differences across doses (p &lt; 0.05). To determine which pairwise comparisons are significant, we next need to apply a post-hoc test. We will apply a pairwise, paired Wilcoxon test due to our experimental design and data distribution, with the Benjamini-Hochberg (BH) correction for multiple testing: dose_wilcox_posthoc_IL1B &lt;- biomarker_data %&gt;% pairwise_wilcox_test(IL1B ~ Dose, paired = TRUE, p.adjust.method = &quot;BH&quot;) dose_wilcox_posthoc_IL1B ## # A tibble: 10 × 9 ## .y. group1 group2 n1 n2 statistic p p.adj p.adj.signif ## * &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;int&gt; &lt;int&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;chr&gt; ## 1 IL1B 0 0.6 16 16 17 0.006 0.025 * ## 2 IL1B 0 1 16 16 11 0.002 0.017 * ## 3 IL1B 0 2 16 16 21 0.013 0.033 * ## 4 IL1B 0 4 16 16 18 0.008 0.025 * ## 5 IL1B 0.6 1 16 16 38 0.13 0.186 ns ## 6 IL1B 0.6 2 16 16 31 0.058 0.096 ns ## 7 IL1B 0.6 4 16 16 25 0.025 0.05 * ## 8 IL1B 1 2 16 16 43 0.211 0.234 ns ## 9 IL1B 1 4 16 16 40 0.159 0.199 ns ## 10 IL1B 2 4 16 16 62 0.782 0.782 ns Here, we can now see whether there are statistically significant differences in IL-1\\(\\beta\\) secretion between each of our doses. To generate pairwise comparison results for each of our inflammatory biomarkers, we can run a for loop similar to the one we ran for our overall test: # Create a vector with the names of the variables you want to run the test on endpoints &lt;- colnames(biomarker_data %&gt;% select(IL1B:VEGF)) # Create data frame to store results dose_wilcox_posthoc &lt;- data.frame() # Run for loop for (i in 1:length(endpoints)) { # Assign a name to the endpoint variable. endpoint &lt;- endpoints[i] # Run wilcox test and store in results data frame. res &lt;- biomarker_data %&gt;% pairwise_wilcox_test(as.formula(paste0(endpoint, &quot;~ Dose&quot;, sep = &quot;&quot;)), paired = TRUE, p.adjust.method = &quot;BH&quot;) dose_wilcox_posthoc &lt;- rbind(dose_wilcox_posthoc, res) } # View results datatable(dose_wilcox_posthoc) We now have a dataframe storing all of our pairwise comparison results. However, this is a lot to scroll through, making it hard to interpret. We can generate a publication-quality table by manipulating the table and joining it with the overall test data. dose_results_cleaned &lt;- dose_wilcox_posthoc %&gt;% unite(comparison, group1, group2, sep = &quot; vs. &quot;) %&gt;% select(c(.y., comparison, p.adj)) %&gt;% pivot_wider(id_cols = &quot;.y.&quot;, names_from = &quot;comparison&quot;, values_from = &quot;p.adj&quot;) %&gt;% left_join(dose_friedmanres, by = &quot;.y.&quot;) %&gt;% relocate(p, .after = &quot;.y.&quot;) %&gt;% rename(&quot;Variable&quot; = &quot;.y.&quot;, &quot;Overall&quot; = &quot;p&quot;) %&gt;% mutate(across(&#39;Overall&#39;:&#39;2 vs. 4&#39;, \\(x) format(x, scientific = TRUE, digits = 3))) datatable(dose_results_cleaned) To more easily see overall significance patterns, we could also make the same table but with significance stars instead of p-values by keeping the p.adjust.signif column instead of the p.adj column in our post-hoc test results dataframe: dose_results_cleaned_2 &lt;- dose_wilcox_posthoc %&gt;% unite(comparison, group1, group2, sep = &quot; vs. &quot;) %&gt;% select(c(.y., comparison, p.adj.signif)) %&gt;% pivot_wider(id_cols = &quot;.y.&quot;, names_from = &quot;comparison&quot;, values_from = &quot;p.adj.signif&quot;) %&gt;% left_join(dose_friedmanres, by = &quot;.y.&quot;) %&gt;% relocate(p, .after = &quot;.y.&quot;) %&gt;% rename(&quot;Variable&quot; = &quot;.y.&quot;, &quot;Overall&quot; = &quot;p&quot;) %&gt;% mutate(across(&#39;Overall&#39;:&#39;2 vs. 4&#39;, \\(x) format(x, scientific = TRUE, digits = 3))) datatable(dose_results_cleaned_2) Answer to Environmental Health Question 1 With this, we can answer Environmental Health Question #1 : Are there significant differences in inflammatory biomarker concentrations between different doses of acrolein? Answer: Yes, there are significant differences in inflammatory biomarker concentrations between different doses of acrolein. The overall p-values for all biomarkers are significant. Within each biomarker, at least one pairwise comparison was significant between doses, with a majority of these significant comparisons being with the highest dose (4 ppm). Visualization of Multi-Group Statistical Results The statistical results we generated are a lot to digest in table format, so it can be helpful to graph the results. As our statistical testing becomes more complicated, so does the code used to generate results. The ggpubr package can perform statistical testing and overlay the results onto graphs for a specific set of tests, such as overall effects tests and unpaired t-tests or Wilcoxon tests. However, for tests that aren’t available by default, the package also contains the helpful stat_pvalue_manual() function that can be added to plots. This is what we will need to use to add the results of our pairwise, paired Wilcoxon test with BH correction, as there is no option for BH correction within the default function we might otherwise use (stat_compare_means()). We will first work through an example of this using one of our endpoints, and then we will demonstrate how to apply it to facet plotting. Single Plot We first need to format our existing statistical results so that they match the format that the function needs as input. Specifically, the dataframe needs to contain the following columns: group1 and group2: the groups being compared A column containing the results you want displayed (p, p.adj, or p.adj.signif typically) y.position, which tells the function where to plot the significance markers Our results dataframe for IL-1\\(\\beta\\) already contains our groups and p-values: datatable(dose_wilcox_posthoc_IL1B) We can add the position columns using the function add_xy_position(): dose_wilcox_posthoc_IL1B &lt;- dose_wilcox_posthoc_IL1B %&gt;% add_xy_position(x = &quot;Dose&quot;, step.increase = 2) datatable(dose_wilcox_posthoc_IL1B) Now, we are ready to make a graph of our results. We will use stat_friedman_test() to add our overall p-value and stat_pvalue_manual() to add our pairwise values. # Set graphing theme theme_set(theme_bw()) # Make plot ggplot(biomarker_data, aes(x = Dose, y = IL1B)) + geom_boxplot(aes(fill = Dose), outlier.shape = NA) + scale_fill_manual(values = c(&quot;#BFBFBF&quot;, &quot;#D5A298&quot;, &quot;#E38273&quot;, &quot;#EB5F4E&quot;, &quot;#EE2B2B&quot;)) + geom_jitter(size = 3, position = position_jitter(0.15)) + stat_friedman_test(wid = &quot;Donor&quot;, p.adjust.method = &quot;none&quot;, label = &quot;p = {p.format}&quot;, label.x.npc = &quot;left&quot;, label.y = 9.5, hjust = 0.5, size = 6) + stat_pvalue_manual(dose_wilcox_posthoc_IL1B, label = &quot;p.adj.signif&quot;, size = 12, hide.ns = TRUE) + ylim(2.5, 10) + labs(y = &quot;Log2(IL-1\\u03B2 (pg/mL))&quot;, x = &quot;Acrolein (ppm)&quot;) + theme(legend.position = &quot;none&quot;, axis.title = element_text(color = &quot;black&quot;, size = 15), axis.title.x = element_text(vjust = -0.75), axis.title.y = element_text(vjust = 2), axis.text = element_text(color = &quot;black&quot;, size = 12)) However, to make room for all of our annotations, our data become compressed, and it makes it difficult to see our data. Although presentation of statistical results is largely a matter of personal preference, we could clean up this plot by making our annotations appear on top of the bars, with indication in the figure legend that the comparison is with a specific dose. We will do this by: Filtering our results to those that are significant. Changing the symbol for comparisons that are not to the 0 dose. Layering this text onto the plot with geom_text() rather than stat_pvalue_manual(). First, let’s filter our results to significant results and change the symbol for comparisons that are not to the 0 dose to a caret (^) instead of stars. We can do this by creating a new column called label that keeps the existing label if group1 is 0, and if not, changes the label to a caret of the same length. We then use the summarize function to paste the labels for each of the groups together, resulting in a final dataframe containing our annotations for our plot. dose_wilcox_posthoc_IL1B_2 &lt;- dose_wilcox_posthoc_IL1B %&gt;% # Filter results to those that are significant filter(p.adj &lt;= 0.05) %&gt;% # Make new symbol mutate(label = ifelse(group1 == &quot;0&quot;, p.adj.signif, strrep(&quot;^&quot;, nchar(p.adj.signif)))) %&gt;% # Select only the columns we need select(c(group1, group2, label)) %&gt;% # Combine symbols for the same group group_by(group2) %&gt;% summarise(label = paste(label, collapse=&quot; &quot;)) %&gt;% # Remove duplicate row distinct(group2, .keep_all = TRUE) %&gt;% # Rename group2 to dose rename(&quot;Dose&quot; = &quot;group2&quot;) dose_wilcox_posthoc_IL1B_2 ## # A tibble: 4 × 2 ## Dose label ## &lt;chr&gt; &lt;chr&gt; ## 1 0.6 * ## 2 1 * ## 3 2 * ## 4 4 * ^ Then, we can use the same code as for our previous plot, but instead of using stat_pvalue_manual(), we will use geom_text() in combination with the dataframe we just created. ggplot(biomarker_data, aes(x = Dose, y = IL1B)) + geom_boxplot(aes(fill = Dose), outlier.shape = NA) + scale_fill_manual(values = c(&quot;#BFBFBF&quot;, &quot;#D5A298&quot;, &quot;#E38273&quot;, &quot;#EB5F4E&quot;, &quot;#EE2B2B&quot;)) + geom_jitter(size = 3, position = position_jitter(0.15)) + stat_friedman_test(wid = &quot;Donor&quot;, p.adjust.method = &quot;none&quot;, label = &quot;p = {p.format}&quot;, label.x.npc = &quot;left&quot;, label.y = 4.85, hjust = 0.5, size = 6) + geom_text(data = dose_wilcox_posthoc_IL1B_2, aes(x = Dose, y = 4.5, label = paste0(label)), size = 10, hjust = 0.5) + ylim(2.5, 5) + labs(y = &quot;Log2(IL-1\\u03B2 (pg/mL))&quot;, x = &quot;Acrolein (ppm)&quot;) + theme(legend.position = &quot;none&quot;, axis.title = element_text(color = &quot;black&quot;, size = 15), axis.title.x = element_text(vjust = -0.75), axis.title.y = element_text(vjust = 2), axis.text = element_text(color = &quot;black&quot;, size = 12)) An appropriate title for this figure could be: “Figure X. Exposure to 0.6-4 ppm acrolein increases IL-1\\(\\beta\\) secretion in primary human bronchial epithelial cells. Groups were compared using the Friedman test to obtain overall p-value and Wilcoxon signed rank test for post-hoc testing. * p &lt; 0.05 in comparison with 0 ppm, ^ p &lt; 0.05 in comparison with 0.6 ppm, n = 16 per group (paired).” Faceted Plot Ideally, we would extend this sort of graphical approach to our faceted plot showing all of our endpoints. However, there are quite a few statistically significant comparisons to graph, including comparisons that are significant between different pairs of doses (not just back to the control). While we could attempt to graph all of them, ultimately, this will lead to a cluttered figure panel. When thinking about how to simplify our plots, some options are: Instead of using the number of symbols to represent p-values, we could use a single symbol to represent any comparison with a p-value with at least p &lt; 0.05, and that symbol could be different depending on which group the significance is in comparison to. Symbols can be difficult to parse in R, so we could use letters or even the group names above the column of interest. For example, if the concentration of an endpoint at 2 ppm was significant in comparison with both 0 and 0.6 ppm, we could annotate “0, 0.6” above the 2 ppm column, or we could choose a letter (“a, b”) or symbol (“*, ^“) to convey these results. If the pattern is the same across many of the endpoints measured, we could graph a subset of the endpoints with the most notable data trends or the most biological meaning for the main body of the manuscript, with data for additional endpoints referred to in the text and shown in the supplemental figures or tables. If most of the significant comparisons are back to the control group, we could choose to only show comparisons with the control group, with textual description of the other significant comparisons and indication that those specific p-values can be viewed in the supplemental table of results. Which approach you decide to take (or maybe another approach altogether) is a matter of both personal preference and your specific study goals. You may also decide that it is important to you to show all significant comparisons, which will require more careful formatting of the plots to ensure that all text and annotations are legible. For this module, we will proceed with option #3 because many of our comparisons to the control dose (0) are significant, and we have enough groups that there likely will not be space to annotate all of them above our data. We will take similar steps here that we did when constructing our single endpoint graph, with a couple of small differences. Specifically, we need to: Create a dataframe of labels/annotations as we did above, but now filtered to only significant comparisons with the 0 group. Add to the label/annotation dataframe what we want the y position for each of the labels to be, which will be different for each endpoint. First, let’s create our annotations dataframe. We will start with the results dataframe from our posthoc testing: datatable(dose_wilcox_posthoc) dose_wilcox_posthoc_forgraph &lt;- dose_wilcox_posthoc %&gt;% filter(p.adj &lt;= 0.05) %&gt;% # Filter for only comparisons to 0 filter(group1 == &quot;0&quot;) %&gt;% # Rename columns rename(&quot;variable&quot; = &quot;.y.&quot;, &quot;Dose&quot; = &quot;group2&quot;) datatable(dose_wilcox_posthoc_forgraph) The Dose column will be used to tell ggplot2 where to place the annotations on the x axis, but we need to also specify where to add the annotations on the y axis. This will be different for each variable because each variable is on a different scale. We can approach this by computing the maximum value of each variable, then increasing that by 20% to add some space on top of the points. sig_labs_y &lt;- biomarker_data %&gt;% summarise(across(IL1B:VEGF, \\(x) max(x))) %&gt;% t() %&gt;% as.data.frame() %&gt;% rownames_to_column(&quot;variable&quot;) %&gt;% rename(&quot;y_pos&quot; = &quot;V1&quot;) %&gt;% mutate(y_pos = y_pos*1.2) sig_labs_y ## variable y_pos ## 1 IL1B 4.967953 ## 2 IL6 14.074658 ## 3 IL8 20.752105 ## 4 IL10 3.224473 ## 5 TNFa 4.478723 ## 6 VEGF 14.148027 Then, we can join these data to our labeling dataframe to complete what we need to make the annotations. dose_wilcox_posthoc_forgraph &lt;- dose_wilcox_posthoc_forgraph %&gt;% left_join(sig_labs_y, by = &quot;variable&quot;) Now, it’s time to graph! Keep in mind that although the plotting script can get long and unweildy, each line is just a new instruction to ggplot about a formatting element or an additional layer to add to the graph. # Pivot data longer biomarker_data_long &lt;- biomarker_data %&gt;% pivot_longer(-c(Donor, Dose), names_to = &quot;variable&quot;, values_to = &quot;value&quot;) # Create clean labels for the graph titles new_labels &lt;- c(&quot;IL10&quot; = &quot;IL-10&quot;, &quot;IL1B&quot; = &quot;IL-1\\u03B2 &quot;, &quot;IL6&quot; = &quot;IL-6&quot;, &quot;IL8&quot; = &quot;IL-8&quot;, &quot;TNFa&quot; = &quot;TNF-\\u03b1&quot;, &quot;VEGF&quot; = &quot;VEGF&quot;) # Make graph ggplot(biomarker_data_long, aes(x = Dose, y = value)) + # outlier.shape = NA removes outliers geom_boxplot(aes(fill = Dose), outlier.shape = NA) + # Changing box plot colors scale_fill_manual(values = c(&quot;#BFBFBF&quot;, &quot;#D5A298&quot;, &quot;#E38273&quot;, &quot;#EB5F4E&quot;, &quot;#EE2B2B&quot;)) + geom_jitter(size = 1.5, position = position_jitter(0.15)) + # Adding a p value from Friedman test stat_friedman_test(wid = &quot;Donor&quot;, p.adjust.method = &quot;none&quot;, label = &quot;p = {p.format}&quot;, label.x.npc = &quot;left&quot;, vjust = -3.5, hjust = 0.1, size = 3.5) + # Add label geom_text(data = dose_wilcox_posthoc_forgraph, aes(x = Dose, y = y_pos, label = p.adj.signif, size = 5, hjust = 0.5)) + # Adding padding y axis scale_y_continuous(expand = expansion(mult = c(0.1, 0.6))) + # Changing y axis label ylab(expression(Log[2]*&quot;(Concentration (pg/ml))&quot;)) + # Changing x axis label xlab(&quot;Acrolein (ppm)&quot;) + # Faceting by each biomarker facet_wrap(~ variable, nrow = 2, scales = &quot;free_y&quot;, labeller = labeller(variable = new_labels)) + # Removing legend theme(legend.position = &quot;none&quot;, axis.title = element_text(color = &quot;black&quot;, size = 12), axis.title.x = element_text(vjust = -0.75), axis.title.y = element_text(vjust = 2), axis.text = element_text(color = &quot;black&quot;, size = 10), strip.text = element_text(size = 12, face = &quot;bold&quot;)) An appropriate title for this figure could be: “Figure X. Exposure to acrolein increases secretion of proinflammatory biomarkers in primary human bronchial epithelial cells. Groups were compared using the Friedman test to obtain overall p-value and Wilcoxon signed rank test for post-hoc testing. * p &lt; 0.05, ** p &lt; 0.01, *** p &lt; 0.001, **** p &lt; 0.0001 for comparison with control. For additional significant comparisons, see Supplemental Table X. n = 16 per group (paired).” Answer to Environmental Health Question 2 With this, we can answer Environmental Health Question #2 : Do TNF-\\(\\alpha\\) concentrations significantly increase with increasing dose of acrolein? Answer: Yes, TNF-\\(\\alpha\\) concentrations significantly increase with increasing dose of acrolein, which we were able to visualize, along with other mediators, in our facet plot. Concluding Remarks In this module, we introduced common multi-group statistical tests, including both overall effects tests and post-hoc testing. We applied these tests to our example dataset and demonstrated how to produce publication-quality tables and figures of our results. Implementing a workflow such as this enables efficient analysis of wet-bench generated data and customization of output figures and tables suited to your personal preferences. Additional Resources STHDA: How to Add P-Values and Significance Levels to ggplots using ggpubr Adding p-values with ggprism Overview of ggsignif Test Your Knowledge Functional endpoints from these cultures were also measured. These endpoints were: 1) Membrane Permeability (MemPerm), 2) Trans-Epithelial Electrical Resistance (TEER), 3) Ciliary Beat Frequency (CBF), and 4) Expression of Mucin (MUC5AC). These data were already processed and tested for normality (see Test Your Knowledge for TAME 2.0 Module 4.2 Data Import, Processing, and Summary Statistics), with results indicating that two of the endpoints are normally distributed and two non-normally distributed. Use the same processes demonstrated in this module and the provided data (“Module4_5_TYKInput.xlsx” (functional data)) to run analyses and make a publication-quality figure panel and table to answer the following question: Are there significant differences in functional endpoints between cells treated with different concentrations of acrolein? For an extra challenge, try also making your faceted plot in the style of option #1 above, with different symbols, letters, or group names above columns to indicate which group that column in significant in comparison with. "],["introduction-to-artificial-intelligence-machine-learning-and-predictive-modeling-for-environmental-health.html", "Introduction to Artificial Intelligence, Machine Learning, and Predictive Modeling for Environmental Health Introduction to Training Module General Historical Context and Taxonomy of Modern AI/ML Application of Machine Learning in Environmental Health Science Concluding Remarks", " Introduction to Artificial Intelligence, Machine Learning, and Predictive Modeling for Environmental Health This training module was developed by David M. Reif, with contributions from Elise Hickman, Alexis Payton, and Julia E. Rager All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module Artificial intelligence (AI), machine learning (ML), and predictive modeling are becoming increasingly popular buzzwords both in the public domain and within research fields, including environmental health. Within environmental health, these computational techniques are implemented to integrate large, high dimensional datasets (e.g., chemical, biological, clinical/medical, model estimates, etc) to better understand links between environmental exposures and biological responses. In this training module, we will: Provide general historical context and taxonomy of modern AI/ML Provide an overview of the intersection between environmental health science ML through discussing… Why there is a need for ML in environmental health science The differences between ML and traditional statistical methods Predictive modeling in the context of environmental health science Additional applications of ML in environmental health science Training Module’s Environmental Health Question This training module was specifically developed to answer the following environmental health question: How and why are artificial intelligence, machine learning, and predictive modeling used in environmental health research? General Historical Context and Taxonomy of Modern AI/ML Before diving in to the applications of AI and ML in environmental health, let’s first establish what these term mean and how they are related. Note that the definitions surrounding AI and ML can be subjective, however the purpose of this module is not to get caught up in semantics, but to broadly understand how AI and ML can be applied to environmental health research. Artificial Intelligence (AI) encompasses computer systems that perform tasks typically associated with human cognition and intelligence. AI is found in our everyday lives, for instance, within face recognition, internet search queries, email spam detection, smart home devices, auto-navigation, and digital assistants. Machine Learning (ML) can be thought of as a subset of AI and describes a computer system that iteratively learns and improves from that experience autonomously. Below is a high level taxonomy of AI. It’s not meant to be an exhaustive depiction of all AI techniques but a simple visualization of how some of these methodologies are nested within each other. Note: AI can be categorized in different ways and may deviate from what is illustrated below. Advantages of AI and ML include the automation of repetitive tasks, complex problem solving, and reducing human error. However, disadvantages include learning from biased datasets or patterns that are reflected in the decisions of AI/ML and the potential limited interpretability of algorithms created by AI/ML. Check out the following resources for… Further explanation on differences in Artificial Intelligence vs. Machine Learning Other subsets of AI that fall outside of the scope of these modules in Types of Artificial Intelligence Additional discussion on the utility of ML approaches for high-dimensional data common in environmental health research in Payton et. al It is important to understand the methodological “roots” of current methods. Otherwise, it seems like every approach is novel! AI and ML methods have been around since the mid- to late- 1900s and continue to evolve in the present day. The earliest conceptual roots for these approaches can be traced from antiquity; however, it is generally thought that the field was named “artificial intelligence” at the “Dartmouth Workshop” in 1956, led by John McCarthy and others. The following schematic demonstrates the general taxonomy (categories, sub-fields, and specific methods) of modern AI and ML: A Brief Detour to Discuss ChatGPT ChatGPT (Chat Generative Pre-trained Transformer) is a publicly available chatbot developed by OpenAI. It was released in November of 2022 and quickly gained popularity due to its accessibility and ability to have human-like conversations with the user across almost any imaginable topic. Language Models (LLMs), including large language models like GPT-3 (a predecessor to ChatGPT), generally fall under the “Connectionist AI” category, which use deep learning techniques and are considered a subset of artificial neural networks. They fall under the deep learning subset due to their use of deep neural networks with many layers, allowing them to learn from large amounts of data and find intricate patterns. LLMs are trained to predict the probability of a word given its context in a dataset (a form of next-word prediction), which is a machine learning methodology. It’s notable that they use architectures like Transformer Networks, which are known for their efficiency in handling sequential data, making them a go-to choice for natural language processing (NLP) tasks. The use of attention mechanisms in these architectures allows the model to focus on different parts of the input sequence when producing an output sequence, offering a substantial improvement in performance for many natural language processing tasks. The role of ChatGPT and similar tools in the environmental health research space is still being explored. Although ChatGPT has the potential to streamline certain parts of the research process, such as text and language polishing, synthesizing existing information, and suggesting custom coding solutions, it is not an intellectual replacement for the expertise and diverse viewpoints of scientists and must be used transparently and with caution. Application of Machine Learning in Environmental Health Science For the rest of this module and chapter, we will focus on machine learning (ML). Generally speaking, ML is considered to encompass the study of computer algorithms that can improve automatically through experience and by the use of data. It is seen as a part of artificial intelligence (AI), discussed broadly above. Why do we need machine learning in environmental health science? There are many avenues to incorporate ML into environmental health research, all aimed at better identifying patterns amongst large datasets spanning medical health records, clinical data, exposure monitoring data, chemistry profiles, and the rapidly expanding realm of biological response data including multiple -omics endpoints. One well-known problem that can be better addressed by incorporating ML is the ‘too many chemicals, too little data’ problem. To detail, there are thousands of chemicals in commerce today. Testing these chemicals one by one for toxicity using comprehensive animal screening experiments would take decades and is not feasible financially. Current efforts to address this problem include using cell-based high throughput screening to efficiently determine biological responses to a variety of chemical exposures and treatment conditions. These screening efforts result in increasing amounts of data, which can be gathered to start building big databases. When many of these datasets and databases are combined, including diversity across different types of screening platforms, technologies, cell types, species, and other experimental variables, the associated dimensionality of the data gets “big.” This presents a problem because these data are diverse and high dimensional (the number of features or endpoints exceeds the number of observations/chemicals). To appropriately analyze and model these data, new approaches beyond traditional statistical methods are needed. Machine Learning vs. Traditional Statistical Methods There is plenty of debate as to where the line(s) between ML and traditional statistics should be drawn. In our opinion, a perfect delineation is not necessary for our purposes. Rather, we will focus on the usual goals/intent of each to help us understand the distinction for environmental health research. Traditional statistics may be able to handle 1:1 or 1:many comparisons of singular quantities (e.g., activity concentrations for two chemicals). However, once the modeling becomes more complex or exploratory, assumptions of most traditional methods will be violated. Furthermore, statistics draws population inferences from a sample, while AI/ML finds generalizable predictive patterns (Bzdok et al 2018). This is particularly helpful in predictive toxicology, in which we leverage high dimensional data to obtain generalizable forecasts for the effects of chemicals on biological systems. This image shows graphical abstractions of how a “problem” is solved using: Traditional statistics ((A) logistic regression and (B) linear regression), OR Machine learning ((C) support vector machines, (D) artificial neural networks, and (E) decision trees) Predictive Modeling in the Context of Environmental Health Science In the previous section, we briefly mentioned predictive toxicology. We often think of predictions as having a forward-time component (i.e. What will happen next?) … what about “prediction” in a different sense as applied to toxicology? Our working definition is that predictive toxicology describes a multidisciplinary approach to chemical toxicity evaluation that more efficiently uses animal test results, when needed, and leverages expanding non-animal test methods to forecast the effects of a chemical on biological systems. Examples of the questions we can answer using predictive toxicology include: Can we more efficiently design animal studies and analyze data from shorter assays using fewer animals to predict long-term health outcomes? Can this suite of in vitro assays predict what would happen in an organism? Can we use diverse, high dimensional data to cluster chemicals into predicted activity classes? Similar logic applies to the field of exposure science. What about “prediction” applied to exposure science? Our working definition is that predictive exposure science describes a multidisciplinary approach to chemical exposure evaluations that more efficiently uses biomonitoring, chemical inventory, and other exposure science-relevant databases to forecast exposure rates in target populations. For example: Can we use existing biomonitoring data from NHANES to predict exposure rates for chemicals that have yet to be measured in target populations? (see ExpoCast program, e.g., Wambagh et al 2014) Can I use chemical product use inventory data to predict the likelihood of a chemical being present in a certain consumer product? (e.g., Phillips et al 2018) There are many different types of ML methods that we can employ in predictive toxicology and exposure science, depending on the data type / purpose of data analysis. A recent review written together with Erin Baker’s lab provides a high-level overview on some of the types of ML methods and challenges to address when analyzing multi-omic data (including chemical signature data). Answer to Environmental Health Question With this, we can now answer our Environmental Health Question: How and why are machine learning, predictive modeling, and artificial intelligence used in environmental health research? Answer: Machine learning, a subcategory of artificial intelligence, can be used in environmental health science to better understand patterns between chemical exposure and biological response in complex, high dimensional datasets. These datasets are often generated as part of efforts to screen many chemicals efficiently. Predictive modeling, which can include machine learning approaches, leverages these data to forecast the effects of a chemical on biological systems. Additional Applications of Machine Learning in Environmental Health Science In addition to the predictive toxicology questions above, ML can also be applied in the analysis of complex, high dimensional data in observational clinical (human subjects) studies in environmental health, such as: Do subjects cluster by chemical exposure? Are there similarities between subjects that cluster together for chemical exposure, suggesting underlying factors relevant to chemical exposure? Are biological signatures in different exposure groups different enough overall that ML can predict which group a subject belongs to based on their signature? Concluding Remarks In conclusion, this training module provides an overview of the field of AI and ML and discusses applications of these tools in environmental health science through predictive modeling. These methods represent common tools that are used in high dimensional data analyses within the field of environmental health sciences. In the following modules, we will provide specific examples detailing how to apply both supervised and unsupervised machine learning methods to environmental health questions and how to interpret the results of these analyses. For a review article on ML, see: Odenkirk MT, Reif DM, Baker ES. Multiomic Big Data Analysis Challenges: Increasing Confidence in the Interpretation of Artificial Intelligence Assessments. Anal Chem. 2021 Jun 8;93(22):7763-7773. PMID: 34029068 For additional case studies that leverage more advanced ML techniques, see the following recent publications that also address environmental health questions from our research groups, with bracketed tags at the end of each citation denoting ML methods used in that study: Clark J, Avula V, Ring C, Eaves LA, Howard T, Santos HP, Smeester L, Bangma JT, O’Shea TM, Fry RC, Rager JE. Comparing the Predictivity of Human Placental Gene, microRNA, and CpG Methylation Signatures in Relation to Perinatal Outcomes. Toxicol Sci. 2021 Sep 28;183(2):269-284. PMID: 34255065 [hierarchical clustering, principal component analysis, random forest] Green AJ, Mohlenkamp MJ, Das J, Chaudhari M, Truong L, Tanguay RL, Reif DM. Leveraging high-throughput screening data, deep neural networks, and conditional generative adversarial networks to advance predictive toxicology. PLoS Comput Biol. 2021 Jul 2;17(7):e1009135. PMID: 3421407 [conditional generative adversarial network, deep neural network, support vector machine, random forest, multilayer perceptron] To KT, Truong L, Edwards S, Tanguay RL, Reif DM. Multivariate modeling of engineered nanomaterial features associated with developmental toxicity. NanoImpact. 2019 Apr;16:10.1016. PMID: 32133425 [random forest] Ring C, Sipes NS, Hsieh JH, Carberry C, Koval LE, Klaren WD, Harris MA, Auerbach SS, Rager JE. Predictive modeling of biological responses in the rat liver using in vitro Tox21 bioactivity: Benefits from high-throughput toxicokinetics. Comput Toxicol. 2021 May;18:100166. PMID: 34013136 [random forest] Hickman E, Payton A, Duffney P, Wells H, Ceppe AS, Brocke S, Bailey A, Rebuli ME, Robinette C, Ring B, Rager JE, Alexis NE, Jaspers I. Biomarkers of Airway Immune Homeostasis Differ Significantly with Generation of E-Cigarettes. Am J Respir Crit Care Med. 2022 Nov 15; 206(10):1248-1258. PMID: 35731626 [hierarchical clustering, quadratic discriminant analysis, multinomial logistic regression] Perryman AN, Kim H-YH, Payton A, Rager JE, McNell EE, Rebuli ME, et al. (2023) Plasma sterols and vitamin D are correlates and predictors of ozone-induced inflammation in the lung: A pilot study. PLoS ONE 18(5): e0285721. PMID: 37186612 [random forest, support vector machine, k nearest neighbor] Payton AD, Perryman AN, Hoffman JR, Avula V, Wells H, Robinette C, Alexis NE, Jaspers I, Rager JE, Rebuli ME. Cytokine signature clusters as a tool to compare changes associated with tobacco product use in upper and lower airway samples. American Journal of Physiology-Lung Cellular and Molecular Physiology 2022 322:5, L722-L736. PMID: 35318855 [k-means clustering, principal component analysis] "],["supervised-machine-learning.html", "Supervised Machine Learning Introduction to Training Module Types of Machine Learning Types of Supervised Machine Learning Algorithms Training Supervised Machine Learning Models Assessing Classification-Based Model Performance Introduction to Activity and Example Dataset Testing for Differences in Predictor Variables across the Outcome Classes Predicting iAs Detection with a Random Forest (RF) Model Class Imbalance Concluding Remarks", " Supervised Machine Learning This training module was developed by Alexis Payton, Oyemwenosa N. Avenbuan, Lauren E. Koval, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module Machine learning is a field that has been around for decades but has exploded in popularity and utility in recent years due to the proliferation of big and/or high dimensional data. Machine learning has the ability to sift through and learn from large volumes of data and use that knowledge to solve problems. The challenges of high dimensional data as they pertain to environmental health and the applications of machine learning to mitigate some of those challenges are discussed further in Payton et. al. In this module, we will introduce different types of machine learning and then focus in on supervised machine learning, including how to train and assess supervised machine learning models. We will then analyze an example dataset with supervised machine learning highlighting an example with random forest modeling. Types of Machine Learning Within the field of machine learning, there are many different types of algorithms that can be leveraged to address environmental health research questions. The two broad categories of machine learning frequently applied to environmental health research are: (1) supervised machine learning and (2) unsupervised machine learning. Supervised machine learning involves training a model using a labeled dataset, where each independent or predictor variable is associated with a dependent variable with a known outcome. This allows the model to learn how to predict the labeled outcome on data it hasn’t “seen” before based on the patterns and relationships it previously identified in the data. For example, supervised machine learning has been used for cancer prediction and prognosis based on variables like tumor size, stage, and age (Lynch et. al, Asadi et. al). Supervised machine learning includes: Classification: Using algorithms to classify a categorical outcome (ie. plant species, disease status, etc.) Regression: Using algorithms to predict a continuous outcome (ie. gene expression, chemical concentration, etc.) Soni, D. (2018, March 22). Supervised vs. Unsupervised Learning. Towards Data Science; Towards Data Science. https://towardsdatascience.com/supervised-vs-unsupervised-learning-14f68e32ea8d Unsupervised machine learning, on the other hand, involves using models to find patterns or associations between variables in a dataset that lacks a known or labeled outcome. For example, unsupervised machine learning has been used to identify new patterns across genes that are co-expressed, informing potential biological pathways mediating human disease (Botía et. al, Pagnuco et. al). Langs, G., Röhrich, S., Hofmanninger, J., Prayer, F., Pan, J., Herold, C., &amp; Prosch, H. (2018). Machine learning: from radiomics to discovery and routine. Der Radiologe, 58(S1), 1–6. PMID: 34013136. Figure regenerated here in alignment with its published Creative Commons Attribution 4.0 International License. Overall, the distinction between supervised and unsupervised learning is an important concept in machine learning, as it can inform the choice of algorithms and techniques used to analyze and make predictions from data. It is worth noting that there are also other types of machine learning, such as semi-supervised learning, reinforcement learning, and deep learning, though we will not further discuss these topics in this module. Types of Supervised Machine Learning Algorithms Although this module’s example will focus on a random forest model in the coding example below, other commonly used algorithms for supervised machine learning include: K-Nearest Neighbors (KNN): Uses distance to classify a data point in the test set based upon the most common class of neighboring data points from the training set. For more information on KNN, see K-Nearest Neighbor. Support Vector Machine (SVM): Creates a decision boundary line (hyperplane) in n-dimensional space to separate the data into each class so that when new data is presented, they can be easily categorized. For more information on SVM, see Support Vector Machine. Random Forest (RF): Uses a multitude of decision trees trained on a subset of different samples from the training set and the resulting classification of a data point in the test set is aggregated from all the decision trees. A decision tree is a hierarchical model that depicts decisions from predictors and their resulting outcomes. It starts with a root node, which represents an initial test from a single predictor. The root node splits into subsequent decision nodes that test another feature. These decision nodes can either feed into more decision nodes or leaf nodes that represent the predicted class label. A branch or a sub-tree refers to a subsection of an entire decision tree. Here is an example decision tree with potential variables and decisions informing a college basketball player’s likelihood of being drafted to the NBA: While decision trees are highly interpretable, they are prone to overfitting, thus they may not always generalize well to data outside of the training set. To address this, random forests are comprised of many different decision trees. Each tree is trained on a subset of the samples in the training data, selected with replacement, and a randomly selected set of predictor variables. For a dataset with p predictors, it is common to test \\(\\sqrt{p}\\), \\(\\frac{p}{2}\\), and p predictors to see which gives the best results. This process decorrelates the trees. For a classification problem, majority vote of the decision trees determines the final class for a prediction. This process loses interpretability inherent to individual trees, but reduces the risk of overfitting. For more information on RF and decision trees, check out Random Forest and Decision Trees. Note: One algorithm is not inherently better than the others with each having their respective advantages and disadvantages. Each algorithm’s predictive ability will be largely dependent on the size of the dataset, the distribution of the data points, and the scenario. Training Supervised Machine Learning Models In supervised machine learning, algorithms need to be trained before they can be used to predict on new data. This involves selecting a smaller portion of the dataset to train the model so it will learn how to predict the outcome as accurately as possible. The process of training an algorithm is essential for enabling the model to learn and improve over time, allowing it to make more accurate predictions and better adapt to new and changing circumstances. Ultimately, the quality and relevance of the training data will have a significant impact on the effectiveness of a machine learning model. Common partitions of the full dataset used to train and test a supervised machine learning model are the following: Training Set: a subset of the data that the algorithm “sees” and uses to identify patterns. Validation Set: a subset of the training set that is used to evaluate the model’s fit in an unbiased way allowing us to fine-tune its parameters and optimize performance. Test Set: a subset of data that is used to evaluate the final model’s fit based on the training and validation sets. This provides an objective assessment of the model’s ability to generalize new data. It is common to split the dataset into a training set that contains 60% of the data and the test set that contains 40% of the data, though other common splits include 70% training / 30% test and 80% training / 20% test. It is important to note that the test set should only be examined after the algorithm has been trained using the training/validation sets. Using the test set during the development process can lead to overfitting, where the model performs well on the test data but poorly on new data. The ideal algorithm is generalizable or flexible enough to accurately predict unseen data. This is known as the bias-variance tradeoff. For further information on the bias-variance tradeoff, see Understanding the Bias-Variance Tradeoff. Cross Validation Finally, we will discuss cross validation, which is an approach used during training to expose the model to more patterns in the data and aid in model evaluation. For example, if a model is trained and tested on a 60:40 split, our model’s accuracy will likely be influenced by where this 60:40 split occurs in the dataset. This will likely bias the data and reduce the algorithm’s ability to predict accurately for data not in the training set. Overall, cross validation (CV) is implemented to fine tune a model’s parameters and improve prediction accuracy and ability to generalize. Although there are a number of cross validation approaches, we will specifically highlight k-fold cross validation. k-fold cross validation works by splitting the samples in the training dataset into k equally sized folds or groups. For example, if we implement 5-fold CV, we start by… Splitting the training data into 5 groups, or “folds”. Five iterations of training/testing are then run where each of the 5 folds serves as the test data once and as part of the training set four times, as seen in the figure below. To measure predictive ability of each of the parameters tested, like the number of features to include, values like accuracy and specificity are calculated for each iteration. The parameters that optimize performance are selected for the final model which will be evaluated against the test set not used in training. Check out these resources for additional information on Cross Validation in Machine Learning and Cross Validation Pros &amp; Cons. Assessing Classification-Based Model Performance Evaluation metrics from a confusion matrix are often used to determine the best model during training and measure model performance during testing for classification-based supervised machine learning models. A confusion matrix consists of a table that displays the numbers of how often the algorithm correctly and incorrectly predicted the outcome. Let’s imagine you’re interested in predicting whether or not a player will be drafted to the National Basketball Association (NBA) based on a dataset that contains variables regarding a player’s assists, points, height etc. Let’s say that this dataset contains information on 253 players with 114 that were actually drafted and 139 that weren’t drafted. The confusion matrix below shows a model’s results where a player that is drafted is the “positive” class and a player that is not drafted is the “negative” class. Helpful confusion matrix terminology: True positive (TP): the number of correctly classified “positive” data points (i.e., the number of correctly classified players to be drafted) True negative (TN): the number of correctly classified “negative” data points (i.e., the number of correctly classified players to be not drafted) False positive (FP): the number of incorrectly classified “positive” data points (i.e., the number of players not drafted incorrectly classified as draft picks) False negative (FN): the number of incorrectly classified “negative” data points (i.e., the number of draft picks incorrectly classified as players not drafted) Some of the metrics that can be obtained from a confusion matrix are listed below: Overall Accuracy: indicates how often the model makes a correct prediction relative to the total number of predictions made and is typically used to assess overall model performance (\\(\\frac{TP+TN}{TP+TN+FP+FN}\\)). Sensitivity or Recall: evaluates how well the model was able to predict the “positive” class. It is calculated as the ratio of correctly classified true positives to the total number of positive cases (\\(\\frac{TP}{TP+FN}\\)). Specificity: evaluates how well the model was able to predict the “negative” class. It is calculated as the ratio of correctly classified true negatives to total number of negatives cases (\\(\\frac{TN}{TN+FP}\\)). Balanced Accuracy: is the mean of sensitivity and specificity and is often used in the case of a class imbalance to gauge how well the model can correctly predict values for both classes (\\(\\frac{sensitivity+specificity}{2}\\)). Positive Predictive Value (PPV) or Precision: evaluates how accurate predictions of the “positive” class are. It is calculated as the ratio of correctly classified true positives to total number of predicted positives (\\(\\frac{TP}{TP+FN}\\)). Negative Predictive Value (NPV): evaluates how accurate predictions of the “negative” class are. It is calculated as the ratio of correctly classified true negatives to total number of predicted negatives (\\(\\frac{TN}{TN+FP}\\)). For the above metrics, values fall between 0 and 1. Instances of 0 indicate that the model was not able to classify any data points correctly, and instances of 1 indicate that the model was able to classify all test data correctly. Although subjective, an overall accuracy of at least 0.7 is considered respectable (Barkved, 2022). Furthermore, a variety of additional metrics exist for evaluating model performance for classification problems (24 Evaluation Metrics for Binary Classification (And When to Use Them)). Selecting a metric for evaluating model performance varies by situation and is dependent not only on the individual dataset, but also the question being answered. Note: For multi-class classification (more than two labeled outcomes to be predicted), the same metrics are often used, but are obtained in a slightly different way. Regression based supervised machine learning models use loss functions to evaluate model performance. For more information regarding confusion matrices and loss functions for regression-based models, see: Additional Confusion Matrix Metrics Precision vs. Recall or Specificity vs. Sensitivity Loss Functions for Machine Learning Regression Introduction to Activity and Example Dataset In this activity, we will analyze an example dataset to see whether we can use environmental monitoring data to predict areas of contamination using random forest (RF). This example model will leverage a dataset of well water variables that span geospatial location, sampling date, and well water attributes, with the goal of predicting whether detectable levels of inorganic arsenic (iAs) are present. This dataset was obtained through the sampling of 713 private wells across North Carolina through the University of North Carolina Superfund Research Program (UNC-SRP) using an analytical method that was capable of detecting levels of iAs greater than 5ppm. As demonstrated through the script below, the algorithm will first be trained and tested, and then resulting model performance will be assessed using the previously detailed confusion matrix and related performance metrics. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: Which well water variables, spanning various geospatial locations, sampling dates, and well water attributes, significantly differ between samples containing detectable levels of iAs vs samples that are not contaminated/ non-detectable? How can we train a random forest (RF) model to predict whether a well might be contaminated with iAs? With this RF model, can we predict if iAs will be detected based on well water information? How could this RF model be improved upon, acknowledging that there is class imbalance? Script Preparations Cleaning the global environment rm(list=ls()) Installing required R packages If you already have these packages installed, you can skip this step, or you can run the below code which checks installation status for you if (!requireNamespace(&quot;readxl&quot;)) install.packages(&quot;readxl&quot;); if (!requireNamespace(&quot;lubridate&quot;)) install.packages(&quot;lubridate&quot;); if (!requireNamespace(&quot;tidyverse&quot;)) install.packages(&quot;tidyverse&quot;); if (!requireNamespace(&quot;gtsummary&quot;)) install.packages(&quot;gtsummary&quot;); if (!requireNamespace(&quot;flextable&quot;)) install.packages(&quot;flextable&quot;); if (!requireNamespace(&quot;caret&quot;)) install.packages(&quot;caret&quot;); if (!requireNamespace(&quot;randomForest&quot;)) install.packages(&quot;randomForest&quot;); Loading R packages required for this session library(readxl); library(lubridate); library(tidyverse); library(gtsummary); library(flextable); library(caret); library(randomForest); Set your working directory setwd(&quot;/filepath to where your input files are&quot;) Importing example dataset # Load the data arsenic_data &lt;- data.frame(read_xlsx(&quot;Module5_2_Input/Module5_2_InputData.xlsx&quot;)) # View the top of the dataset head(arsenic_data) ## Well_ID Water_Sample_Date Casing_Depth Well_Depth Static_Water_Depth ## 1 W_1 9/24/12 52 165 41 ## 2 W_2 12/17/15 40 445 42 ## 3 W_3 2/2/15 45 160 40 ## 4 W_4 10/22/12 42 440 57 ## 5 W_5 1/3/11 48 120 42 ## 6 W_6 12/15/15 60 280 32 ## Flow_Rate pH Detect_Concentration ## 1 60.0 7.7 ND ## 2 2.0 7.3 ND ## 3 40.0 7.4 ND ## 4 1.5 8.0 D ## 5 25.0 7.1 ND ## 6 10.0 8.2 D The columns in this dataset are described below: Well_ID: Unique id for each well (This is the sample identifier and not a predictive feature) Water_Sample_Date: Date that the well was sampled Casing_Depth: Depth of the casing of the well (ft) Well_Depth: Depth of the well (ft) Static_Water_Depth: Static water depth in the well (ft) Flow_Rate: Well flow rate (gallons per minute) pH: pH of water sample Detect_Concentration: Binary identifier (either non-detect “ND” or detect “D”) if iAs concentration detected in water sample Changing Data Types First, Detect_Concentration needs to be converted from a character to a factor so that Random Forest knows that the non-detect class is the baseline or “negative” class, while the detect class will be the “positive” class. Water_Sample_Date will be converted from a character to a date type using the mdy() function from the lubridate package. This is done so that the model understands this column contains dates. arsenic_data &lt;- arsenic_data %&gt;% # Converting `Detect_Concentration` from a character to a factor mutate(Detect_Concentration = relevel(factor(Detect_Concentration), ref = &quot;ND&quot;), # Converting water sample date from a character to a date type Water_Sample_Date = mdy(Water_Sample_Date)) %&gt;% # Removing tax id and only keeping the predictor and outcome variables in the dataset # This allows us to put the entire dataframe as is into RF select(-Well_ID) # Look at the top of the revised dataset head(arsenic_data) ## Water_Sample_Date Casing_Depth Well_Depth Static_Water_Depth Flow_Rate pH ## 1 2012-09-24 52 165 41 60.0 7.7 ## 2 2015-12-17 40 445 42 2.0 7.3 ## 3 2015-02-02 45 160 40 40.0 7.4 ## 4 2012-10-22 42 440 57 1.5 8.0 ## 5 2011-01-03 48 120 42 25.0 7.1 ## 6 2015-12-15 60 280 32 10.0 8.2 ## Detect_Concentration ## 1 ND ## 2 ND ## 3 ND ## 4 D ## 5 ND ## 6 D Testing for Differences in Predictor Variables across the Outcome Classes It is useful to run summary statistics on the variables that will be used as predictors in the algorithm to see if there are differences in distributions between the outcomes classes (either non-detect or detect in this case). Typically, greater significance often leads to better predictivity for a certain variable, since the model is better able to separate the classes. We’ll use the tbl_summary() function from the gtsummary package. Note, this may only be practical with smaller datasets or for a subset of predictors if there are many. For more information on the tbl_summary() function, check out this helpful Tutorial. arsenic_data %&gt;% # Displaying the mean and standard deviation in parentheses for all continuous variables tbl_summary(by = Detect_Concentration, statistic = list(all_continuous() ~ &quot;{mean} ({sd})&quot;)) %&gt;% # Adding a column that displays the total number of samples for each variable. This will be 713 for all variables since we have no missing data add_n() %&gt;% # Adding a column that displays the p-value from an anova test add_p(test = list(all_continuous() ~ &quot;aov&quot;)) %&gt;% as_flex_table() %&gt;% bold(bold = TRUE, part = &quot;header&quot;) ## The following errors were returned during `as_flex_table()`: ## ✖ For variable `Casing_Depth` (`Detect_Concentration`) and &quot;p.value&quot; ## statistic: The package &quot;cardx&quot; (&gt;= 0.2.1) is required. ## ✖ For variable `Flow_Rate` (`Detect_Concentration`) and &quot;p.value&quot; ## statistic: The package &quot;cardx&quot; (&gt;= 0.2.1) is required. ## ✖ For variable `Static_Water_Depth` (`Detect_Concentration`) and ## &quot;p.value&quot; statistic: The package &quot;cardx&quot; (&gt;= 0.2.1) is required. ## ✖ For variable `Water_Sample_Date` (`Detect_Concentration`) and ## &quot;p.value&quot; statistic: The package &quot;cardx&quot; (&gt;= 0.2.1) is required. ## ✖ For variable `Well_Depth` (`Detect_Concentration`) and &quot;p.value&quot; ## statistic: The package &quot;cardx&quot; (&gt;= 0.2.1) is required. ## ✖ For variable `pH` (`Detect_Concentration`) and &quot;p.value&quot; statistic: ## The package &quot;cardx&quot; (&gt;= 0.2.1) is required. .cl-0244e6de{}.cl-023f84c8{font-family:'Arial';font-size:11pt;font-weight:bold;font-style:normal;text-decoration:none;color:rgba(0, 0, 0, 1.00);background-color:transparent;}.cl-023f84d2{font-family:'Arial';font-size:6.6pt;font-weight:bold;font-style:normal;text-decoration:none;color:rgba(0, 0, 0, 1.00);background-color:transparent;position: relative;bottom:3.3pt;}.cl-023f84dc{font-family:'Arial';font-size:11pt;font-weight:normal;font-style:normal;text-decoration:none;color:rgba(0, 0, 0, 1.00);background-color:transparent;}.cl-023f84dd{font-family:'Arial';font-size:6.6pt;font-weight:normal;font-style:normal;text-decoration:none;color:rgba(0, 0, 0, 1.00);background-color:transparent;position: relative;bottom:3.3pt;}.cl-02418cfa{margin:0;text-align:left;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);padding-bottom:2pt;padding-top:2pt;padding-left:5pt;padding-right:5pt;line-height: 1;background-color:transparent;}.cl-02418cfb{margin:0;text-align:center;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);padding-bottom:2pt;padding-top:2pt;padding-left:5pt;padding-right:5pt;line-height: 1;background-color:transparent;}.cl-02418d04{margin:0;text-align:left;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);padding-bottom:5pt;padding-top:5pt;padding-left:5pt;padding-right:5pt;line-height: 1;background-color:transparent;}.cl-02418d05{margin:0;text-align:center;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);padding-bottom:5pt;padding-top:5pt;padding-left:5pt;padding-right:5pt;line-height: 1;background-color:transparent;}.cl-02418d06{margin:0;text-align:left;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);padding-bottom:5pt;padding-top:5pt;padding-left:5pt;padding-right:5pt;line-height: 1;background-color:transparent;}.cl-0241a1b8{width:1.703in;background-color:transparent;vertical-align: middle;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 1pt solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1b9{width:0.54in;background-color:transparent;vertical-align: middle;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 1pt solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1c2{width:2.528in;background-color:transparent;vertical-align: middle;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 1pt solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1c3{width:0.82in;background-color:transparent;vertical-align: middle;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 1pt solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1cc{width:1.703in;background-color:transparent;vertical-align: top;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1cd{width:0.54in;background-color:transparent;vertical-align: top;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1ce{width:2.528in;background-color:transparent;vertical-align: top;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1cf{width:0.82in;background-color:transparent;vertical-align: top;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1d6{width:1.703in;background-color:transparent;vertical-align: top;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1d7{width:0.54in;background-color:transparent;vertical-align: top;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1d8{width:2.528in;background-color:transparent;vertical-align: top;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1d9{width:0.82in;background-color:transparent;vertical-align: top;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1da{width:1.703in;background-color:transparent;vertical-align: middle;border-bottom: 0 solid rgba(255, 255, 255, 0.00);border-top: 0 solid rgba(255, 255, 255, 0.00);border-left: 0 solid rgba(255, 255, 255, 0.00);border-right: 0 solid rgba(255, 255, 255, 0.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1e0{width:0.54in;background-color:transparent;vertical-align: middle;border-bottom: 0 solid rgba(255, 255, 255, 0.00);border-top: 0 solid rgba(255, 255, 255, 0.00);border-left: 0 solid rgba(255, 255, 255, 0.00);border-right: 0 solid rgba(255, 255, 255, 0.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1e1{width:2.528in;background-color:transparent;vertical-align: middle;border-bottom: 0 solid rgba(255, 255, 255, 0.00);border-top: 0 solid rgba(255, 255, 255, 0.00);border-left: 0 solid rgba(255, 255, 255, 0.00);border-right: 0 solid rgba(255, 255, 255, 0.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1e2{width:0.82in;background-color:transparent;vertical-align: middle;border-bottom: 0 solid rgba(255, 255, 255, 0.00);border-top: 0 solid rgba(255, 255, 255, 0.00);border-left: 0 solid rgba(255, 255, 255, 0.00);border-right: 0 solid rgba(255, 255, 255, 0.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}CharacteristicNND N = 5151D N = 1981p-valueWater_Sample_Date7132013-06-05 (979.174260670888)2013-03-05 (957.843005291701)Casing_Depth71374 (33)55 (23)Well_Depth713301 (144)334 (128)Static_Water_Depth71335 (12)36 (13)Flow_Rate71325 (33)14 (16)pH7137.45 (0.55)7.82 (0.40)1Mean (SD) Note that N refers to the total sample number; ND refers to the samples that contained non-detectable levels of iAs; and D refers to the samples that contained detectable levels of iAs. Answer to Environmental Health Question 1 With this, we can answer Environmental Health Question #1: Which well water variables, spanning various geospatial locations, sampling dates, and well water attributes, significantly differ between samples containing detectable levels of iAs vs samples that are not contaminated/ non-detect? Answer: All of the evaluated descriptor variables are significantly different, with p&lt;0.05 between detect and non-detect iAs samples, with the exception of the sample date and the static water depth. With these findings, we feel comfortable moving forward with these well water descriptive variables as predictors in our model. Setting up Cross Validation At this point, we can move forward with training and testing a RF model aimed at predicting whether or not detectable levels of iAs are present in well water samples. We’ll take a glance at the distribution of Detect_Concentration between the two classes. # Set seed for reproducibility set.seed(17) # Establish a list of indices that will used to identify our training and testing data with a 60-40 split tt_indices &lt;- createDataPartition(y = arsenic_data$Detect_Concentration, p = 0.6, list = FALSE) # Use indices to make our training and testing datasets and view the number of Ds and NDs iAs_train &lt;- arsenic_data[tt_indices,] table(iAs_train$Detect_Concentration) ## ## ND D ## 309 119 iAs_test &lt;- arsenic_data[-tt_indices,] table(iAs_test$Detect_Concentration) ## ## ND D ## 206 79 We can see that there are notably more non-detects (ND) than detects (D) in both our training and testing sets. This is something important to consider when evaluating our model’s performance. Now we can set up our cross validation and train our model. We will be using the trainControl() function from the caret package for this task. It is one of the most commonly used libraries for supervised machine learning in R and can be leveraged for a variety algorithms including RF, SVM, KNN, and others. This model will be trained with 5-fold cross validation. Additionally, we will test 2, 3, and 6 predictors through the mtry parameter. See the caret documentation here. # Establish the parameters for our cross validation with 5 folds control &lt;- trainControl(method = &#39;cv&#39;, number = 5, search = &#39;grid&#39;, classProbs = TRUE) # Establish grid of predictors to test in our model as part of hyperparameter tuning p &lt;- ncol(arsenic_data) - 1 # p is the total number of predictors in the dataset tunegrid_rf &lt;- expand.grid(mtry = c(floor(sqrt(p)), p/2, p)) # We will test sqrt(p), p/2, and p predictors (2,3,&amp; 6 predictors, respectively) to see which performs best Predicting iAs Detection with a Random Forest (RF) Model # Look at the column names in training dataset colnames(iAs_train) ## [1] &quot;Water_Sample_Date&quot; &quot;Casing_Depth&quot; &quot;Well_Depth&quot; ## [4] &quot;Static_Water_Depth&quot; &quot;Flow_Rate&quot; &quot;pH&quot; ## [7] &quot;Detect_Concentration&quot; # Train model rf_train &lt;- train(x = iAs_train[,1:6], # Our predictor variables are in columns 1-6 of the dataframe y = iAs_train[,7], # Our outcome variable is in column 7 of the dataframe trControl = control, # Specify the cross-validation parameters we defined above method = &#39;rf&#39;, # Specify we want to train a Random Forest importance = TRUE, # This parameter calculates the variable importance for RF models specifically which can help with downstream analyses tuneGrid = tunegrid_rf, # Specify the number of predictors we want to test as defined above metric = &quot;Accuracy&quot;, ) # Specify what evaluation metric we want to use to decide which model is the best # Look at the results of training rf_train ## Random Forest ## ## 428 samples ## 6 predictor ## 2 classes: &#39;ND&#39;, &#39;D&#39; ## ## No pre-processing ## Resampling: Cross-Validated (5 fold) ## Summary of sample sizes: 344, 342, 342, 342, 342 ## Resampling results across tuning parameters: ## ## mtry Accuracy Kappa ## 2 0.7548726 0.3329518 ## 3 0.7361019 0.2891243 ## 6 0.7501661 0.3257679 ## ## Accuracy was used to select the optimal model using the largest value. ## The final value used for the model was mtry = 2. # Save the best model from our training. The best performing model is determined by the number of predictor variables we tested that resulted in the highest accuracy during the cross validation step. rf_final &lt;- rf_train$finalModel # View confusion matrix for best model rf_final ## ## Call: ## randomForest(x = x, y = y, mtry = param$mtry, importance = TRUE) ## Type of random forest: classification ## Number of trees: 500 ## No. of variables tried at each split: 2 ## ## OOB estimate of error rate: 24.77% ## Confusion matrix: ## ND D class.error ## ND 275 34 0.1100324 ## D 72 47 0.6050420 Answer to Environmental Health Question 2 With this, we can answer Environmental Health Question #2: How can we train a random forest (RF) model to predict whether a well might be contaminated with iAs? Answer: As is standard practice with supervised ML, we split our full dataset into a training dataset and a test dataset using a 60-40 split. Using the caret package, we implemented 5-fold cross validation to train a RF while also testing different numbers of predictors to see which optimized performance. The model that resulted in the greatest accuracy was selected as the final model. Now we can see how well our model does on data it hasn’t seen before by applying it to our testing data. # Use our best model to predict the classes for our test data. We need to make sure we remove the column of Ds/NDs from our test data. rf_res &lt;- predict(rf_final, iAs_test %&gt;% select(!Detect_Concentration)) # View a confusion matrix of the results and gauge model performance # Be sure to include the &#39;positive&#39; parameter to specify the correct positive class confusionMatrix(rf_res, iAs_test$Detect_Concentration, positive = &quot;D&quot;) ## Confusion Matrix and Statistics ## ## Reference ## Prediction ND D ## ND 178 47 ## D 28 32 ## ## Accuracy : 0.7368 ## 95% CI : (0.6817, 0.787) ## No Information Rate : 0.7228 ## P-Value [Acc &gt; NIR] : 0.32445 ## ## Kappa : 0.2907 ## ## Mcnemar&#39;s Test P-Value : 0.03767 ## ## Sensitivity : 0.4051 ## Specificity : 0.8641 ## Pos Pred Value : 0.5333 ## Neg Pred Value : 0.7911 ## Prevalence : 0.2772 ## Detection Rate : 0.1123 ## Detection Prevalence : 0.2105 ## Balanced Accuracy : 0.6346 ## ## &#39;Positive&#39; Class : D ## Answer to Environmental Health Question 3 With this, we can answer Environmental Health Question #3: With this RF model, can we predict if iAs will be detected based on well water information? Answer: We can use this model to predict if iAs can be detected in well water given that an overall accuracy of ~0.72 is decent, however we should consider other metrics that may influence how good we feel about this model depending on what is important to the question we are trying to answer. For example, the model did a good job at predicting non-detect data based on a sensitivity of ~0.85 and a NPV ~0.78, but struggled at predicting detect data based on a specificity of ~0.39 and a PPV of ~0.50. Additionally, the balanced accuracy of ~0.62 further emphasizes the difference in predictive ability of the model for non-detects and detects. If it is highly important to us that detects are classified correctly, we may want to improve this model before implementing it. Class Imbalance It is worth noting this discrepancy in predictive capabilities for detects vs. non-detects makes sense due to the observed class imbalance in our training data. There were notably more non-detects than detects in the training set, so the model was exposed to more of these data points and struggles to distinguish unique characteristics of detects when compared to non-detects. Additionally, we told the training algorithm to prioritize selecting a final model based on its overall accuracy. In the instances of a heavy class imbalance, it is common for a high accuracy to be achieved as the more prevalent class is predicted more often, though this doesn’t give the full picture of the model’s predictive capabilities. For example, if you consider a dog/cat case with a set of 90 dogs and 10 cats, a model could achieve 90% accuracy by predicting dog every time, which isn’t at all helpful in predicting cats. This is particularly important, because for toxicology related datasets, the “positive” class often represents the class with greater public health risk/ interest but can have less data. For example, when you classify subjects based upon whether or not they have asthma based on gene expression data. Asthmatics would likely be the “positive” class, but given that asthmatics are less prevalent than non-asthmatics in the general population, they would likely represent the minority class too. To address this issue, a few methods can be considered. Full implementation of these approaches is beyond the scope of this module, but relevant resources for further exploration are given. Synthetic Minority Oversampling Technique (SMOTE)- increases the number of minority classes in the training data, thereby reducing the class imbalance by synthetically generating additional samples derived from the existing minority class samples. SMOTE Oversampling &amp; Tutorial On How To Implement In Python And R How to Use SMOTE for Imbalanced Data in R (With Example) Adjusting the loss function- Loss functions in machine learning quantify the penalty for a bad prediction. They can be adjusted to where the minority class is penalized more forcing the model to learn to make fewer mistakes when predicting the minority class. Alternative Performance Metrics- When training the model, alternative metrics to overall accuracy may yield a more robust model capable of better predicting the minority class. Example alternatives may include balanced accuracy or an F1-score. The caret package further allows for custom, user-defined metrics to be evaluated during training by specifying the summaryFunction parameter in the trainControl() function, as seen below, in addition to the defaultSummary() and twoClassSummary() functions. In the example code below, we’re creating a function (f1) that will calculate the F1 score and find the optimal model with the highest F1 score as opposed to the highest accuracy as we did above. install.packages(&quot;MLmetrics&quot;) library(MLmetrics) f1 &lt;- function(data, lev = NULL, model = NULL) { # Creating a function to calculate the F1 score f1_val &lt;- F1_Score(y_pred = data$pred, y_true = data$obs, positive = lev[1]) c(F1 = f1_val) } # 5 fold CV ctrl &lt;- trainControl( method = &quot;cv&quot;, number = 5, classProbs = TRUE, summaryFunction = f1 ) # Training the RF model mod &lt;- train(x = X, y = Y, trControl = ctrl, method = &quot;rf&quot;, tuneGrid = tunegrid_rf, importance = TRUE, # Basing the best model performance off of the F1 score within 5 CV metric = &quot;F1&quot;) For more in-depth information and additional ways to address class imbalance check out How to Deal with Imbalanced Data in Classification. Answer to Environmental Health Question 4 With this, we can answer Environmental Health Question #4: How could this RF model be improved upon, acknowledging that there is class imbalance? Answer: We can implement SMOTE to increase the number of training data points for the minority class thereby reducing the class imbalance. In conjunction with using SMOTE, another approach includes selecting an alternative performance metric during training that does a better job taking the existing class imbalance into consideration, such as balanced accuracy or an F1-score, improves our predictive ability for the minority class. Concluding Remarks In conclusion, this training module has provided an introduction to supervised machine learning using classification techniques in R. Machine learning is a powerful tool that can help researchers gain new insights and improve models to analyze complex datasets faster and in a more comprehensive way. The example we’ve explored demonstrates the utility of supervised machine learning models on an environmentally relevant dataset. Additional Resources To learn more check out the following resources: IBM - What is Machine Learning Curate List of AI and Machine Learning Resources Introduction to Machine Learning in R Machine Learning by Mueller, J. P. (2021). Machine learning for dummies. John Wiley &amp; Sons. Test Your Knowledge Using the “Module5_2TYKInput.xlsx”, use RF to determine if well water data can be accurate predictors of Manganese detection. The data is structured similarly to the “Module5_2_InputData.xlsx” used in this module, however it now includes 4 additional features: Longitude: Longitude of address (decimal degrees) Latitude: Latitude of address (decimal degrees) Stream_Distance: Euclidean distance to the nearest stream (feet) Elevation: Surface elevation of the sample location (feet) "],["supervised-machine-learning-model-interpretation.html", "Supervised Machine Learning Model Interpretation Introduction to Training Module Variable Importance Decision Boundary Introduction to Example Dataset and Activity Predicting iAs Detection with a Random Forest (RF) Model Variable Importance Plot Decision Boundary Plot Decision Boundary Plot Incorporating SMOTE Concluding Remarks", " Supervised Machine Learning Model Interpretation This training module was developed by Alexis Payton, Lauren E. Koval, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module Supervised machine learning (ML) represents a subset of ML methods wherein the outcome variable is known or assigned prior to training a model to be able to predict said outcome. As we discussed in previous modules, ML methods are advantageous in that they easily incorporate a multitude of potential predictor variables, which allows these models to more closely consider real-world, complex environmental health scenarios and offer new insights through a more holistic consideration of available data inputs. However, one disadvantage of ML is that it is often not as easily interpretable as traditional statistics (e.g., regression based methods with defined beta coefficients for each input predictor variable). With this limitation in mind, there are methods and concepts that can be applied to supervised ML algorithms to aid in the understanding of their predictions including variable (feature) importance and decision boundaries, which we will cover in this module. We will also include example visualization techniques of these methods, representing important aspects contributing to model interpretability, since visualizing helps convey concepts faster and across a broader target audience. In addition, this module addresses methods to communicate these findings in a paper so that a wider span of readers can understand overall take-home points. As with other data analyses, we advise to focus just as much on the why components of a study’s research question(s) as opposed to only focusing on the what or how. To elaborate, we explain through this module that it is not as important to explain all the intricacies of how a model works and how its parameters were tuned; rather, it is more important to focus on why a particular model was selected and how it will be leveraged to answer your research questions. This can all be a bit subjective and requires expertise within your research field. As a first step, let’s first learn about some model interpretation methodologies highlighting Variable Importance and Decision Boundaries as important examples relevant to environmental health research. Then, this training module will further describe approaches to summarize these methods and communicate supervised ML findings to a broader audience. Variable Importance When a supervised ML algorithm makes predictions, it relies more heavily on some variables than others. How much a variable contributes to classifying data is known as variable (feature) importance. Often times, this is thought of as the impact on overall model performance if a variable were to be removed from the model. There are many methods that are used to measure feature importance, including… SHapley Additive exPlanations (SHAP): based on game theory where each variable is considered a “player” where we’re seeking to determine each player’s contribution to the outcome of a “game” or overall model performance. It divides the model performance metric amongst all the variables, so that the sum of the shapley values for all the predictors is equal to the overall model performance. For more information on SHAP, see A Novel Approach to Feature Importance. Mean decrease gini (gini impurity): quantifies the improvement of predictivity with the addition of each predictor in a decision tree, which is then averaged over all the decision trees tested. The higher the value the greater the importance on the algorithm. This metric can easily be extracted from classification-based models, including random forest (RF) classifications, which is what we will focus on in this module. Note for RF regression-based models, node purity can be extracted as a measure of feature importance. For more information, please see the following resources regarding Feature Importance and Mean Decrease Gini. Decision Boundary Another concept that is pertinent to a model’s interpretability is understanding a decision boundary and how visualizing it can further aid in understanding how the model classifies new data points. A decision boundary is a line (or a hyperplane) that seeks to separate the training data by class. This line can be linear or non-linear and is formed in n-dimensional space. To clarify, although support vector machine (SVM) specifically uses decision boundaries to classify training data and make predictions on test data, decision boundaries can still be drawn for other algorithms. A decision boundary can be visualized to convey how well an algorithm is able to classify an outcome based on the data given. It is important to note that most ML models make use of datasets that contain three or more predictors, and it is difficult to visualize a plot in more than three dimensions. Therefore, the number of features and which features to plot need to be narrowed down to two variables. For this reason, the resulting visualization is not a true representation of the decision boundary from the initial model using all predictors, since the visualization only relies on prediction results from two variables. Nevertheless, decision boundary plots can be powerful visualizations to determine thresholds between the outcome classes. When choosing variables for decision boundary plots, features that have the most influence on the model are often selected, but that is not always the case. Sometimes predictors are selected based upon the environmental health implications relevant to the research question. For example in Perryman et. al, lung response following ozone exposure was investigated by sampling derivatives of cholesterol biosynthesis in human subjects. In this paper, these sterol metabolites were used to predict whether a subject would be classified as having a lung response that was considered non-responsive or responsive. A decision boundary plot was made using two predictors: Cholesterol, given that it had the highest variable importance and Vitamin D, given its synthesis can be affected by ozone despite it having a lower variable importance in the paper’s models. Figure 5. Decision boundary plot for SVM model predicting lung response class. Cholesterol and 25-hydroxyvitamin D were used as predictors visualizing responder status [non-responders(green) and responders (yellow)] and disease status [non-asthmatics (triangles) and asthmatics (circles)]. The shaded regions are the model’s prediction of a subject’s lung response class at a given cholesterol and 25-hydroxyvitamin D concentration. Takeaways from this decision boundary plot: Subjects with more lung inflammation (“responders”) after ozone exposure tended to have higher Vitamin D levels (&gt; 35pmol/mL) and lower Cholesterol levels (&lt; 675nmol/mL). These “responder” subjects were more likely to be non-asthmatics. Introduction to Example Dataset and Activity In the previous module, we investigated whether a classification-based RF model using well water variables would be accurate predictors of inorganic arsenic (iAs) contamination. While it is helpful to know if certain variables are able to be used to construct a model that accurately predict detectability, from a public health standpoint, it is also helpful to know which of those features contribute the most to a model’s accuracy. Therefore, if we can identify the features that are associated with having lower arsenic detection, we can use that information to inform policies when new wells are constructed. In addition to identifying variables with the greatest importance to the algorithm, it is also pertinent to understand the ranges of when a well is more or less likely to have arsenic detected. For example, are wells with a lower flow rate more likely to have arsenic detected? In this module, this will be addressed by extracting variable importance from the same algorithm and plotting it. The two features with the highest variable importance will be identified and used to construct a decision boundary plot to determine how features are associated with iAs detection. The data to be used in this module was described and referenced previously in TAME 2.0 Module 5.2 Supervised Machine Learning. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: After plotting variable importance from highest to lowest, which two predictors have the highest variable importance on the predictive accuracy of iAs detection from a RF algorithm? Using the two features with the highest variable importance, under what conditions are we more likely to predict detectable iAs in wells based on a decision boundary plot? How do the decision boundaries shift after incorporating SMOTE to address class imbalance? Script Preparations Cleaning the global environment rm(list=ls()) Installing required R packages If you already have these packages installed, you can skip this step, or you can run the below code which checks installation status for you if (!requireNamespace(&quot;readxl&quot;)) install.packages(&quot;readxl&quot;); if (!requireNamespace(&quot;lubridate&quot;)) install.packages(&quot;lubridate&quot;); if (!requireNamespace(&quot;tidyverse&quot;)) install.packages(&quot;tidyverse&quot;); if (!requireNamespace(&quot;caret&quot;)) install.packages(&quot;caret&quot;); if (!requireNamespace(&quot;randomForest&quot;)) install.packages(&quot;randomForest&quot;); if (!requireNamespace(&quot;themis&quot;)) install.packages(&quot;themis&quot;); Loading R packages required for this session library(readxl) library(lubridate) library(tidyverse) library(caret) library(randomForest) library(e1071) library(ggsci) library(themis) Set your working directory setwd(&quot;/filepath to where your input files are&quot;) Importing example dataset # Load the data arsenic_data &lt;- data.frame(read_excel(&quot;Module5_3_Input/Module5_3_InputData.xlsx&quot;)) # View the top of the dataset head(arsenic_data) ## Well_ID Water_Sample_Date Casing_Depth Well_Depth Static_Water_Depth ## 1 W_1 9/24/12 52 165 41 ## 2 W_2 12/17/15 40 445 42 ## 3 W_3 2/2/15 45 160 40 ## 4 W_4 10/22/12 42 440 57 ## 5 W_5 1/3/11 48 120 42 ## 6 W_6 12/15/15 60 280 32 ## Flow_Rate pH Detect_Concentration ## 1 60.0 7.7 ND ## 2 2.0 7.3 ND ## 3 40.0 7.4 ND ## 4 1.5 8.0 D ## 5 25.0 7.1 ND ## 6 10.0 8.2 D Changing Data Types First, Detect_Concentration needs to be converted from a character to a factor so that Random Forest knows that the non-detect class is the baseline or “negative” class, while the detect class will be the “positive” class. Water_Sample_Date will be converted from a character to a date type using the mdy() function from the lubridate package. This is done so that the model understands this column contains dates. arsenic_data &lt;- arsenic_data %&gt;% # Converting `Detect_Concentration` from a character to a factor mutate(Detect_Concentration = relevel(factor(Detect_Concentration), ref = &quot;ND&quot;), # Converting water sample date from a character to a date type Water_Sample_Date = mdy(Water_Sample_Date)) %&gt;% # Removing well id and only keeping the predictor and outcome variables in the dataset # This allows us to put the entire dataframe as is into RF select(-Well_ID) # View the top of the current dataset head(arsenic_data) ## Water_Sample_Date Casing_Depth Well_Depth Static_Water_Depth Flow_Rate pH ## 1 2012-09-24 52 165 41 60.0 7.7 ## 2 2015-12-17 40 445 42 2.0 7.3 ## 3 2015-02-02 45 160 40 40.0 7.4 ## 4 2012-10-22 42 440 57 1.5 8.0 ## 5 2011-01-03 48 120 42 25.0 7.1 ## 6 2015-12-15 60 280 32 10.0 8.2 ## Detect_Concentration ## 1 ND ## 2 ND ## 3 ND ## 4 D ## 5 ND ## 6 D Setting up Cross Validation Note that the code below is different than the code presented in the previous module, TAME 2.0 Module 5.2 Supervised Machine Learning. Both coding methods are valid and produce comparable results, however we wanted to present another way to run k-fold cross validation and random forest. In 5-fold cross validation (CV), there are 5 equally-sized folds (ideally!). This means that 80% of the original dataset is split into the 4 folds that comprise the training set and the remaining 20% in the last fold is reserved for the test set. Previously, the trainControl() function was used for CV. This time we’ll use the createFolds() function also from the caret package. # Setting seed for reproducibility set.seed(12) # 5-fold cross validation arsenic_index = createFolds(arsenic_data$Detect_Concentration, k = 5) # Seeing if about 20% of the records are in the testing set kfold1 = arsenic_index[[1]] length(kfold1)/nrow(arsenic_data) ## [1] 0.1991585 # Creating vectors for parameters to be tuned ntree_values = c(50, 250, 500) # number of decision trees p = dim(arsenic_data)[2] - 1 # number of predictor variables in the dataset mtry_values = c(sqrt(p), p/2, p) # number of predictors to be used in the model Predicting iAs Detection with a Random Forest (RF) Model Notice that in the code below we are choosing the final RF model to be the one with the lowest out of bag (OOB) error. In the previous module, the final model was chosen based on the highest accuracy, however this is a similar approach here given that OOB error = 1 - Accuracy. # Setting the seed again so the predictions are consistent set.seed(12) # Creating an empty dataframe to save the confusion matrix metrics and variable importance metrics = data.frame() variable_importance_df = data.frame() # Iterating through the cross validation folds for (i in 1:length(arsenic_index)){ # Training data data_train = arsenic_data[-arsenic_index[[i]],] # Test data data_test = arsenic_data[arsenic_index[[i]],] # Creating empty lists and dataframes to store errors reg_rf_pred_tune = list() rf_OOB_errors = list() rf_error_df = data.frame() # Tuning parameters: using ntree and mtry values to determine which combination yields the smallest OOB error # from the validation datasets for (j in 1:length(ntree_values)){ for (k in 1:length(mtry_values)){ # Running RF to tune parameters reg_rf_pred_tune[[k]] = randomForest(Detect_Concentration ~ ., data = data_train, ntree = ntree_values[j], mtry = mtry_values[k]) # Obtaining the OOB error rf_OOB_errors[[k]] = data.frame(&quot;Tree Number&quot; = ntree_values[j], &quot;Variable Number&quot; = mtry_values[k], &quot;OOB_errors&quot; = reg_rf_pred_tune[[k]]$err.rate[ntree_values[j],1]) # Storing the values in a dataframe rf_error_df = rbind(rf_error_df, rf_OOB_errors[[k]]) } } # Finding the lowest OOB error from the 5 folds using best number of predictors at split best_oob_errors &lt;- which(rf_error_df$OOB_errors == min(rf_error_df$OOB_errors)) # Now running RF on the entire training set with the tuned parameters # This will be done 5 times for each fold reg_rf &lt;- randomForest(Detect_Concentration ~ ., data = data_train, ntree = rf_error_df$Tree.Number[min(best_oob_errors)], mtry = rf_error_df$Variable.Number[min(best_oob_errors)]) # Predicting on test set and adding the predicted values as an additional column to the test data data_test$Pred_Detect_Concentration = predict(reg_rf, newdata = data_test, type = &quot;response&quot;) matrix = confusionMatrix(data = data_test$Pred_Detect_Concentration, reference = data_test$Detect_Concentration, positive = &quot;D&quot;) # Extracting accuracy, sens, spec, PPV, NPV and adding to the dataframe to take mean later matrix_values = data.frame(t(c(matrix$byClass[11])), t(c(matrix$byClass[1:4]))) metrics = rbind(metrics, matrix_values) # Extracting variable importance variable_importance_values = data.frame(importance(reg_rf)) %&gt;% rownames_to_column(var = &quot;Predictor&quot;) variable_importance_df = rbind(variable_importance_df, variable_importance_values) } # Taking average across the 5 folds metrics = metrics %&gt;% summarise(`Balanced Accuracy` = mean(Balanced.Accuracy), Sensitivity = mean(Sensitivity), Specificity = mean(Specificity), PPV = mean(Pos.Pred.Value), NPV = mean(Neg.Pred.Value)) variable_importance_df = variable_importance_df %&gt;% group_by(Predictor) %&gt;% summarise(MeanDecreaseGini = mean(MeanDecreaseGini)) %&gt;% # Sorting from highest to lowest arrange(-MeanDecreaseGini) The confusion matrix results from the previous module are shown below. Now let’s double check that when using this new method, our results are still comparable. # First comparing results to the previous module round(metrics, 2) ## Balanced Accuracy Sensitivity Specificity PPV NPV ## 1 0.64 0.41 0.87 0.55 0.79 They are! Now we’ll take a look at the model’s variable importance. variable_importance_df ## # A tibble: 6 × 2 ## Predictor MeanDecreaseGini ## &lt;chr&gt; &lt;dbl&gt; ## 1 Casing_Depth 50.6 ## 2 pH 42.3 ## 3 Water_Sample_Date 37.6 ## 4 Flow_Rate 33.5 ## 5 Well_Depth 32.2 ## 6 Static_Water_Depth 31.7 Although we have the results we need, let’s take it a step further and plot the data. Reformatting the dataframe for plotting First, the dataframe will be transformed so that the figure is more legible. Specifically, spaces will be added between the variables, and the Predictor column will be put into a factor to rearrange the order of the variables from lowest to highest mean decrease gini. For additional information on tricks like this to make visualizations easier to read, see TAME 2.0 Module 3.2 Improving Data Visualizations. # Adding spaces between the variables that need the space modified_variable_importance_df = variable_importance_df %&gt;% mutate(Predictor = gsub(&quot;_&quot;, &quot; &quot;, Predictor)) # Saving the order of the variables from lowest to highest mean decrease gini by putting into a factor predictor_order = rev(modified_variable_importance_df$Predictor) modified_variable_importance_df$Predictor = factor(modified_variable_importance_df$Predictor, levels = predictor_order) head(modified_variable_importance_df) ## # A tibble: 6 × 2 ## Predictor MeanDecreaseGini ## &lt;fct&gt; &lt;dbl&gt; ## 1 Casing Depth 50.6 ## 2 pH 42.3 ## 3 Water Sample Date 37.6 ## 4 Flow Rate 33.5 ## 5 Well Depth 32.2 ## 6 Static Water Depth 31.7 Variable Importance Plot ggplot(data = modified_variable_importance_df , aes(x = MeanDecreaseGini, y = Predictor, size = 2)) + geom_point() + theme_light() + theme(axis.line = element_line(color = &quot;black&quot;), #making x and y axes black axis.text = element_text(size = 12), #changing size of x axis labels axis.title = element_text(face = &quot;bold&quot;, size = rel(1.7)), #changes axis titles legend.title = element_text(face = &#39;bold&#39;, size = 14), #changes legend title legend.text = element_text(size = 12), #changes legend text strip.text.x = element_text(size = 15, face = &quot;bold&quot;), #changes size of facet x axis strip.text.y = element_text(size = 15, face = &quot;bold&quot;)) + #changes size of facet y axis labs(x = &#39;Variable Importance&#39;, y = &#39;Predictor&#39;) + #changing axis labels guides(size = &quot;none&quot;)#removing size legend An appropriate title for this figure could be: “Figure X. Variable importance from random forest models predicting iAs detection. Variable importance is derived from mean decrease gini values extracted from random forest models. Features are listed on the y axis from greatest (top) to least (bottom) mean decrease gini.” Answer to Environmental Health Question 1 With this, we can answer Environmental Health Question #1: After plotting variable importance from highest to lowest, which two predictors have the highest variable importance on the predictive accuracy of iAs detection from a RF algorithm? Answer: From the variable importance dataframe and plot, we can see that casing depth and pH had the greatest impact on RF followed by water sample date, flow rate, static water depth, and well depth in descending order. Since casing depth and pH have been identified as the predictors with the highest variable importance, they will be prioritized as the two predictors included in the decision boundary plot example below. Decision Boundary Calculation First, models will be trained using only casing depth and pH as variables. Since, the decision boundary plot will be used for visualization purposes, and a 2-D figure can only plot two variables, we will not worry about tuning the parameters as was previously done. In this module, we’re creating a decision boundary based on a random forest model, however we’ll also explore what decision boundaries look like for other algorithms including support vector machine (SVM), and k nearest neighbor (KNN), logistic regression. Each supervised ML method has its advantages and performance is dependent upon the situation and the dataset. Therefore, it is common to see multiple models used to predict an outcome of interest in a publication. Let’s create additional boundary plots still using casing depth and pH, but this time we will use logistic regression, SVM, and KNN as comparisons to RF. # Creating a dataframe with variables based on the highest predictors highest_pred_data = data.frame(arsenic_data[,c(&quot;Casing_Depth&quot;, &quot;pH&quot;, &quot;Detect_Concentration&quot;)]) # Training RF rf_detect_arsenic = randomForest(Detect_Concentration~., data = highest_pred_data) # Logistic regression lr_detect_arsenic = glm(Detect_Concentration~., data = highest_pred_data, family = binomial(link = &#39;logit&#39;)) # SVM with a radial kernel (hyperplane) svm_detect_arsenic = svm(Detect_Concentration~., data = highest_pred_data, kernel = &quot;radial&quot;) # KNN knn_detect_arsenic = knn3(Detect_Concentration~., data = highest_pred_data) # specifying 2 classes From these predictions, decision boundaries will be calculated. This will be done by predicting Detect_Concentration between a grid of values - specifically the minimum and maximum of the two predictors (casing depth and pH). A non-linear line will be drawn on the plot to separate the two classes. get_grid_df &lt;- function(classification_model, data, resolution = 100, predict_type) { # This function predicts the outcome (Detect_Concentration) at evenly spaced data points using the two variables (pH and casing depth) # to create a decision boundary between the outcome classes (detect and non-detect samples). # :parameters: a classification-based supervised machine learning model, dataset containing the predictors and outcome variable, # specifies the number of data points to make between the minimum and maximum predictor values, prediction type # :output: a grid of values for both predictors and their corresponding predicted outcome class # Grabbing only the predictor data predictor_data &lt;- data[,1:2] # Creating a dataframe that contains the min and max for both features min_max_df &lt;- sapply(predictor_data, range, na.rm = TRUE) # Creating a vector of evenly spaced points between the min and max for the first variable (casing depth) variable1_vector &lt;- seq(min_max_df[1,1], min_max_df[2,1], length.out = resolution) # Creating a vector of evenly spaced points between the min and max for the second variable (pH) variable2_vector &lt;- seq(min_max_df[1,2], min_max_df[2,2], length.out = resolution) # Creating a dataframe of grid values by combining the two vectors grid_df &lt;- data.frame(cbind(rep(variable1_vector, each = resolution), rep(variable2_vector, time = resolution))) colnames(grid_df) &lt;- colnames(min_max_df) # Predicting class label based on all the predictor pairs of data grid_df$Pred_Class = predict(classification_model, grid_df, type = predict_type) return(grid_df) } # calling function # RF grid_df_rf = get_grid_df(rf_detect_arsenic, highest_pred_data, predict_type = &quot;class&quot;) %&gt;% # Adding in a column that indicates the model so all the dataframes can be combined mutate(Model = &quot;A. Random Forest&quot;) # SVM with a radial kernel (hyperplane) grid_df_svm = get_grid_df(svm_detect_arsenic, highest_pred_data, predict_type = &quot;class&quot;) %&gt;% mutate(Model = &quot;B. Support Vector Machine&quot;) # KNN grid_df_knn = get_grid_df(knn_detect_arsenic, highest_pred_data, predict_type = &quot;class&quot;) %&gt;% mutate(Model = &quot;C. K Nearest Neighbor&quot;) # Logistic regression grid_df_lr = get_grid_df(lr_detect_arsenic, highest_pred_data, predict_type = &quot;response&quot;) %&gt;% # First specifying the cutoff point for logistic regression predictions # If the response is &gt;= 0.5 it will be classified as a detect prediction mutate(Pred_Class = relevel(factor(ifelse(Pred_Class &gt;= 0.5, &quot;D&quot;, &quot;ND&quot;)), ref = &quot;ND&quot;), Model = &quot;D. Logistic Regression&quot;) # Creating 1 dataframe grid_df = rbind(grid_df_rf, grid_df_lr, grid_df_svm, grid_df_knn) # Viewing the dataframe to be plotted head(grid_df) ## Casing_Depth pH Pred_Class Model ## 1 27 5.400000 ND A. Random Forest ## 2 27 5.433333 ND A. Random Forest ## 3 27 5.466667 ND A. Random Forest ## 4 27 5.500000 ND A. Random Forest ## 5 27 5.533333 ND A. Random Forest ## 6 27 5.566667 ND A. Random Forest Decision Boundary Plot Now let’s plot the grid of predictions with the sampled data. # choosing palette from package ggsci_colors = pal_npg()(5) ggplot() + geom_point(data = arsenic_data, aes(x = pH, y = Casing_Depth, color = Detect_Concentration), position = position_jitter(w = 0.1, h = 0.1), size = 4, alpha = 0.8) + geom_contour(data = grid_df, aes(x = pH, y = Casing_Depth, z = as.numeric(Pred_Class == &quot;D&quot;)), color = &quot;black&quot;, breaks = 0.5) + # adds contour line geom_point(data = grid_df, aes(x = pH, y = Casing_Depth, color = Pred_Class), size = 0.1) + # shades plot xlim(5.9, NA) + # changes the limits of the x axis facet_wrap(~Model, scales = &#39;free&#39;) + theme_light() + theme(axis.line = element_line(color = &quot;black&quot;), #making x and y axes black axis.text = element_text(size = 10), #changing size of x axis labels axis.title = element_text(face = &quot;bold&quot;, size = rel(1.7)), #changes axis titles legend.title = element_text(face = &#39;bold&#39;, size = 12), #changes legend title legend.text = element_text(size = 12), #changes legend text legend.position = &quot;bottom&quot;, # move legend to top left corner legend.background = element_rect(color = &#39;black&#39;, fill = &#39;white&#39;, linetype = &#39;solid&#39;), # changes legend background strip.text = element_text(size = 15, face = &quot;bold&quot;)) + #changes size of facet x axis labs(y = &#39;Casing Depth (ft)&#39;) + #changing axis labels scale_color_manual(name = &quot;Arsenic Detection&quot;, # renaming the legend values = ggsci_colors[c(4,5)], labels = c(&#39;Non-Detect&#39;,&#39;Detect&#39;)) # renaming the classes Answer to Environmental Health Question 2 With this, we can answer Environmental Health Question #2: Using the two features with the highest variable importance, under what conditions are we more likely to predict detectable iAs in wells based on a decision boundary plot? Answer: There is some overlap between detect and non-detect iAs samples; however, it is evident that wells with detectable levels of iAs were more likely to have lower (&lt;80 ft) casing depths and a more basic pH (&gt; 7) based on RF and KNN models. It seems like SVM and logistic regression could have potentially captured a greater “detect” region indicating that the models likely struggled to predict “detect” values. In the next section, SMOTE will be used to see if these decision boundaries can be improved. Decision Boundary Plot Incorporating SMOTE Here, we will create a decision boundary plot still using casing depth and pH, but this time we will make our dataset more balance to see how improve model performance visually. The Synthetic Minority Oversampling Technique (SMOTE) was introduced in TAME 2.0 Module 5.2 Supervised Machine Learning and will be used to make the dataset more balanced by oversampling the minority class (detect values) and undersampling the majority class (non-detect values). Starting by training each model: # Using SMOTE first to balance classes balanced_highest_pred_data = smotenc(highest_pred_data, &quot;Detect_Concentration&quot;) # Training RF rf_detect_arsenic = randomForest(Detect_Concentration~., data = balanced_highest_pred_data) # Logistic regression lr_detect_arsenic = glm(Detect_Concentration~., data = balanced_highest_pred_data, family = binomial(link = &#39;logit&#39;)) # SVM with a radial kernel (hyperplane) svm_detect_arsenic = svm(Detect_Concentration~., data = balanced_highest_pred_data, kernel = &quot;radial&quot;) # KNN knn_detect_arsenic = knn3(Detect_Concentration~., data = balanced_highest_pred_data) # specifying 2 classes Now calling the get_grid_df() function we created above to create a grid of predictions. # Calling function # RF balanced_grid_df_rf = get_grid_df(rf_detect_arsenic, balanced_highest_pred_data, predict_type = &quot;class&quot;) %&gt;% # Adding in a column that indicates the model so all the dataframes can be combined mutate(Model = &quot;A. Random Forest&quot;) # SVM with a radial kernel (hyperplane) balanced_grid_df_svm = get_grid_df(svm_detect_arsenic, balanced_highest_pred_data, predict_type = &quot;class&quot;) %&gt;% mutate(Model = &quot;B. Support Vector Machine&quot;) # KNN balanced_grid_df_knn = get_grid_df(knn_detect_arsenic, balanced_highest_pred_data, predict_type = &quot;class&quot;) %&gt;% mutate(Model = &quot;C. K Nearest Neighbor&quot;) # Logistic regression balanced_grid_df_lr = get_grid_df(lr_detect_arsenic, balanced_highest_pred_data, predict_type = &quot;response&quot;) %&gt;% # First specifying the cutoff point for logistic regression predictions # If the response is &gt;= 0.5 it will be classified as a detect prediction mutate(Pred_Class = relevel(factor(ifelse(Pred_Class &gt;= 0.5, &quot;D&quot;, &quot;ND&quot;)), ref = &quot;ND&quot;), Model = &quot;D. Logistic Regression&quot;) # Creating 1 dataframe balanced_grid_df = rbind(balanced_grid_df_rf, balanced_grid_df_lr, balanced_grid_df_svm, balanced_grid_df_knn) # Viewing the dataframe to be plotted head(balanced_grid_df) ## Casing_Depth pH Pred_Class Model ## 1 27 5.400000 ND A. Random Forest ## 2 27 5.433333 ND A. Random Forest ## 3 27 5.466667 ND A. Random Forest ## 4 27 5.500000 ND A. Random Forest ## 5 27 5.533333 ND A. Random Forest ## 6 27 5.566667 ND A. Random Forest # choosing palette from package ggsci_colors = pal_npg()(5) ggplot() + geom_point(data = arsenic_data, aes(x = pH, y = Casing_Depth, color = Detect_Concentration), position = position_jitter(w = 0.1, h = 0.1), size = 4, alpha = 0.8) + geom_contour(data = balanced_grid_df, aes(x = pH, y = Casing_Depth, z = as.numeric(Pred_Class == &quot;D&quot;)), color = &quot;black&quot;, breaks = 0.5) + # adds contour line geom_point(data = balanced_grid_df, aes(x = pH, y = Casing_Depth, color = Pred_Class), size = 0.1) + # shades plot xlim(5.9, NA) + # changes the limits of the x axis facet_wrap(~Model, scales = &#39;free&#39;) + theme_light() + theme(axis.line = element_line(color = &quot;black&quot;), #making x and y axes black axis.text = element_text(size = 10), #changing size of x axis labels axis.title = element_text(face = &quot;bold&quot;, size = rel(1.7)), #changes axis titles legend.title = element_text(face = &#39;bold&#39;, size = 12), #changes legend title legend.text = element_text(size = 12), #changes legend text legend.position = &quot;bottom&quot;, # move legend to top left corner legend.background = element_rect(color = &#39;black&#39;, fill = &#39;white&#39;, linetype = &#39;solid&#39;), # changes legend background strip.text = element_text(size = 15, face = &quot;bold&quot;)) + #changes size of facet x axis labs(y = &#39;Casing Depth (ft)&#39;) + #changing axis labels scale_color_manual(name = &quot;Arsenic Detection&quot;, # renaming the legend values = ggsci_colors[c(4,5)], labels = c(&#39;Non-Detect&#39;,&#39;Detect&#39;)) # renaming the classes An appropriate title for this figure could be: “Figure X. Decision boundary plots from supervised machine learning models predicting iAs detection. The top two predictors on model performance, casing depth and pH, were used to visualize arsenic detection [non-detect (red) and detect (blue)]. The shaded regions represent prediction of a well’s detection class based on varying casing depth and pH values using (A) Random Forest, (B) Support Vector Machine, (C) K Nearest Neighbor, and (D) Logistic Regression. Answer to Environmental Health Question 3 With this, we can answer Environmental Health Question #3: How do the decision boundaries shift after incorporating SMOTE to address class imbalance? Answer: It is still evident that wells with detectable levels of iAs were more likely to have lower (&lt;80 ft) casing depths and a more basic pH (&gt; 7). However, we see the greatest shifts in the decision boundaries of SVM and logistic regression with both models now predicting greater regions to detectable iAs levels. Concluding Remarks In conclusion, this training module provided methodologies to aid in the interpretation of supervised ML with variable importance and decision boundary plots. Variable importance helps quantify the impact of each feature’s importance on an algorithm’s predictivity. The most important or environmentally-relevant predictors can be selected in a decision boundary plot to further understand and visualize the features impact on the model’s classification. Additional Resources Christoph Molnar. (2019, August 27). Interpretable Machine Learning. Github.io. https://christophm.github.io/interpretable-ml-book/ Variable Importance Decision Boundary Test Your Knowledge Using the “Module5_2_TYKInput.xlsx”, use RF to determine if well water data can be accurate predictors of manganese detection as was done in the previous module. However, this time, incorporate SMOTE in the model. Feel free to use either the trainControl() or createFolds() function for CV. Extract the variable importance for each predictor on a RF model. What two features have the highest variable importance? Hint: Regardless of the cross validation function you choose, run SMOTE on the training dataset only to create a more balanced training set while the test set will remain unchanged. Using casing depth and the feature with the highest variable importance, construct a decision boundary plot. Under what conditions are a well more likely to predict detectable manganese levels based on a decision boundary plot? "],["unsupervised-machine-learning-part-1-k-means-clustering-pca.html", "Unsupervised Machine Learning Part 1: K-Means Clustering &amp; PCA Introduction to Training Module Introduction to Unsupervised Machine Learning Introduction to Example Data Identifying Clusters of Chemicals through K-Means Principal Component Analysis (PCA) Incorporating K-Means into PCA for Predictive Modeling Concluding Remarks", " Unsupervised Machine Learning Part 1: K-Means Clustering &amp; PCA This training module was developed by David M. Reif with contributions from Alexis Payton, Lauren E. Koval, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module To reiterate what has been discussed in the previous module, machine learning is a field that has great utility in environmental health sciences, often to investigate high-dimensional datasets. The two main classifications of machine learning discussed throughout the TAME Toolkit are supervised and unsupervised machine learning, though additional classifications exist. Previously, we discussed artificial intelligence and supervised machine learning in TAME 2.0 Module 5.1 Introduction to Machine Learning &amp; Artificial Intelligence, TAME 2.0 Module 5.2 Supervised Machine Learning, and TAME 2.0 Module 5.3 Supervised Machine Learning Model Interpretation. In this module, we’ll cover background information on unsupervised machine learning and then work through a scripted example of an unsupervised machine learning analysis. Introduction to Unsupervised Machine Learning Unsupervised machine learning, as opposed to supervised machine learning, involves training a model on a dataset lacking ground truths or response variables. In this regard, unsupervised approaches are often used to identify underlying patterns amongst data in a more unbiased manner. This can provide the analyst with insights into the data that may not otherwise be apparent. Unsupervised machine learning has been used for understanding differences in gene expression patterns of breast cancer patients (Jezequel et. al, 2015) and evaluating metabolomic signatures of patients with and without cystic fibrosis (Laguna et. al, 2015). Note: Unsupervised machine learning is used for exploratory purposes, and just because it can find relationships between data points, that doesn’t necessarily mean that those relationships have merit, are indicative of causal relationships, or have direct biological implications. Rather, these methods can be used to find new patterns that can also inform future studies testing direct relationships. Langs, G., Röhrich, S., Hofmanninger, J., Prayer, F., Pan, J., Herold, C., &amp; Prosch, H. (2018). Machine learning: from radiomics to discovery and routine. Der Radiologe, 58(S1), 1–6. PMID: 34013136. Figure regenerated here in alignment with its published Creative Commons Attribution 4.0 International License Unsupervised machine learning includes: Clustering: Involves grouping elements in a dataset such that the elements in the same group are more similar to each other than to the elements in the other groups. Exclusive (K-means) Overlapping Hierarchical Probabilistic Dimensionality reduction: Focuses on taking high-dimensional data and transforming it into a lower-dimensional space that has fewer features while preserving important information inherent to the original dataset. This is useful because reducing the number of features makes the data easier to visualize while trying to maintain the initial integrity of the dataset. Principal Component Analysis (PCA) Singular Value Decomposition (SVD) t-Distributed Stochastic Neighbor Embedding (t-SNE) Uniform Manifold Approximation and Projection (UMAP) Partial Least Squares-Discriminant Analysis (PLS-DA) In this module, we’ll focus on methods for K-means clustering and Principal Component Analysis described in more detail in the following sections. In the next module, TAME 2.0 Module 5.5 Unsupervised Machine Learning II: Hierarchical Clustering, we’ll focus on hierarchical clustering. For further information on types of unsupervised machine learning, check out Unsupervised Learning. K-Means Clustering K-means is a common clustering algorithm used to partition quantitative data. This algorithm works by first randomly selecting a pre-specified number of clusters, k, across the data space with each cluster having a data centroid. When using a standard Euclidean distance metric, the distance is calculated from an observation to each centroid, then the observation is assigned to the cluster of the closest centroid. After all observations have been assigned to one of the k clusters, the average of all observations in a cluster is calculated, and the centroid for the cluster is moved to the location of the mean. The process then repeats, with the distance computed between the observations and the updated centroids. Observations may be reassigned to the same cluster or moved to a different cluster if it is closer to another centroid. These iterations continue until there are no longer changes between cluster assignments for observations, resulting in the final cluster assignments that are then carried forward for analysis/interpretation. Helpful resources on k-means clustering include the following: The Elements of Statistical Learning &amp; Towards Data Science. Principal Component Analysis (PCA) Principal Component Analysis, or PCA, is a dimensionality-reduction technique used to transform high-dimensional data into a lower dimensional space while trying to preserve as much of the variability in the original data as possible. PCA has strong foundations in linear algebra, so background knowledge of eigenvalues and eigenvectors is extremely useful. Though the mathematics of PCA is beyond the scope of this module, a variety of more in-depth resources on PCA exist including this Towards Data Science Blog, and this Sartorius Blog. At a higher level, important concepts in PCA include: PCA partitions variance in a dataset into linearly uncorrelated principal components (PCs), which are weighted combinations of the original features. Each PC (starting from the first one) summarizes a decreasing percentage of variance. Every instance (e.g. chemical) in the original dataset has a “weight” or score” on each PC. Any combination of PCs can be compared to summarize relationships amongst the instances (e.g. chemicals), but typically it’s the first two eigenvectors that capture a majority of the variance. Introduction to Example Data In this activity, we are going to analyze an example dataset of physicochemical property information for chemicals spanning per- and polyfluoroalkyl substances (PFAS) and statins. PFAS represent a ubiquitous and pervasive class of man-made industrial chemicals that are commonly used in food packaging, commercial household products such as Teflon, cleaning products, and flame retardants. PFAS are recognized as highly stable compounds that, upon entering the environment, can persist for many years and act as harmful sources of exposure. Statins represent a class of lipid-lowering compounds that are commonly used as pharmaceutical treatments for patients at risk of cardiovascular disease. Because of their common use amongst patients, statins can also end up in water and wastewater effluent, making them environmentally relevant as well. This example analysis was designed to evaluate the chemical space of these diverse compounds and to illustrate the utility of unsupervised machine learning methods to differentiate chemical class and make associations between chemical groupings that can inform a variety of environmental and toxicological applications. The two types of machine learning methods that will be employed are k-means and PCA (as described in the introduction). Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: Can we differentiate between PFAS and statin chemical classes when considering just the raw physicochemical property variables without applying unsupervised machine learning techniques? If substances are able to be clustered, what are some of the physicochemical properties that seem to be driving chemical clustering patterns derived through k-means? How do the data compare when physicochemical properties are reduced using PCA? Upon reducing the data through PCA, which physicochemical property contributes the most towards informing data variance captured in the primary principal component? If we did not have information telling us which chemical belonged to which class, could we use PCA and k-means to inform whether a chemical is more similar to a PFAS or a statin? What kinds of applications/endpoints can be better understood and/or predicted because of these derived chemical groupings? Script Preparations Cleaning the global environment rm(list=ls()) Installing required R packages If you already have these packages installed, you can skip this step, or you can run the below code which checks installation status for you if (!requireNamespace(&quot;factoextra&quot;)) install.packages(&quot;factoextra&quot;); if (!requireNamespace(&quot;pheatmap&quot;)) install.packages(&quot;pheatmap&quot;); if (!requireNamespace(&quot;cowplot&quot;)) install.packages(&quot;cowplot&quot;); Loading required R packages library(tidyverse) library(factoextra) library(pheatmap) #used to make heatmaps library(cowplot) Getting help with packages and functions ?tidyverse # Package documentation for tidyverse ?kmeans # Package documentation for kmeans (a part of the standard stats R package, automatically uploaded) ?prcomp # Package documentation for deriving principal components within a PCA (a part of the standard stats R package, automatically uploaded) ?pheatmap # Package documentation for pheatmap Set your working directory setwd(&quot;/filepath to where your input files are&quot;) Loading the Example Dataset Let’s start by loading the datasets needed for this training module. We are going to use a dataset of substances that have a diverse chemical space of PFAS and statin compounds. This list of chemicals will be uploaded alongside physicochemical property data. The chemical lists for ‘PFAS’ and ‘Statins’ were obtained from the EPA’s Computational Toxicology Dashboard Chemical Lists. The physicochemical properties were obtained by uploading these lists into the National Toxicology Program’s Integrated Chemical Environment (ICE). dat &lt;- read.csv(&quot;Module5_4_Input/Module5_4_InputData.csv&quot;, fileEncoding = &quot;UTF-8-BOM&quot;) Data Viewing Starting with the overall dimensions: dim(dat) ## [1] 144 14 Then looking at the first four rows and five columns of data: dat[1:4,1:5] ## List Substance.Name CASRN ## 1 PFAS Perfluoro-2-(trifluoromethyl)propanesulphonic acid 93762-09-5 ## 2 PFAS Potassium perfluoroheptanesulfonate 60270-55-5 ## 3 PFAS Bis(2-hydroxyethyl)ammonium perfluoroheptanesulfonate 70225-15-9 ## 4 PFAS Potassium perfluoro-p-ethylcyclohexanesulfonate 335-24-0 ## DTXSID Molecular.Weight ## 1 DTXSID90239569 300.100 ## 2 DTXSID9069392 488.212 ## 3 DTXSID60880946 555.258 ## 4 DTXSID50880117 500.223 Note that the first column, List, designates the following two larger chemical classes: unique(dat$List) ## [1] &quot;PFAS&quot; &quot;Statins&quot; Let’s lastly view all of the column headers: colnames(dat) ## [1] &quot;List&quot; ## [2] &quot;Substance.Name&quot; ## [3] &quot;CASRN&quot; ## [4] &quot;DTXSID&quot; ## [5] &quot;Molecular.Weight&quot; ## [6] &quot;OPERA..Boiling.Point&quot; ## [7] &quot;OPERA..Henry.s.Law.Constant&quot; ## [8] &quot;OPERA..Melting.Point&quot; ## [9] &quot;OPERA..Negative.Log.of.Acid.Dissociation.Constant&quot; ## [10] &quot;OPERA..Octanol.Air.Partition.Coefficient&quot; ## [11] &quot;OPERA..Octanol.Water.Distribution.Coefficient&quot; ## [12] &quot;OPERA..Octanol.Water.Partition.Coefficient&quot; ## [13] &quot;OPERA..Vapor.Pressure&quot; ## [14] &quot;OPERA..Water.Solubility&quot; In the data file, the first four columns represent chemical identifier information. All remaining columns represent different physicochemical properties derived from OPERA via Integrated Chemical Environment (ICE). Because the original titles of these physicochemical properties contained commas and spaces, R automatically converted these into periods. Hence, titles like OPERA..Boiling.Point. For ease of downstream data analyses, let’s create a more focused dataframe option containing only one chemical identifier (CASRN) as row names and then just the physicochemical property columns. # Creating a new dataframe that contains the physiocochemical properties chemical_prop_df &lt;- dat[,5:ncol(dat)] rownames(chemical_prop_df) &lt;- dat$CASRN Now explore this data subset: dim(chemical_prop_df) # overall dimensions ## [1] 144 10 chemical_prop_df[1:4,1:5] # viewing the first four rows and five columns ## Molecular.Weight OPERA..Boiling.Point OPERA..Henry.s.Law.Constant ## 93762-09-5 300.100 213.095 -3.60 ## 60270-55-5 488.212 223.097 -9.75 ## 70225-15-9 555.258 223.097 -9.75 ## 335-24-0 500.223 220.578 -7.56 ## OPERA..Melting.Point ## 93762-09-5 96.455 ## 60270-55-5 273.228 ## 70225-15-9 182.152 ## 335-24-0 231.827 ## OPERA..Negative.Log.of.Acid.Dissociation.Constant ## 93762-09-5 0.175 ## 60270-55-5 -1.810 ## 70225-15-9 1.000 ## 335-24-0 1.000 colnames(chemical_prop_df) ## [1] &quot;Molecular.Weight&quot; ## [2] &quot;OPERA..Boiling.Point&quot; ## [3] &quot;OPERA..Henry.s.Law.Constant&quot; ## [4] &quot;OPERA..Melting.Point&quot; ## [5] &quot;OPERA..Negative.Log.of.Acid.Dissociation.Constant&quot; ## [6] &quot;OPERA..Octanol.Air.Partition.Coefficient&quot; ## [7] &quot;OPERA..Octanol.Water.Distribution.Coefficient&quot; ## [8] &quot;OPERA..Octanol.Water.Partition.Coefficient&quot; ## [9] &quot;OPERA..Vapor.Pressure&quot; ## [10] &quot;OPERA..Water.Solubility&quot; Evaluating the Original Physicochemical Properties across Substances Let’s first plot two physicochemical properties to determine if and how substances group together without any fancy data reduction or other machine learning techniques. This will answer Environmental Health Question #1: Can we differentiate between PFAS and statin chemical classes when considering just the raw physicochemical property variables without applying unsupervised machine learning techniques? Let’s put molecular weight (Molecular.Weight) as one axis and boiling point (OPERA..Boiling.Point) on the other. We’ll also color by the chemical classes using the List column from the original dataframe. ggplot(chemical_prop_df[,1:2], aes(x = Molecular.Weight, y = OPERA..Boiling.Point, color = dat$List)) + geom_point(size = 2) + theme_bw() + ggtitle(&#39;Version A: Bivariate Plot of Two Original Physchem Variables&#39;) + xlab(&quot;Molecular Weight&quot;) + ylab(&quot;Boiling Point&quot;) Let’s plot two other physicochemical property variables, Henry’s Law constant (OPERA..Henry.s.Law.Constant) and melting point (OPERA..Melting.Point), to see if the same separation of chemical classes is apparent. ggplot(chemical_prop_df[,3:4], aes(x = OPERA..Henry.s.Law.Constant, y = OPERA..Melting.Point, color = dat$List)) + geom_point(size = 2) + theme_bw() + ggtitle(&#39;Version B: Bivariate Plot of Two Other Original Physchem Variables&#39;) + xlab(&quot;OPERA..Henry.s.Law.Constant&quot;) + ylab(&quot;OPERA..Melting.Point&quot;) Answer to Environmental Health Question 1 With these, we can answer Environmental Health Question #1: Can we differentiate between PFAS and statin chemical classes when considering just the raw physicochemical property variables without applying machine learning techniques? Answer: Only in part. From the first plot, we can see that PFAS tend to have lower molecular weight ranges in comparison to the statins, though other property variables clearly overlap in ranges of values making the groupings not entirely clear. Identifying Clusters of Chemicals through K-Means Let’s turn our attention to Environmental Health Question #2: If substances are able to be clustered, what are some of the physicochemical properties that seem to be driving chemical clustering patterns derived through k-means? This will be done deriving clusters of chemicals based on ALL underlying physicochemical property data using k-means clustering. For this example, let’s coerce the k-means algorithms to calculate 2 distinct clusters (based on their corresponding mean centered values). Here, we choose to derive two distinct clusters, because we are ultimately going to see if we can use this information to predict each chemical’s classification into two distinct chemical classes (i.e., PFAS vs statins). Note that we can derive more clusters using similar code depending on the question being addressed. We can give a name to this variable to easily provide the number of clusters in the next lines of code, num.centers: num.centers &lt;- 2 Here we derive chemical clusters using k-means: clusters &lt;- kmeans(chemical_prop_df, # input dataframe centers = num.centers, # number of cluster centers to calculate iter.max = 1000, # the maximum number of iterations allowed nstart = 50) # the number of rows used as the random set for the initial centers (during the first iteration) The resulting property values that were derived as the final cluster centers can be pulled using: clusters$centers ## Molecular.Weight OPERA..Boiling.Point OPERA..Henry.s.Law.Constant ## 1 690.1443 233.0402 -9.589444 ## 2 395.0716 281.4445 -8.655185 ## OPERA..Melting.Point OPERA..Negative.Log.of.Acid.Dissociation.Constant ## 1 183.7980 0.01658333 ## 2 157.5036 1.33226852 ## OPERA..Octanol.Air.Partition.Coefficient ## 1 5.940861 ## 2 6.629556 ## OPERA..Octanol.Water.Distribution.Coefficient ## 1 -2.541750 ## 2 -1.271315 ## OPERA..Octanol.Water.Partition.Coefficient OPERA..Vapor.Pressure ## 1 4.000639 -5.538889 ## 2 3.010302 -6.762009 ## OPERA..Water.Solubility ## 1 -3.760222 ## 2 -3.450750 Let’s add the cluster assignments to the physicochemical data and create a new dataframe, which can then be used in a heatmap visualization to see how these physicochemical data distributions clustered according to k-means. These cluster assignments can be pulled from the cluster list output, where chemicals are designated to each cluster with either a 1 or 2. You can view these using: clusters$cluster ## 93762-09-5 60270-55-5 70225-15-9 335-24-0 647-29-0 68259-12-1 ## 2 2 1 2 2 1 ## 68259-09-6 68259-07-4 60453-92-1 357-31-3 441296-91-9 749786-16-1 ## 2 2 2 2 1 1 ## 93762-10-8 135524-36-6 93894-55-4 34642-43-8 2706-91-4 791563-89-8 ## 2 2 1 2 2 1 ## 742-73-4 29420-49-3 3871-99-6 29359-39-5 3872-25-1 126105-34-8 ## 2 2 2 1 2 1 ## 630402-22-1 2274731-07-4 98789-57-2 85963-79-7 375-73-5 108427-53-8 ## 2 1 1 2 2 2 ## 4021-47-0 117806-54-9 67906-42-7 68555-66-8 92982-03-1 375-92-8 ## 2 2 1 2 1 2 ## 175905-36-9 102061-82-5 134615-58-0 174675-49-1 79780-39-5 91036-71-4 ## 2 2 1 1 1 1 ## 70225-17-1 6401-03-2 374-58-3 646-83-3 86525-30-6 3916-24-3 ## 2 2 2 2 1 2 ## 42409-05-2 474511-07-4 2795-39-3 45187-15-3 82382-12-5 79963-95-4 ## 2 1 1 2 2 2 ## 45298-90-6 134615-57-9 927670-12-0 2806-15-7 70225-14-8 131651-65-5 ## 2 2 2 1 1 2 ## 343629-46-9 144797-51-3 29081-56-9 80988-54-1 1379460-39-5 343629-43-6 ## 1 2 2 2 1 1 ## 146689-46-5 29457-72-5 355-46-4 3107-18-4 70259-86-8 1036375-28-6 ## 2 2 2 2 2 2 ## 70225-18-2 70225-16-0 84224-48-6 507453-86-3 40365-28-4 110676-15-8 ## 2 2 2 1 1 2 ## 70259-85-7 2106-55-0 1997344-07-6 423-41-6 115416-68-7 17202-41-4 ## 2 2 2 2 2 1 ## 93894-73-6 134615-56-8 134615-59-1 68259-08-5 68259-10-9 374-62-9 ## 1 2 1 2 2 2 ## 68555-67-9 2806-16-8 36913-91-4 85187-17-3 803688-15-5 55120-77-9 ## 2 1 1 1 2 2 ## 335-77-3 141263-54-9 95465-60-4 130200-44-1 144535-22-8 130468-11-0 ## 1 2 2 2 2 2 ## 93957-54-1 126059-69-6 153463-20-8 154417-69-3 147511-69-1 141263-69-6 ## 2 2 2 2 2 2 ## 77517-29-4 80799-31-1 73390-02-0 503-49-1 117678-63-4 145599-86-6 ## 2 2 2 2 2 2 ## 147098-20-2 85798-96-5 120551-59-9 13552-81-3 90761-31-2 79691-18-2 ## 1 2 2 1 2 2 ## 73573-88-3 114801-27-3 151106-12-6 129443-92-1 134523-03-8 122254-45-9 ## 2 2 2 2 1 2 ## 75330-75-5 137023-81-5 136320-61-1 87770-13-6 85551-06-0 144501-27-9 ## 2 2 2 2 2 2 ## 159014-70-7 153321-50-7 133983-25-2 78366-44-6 148750-02-1 79902-63-9 ## 2 2 2 2 2 2 ## 120185-34-4 120171-12-2 141267-47-2 94061-80-0 141240-46-2 81093-37-0 ## 2 2 2 2 2 2 Because these results are listed in the exact same order as the inputted dataframe, we can simply add these assignments to the chemical_prop_df dataframe. dat_wclusters &lt;- cbind(chemical_prop_df,clusters$cluster) colnames(dat_wclusters)[11] &lt;- &quot;Cluster&quot; # renaming this new column &quot;Custer&quot; dat_wclusters &lt;- dat_wclusters[order(dat_wclusters$Cluster),] # sorting data by cluster assignments To generate a heatmap, we need to first create a separate dataframe for the cluster assignments, ordered in the same way as the physicochemical data: hm_cluster &lt;- data.frame(dat_wclusters$Cluster, row.names = row.names(dat_wclusters)) # creating the dataframe colnames(hm_cluster) &lt;- &quot;Cluster&quot; # reassigning the column name hm_cluster$Cluster &lt;- as.factor(hm_cluster$Cluster) # coercing the cluster numbers into factor variables, to make the heatmap prettier head(hm_cluster) # viewing this new cluster assignment dataframe ## Cluster ## 70225-15-9 1 ## 68259-12-1 1 ## 441296-91-9 1 ## 749786-16-1 1 ## 93894-55-4 1 ## 791563-89-8 1 We’re going to go ahead and clean up the physiocochemical property names to make the heatmap a bit tidier. clean_names1 = gsub(&quot;OPERA..&quot;, &quot;&quot;, colnames(dat_wclusters)) # &quot;\\\\.&quot; denotes a period clean_names2 = gsub(&quot;\\\\.&quot;, &quot; &quot;, clean_names1) # Reassigning the cleaner names back to the df colnames(dat_wclusters) = clean_names2 # Going back to add in the apostrophe in &quot;Henry&#39;s Law Constant&quot; colnames(dat_wclusters)[3] = &quot;Henry&#39;s Law Constant&quot; Then we can call this dataframe (data_wclusters) into the following heatmap visualization code leveraging the pheatmap() function. This function was designed specifically to enable clustered heatmap visualizations. Check out pheatmap Documenation for additional information. Heatmap Visualization of the Resulting K-Means Clusters pheatmap(dat_wclusters[,1:10], cluster_rows = FALSE, cluster_cols = FALSE, # no further clustering, for simplicity scale = &quot;column&quot;, # scaling the data to make differences across chemicals more apparent annotation_row = hm_cluster, # calling the cluster assignment dataframe as a separate color bar annotation_names_row = FALSE, # adding removing the annotation name (&quot;Cluster&quot;) from the x axis angle_col = 45, fontsize_col = 7, fontsize_row = 3, # adjusting size/ orientation of axes labels cellheight = 3, cellwidth = 25, # setting height and width for cells border_color = FALSE # specify no border surrounding the cells ) An appropriate title for this figure could be: “Figure X. Heatmap of physicochemical properties with k-means cluster assignments. Shown are the relative values for each physicochemical property labeled on the x axis. Individual chemical names are listed on the y axis. The chemicals are grouped based on their k-means cluster assignment as denoted by the color bar on the left.” Notice that the pheatmap() function does not add axes or legend titles. Adding those can provide clarity, however those can be added to the figure after exporting from R in MS Powerpoint or Adobe. Answer to Environmental Health Question 2 With this, we can answer Environmental Health Question #2: What are some of the physicochemical properties that seem to be driving chemical clustering patterns derived through k-means? Answer: Properties with values that show obvious differences between resulting clusters including molecular weight, boiling point, negative log of acid dissociation constant, octanol air partition coefficient, and octanol water distribution coefficient. Principal Component Analysis (PCA) Next, we will run through some example analyses applying the common data reduction technique of PCA. We’ll start by determining how much of the variance is able to be captured within the first two principal components to answer Environmental Health Question #3: How do the data compare when physicochemical properties are reduced using PCA? We can calculate the principal components across ALL physicochemical data across all chemicals using the prcomp() function. Always make sure your data is centered and scaled prior to running to PCA, since it’s sensitive to variables having different scales. my.pca &lt;- prcomp(chemical_prop_df, # input dataframe of physchem data scale = TRUE, center = TRUE) We can see how much of the variance was able to be captured in each of the eigenvectors or dimensions using a scree plot. fviz_eig(my.pca, addlabels = TRUE) We can also calculate these values and pull them into a dataframe for future use. For example, to pull the percentage of variance explained by each principal component, we can run the following calculations, where first eigenvalues (eigs) are calculated and then used to calculate percent of variance per principal component: eigs &lt;- my.pca$sdev^2 Comp.stats &lt;- data.frame(eigs, eigs/sum(eigs), row.names = names(eigs)) colnames(Comp.stats) &lt;- c(&quot;Eigen_Values&quot;, &quot;Percent_of_Variance&quot;) head(Comp.stats) ## Eigen_Values Percent_of_Variance ## 1 4.0925925 0.40925925 ## 2 2.3840022 0.23840022 ## 3 1.5415308 0.15415308 ## 4 0.5873057 0.05873057 ## 5 0.4769190 0.04769190 ## 6 0.3659181 0.03659181 Answer to Environmental Health Question 3 With this, we can answer Environmental Health Question #3: How do the data compare when physicochemical properties are reduced using PCA? Answer: Principal Component 1 captures ~41% of the variance and Principal Component 2 captures ~24% across all physicochemical property values across all chemicals. These two components together describe ~65% of data. Next, we’ll use PCA to answer Environmental Health Question #4: Upon reducing the data through PCA, which physicochemical property contributes the most towards informing data variance captured in the primary principal component (Comp.1)? Here are the resulting scores for each chemical’s contribution towards each principal component (shown here as components PC1-PC10). head(my.pca$x) ## PC1 PC2 PC3 PC4 PC5 ## 93762-09-5 -2.0354310 -1.4824239 1.29328367 0.04865893 0.25305471 ## 60270-55-5 -1.2249015 2.2857092 -0.24261251 -0.40140814 0.62781108 ## 70225-15-9 -1.0944361 1.3915069 -0.03340358 -0.90391749 0.05735985 ## 335-24-0 -1.1334897 1.0675553 0.14299978 -1.08713268 0.21172964 ## 647-29-0 -0.4830620 0.1259827 -1.16147937 1.11383217 -0.29571642 ## 68259-12-1 -0.3264762 0.2369031 -1.31984895 0.47512052 -1.17555774 ## PC6 PC7 PC8 PC9 PC10 ## 93762-09-5 -0.67755281 -0.14546495 1.25520979 0.230936852 0.14075496 ## 60270-55-5 -1.01515519 -0.11316504 -0.34335060 -0.383682123 0.09849160 ## 70225-15-9 0.14334979 0.38355763 0.01717491 -0.006702116 -0.02715723 ## 335-24-0 -0.93947357 -0.22877517 0.11457280 -0.107720120 0.17701035 ## 647-29-0 0.19174355 -0.18228398 -0.48013543 0.074967840 0.22750496 ## 68259-12-1 0.05910151 0.01399123 0.03790455 0.043309249 -0.18032083 And the resulting loading factors of each property’s contribution towards each principal component. my.pca$rotation ## PC1 PC2 ## Molecular.Weight 0.09825313 0.108454961 ## OPERA..Boiling.Point 0.46350428 0.029650863 ## OPERA..Henry.s.Law.Constant -0.17856542 -0.502116638 ## OPERA..Melting.Point 0.20645719 0.474473487 ## OPERA..Negative.Log.of.Acid.Dissociation.Constant 0.32172963 -0.119465105 ## OPERA..Octanol.Air.Partition.Coefficient 0.45329804 -0.008918089 ## OPERA..Octanol.Water.Distribution.Coefficient 0.33026147 -0.437053508 ## OPERA..Octanol.Water.Partition.Coefficient 0.16152270 -0.342905461 ## OPERA..Vapor.Pressure -0.35205065 -0.349618523 ## OPERA..Water.Solubility -0.36456684 0.254809066 ## PC3 PC4 ## Molecular.Weight -0.679740448 -0.48432419 ## OPERA..Boiling.Point 0.199365885 0.03108544 ## OPERA..Henry.s.Law.Constant 0.179876660 -0.27695374 ## OPERA..Melting.Point -0.214857945 -0.09449999 ## OPERA..Negative.Log.of.Acid.Dissociation.Constant 0.286239544 -0.58268278 ## OPERA..Octanol.Air.Partition.Coefficient 0.132157694 0.04820475 ## OPERA..Octanol.Water.Distribution.Coefficient -0.005904541 0.15115855 ## OPERA..Octanol.Water.Partition.Coefficient -0.466833374 0.48460961 ## OPERA..Vapor.Pressure -0.195008167 -0.24988717 ## OPERA..Water.Solubility 0.253557563 0.12980079 ## PC5 PC6 ## Molecular.Weight 0.17351578 0.357367954 ## OPERA..Boiling.Point 0.22224554 -0.018507534 ## OPERA..Henry.s.Law.Constant 0.30566003 -0.470666686 ## OPERA..Melting.Point -0.08063905 -0.686723561 ## OPERA..Negative.Log.of.Acid.Dissociation.Constant -0.66998767 0.029248038 ## OPERA..Octanol.Air.Partition.Coefficient 0.20778895 0.205757894 ## OPERA..Octanol.Water.Distribution.Coefficient 0.06378821 0.169005123 ## OPERA..Octanol.Water.Partition.Coefficient -0.48515616 -0.162279640 ## OPERA..Vapor.Pressure -0.06782464 -0.007031883 ## OPERA..Water.Solubility -0.29062543 0.284060941 ## PC7 PC8 ## Molecular.Weight -0.11763362 0.32938640 ## OPERA..Boiling.Point -0.12503355 0.09718690 ## OPERA..Henry.s.Law.Constant -0.21138163 0.44526650 ## OPERA..Melting.Point -0.34342931 -0.10233816 ## OPERA..Negative.Log.of.Acid.Dissociation.Constant 0.09083446 0.03113686 ## OPERA..Octanol.Air.Partition.Coefficient -0.44434707 -0.29734602 ## OPERA..Octanol.Water.Distribution.Coefficient -0.22039378 0.05052653 ## OPERA..Octanol.Water.Partition.Coefficient -0.08082351 0.17794120 ## OPERA..Vapor.Pressure -0.34958097 -0.65367959 ## OPERA..Water.Solubility -0.65183980 0.34989539 ## PC9 PC10 ## Molecular.Weight 0.03295675 -0.02698233 ## OPERA..Boiling.Point 0.03336277 -0.81709497 ## OPERA..Henry.s.Law.Constant 0.19706729 0.10099077 ## OPERA..Melting.Point -0.24532148 0.10229774 ## OPERA..Negative.Log.of.Acid.Dissociation.Constant 0.02576652 0.03380215 ## OPERA..Octanol.Air.Partition.Coefficient 0.49672303 0.39565984 ## OPERA..Octanol.Water.Distribution.Coefficient -0.74417871 0.19866516 ## OPERA..Octanol.Water.Partition.Coefficient 0.30648450 -0.09852386 ## OPERA..Vapor.Pressure -0.03202641 -0.31333091 ## OPERA..Water.Solubility -0.05123040 -0.07696401 Answer to Environmental Health Question 4 With these results, we can answer Environmental Health Question #4: Upon reducing the data through PCA, which physicochemical property contributes the most towards informing data variance captured in the primary principal component (Comp.1)? Answer: Boiling point contributes the most towards principal component 1, as it has the largest magnitude (0.464). Visualizing PCA Results Let’s turn our attention to Environmental Health Question #5: If we did not have information telling us which chemical belonged to which class, could we use PCA and k-means to inform whether a chemical is more similar to a PFAS or a statin? We can start by answering this question by visualizing the first two principal components and coloring each chemical according to class (i.e. PFAS vs statins). ggplot(data.frame(my.pca$x), aes(x = PC1, y = PC2, color = dat$List)) + geom_point(size = 2) + theme_bw() + ggtitle(&#39;Version C: PCA Plot of the First 2 PCs, colored by Chemical Class&#39;) + # it&#39;s good practice to put the percentage of the variance captured in the axes titles xlab(&quot;Principal Component 1 (40.9%)&quot;) + ylab(&quot;Principal Component 2 (23.8%)&quot;) Answer to Environmental Health Question 5 With this, we can answer Environmental Health Question #5: If we did not have information telling us which chemical belonged to which class, could we use PCA and k-means to inform whether a chemical is more similar to a PFAS or a statin? Answer: Data become more compressed and variables reduce across principal components capturing the majority of the variance from the original dataset (~65%). This results in improved data visualizations, where all dimensions of the physiochemical dataset are compressed and captured across the displayed components. In addition, the figure above shows a clear separation between PFAS and statin chemical when visualizing the reduced dataset. Incorporating K-Means into PCA for Predictive Modeling We can also identify cluster-based trends within data that are reduced after running PCA. This example analysis does so, expanding upon the previously generated PCA results. Estimate K-Means Clusters from PCA Results Let’s first run code similar to the previous k-means analysis and associated parameters, though instead here we will use data reduced values from the PCA analysis. Specifically, clusters across PCA “scores” values will be derived, where scores represent the relative amount each chemical contributed to each principal component. clusters_PCA &lt;- kmeans(my.pca$x, centers = num.centers, iter.max = 1000, nstart = 50) The resulting PCA score values that were derived as the final cluster centers can be pulled using: clusters_PCA$centers ## PC1 PC2 PC3 PC4 PC5 PC6 ## 1 2.612553 -0.3054641 0.6219575 -0.008818739 0.18490367 0.05041185 ## 2 -1.187524 0.1388473 -0.2827080 0.004008518 -0.08404712 -0.02291448 ## PC7 PC8 PC9 PC10 ## 1 -0.03417184 -0.05270363 -0.018330999 -0.011904302 ## 2 0.01553265 0.02395620 0.008332272 0.005411046 Viewing the final cluster assignment per chemical: head(cbind(rownames(chemical_prop_df),clusters_PCA$cluster)) ## [,1] [,2] ## 93762-09-5 &quot;93762-09-5&quot; &quot;2&quot; ## 60270-55-5 &quot;60270-55-5&quot; &quot;2&quot; ## 70225-15-9 &quot;70225-15-9&quot; &quot;2&quot; ## 335-24-0 &quot;335-24-0&quot; &quot;2&quot; ## 647-29-0 &quot;647-29-0&quot; &quot;2&quot; ## 68259-12-1 &quot;68259-12-1&quot; &quot;2&quot; Visualizing K-Means Clusters from PCA Results Let’s now view, again, the results of the main PCA focusing on the first two principal components; though this time let’s color each chemical according to k-means cluster. ggplot(data.frame(my.pca$x), aes(x = PC1, y = PC2, color = as.factor(clusters_PCA$cluster))) + geom_point(size = 2) + theme_bw() + ggtitle(&#39;Version D: PCA Plot of the First 2 PCs, colored by k-means Clustering&#39;) + # it&#39;s good practice to put the percentage of the variance capture in the axes titles xlab(&quot;Principal Component 1 (40.9%)&quot;) + ylab(&quot;Principal Component 2 (23.8%)&quot;) Let’s put these two PCA plots side by side to compare them more easily. We’ll also tidy up the figures a bit so they’re closer to publication-ready. # PCA plot colored by chemical class pcaplot1 = ggplot(data.frame(my.pca$x), aes(x = PC1, y = PC2, color = dat$List)) + geom_point(size = 2) + theme_light() + theme(axis.text = element_text(size = 9), # changing size of axis labels axis.title = element_text(face = &quot;bold&quot;, size = rel(1.3)), # changes axis titles legend.title = element_text(face = &#39;bold&#39;, size = 10), # changes legend title legend.text = element_text(size = 9)) + # changes legend text labs(x = &#39;Principal Component 1 (40.9%)&#39;, y = &#39;Principal Component 2 (23.8%)&#39;, color = &quot;Chemical Class&quot;) # changing axis labels # PCA Plot by k means clusters pcaplot2 = ggplot(data.frame(my.pca$x), aes(x = PC1, y = PC2, color = as.factor(clusters_PCA$cluster))) + geom_point(size = 2) + theme_light() + theme(axis.text = element_text(size = 9), # changing size of axis labels axis.title = element_text(face = &quot;bold&quot;, size = rel(1.3)), # changes axis titles legend.text = element_text(size = 9)) + # changes legend text labs(x = &#39;Principal Component 1 (40.9%)&#39;, y = &#39;Principal Component 2 (23.8%)&#39;, color = expression(bold(bolditalic(K)-Means~Cluster))) # changing axis labels # Creating 1 figure plot_grid(pcaplot1, pcaplot2, # Adding labels, changing size their size and position labels = &quot;AUTO&quot;, label_size = 15, label_x = 0.03) An appropriate title for this figure could be: “Figure X. Principal Component Analysis (PCA) plots highlight similarities between chemical class and k-means clusters. These PCA plots are based on physiochemical properties and compare (A) chemical class categories and the (B) K-means derived cluster assignments.” Answer to Environmental Health Question 6 With this we can answer Environmental Health Question #6: What kinds of applications/endpoints can be better understood and/or predicted because of these derived chemical groupings? Answer: With these well-informed chemical groupings, we can now better understand the variables that attribute to the chemical classifications. We can also use this information to better understand data trends and predict environmental fate and transport for these chemicals. The reduced variables derived through PCA, and/or k-means clustering patterns can also be used as input variables to predict toxicological outcomes. Concluding Remarks In conclusion, this training module provides an example exercise on organizing physicochemical data and analyzing trends within these data to determine chemical groupings. Results are compared from those produced using just the original data vs. clustered data from k-means vs. reduced data from PCA. These methods represent common tools that are used in high dimensional data analyses within the field of environmental health sciences. Additional Resources Detailed study of Principal Component Analysis Practical Guide to Cluster Analysis in R Test Your Knowledge In this training module, we presented an unsupervised machine learning example that was based on defining k-means clusters based on chemical class where k = 2. Often times, analyses are conducted to explore potential clustering relationships without a preexisting idea of what k or the number of clusters should be. In this test your knowledge section, we’ll go through an example like that. Using the accompanying flame retardant and pesticide physicochemical property variables found in the file (“Module5_4_TYKInput.csv”), answer the following questions: What are some of the physicochemical properties that seem to be driving chemical clustering patterns derived through k-means? Upon reducing the data through PCA, which physicochemical property contributes the most towards informing data variance captured in the primary principal component? If we did not have information telling us which chemical belonged to which class, could we use PCA and k-means to accurately predict whether a chemical is a PFAS or a statin? "],["unsupervised-machine-learning-part-2-additional-clustering-applications.html", "Unsupervised Machine Learning Part 2: Additional Clustering Applications Introduction to Training Module K-Means Clustering Hierarchical Clustering Introduction to Example Data Hierarchical Clustering Clustering Plot Hierarchical Clustering Visualization Variable Contributions Concluding Remarks", " Unsupervised Machine Learning Part 2: Additional Clustering Applications This training module was developed by Alexis Payton, Lauren E. Koval, David M. Reif, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module The previous module TAME 2.0 Module 5.4 Unsupervised Machine Learning, served as introduction to unsupervised machine learning (ML). Unsupervised ML involves training a model on a dataset lacking ground truths or response variables. However, in the previous module, the number of clusters was selected based on prior information (i.e., chemical class), but what if you’re in a situation where you don’t know how many clusters to investigate a priori? This commonly occurs, particularly in the field of environmental health research in instances when investigators want to take a more unbiased view of their data and/or do not have information that can be used to inform the optimal number of clusters to select. In these instances,unsupervised ML techniques can be very helpful, and in this module, we’ll explore the following concepts to further understand unsupervised ML: K-Means and hierarchical clustering Deriving the optimal number of clusters Visualizing clusters through a PCA-based plot, dendrograms, and heatmaps Determining each variable’s contribution to the clusters K-Means Clustering As mentioned in the previous module, K-means is a common clustering algorithm used to partition quantitative data. This algorithm works by first, randomly selecting a pre-specified number of clusters, k, across the data space, with each cluster having a data centroid. When using a standard Euclidean distance metric, the distance is calculated from an observation to each centroid, then the observation is assigned to the cluster of the closest centroid. After all observations have been assigned to one of the k clusters, the average of all observations in a cluster is calculated, and the centroid for the cluster is moved to the location of the mean. The process then repeats, with the distance computed between the observations and the updated centroids. Observations may be reassigned to the same cluster or moved to a different cluster if it is closer to another centroid. These iterations continue until there are no longer changes between cluster assignments for observations, resulting in the final cluster assignments that are then carried forward for analysis/interpretation. Helpful resources on k-means clustering include the following: The Elements of Statistical Learning &amp; Towards Data Science. Hierarchical Clustering Hierarchical clustering groups objects into clusters by repetitively joining similar observations until there is one large cluster (aka agglomerative or bottom-up) or repetitively splitting one large cluster until each observation stands alone (aka divisive or top-down). Regardless of whether agglomerative or divisive hierarchical clustering is used, the results can be visually represented in a tree-like figure called a dendrogram. The dendrogram below is based on the USArrests dataset available in R. The datset contains statistics on violent crimes rates (murder, assault, and rape) per capita (per 100,000 residents) for each state in the United States in 1973. For more information on the USArrests dataset, check out its associated RDocumentation. An appropriate title for this figure could be: “Figure X. Hierarchical clustering of states based on violent crime. The dendogram uses violent crime data including murder, assault, and rape rates per 100,000 residents for each state in 1973.” Takeaways from this dendogram: The 50 states can be grouped into 4 clusters based on violent crime statistics from 1973 The dendogram can only show us clusters of states but not the data trends that led to the clustering patterns that we see. Yes, it is useful to know what states have similar violent crime patterns overall, but it is important to pinpoint the variables (ie. murder, assault, and rape) that are responsible for the clustering patterns we’re seeing. This idea will be explored later in the module with an environmentally-relevant dataset. Going back to hiearchical clustering, during the repetitive splitting or joining of observations, the similarity between existing clusters is calculated after each iteration. This value informs the formation of subsequent clusters. Different methods, or linkage functions, can be considered when calculating this similarity, particularly for agglomerative clustering which is often the preferred approach. Some example methods include: Complete Linkage: the maximum distance between two data points located in separate clusters. Single Linkage: the minimum distance between two data points located in separate clusters. Average Linkage: the average pairwise distance between all pairs of data points in separate clusters. Centroid Linkage: the distance between the centroids or centers of each cluster. Ward Linkage: seeks to minimize the variance between clusters. Each method has its advantages and disadvantages and more information on all distance calculations between clusters can be found at the following resource: Hierarchical Clustering. Deriving the Optimal Number of Clusters Before clustering can be performed, the function needs to be informed of the number of clusters to group the objects into. In the previous module, an example was explored to see if k-means clustering would group the chemicals similarly to their chemical class (either a PFAS or statin). Therefore, we told the k-means function to cluster into 2 groups. In situations where there is little to no prior knowledge regarding the “correct” number of clusters to specify, methods exist for deriving the optimal number of clusters. Three common methods to find the optimal k, or number of clusters, for both k-means and hierarchical clustering include: the elbow method, silhouette method, and the gap statistic method. These techniques help us in determining the optimal k using visual inspection. Elbow Method: uses a plot of the within cluster sum of squares (WCSS) on the y axis and different values of k on the x axis. The location where we no longer observe a significant reduction in WCSS, or where an “elbow” can be seen, is the optimal k value. As we can see, after a certain point, having more clusters does not lead to a significant reduction in WCSS. Looking at the figures above, the elbow point is much clearer in the first plot versus the second, however, elbow curves from real-world datasets typically resemble the second figure. This is why it’s recommended to consider more than one method to determine the optimal number of clusters. Silhouette Method: uses a plot of the average silhouette width (score) on the y axis and different values of k on the x axis. The silhouette score is measure of each object’s similarity to its own cluster and how dissimilar it is to other clusters. The location where the average silhouette width is maximized is the optimal k value. Based on the figure above, the optimal number of clusters is 2 using the silhouette method. Gap Statistic Method: uses a plot of the gap statistic on the y axis and different values of k on the x axis. The gap statistic evaluates the intracluster variation in comparison to expected values derived from a Monte Carlo generated, null reference data distribution for varying values of k. The optimal number of clusters is the smallest value where the gap statistic of k is greater than or equal to the gap statistic of k+1 minus the standard deviation of k+1. More details can be found here. Based on the figure above, the optimal number of clusters is 2 using the gap statistic method. For additional information and code on all three methods, check out Determining the Optimal Number of Clusters: 3 Must Know Methods. It is also worth mentioning that while these methods are useful, further interpreting the result in the context of your problem can be beneficial, for example, checking whether clusters make biological sense when working with a genomic dataset. Introduction to Example Data We will apply these techniques using an example dataset from a previously published study where 22 cytokine concentrations were derived from 44 subjects with varying smoking statuses (14 non-smokers, 17 e-cigarette users, and 13 cigarette smokers) from 4 different sampling regions in the body. These samples were derived from nasal lavage fluid (NLF), nasal epithelieum fluid (NELF), sputum, and serum as pictured below. Samples were taken from different regions in the body to compare cytokine expression in the upper respiratory tract, lower respiratory tract, and systemic circulation. A research question that we had was “Does cytokine expression change based on a subject’s smoking habits? If so, does cigarette smoke or e-cigarette vapor induce cytokine suppression or proliferation?” Traditionally these questions would have been answered by analyzing each biomarker individually using a two-group comparison test like a t test (which we completed in this study). However, biomarkers do not work in isolation in the body, suggesting that individual biomarker statistical approaches may not capture the full biological responses occurring. Therefore we used a clustering approach to group cytokines as an attempt to more closely simulate interactions that occur in vivo. From there, statistical tests were run to assess the effects of smoking status on each cluster. For the purposes of this training exercise, we will focus solely on the nasal epithelieum lining fluid, or NELF, samples. In addition, we’ll use k-means and hierarchical clustering to compare how cytokines cluster at baseline. Full methods are further described in the publication below: Payton AD, Perryman AN, Hoffman JR, Avula V, Wells H, Robinette C, Alexis NE, Jaspers I, Rager JE, Rebuli ME. Cytokine signature clusters as a tool to compare changes associated with tobacco product use in upper and lower airway samples. American Journal of Physiology-Lung Cellular and Molecular Physiology 2022 322:5, L722-L736. PMID: 35318855 Let’s read in and view the dataset we’ll be working with. Script Preparations Cleaning the global environment rm(list=ls()) Installing required R packages If you already have these packages installed, you can skip this step, or you can run the below code which checks installation status for you if (!requireNamespace(&quot;vegan&quot;)) install.packages(&quot;vegan&quot;); if (!requireNamespace(&quot;ggrepel&quot;)) install.packages(&quot;ggrepel&quot;); if (!requireNamespace(&quot;dendextend&quot;)) install.packages(&quot;dendextend&quot;); if (!requireNamespace(&quot;ggsci&quot;)) install.packages(&quot;ggsci&quot;); if (!requireNamespace(&quot;FactoMineR&quot;)) install.packages(&quot;FactoMineR&quot;); Loading required R packages library(readxl) library(factoextra) library(FactoMineR) library(tidyverse) library(vegan) library(ggrepel) library(reshape2) library(pheatmap) library(ggsci) suppressPackageStartupMessages(library(dendextend)) Set your working directory setwd(&quot;/filepath to where your input files are&quot;) Importing example dataset Then let’s read in our example dataset. As mentioned in the introduction, this example dataset contains cytokine concentrations derived from 44 subjects. Let’s import and view these data: # Reading in file cytokines_df &lt;- data.frame(read_excel(&quot;Module5_5_Input/Module5_5_InputData.xlsx&quot;, sheet = 2)) # Viewing data head(cytokines_df) ## Original_Identifier Group SubjectNo SubjectID Compartment Protein Conc ## 1 E_C_F_002 NS 1 NS_1 NELF IFNg 17.642316 ## 2 E_C_F_002 NS 1 NS_1 NELF IL10 2.873724 ## 3 E_C_F_002 NS 1 NS_1 NELF IL12p70 1.625272 ## 4 E_C_F_002 NS 1 NS_1 NELF IL13 36.117692 ## 5 E_C_F_002 NS 1 NS_1 NELF IL1b 104.409217 ## 6 E_C_F_002 NS 1 NS_1 NELF IL6 21.159536 ## Conc_pslog2 ## 1 4.220509 ## 2 1.953721 ## 3 1.392467 ## 4 5.214035 ## 5 6.719857 ## 6 4.469856 These data contain the following information: Original_Identifier: initial identifier given to each subject by our wet bench colleagues Group: denotes the smoking status of the subject (“NS” = “non-smoker”, “Ecig” = “E-cigarette user”, “CS” = “cigarette smoker”) SubjectNo: ordinal subject number assigned to each subject after the dataset was wrangled (1-44) SubjectID: unique subject identifier that combines the group and subject number Compartment: region of the body from which the sample was taken (“NLF” = “nasal lavage fluid sample”, “NELF” = “nasal epithelieum lining fluid sample”, “Sputum” = “induced sputum sample”, “Serum” = “blood serum sample”) Protein: cytokine name Conc: concentration (pg/mL) Conc_pslog2: psuedo-log2 concentration Now that the data has been read in, we can start by asking some initial questions about the data. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: What is the optimal number of clusters the cytokines can be grouped into that were derived from nasal epithelium fluid (NELF) in non-smokers using k-means clustering? After selecting a cluster number, which cytokines were assigned to each k-means cluster? What is the optimal number of clusters the cytokines can be grouped into that were derived from nasal epithelium fluid (NELF) in non-smokers using hierarchical clustering? How do the hierarchical cluster assignments compare to the k-means cluster assignments? Which cytokines have the greatest contributions to the first two eigenvectors? To answer the first environmental health question, let’s start by filtering to include only NELF derived samples and non-smokers. baseline_df &lt;- cytokines_df %&gt;% filter(Group == &quot;NS&quot;, Compartment == &quot;NELF&quot;) head(baseline_df) ## Original_Identifier Group SubjectNo SubjectID Compartment Protein Conc ## 1 E_C_F_002 NS 1 NS_1 NELF IFNg 17.642316 ## 2 E_C_F_002 NS 1 NS_1 NELF IL10 2.873724 ## 3 E_C_F_002 NS 1 NS_1 NELF IL12p70 1.625272 ## 4 E_C_F_002 NS 1 NS_1 NELF IL13 36.117692 ## 5 E_C_F_002 NS 1 NS_1 NELF IL1b 104.409217 ## 6 E_C_F_002 NS 1 NS_1 NELF IL6 21.159536 ## Conc_pslog2 ## 1 4.220509 ## 2 1.953721 ## 3 1.392467 ## 4 5.214035 ## 5 6.719857 ## 6 4.469856 The functions we use will require us to cast the data wider. We will accomplish this using the dcast() function from the reshape2 package. wider_baseline_df &lt;- reshape2::dcast(baseline_df, Protein ~ SubjectID, value.var = &quot;Conc_pslog2&quot;) %&gt;% column_to_rownames(&quot;Protein&quot;) head(wider_baseline_df) ## NS_1 NS_10 NS_11 NS_12 NS_13 NS_14 ## Eotaxin 7.215729 8.016193 6.264252 6.938344 5.152683 7.937892 ## Eotaxin3 4.396783 4.006488 5.289509 3.597945 3.051951 3.063025 ## Fractalkine 13.584554 13.679882 12.952220 13.464652 12.110742 12.785656 ## I309 0.000000 0.000000 0.000000 0.000000 0.000000 1.617372 ## IFNg 4.220509 1.990018 2.656691 3.439682 2.318340 3.439850 ## IL10 1.953721 1.145162 1.348623 1.801954 1.176498 1.800412 ## NS_2 NS_3 NS_4 NS_5 NS_6 NS_7 ## Eotaxin 9.773407 6.12290913 5.758602 4.19868540 6.832578 7.998467 ## Eotaxin3 9.022390 2.50161364 2.721195 2.21084320 3.820379 5.549164 ## Fractalkine 14.961872 12.71538464 12.057105 10.38711624 12.917180 14.636190 ## I309 1.103140 0.00000000 0.000000 0.00000000 0.000000 0.000000 ## IFNg 4.394705 0.00000000 2.638213 0.04830196 1.914092 3.251501 ## IL10 2.652120 0.08945805 2.020155 0.84649523 1.119272 2.007941 ## NS_8 NS_9 ## Eotaxin 9.242464 7.127746 ## Eotaxin3 13.724108 3.553097 ## Fractalkine 14.047382 13.146599 ## I309 0.000000 0.000000 ## IFNg 4.104820 2.940248 ## IL10 2.576814 1.970627 Now we can derive clusters using the fviz_nbclust() function to determine the optimal k based on suggestions from the elbow, silhouette, and gap statistic methods. We can use this code for both k-means and hierarchical clustering by changing the FUNcluster parameter. Lets start with k-means:. # Elbow method fviz_nbclust(wider_baseline_df, FUNcluster = kmeans, method = &quot;wss&quot;) + labs(subtitle = &quot;Elbow method&quot;) # Silhouette method fviz_nbclust(wider_baseline_df, FUNcluster = kmeans, method = &quot;silhouette&quot;) + labs(subtitle = &quot;Silhouette method&quot;) # Gap statistic method fviz_nbclust(wider_baseline_df, FUNcluster = kmeans, method = &quot;gap_stat&quot;) + labs(subtitle = &quot;Gap Statisitc method&quot;) The elbow method is suggesting 2 or 3 clusters, the silhouette method is suggesting 2, and the gap statistic method is suggesting 1. Since each of these methods is recommending different k values, we can go ahead and run k-means to visualize the clusters and test those different k’s. K-means clusters will be visualized using the fviz_cluster() function. # Choosing to iterate through 2 or 3 clusters using i as our iterator for (i in 2:3){ # nstart = number of random starting partitions, it&#39;s recommended for nstart &gt; 1 cluster_k &lt;- kmeans(wider_baseline_df, centers = i, nstart = 25) cluster_plot &lt;- fviz_cluster(cluster_k, data = wider_baseline_df) + ggtitle(paste0(&quot;k = &quot;, i)) print(cluster_plot) } Answer to Environmental Health Question 1 With this, we can answer Environmental Health Question #1: What is the optimal number of clusters the cytokines can be grouped into that were derived from nasal epithelium fluid (NELF) in non-smokers using k-means clustering? Answer: 2 or 3 clusters can be justified here, based on using the elbow or silhouette method or if k-means happens to group cytokines together that were implicated in similar biological pathways. In the final paper, we moved forward with 3 clusters, because it was justifiable from the methods and provided more granularity in the clusters. The final cluster assignments can easily be obtained using the kmeans() function from the stats package. cluster_kmeans_3 &lt;- kmeans(wider_baseline_df, centers = 3, nstart = 25) cluster_kmeans_df &lt;- data.frame(cluster_kmeans_3$cluster) %&gt;% rownames_to_column(&quot;Cytokine&quot;) %&gt;% rename(`K-Means Cluster` = cluster_kmeans_3.cluster) %&gt;% # Ordering the dataframe for easier comparison arrange(`K-Means Cluster`) cluster_kmeans_df ## Cytokine K-Means Cluster ## 1 Fractalkine 1 ## 2 IL8 1 ## 3 IP10 1 ## 4 MIG 1 ## 5 Eotaxin 2 ## 6 Eotaxin3 2 ## 7 IL13 2 ## 8 IL17B 2 ## 9 IL1b 2 ## 10 IL6 2 ## 11 ITAC 2 ## 12 MCP1 2 ## 13 MIP1a 2 ## 14 MIP1b 2 ## 15 TARC 2 ## 16 I309 3 ## 17 IFNg 3 ## 18 IL10 3 ## 19 IL12p70 3 ## 20 IL17D 3 ## 21 IL4 3 ## 22 TNFa 3 Answer to Environmental Health Question 2 With this, we can answer Environmental Health Question #2: After selecting a cluster number, which cytokines were assigned to each k-means cluster? Answer: After choosing the number of clusters to be 3, the cluster assignments are as follows: Hierarchical Clustering Next, we’ll turn our attention to answering environmental health questions 3 and 4: What is the optimal number of clusters the cytokines can be grouped into that were derived from nasal epithelium fluid (NELF) in non-smokers using hierarchical clustering? How do the hierarchical cluster assignments compare to the k-means cluster assignments? Just as we used the elbow method, silhouette profile, and gap statistic to determine the optimal number of clusters for k-means, we can leverage the same approaches for hierarchical by changing the FUNcluster parameter. # Elbow method fviz_nbclust(wider_baseline_df, FUNcluster = hcut, method = &quot;wss&quot;) + labs(subtitle = &quot;Elbow method&quot;) # Silhouette method fviz_nbclust(wider_baseline_df, FUNcluster = hcut, method = &quot;silhouette&quot;) + labs(subtitle = &quot;Silhouette method&quot;) # Gap statistic method fviz_nbclust(wider_baseline_df, FUNcluster = hcut, method = &quot;gap_stat&quot;) + labs(subtitle = &quot;Gap Statisitc method&quot;) We can see the results are quite similar with 2-3 clusters appearing optimal. Answer to Environmental Health Question 3 With this, we can answer Environmental Health Question #3: What is the optimal number of clusters the cytokines can be grouped into that were derived from nasal epithelium fluid (NELF) in non-smokers using hierarchical clustering? Answer: Again, 2 or 3 clusters can be justified here, but for the same reasons mentioned for the first environmental health question, we landed on 3 clusters. Now we can perform the clustering and visualize and extract the results. We’ll start by using the dist() function to calculate the euclidean distance between the clusters followed by the hclust() function to obtain the hierarchical clustering assignments. # Viewing the wider dataframe we&#39;ll be working with head(wider_baseline_df) ## NS_1 NS_10 NS_11 NS_12 NS_13 NS_14 ## Eotaxin 7.215729 8.016193 6.264252 6.938344 5.152683 7.937892 ## Eotaxin3 4.396783 4.006488 5.289509 3.597945 3.051951 3.063025 ## Fractalkine 13.584554 13.679882 12.952220 13.464652 12.110742 12.785656 ## I309 0.000000 0.000000 0.000000 0.000000 0.000000 1.617372 ## IFNg 4.220509 1.990018 2.656691 3.439682 2.318340 3.439850 ## IL10 1.953721 1.145162 1.348623 1.801954 1.176498 1.800412 ## NS_2 NS_3 NS_4 NS_5 NS_6 NS_7 ## Eotaxin 9.773407 6.12290913 5.758602 4.19868540 6.832578 7.998467 ## Eotaxin3 9.022390 2.50161364 2.721195 2.21084320 3.820379 5.549164 ## Fractalkine 14.961872 12.71538464 12.057105 10.38711624 12.917180 14.636190 ## I309 1.103140 0.00000000 0.000000 0.00000000 0.000000 0.000000 ## IFNg 4.394705 0.00000000 2.638213 0.04830196 1.914092 3.251501 ## IL10 2.652120 0.08945805 2.020155 0.84649523 1.119272 2.007941 ## NS_8 NS_9 ## Eotaxin 9.242464 7.127746 ## Eotaxin3 13.724108 3.553097 ## Fractalkine 14.047382 13.146599 ## I309 0.000000 0.000000 ## IFNg 4.104820 2.940248 ## IL10 2.576814 1.970627 # First scaling data with each subject (down columns) scaled_df &lt;- data.frame(apply(wider_baseline_df, 2, scale)) rownames(scaled_df) = rownames(wider_baseline_df) head(scaled_df) ## NS_1 NS_10 NS_11 NS_12 NS_13 NS_14 ## Eotaxin 0.1563938 0.5913460 0.20280481 0.3812918 0.04218372 0.3889166 ## Eotaxin3 -0.4836295 -0.3751558 -0.03031899 -0.4135856 -0.51644074 -0.7745030 ## Fractalkine 1.6023933 1.9565250 1.80232784 1.9342844 1.89246390 1.5458678 ## I309 -1.4818900 -1.3408820 -1.29538062 -1.2697485 -1.32801259 -1.1195175 ## IFNg -0.5236512 -0.8612069 -0.65999509 -0.4512458 -0.71152186 -0.6845711 ## IL10 -1.0383105 -1.0648515 -0.97283813 -0.8409577 -1.01515934 -1.0758338 ## NS_2 NS_3 NS_4 NS_5 NS_6 NS_7 ## Eotaxin 0.5142129 0.5748344 0.3352251 0.1297382 0.6605088 0.4509596 ## Eotaxin3 0.3434700 -0.3390521 -0.4516462 -0.3710044 -0.1481961 -0.1213017 ## Fractalkine 1.6938060 2.2385415 1.9669167 1.6886200 2.2940816 2.0018134 ## I309 -1.4569648 -0.9703707 -1.1565995 -0.9279215 -1.1738781 -1.4178220 ## IFNg -0.7086302 -0.9703707 -0.4731435 -0.9157541 -0.6599894 -0.6581332 ## IL10 -1.1048056 -0.9477947 -0.6332576 -0.7146872 -0.8733797 -0.9486817 ## NS_8 NS_9 ## Eotaxin 0.5504372 0.6258899 ## Eotaxin3 1.4999690 -0.3063521 ## Fractalkine 1.5684615 2.1955622 ## I309 -1.4077759 -1.2329735 ## IFNg -0.5380823 -0.4661787 ## IL10 -0.8618228 -0.7190487 The dist() function is initially used to calculate the Euclidean distance between each cytokine. Next, the hclust() function is used to run the hierarchical clustering analysis using the complete method by default. The method can be changed in the function using the method parameter. # Calculating euclidean dist dist_matrix &lt;- dist(scaled_df, method = &#39;euclidean&#39;) # Hierarchical clustering cytokines_hc &lt;- hclust(dist_matrix) Now we can generate a dendrogram to help us evaluate the results using the fviz_dend() function from the factoextra package. We use k=3 to be consistent with the k-means analysis. fviz_dend(cytokines_hc, k = 3, # Specifying k cex = 0.85, # Label size palette = &quot;futurama&quot;, # Color palette see ?ggpubr::ggpar rect = TRUE, rect_fill = TRUE, # Add rectangle around groups horiz = TRUE, # Changes the orientation of the dendogram rect_border = &quot;futurama&quot;, # Rectangle color labels_track_height = 0.8 # Changes the room for labels ) We can also extract those cluster assignments using the cutree() function from the stats package. hc_assignments_df &lt;- data.frame(cutree(cytokines_hc, k = 3)) %&gt;% rownames_to_column(&quot;Cytokine&quot;) %&gt;% rename(`Hierarchical Cluster` = cutree.cytokines_hc..k...3.) %&gt;% # Ordering the dataframe for easier comparison arrange(`Hierarchical Cluster`) # Combining the dataframes to compare the cluster assignments from each approach comp &lt;- full_join(cluster_kmeans_df, hc_assignments_df, by = &quot;Cytokine&quot;) comp ## Cytokine K-Means Cluster Hierarchical Cluster ## 1 Fractalkine 1 2 ## 2 IL8 1 2 ## 3 IP10 1 2 ## 4 MIG 1 2 ## 5 Eotaxin 2 1 ## 6 Eotaxin3 2 1 ## 7 IL13 2 1 ## 8 IL17B 2 1 ## 9 IL1b 2 1 ## 10 IL6 2 1 ## 11 ITAC 2 1 ## 12 MCP1 2 1 ## 13 MIP1a 2 1 ## 14 MIP1b 2 1 ## 15 TARC 2 1 ## 16 I309 3 3 ## 17 IFNg 3 3 ## 18 IL10 3 3 ## 19 IL12p70 3 3 ## 20 IL17D 3 3 ## 21 IL4 3 3 ## 22 TNFa 3 3 For additional resources on running hierarchical clustering in R, see Visualizing Clustering Dendrogram in R and Hiearchical Clustering on Principal Components. Answer to Environmental Health Question 4 With this, we can answer Environmental Health Question #4: How do the hierarchical cluster assignments compare to the k-means cluster assignments? Answer: Though this may not always be the case, in this instance, we see that k-means and hierarchical clustering with k=3 clusters yield the same groupings despite the clusters being presented in a different order. Clustering Plot One additional way to visualize clustering is to plot the first two principal components on the axes and color the data points based on their corresponding cluster. This visualization can be used for both k-means and hierarchical clustering using the fviz_cluster() function. This figure is essentially a PCA plot with shapes drawn around each cluster to make them distinct from each other. fviz_cluster(cluster_kmeans_3, data = wider_baseline_df) Rather than using the fviz_cluster() function as shown in the figure above, we’ll extract the data to recreate the sample figure using ggplot(). For the manuscript this was necessary, since it was important to facet the plots for each compartment (i.e., NLF, NELF, sputum, and serum). For a single plot, this data extraction isn’t required, and the figure above can be further customized within the fviz_cluster() function. However, we’ll go through the steps of obtaining the indices to recreate the same polygons in ggplot() directly. K-means actually uses principal component analysis (PCA) to reduce a dataset’s dimensionality prior to obtaining the cluster assignments and plotting those clusters. Therefore, to obtain the coordinates of each cytokine within their respective clusters, PCA will need to be run first. # First running PCA pca_cytokine &lt;- prcomp(wider_baseline_df, scale = TRUE, center = TRUE) # Only need PC1 and PC2 for plotting, so selecting the first two columns baseline_scores_df &lt;- data.frame(scores(pca_cytokine)[,1:2]) baseline_scores_df$Cluster &lt;- cluster_kmeans_3$cluster baseline_scores_df$Protein &lt;- rownames(baseline_scores_df) # Changing cluster to a character for plotting baseline_scores_df$Cluster = as.character(baseline_scores_df$Cluster) head(baseline_scores_df) ## PC1 PC2 Cluster Protein ## Eotaxin -1.4837936 -0.37325290 2 Eotaxin ## Eotaxin3 0.7505718 -1.04251406 2 Eotaxin3 ## Fractalkine -7.0500961 -0.49034208 1 Fractalkine ## I309 4.6869896 0.42263944 3 I309 ## IFNg 2.4935464 -0.21910385 3 IFNg ## IL10 3.4361668 -0.04689913 3 IL10 Within each cluster, the chull() function is used to compute the indices of the points on the convex hull. These are needed for ggplot() to create the polygon shapes of each cluster. # hull values for cluster 1 cluster_1 &lt;- baseline_scores_df[baseline_scores_df$Cluster == 1, ][chull(baseline_scores_df %&gt;% filter(Cluster == 1)),] # hull values for cluster 2 cluster_2 &lt;- baseline_scores_df[baseline_scores_df$Cluster == 2, ][chull(baseline_scores_df %&gt;% filter(Cluster == 2)),] # hull values for cluster 3 cluster_3 &lt;- baseline_scores_df[baseline_scores_df$Cluster == 3, ][chull(baseline_scores_df %&gt;% filter(Cluster == 3)),] all_hulls_baseline &lt;- rbind(cluster_1, cluster_2, cluster_3) # Changing cluster to a character for plotting all_hulls_baseline$Cluster = as.character(all_hulls_baseline$Cluster) head(all_hulls_baseline) ## PC1 PC2 Cluster Protein ## MIG -4.7353654 -0.2498662 1 MIG ## IL8 -7.7200078 -0.8317172 1 IL8 ## IP10 -3.8660925 4.2481655 1 IP10 ## Eotaxin3 0.7505718 -1.0425141 2 Eotaxin3 ## MIP1b -0.8699488 -0.8834532 2 MIP1b ## MCP1 -2.4837364 -0.3652235 2 MCP1 Now plotting the clusters using ggplot(). ggplot() + geom_point(data = baseline_scores_df, aes(x = PC1, y = PC2, color = Cluster, shape = Cluster), size = 4) + # Adding cytokine names geom_text_repel(data = baseline_scores_df, aes(x = PC1, y = PC2, color = Cluster, label = Protein), show.legend = FALSE, size = 4.5) + # Creating polygon shapes of the clusters geom_polygon(data = all_hulls_baseline, aes(x = PC1, y = PC2, group = as.factor(Cluster), fill = Cluster, color = Cluster), alpha = 0.25, show.legend = FALSE) + theme_light() + theme(axis.text.x = element_text(vjust = 0.5), #rotating x labels/ moving x labels slightly to the left axis.line = element_line(colour=&quot;black&quot;), #making x and y axes black axis.text = element_text(size = 13), #changing size of x axis labels axis.title = element_text(face = &quot;bold&quot;, size = rel(1.7)), #changes axis titles legend.title = element_text(face = &#39;bold&#39;, size = 17), #changes legend title legend.text = element_text(size = 14), #changes legend text legend.position = &#39;bottom&#39;, # moving the legend to the bottom legend.background = element_rect(colour = &#39;black&#39;, fill = &#39;white&#39;, linetype = &#39;solid&#39;), #changes the legend background strip.text.x = element_text(size = 18, face = &quot;bold&quot;), #changes size of facet x axis strip.text.y = element_text(size = 18, face = &quot;bold&quot;)) + #changes size of facet y axis xlab(&#39;Dimension 1 (85.1%)&#39;) + ylab(&#39;Dimension 2 (7.7%)&#39;) + #changing axis labels # Using colors from the startrek palette from ggsci scale_color_startrek(name = &#39;Cluster&#39;) + scale_fill_startrek(name = &#39;Cluster&#39;) An appropriate title for this figure could be: “Figure X. K-means clusters of cytokines at baseline. Cytokines samples are derived from nasal epithelium (NELF) samples in 14 non-smoking subjects. Cytokine concentration values were transformed using a data reduction technique known as Principal Component Analysis (PCA). The first two eigenvectors plotted on the axes were able to capture a majority of the variance across all samples from the original dataset.” Takeaways from this clustering plot: PCA was able to capture almost all (~93%) of the variance from the original dataset The 22 cytokines were able to be clustered into 3 distinct clusters using k-means Hierarchical Clustering Visualization We can also build a heatmap using the pheatmap() function that has the capability to display hierarchical clustering dendrograms. To do so, we’ll need to go back and use the wider_baseline_df dataframe. pheatmap(wider_baseline_df, cluster_cols = FALSE, # hierarchical clustering of cytokines scale = &quot;column&quot;, # scaling the data to make differences across cytokines more apparent cutree_row = 3, # adds a space between the 3 largest clusters display_numbers = TRUE, number_color = &quot;black&quot;, fontsize = 12, # adding average concentration values angle_col = 45, fontsize_col = 12, fontsize_row = 12, # adjusting size/ orientation of axes labels cellheight = 17, cellwidth = 30 # setting height and width for cells ) An appropriate title for this figure could be: “Figure X. Hierarchical clustering of cytokines at baseline. Cytokines samples are derived from nasal epithelium (NELF) samples in 14 non-smoking subjects. The heatmap visualizes psuedo log2 cytokine concentrations that were scaled within each subject.” It may be helpful to add axes titles like “Subject ID” for the x axis, “Cytokine” for the y axis, and “Scaled pslog2 Concentration” for the legend after exporting from R. The pheatmap() function does not have the functionality to add those titles. Nevertheless, let’s identify some key takeaways from this heatmap: The 22 cytokines were able to be clustered into 3 distinct clusters using hierarchical clustering These clusters are based on cytokine concentration levels with the first cluster having the highest expression, the second cluster having the lowest expression, and the last cluster having average expression Variable Contributions To answer our final environmental health question: Which cytokines have the greatest contributions to the first two eigenvectors, we’ll use the fviz_contrib() function that plots the percentage of each variable’s contribution to the principal component(s). It also displays a red dashed line, and variables that fall above are considered to have significant contributions to those principal components. For a refresher on PCA and variable contributions, see the previous module, TAME 2.0 Module 5.4 Unsupervised Machine Learning. # kmeans contributions fviz_contrib(pca_cytokine, choice = &quot;ind&quot;, addlabels = TRUE, axes = 1:2) # specifies to show contribution percentages for first 2 PCs An appropriate title for this figure could be: “Figure X. Cytokine contributions to principal components. The bar chart displays each cytokine’s contribution to the first two eigenvectors in descending order from left to right. The red dashed line represents the expected contribution of each cytokine if all inputs were uniform, therefore the seven cytokines that fall above this reference line are considered to have significant contributions to the first two principal components.” Answer to Environmental Health Question 5 With this, we can answer Environmental Health Question #5: Which cytokines have the greatest contributions to the first two eigenvectors? Answer: The cytokines that have significant contributions to the first two principal components include IL-8, Fractalkine, IP-10, IL-4, MIG, I309, and IL-12p70. Concluding Remarks In this module, we explored scenarios where clustering would be appropriate but lack contextual details informing the number of clusters that should be considered, thus resulting in the need to derive such a number. In addition, methodology for k-means and hierarchical clustering was presented, along with corresponding visualizations. Lastly, variable contributions to the eigenvectors were introduced as a means to determine the most influential variables on the principal components’ composition. Additional Resources K-Means Cluster Analysis K-Means Clustering in R Hierarchical Clustering in R Test Your Knowledge Using the same dataset, answer the questions below. Determine the optimal number of k-means clusters of cytokines derived from the nasal epithelieum lining fluid of e-cigarette users. How do those clusters compare to the ones that were derived at baseline (in non-smokers)? Which cytokines have the greatest contributions to the first two eigenvectors? "]]
+[["index.html", "TAME 2.0: An Update to the TAME Toolkit for Introductory Data Science, Chemical-Biological Analyses, Machine Learning and Predictive Modeling, and Database Mining for Environmental Health Research Preface", " TAME 2.0: An Update to the TAME Toolkit for Introductory Data Science, Chemical-Biological Analyses, Machine Learning and Predictive Modeling, and Database Mining for Environmental Health Research Rager Lab 2024-10-01 Preface Background Research in exposure science, toxicology, and environmental health is becoming increasingly reliant upon data science and computational methods that can more efficiently extract information from complex datasets. These methods can be used to better identify relationships between exposures to chemicals in the environment and human disease outcomes. Still, there remains a critical gap surrounding the training of researchers on these in silico methods. Objectives We aimed to address this critical gap by developing the inTelligence And Machine lEarning (TAME) Toolkit, promoting trainee-driven data generation, management, and analysis methods to “TAME” data in environmental health studies. This toolkit encompasses training modules, organized as chapters within this Github Bookdown site. TAME site users are welcome to participate in training modules by viewing them online and/or downloading all underlying script, data, and figure input files through the UNC-SRP TAME2 GitHub website. Contributors to TAME 2.0 Julia E. Rager, PhD, MSEE (UNC-Chapel Hill) directed the developed of TAME 2.0. All training resources were developed by combining modules drafted by multiple authors and experts in the field of environmental health and data science. Authors are specifically recognized at the beginning of each module section. Contributors specifically include the following, listed in alphabetical order: Oyemwenosa N. Avenbuan, University of North Carolina at Chapel Hill (UNC-Chapel Hill) Rebecca Boyles, MSPH, Research Triangle Institute (RTI) International Jessie Chappel, PhD, University of North Carolina at Chapel Hill (UNC-Chapel Hill) Chloe K. Chou, PhD Candidate, University of North Carolina at Chapel Hill (UNC-Chapel Hill) Lauren Eaves, PhD, University of North Carolina at Chapel Hill (UNC-Chapel Hill) Rebecca C. Fry, PhD, University of North Carolina at Chapel Hill (UNC-Chapel Hill) Elise Hickman, PhD, University of North Carolina at Chapel Hill (UNC-Chapel Hill) Ilona Jaspers, PhD, University of North Carolina at Chapel Hill (UNC-Chapel Hill) Lauren E. Koval, PhD Candidate, University of North Carolina at Chapel Hill (UNC-Chapel Hill) Paul Kruse, PhD, U.S. Environmental Protection Agency (U.S. EPA) Sarah Miller, PhD Candidate, University of North Carolina at Chapel Hill (UNC-Chapel Hill) David M. Reif, PhD, Division of Translational Toxicology (DTT), National Institutes of Environmental Health Science (NIEHS) Cynthia Rider, PhD, Division of Translational Toxicology (DTT), National Institutes of Environmental Health Science (NIEHS) Kyle Roell, PhD, University of North Carolina at Chapel Hill (UNC-Chapel Hill) Alexis Payton, MS, University of North Carolina at Chapel Hill (UNC-Chapel Hill) Grace Patlewicz, PhD, U.S. Environmental Protection Agency (U.S. EPA) Caroline Ring, PhD, U.S. Environmental Protection Agency (U.S. EPA) Cavin Ward-Caviness, PhD, U.S. Environmental Protection Agency (U.S. EPA) Module Development Overview Training modules were developed to provide applications-driven examples of data organization and analysis methods that can be used to address environmental health questions. Target audiences for these modules include students and professionals in academia, government, and industry that are interested in expanding their skillset. Modules were developed by study coauthors using annotated script formatted for R/RStudio coding language and interface and are now organized into seven chapters. The first group of modules focuses on introductory data science, which included the following topics: data management and sharing practices, file management, and basic data wrangling. The second chapter of modules provides an overview of scripting, using R as the example language. Topics span downloading and programming in R, coding best practices, data manipulation and reshaping, and coding efficiencies. The third chapter demonstrates methods for basic data analysis and visualizations, spanning an introduction to data visualizations, methods to improve data visualizations, normality tests and data transformations, and introduction to statistical tests. The fourth chapter focuses on the translation of wet lab data into dry lab analyses. Topics in this chapter include an introduction to key experimental features and terms; data importing, cleaning and summary statistics from wet lab sources; two-group comparisons and visualizations; multigroup comparisons and visualizations; advanced multigroup comparisons; and advanced data export and sharing. The fifth chapter provides an introduction to machine learning and artifical intelligence (ML/AI). Topics in this chapter include a general introduction of ML/AI and then provides examples of unsupervised machine learning methods, supervised machine learning methods, and interpretation of developed ML models. The sixth chapter provides applications in toxicology and exposure science, spanning descriptive cohort analyses; -omics analyses; mixtures methodologies; toxicokinetic modeling; and chemical read-across.The seventh chapter delves into examples of environmental health database mining, and highlights the helpful databases of Comparative Toxicogenomics Database, Gene Expression Omnibus, Computational Toxicology Dashboard access through APIs, and database integration using geospatial methods. The overall organization of this TAME toolkit is summarized below. Modules are organized into seven chapters that are listed on the left side of this website. Ways to use TAME 2.0 and Underlying Site License Site viewers are encouraged to follow along each module, either starting at the very beginning and reading through the material like a classical book or clicking on specific modules of interest and learning topics of greatest interest to their data science development. Each module is designed to stand-alone. Site license This site is covered by the BY-NC-ND Creative Commons license. This license allows reusers to copy and distribute the material in any medium or format in unadapted form only, for noncommercial purposes only, and only so long as attribution is given to the creator. Figure credit Figures included in TAME 2.0 were either generated in R, created within Biorender.com, and/or pulled from online publications with site licensing and author attribution throughout. Concluding Remarks Together, this valuable resource provides unique opportunities to obtain introductory-level training on current data analysis methods applicable to 21st century exposure science, toxicology, and environmental health. These modules serve as applications-based examples on how to “TAME” data within the environmental health research field, expanding the toolbox for career development and cross-training of scientists in multiple specialties, as well as supporting the next generation of data scientists. Funding This study was supported by the National Institutes of Health (NIH) from the National Institute of Environmental Health Sciences, including the following grant funds and associated programs: P42ES031007: The University of North Carolina (UNC)-Superfund Research Program (SRP) seeks to develop new solutions for reducing exposure to inorganic arsenic and prevent arsenic-induced diabetes through mechanistic and translational research. The UNC-SRP Data Analysis and Management Core (UNC-SRP-DMAC) provides the UNC-SRP with critical expertise in bioinformatics, statistics, data management, and data integration. T32ES007126: The UNC Curriculum in Toxicology and Environmental Medicine (CiTEM) seeks to provide a cutting edge research and mentoring environment to train students and postdoctoral fellows in environmental health and toxicology. Towards this goal, the CiTEM has a T32 Training Program for Pre- and Postdoctoral Training in Toxicology to support the development of future investigators in environmental health and toxicology. Support was additionally provided through the Institute for Environmental Health Solutions (IEHS) at the University of North Carolina (UNC) Gillings School of Global Public Health. The IEHS is aimed at protecting those who are particularly vulnerable to diseases caused by environmental factors, putting solutions directly into the hands of individuals and communities of North Carolina and beyond. "],["fair-data-management-practices.html", "FAIR Data Management Practices Introduction to Training Module Introduction to FAIR Breaking Down FAIR, Letter-by-Letter What Does This Mean for You? Data Repositories for Sharing of Data Recent Shifts in Regulatory Policies for Data Sharing Additional Training Resources on FAIR", " FAIR Data Management Practices This training module was developed by Rebecca Boyles, with contributions from Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module This training module provides a description of FAIR data management practices, and points participants to important resources to help ensure generated data meet current FAIR guidelines. This training module is descriptive content-based (as opposed to coding-based), in order to present information clearly and serve as an important resource alongside the other scripted training activities. Training Module’s Environmental Heatlh Questions This training module was specifically developed to answer the following questions: What is FAIR? When was FAIR first developed? When making data ‘Findable’, who and what should be able to find your data? When saving/formatting your data, which of the following formats is preferred to meet FAIR principles: .pdf, .csv, or a proprietary output file from your lab instrument? How can I find a suitable data repository for my data? Introduction to FAIR Proper data management is of utmost importance while leading data analyses within the field of environmental health science. A method to ensure proper data management is the implementation of Findability, Accessibility, Interoperability, and Reusability (FAIR) practices. A landmark paper that describes FAIR practices in environmental health research is the following: Wilkinson MD, Dumontier M, Aalbersberg IJ, et al. The FAIR Guiding Principles for scientific data management and stewardship. Sci Data. 2016 Mar 15. PMID: 26978244. The FAIR principles describe a framework for data management and stewardship aimed at increasing the value of data by enabling sharing and reuse. These principles were originally developed from discussions during the Jointly Designing a Data FAIRport meeting at the Lorentz Center in Leiden, The Netherlands in 2014, which brought together stakeholders to discuss the creation of an environment for virtual computational science. The resulting principles are technology agnostic, discipline independent, community driven, and internationally adopted. Below is a schematic providing an overview of this guiding principle: Answer to Environmental Health Question 1 &amp; 2 With this background, we can answer Environmental Health Question #1 and #2: What is FAIR and when was it first developed? Answer: FAIR is guiding framework that was recently established to promote best data management practices, to ensure that data are Findable, Accessibility, Interoperable, and Reusable. It was first developed in 2014- which means that these principles are very new and continuing to evolve! Breaking Down FAIR, Letter-by-Letter The aspects of the FAIR principles apply to data and metadata with the aim of making the information available to people and computers as described in the seminal paper by Wilkinson et al., 2016. F (Findable) in FAIR The F in FAIR identifies components of the principles needed to make the meta(data) findable through the application of unique persistent identifiers, thoroughly described, reference the unique identifiers, and that the descriptive information (i.e., metadata) could be searched by both humans and computer systems. F1. (Meta)data are assigned a globally unique and persistent identifier Each dataset is assigned a globally unique and persistent identifier (PID), for example a DOI. These identifiers allow to find, cite and track (meta)data. A DOI looks like: https://doi.org/10.1109/5.771073 Action: Ensure that each dataset is assigned a globally unique and persistent identifier. Certain repositories automatically assign identifiers to datasets as a service. If not, obtain a PID via a PID registration service. F2. Data are described with rich metadata Each dataset is thoroughly (see R1) described: these metadata document how the data was generated, under what term (license) and how it can be (re)used and provide the necessary context for proper interpretation. This information needs to be machine-readable. Action: Fully document each dataset in the metadata, which may include descriptive information about the context, quality and condition, or characteristics of the data. Another researcher in any field, or their computer, should be able to properly understand the nature of your dataset. Be as generous as possible with your metadata (see R1). F3. Metadata clearly and explicitly include the identifier of the data it describes Explanation: The metadata and the dataset they describe are separate files. The association between a metadata file and the dataset is obvious thanks to the mention of the dataset’s PID in the metadata. Action: Make sure that the metadata contains the dataset’s PID. F4. (Meta)data are registered or indexed in a searchable resource Explanation: Metadata are used to build easily searchable indexes of datasets. These resources will allow to search for existing datasets similarly to searching for a book in a library. Action: Provide detailed and complete metadata for each dataset (see F2). Answer to Environmental Health Question 3 With this, we can answer Environmental Health Question #3: When making data ‘Findable’, who and what should be able to find your data? Answer: Both humans and computer systems should be able to find your data. A (Accessible) in FAIR The A components are designed to enable meta(data) be available long-term, accessed by humans and machines using standard communication protocols with clearly described limitations on reuse. A1. (Meta)data are retrievable by their identifier using a standardized communications protocol Explanation: If one knows a dataset’s identifier and the location where it is archived, one can access at least the metadata. Furthermore, the user knows how to proceed to get access to the data. Action: Clearly define who can access the actual data and specify how. It is possible that data will not be downloaded, but rather reused in situ. If so, the metadata must specify the conditions under which this is allowed (sometimes versus the conditions needed to fulfill for external usage/“download”). A1.1 The protocol is open, free, and universally implementable Explanation: Anyone with a computer and an internet connection can access at least the metadata. A1.2 The protocol allows for an authentication and authorization procedure, where necessary Explanation: It often makes sense to request users to create a user account on a repository. This allows to authenticate the owner (or contributor) of each dataset, and to potentially set user specific rights. A2. Metadata are accessible, even when the data are no longer available Explanation: Maintaining all datasets in a readily usable state eternally would require an enormous amount of curation work (adapting to new standards for formats, converting to different format if specifically needed software is discontinued, etc). Keeping the metadata describing each dataset accessible, however, can be done with fewer resources. This allows to build comprehensive data indexes including all current, past, and potentially arising datasets. Action: Provide detailed and complete metadata for each dataset (see R1). I (Interoperable) in FAIR The I components of the principles address needs for data exchange and interpretation by humans and machines which includes the use of controlled vocabularies or ontologies to describe meta(data) and to describe provenance relationships through appropriate data citation. I1. (Meta)data use a formal, accessible, shared, and broadly applicable language Explanation: Interoperability typically means that each computer system has at least knowledge of the other system’s formats in which data is exchanged. If (meta)data are to be searchable and if compatible data sources should be combinable in a (semi)automatic way, computer systems need to be able to decide if the content of datasets are comparable. Action: Provide machine readable data and metadata in an accessible language, using a well-established formalism. Data and metadata are annotated with resolvable vocabularies/ontologies/thesauri that are commonly used in the field (see I2). I2. (Meta)data use vocabularies that follow FAIR principles Explanation: The controlled vocabulary (e.g., MESH) used to describe datasets needs to be documented. This documentation needs to be easily findable and accessible by anyone who uses the dataset. Action: The vocabularies/ontologies/thesauri are themselves findable, accessible, interoperable and thoroughly documented, hence FAIR. Lists of these standards can be found at: NCBO BioPortal, FAIRSharing, OBO Foundry. I3. (Meta)data include qualified references to other (meta)data Explanation: If the dataset builds on another dataset, if additional datasets are needed to complete the data, or if complementary information is stored in a different dataset, this needs to be specified. In particular, the scientific link between the datasets needs to be described. Furthermore, all datasets need to be properly cited (i.e. including their persistent identifiers). Action: Properly cite relevant/associated datasets, by providing their persistent identifiers, in the metadata, and describe the scientific link/relation to your dataset. R (Reusable) in FAIR The R components highlight needs for the meta(data) to be reused and support integration such as sufficient description of the data and data use limitations. R1. Meta(data) are richly described with a plurality of accurate and relevant attributes Explanation: Description of a dataset is required at two different levels: Metadata describing the dataset: what does the dataset contain, how was the data generated, how has it been processed, how can it be reused. Metadata describing the data: any needed information to properly use the data, such as definitions of the variable names Action: Provide complete metadata for each data file. Scope of your data: for what purpose was it generated/collected? Particularities or limitations about the data that other users should be aware of. Date of the dataset generation, lab conditions, who prepared the data, parameter settings, name and version of the software used. Variable names are explained or self-explanatory. Version of the archived and/or reused data is clearly specified and documented. What Does This Mean for You? We advise the following as ‘starting-points’ for participants to start meeting FAIR guidances: Learn how to create a Data Management Plan Keep good documentation (project &amp; data-level) while working Do not use proprietary file formats (.csv is a great go-to formats for your data!) When able, use a domain appropriate metadata standard or ontology Ruthlessly document any steps in a project Most of FAIR can be handled by selecting a good data or software repository Don’t forget to include a license! Answer to Environmental Health Question 4 With these, we can answer Environmental Health Question #4: When saving/formatting your data, which of the following formats is preferred to meet FAIR principles: .pdf, .csv, or a proprietary output file from your lab instrument? Answer: A .csv file is preferred to enhance data sharing. Data Repositories for Sharing of Data When you are organizing your data to deposit online, it is important to identify an appropriate repository to publish your dataset it. A good starting place is a repository registry such as FAIRsharing.org or re3data.org. Journals can also provide helpful resources and starting repository lists, such as Nature and PLOS, which both have published a list of recommended repositories. Funding agencies, including the NIH, can also inform specific repositories. Below are some examples of two main categories of data repositories: 1. Domain Agnostic Data Repositories Domain agnostic repositories allow the deposition of any data type. Some examples include the following: Data in Brief Articles (e.g., Elsevier’s Data in Brief Journal) Dryad Figshare The Dataverse Project Zenodo 2. Domain Specific Data Repositories Domain specific repositories allow the deposition of specific types of data, produced from specific types of technologies or within specific domains. Some examples include the following: Database of Genotypes and Phenotypes Gene Expression Omnibus The Immunology Database and Analysis Portal Metabolomics Workbench (National Metabolomics Data Repository) Microphysiology Systems Database Mouse Genome Informatics Mouse Phenome Database OpenNeuro Protein Data Bank ProteomeXchange Rat Genome Database The Database of Genotypes and Phenotypes Zebrafish Model Organism Database and many, many, many others… Answer to Environmental Health Question 5 With these, we can answer Environmental Health Question #5: How can I find a suitable data repository for my data? Answer: I can search through a data repository registry service or look for recommendations from NIH or other funding agencies. Recent Shifts in Regulatory Policies for Data Sharing The NIH Data Management and Sharing Policy NIH’s data management and sharing (DMS) policy became effective January 2023. This policy specifically lists the expectations that investigators must comply with in order to promote the sharing of scientific data. Information about this recent policy can be found through updated NIH websites. Information about writing an official Data Management and Sharing (DMS) plan for your research can be found through NIH’s Guidance on Writing a Data Management &amp; Sharing Plan. The 2018 Evidence Act The Evidence Act, or Foundations for Evidence-Based Policymaking Act of 2018, was signed into U.S. law on January 14, 2019. The Act requires federal agencies to build the capacity to use evidence and data in their decision-making and policymaking. It also requires agencies to: Develop an evidence-building plan as part of their quadrennial strategic plan &amp; Develop an evaluation plan concurrent with their annual performance plan. The Evidence Act also: Mandates that data be “open by default” Specifies that a comprehensive data inventory should be created for each agency’s open data assets How Does the NIH Data Management and Sharing Policy Intersect with the 2018 Evidence Act? Making your data FAIR, by definition, makes it more shareable and reusable. Many of the requirements in the NIH DMS and the Evidence Act policy overlap with the FAIR principles. The CARE Principles for Indigenous Data Governance While we are experiencing increased requirements for the open sharing of data, it is important to recognize that there are circumstances and populations that should, at the same time, be carefully protected. Examples include human clinical or epidemiological data that may become identifiable upon the sharing of sensitive data. Another example includes the consideration of Indigenous populations. A recent article by Carroll et al. 2021 describes in their abstract: As big data, open data, and open science advance to increase access to complex and large datasets for innovation, discovery, and decision-making, Indigenous Peoples’ rights to control and access their data within these data environments remain limited. Operationalizing the FAIR Principles for scientific data with the CARE Principles for Indigenous Data Governance enhances machine actionability and brings people and purpose to the fore to resolve Indigenous Peoples’ rights to and interests in their data across the data lifecycle. Additional Training Resources on FAIR Many organizations, from specific programs to broad organizations, provide training and resources for scientists in FAIR principles. Some of the notable global organizations organizing and providing training that offer opportunities for community involvement are: Committee on Data for Science and Technology (CODATA) Global Alliance for Genomics &amp; Health GoFAIR Force11 Research Data Alliance Example Workshops discussing FAIR: NAS Implementing FAIR Data for People and Machines: Impacts and Implications (2019). Available at: https://www.nationalacademies.org/our-work/implementing-fair-data-for-people-and-machines-impacts-and-implications NIH Catalyzing Knowledge-driven Discovery in Environmental Health Sciences Through a Harmonized Language, Virtual Workshop (2021). Available at: https://www.niehs.nih.gov/news/events/pastmtg/2021/ehslanguage/index.cfm NIH Trustworthy Data Repositories Workshop (2019). Available at: https://datascience.nih.gov/data-ecosystem/trustworthy-data-repositories-workshop NIH Virtual Workshop on Data Metrics (2020). Available at: https://datascience.nih.gov/data-ecosystem/nih-virtual-workshop-on-data-metrics NIH Workshop on the Role of Generalist Repositories to Enhance Data Discoverability and Reuse: Workshop Summary (2020). Available at: https://datascience.nih.gov/data-ecosystem/nih-data-repository-workshop-summary Example Government Report Documents on FAIR: Collins S, Genova F, Harrower N, Hodson S, Jones S, Laaksonen L, Mietchen D, Petrauskaite R, Wittenburg P. Turning FAIR into reality: Final report and action plan from the European Commission expert group on FAIR data: European Union; 2018. Available at: https://www.vdu.lt/cris/handle/20.500.12259/103794. EU. FAIR Data Advanced Use Cases: From Principles to Practice in the Netherlands. 2018. European Union. Available at: doi:10.5281/zenodo.1250535. NIH. Final NIH Policy for Data Management and Sharing and Supplemental Information. National Institutes of Health. Federal Register, vol. 85, 2020-23674, 30 Oct. 2020, pp. 68890–900. Available at: https://www.federalregister.gov/d/2020-23674. NIH. NIH Strategic Plan for Data Science 2018. National Institutes of Health. Available at: https://datascience.nih.gov/strategicplan. NLM. NLM Strategic Plan 2017 to 2027. U.S. National Library of Medicine, Feb. 2018. Available at: https://www.nlm.nih.gov/about/strategic-plan.html. Example Related Publications on FAIR: Comess S, Akbay A, Vasiliou M, Hines RN, Joppa L, Vasiliou V, Kleinstreuer N. Bringing Big Data to Bear in Environmental Public Health: Challenges and Recommendations. Front Artif Intell. 2020 May;3:31. doi: 10.3389/frai.2020.00031. Epub 2020 May 15. PMID: 33184612; PMCID: PMC7654840. Koers H, Bangert D, Hermans E, van Horik R, de Jong M, Mokrane M. Recommendations for Services in a FAIR Data Ecosystem. Patterns (N Y). 2020 Jul 7;1(5):100058. doi: 10.1016/j.patter.2020.100058. Erratum in: Patterns (N Y). 2020 Sep 11;1(6):100104. PMID: 33205119. Kush RD, Warzel D, Kush MA, Sherman A, Navarro EA, Fitzmartin R, Pétavy F, Galvez J, Becnel LB, Zhou FL, Harmon N, Jauregui B, Jackson T, Hudson L. FAIR data sharing: The roles of common data elements and harmonization. J Biomed Inform. 2020 Jul;107:103421. doi: 10.1016/j.jbi.2020.103421. Epub 2020 May 12. PMID: 32407878. Lin D, Crabtree J, Dillo I, Downs RR, Edmunds R, Giaretta D, De Giusti M, L’Hours H, Hugo W, Jenkyns R, Khodiyar V, Martone ME, Mokrane M, Navale V, Petters J, Sierman B, Sokolova DV, Stockhause M, Westbrook J. The TRUST Principles for digital repositories. Sci Data. 2020 May 14;7(1):144. PMID: 32409645. Thessen AE, Grondin CJ, Kulkarni RD, Brander S, Truong L, Vasilevsky NA, Callahan TJ, Chan LE, Westra B, Willis M, Rothenberg SE, Jarabek AM, Burgoon L, Korrick SA, Haendel MA. Community Approaches for Integrating Environmental Exposures into Human Models of Disease. Environ Health Perspect. 2020 Dec;128(12):125002. PMID: 33369481. Roundtable on Environmental Health Sciences, Research, and Medicine; Board on Population Health and Public Health Practice; Health and Medicine Division; National Academies of Sciences, Engineering, and Medicine. Principles and Obstacles for Sharing Data from Environmental Health Research: Workshop Summary. Washington (DC): National Academies Press (US); 2016 Apr 29. PMID: 27227195. Test Your Knowledge Let’s imagine that you’re a researcher who is planning on gathering a lot of data using the zebrafish model. In order to adequately prepare your studies and steps to ensure data are deposited into proper repositories, you have the idea to check repository information obtained in FAIRsharing.org. What are some example repositories and relevant ontology resources that you could use to organize, deposit, and share your zebrafish data (hint: use the search tool)? "],["data-sharing-through-online-repositories.html", "Data Sharing through Online Repositories An Overview and Example with the Dataverse Repository Introduction to Training Module Data Repositories The Dataverse Project What is a Dataverse? Metadata Creating a Dataverse Creating a Dataset Concluding Remarks", " Data Sharing through Online Repositories An Overview and Example with the Dataverse Repository This training module was developed by Kyle R. Roell, Alexis Payton, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module Submitting data to publicly available repositories is an essential part of ensuring data meet FAIR guidelines, as discussed in detail in the previous training module. There are many benefits to sharing and submitting your researching, such as: Making more use out of data that are generated in your lab More easily sharing and integrating across datasets Ensuring reproducibility in analysis findings and conclusions Improving the tracking and archiving of data sources, and data updates Increasing the awareness and attention surrounding your research as others locate your data through additional online queries Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: How should I structure my data for upload into online repositories? What does the term ‘metadata’ mean and what does it look like? This module will introduce some of the repositories that are commonly used to deposit data, how to set up metadata files, and how to organize example data in preparation for sharing. We will also provide information surrounding best practices for data organization and sharing through these repositories. Additional resources are also provided throughout, as there are many ways to organize, share, and deposit data depending on your data types and structures and overall research goals. Data Repositories There are many publicly available repositories that we should consider when depositing data. Some general repository registries that are helpful to search through include FAIRsharing.org or re3data.org. Journals can also provide helpful resources and starting repository lists, such as Nature and PLOS, which both have published a list of recommended repositories. As detailed in the FAIR training module, there are two main categories of data repositories: 1. Domain Agnostic Data Repositories Domain agnostic repositories allow the deposition of any data type. Some examples include: Data in Brief Articles (e.g., Elsevier’s Data in Brief Journal) Dryad Figshare The Dataverse Project Zenodo 2. Domain Specific Data Repositories Domain specific repositories allow the deposition of specific types of data, produced from specific types of technologies or within specific domains. Some examples include: Database of Genotypes and Phenotypes Gene Expression Omnibus The Immunology Database and Analysis Portal Metabolomics Workbench (National Metabolomics Data Repository) Microphysiology Systems Database Mouse Genome Informatics Mouse Phenome Database OpenNeuro Protein Data Bank ProteomeXchange Rat Genome Database The Database of Genotypes and Phenotypes Zebrafish Model Organism Database and many, many, many others… This training module focuses on providing an example of how to organize and upload data into the Dataverse; though many of the methods described below pertain to other data repositories as well, and also incorporate general data organization and sharing best practices. The Dataverse Project Dataverse, organized through The Dataverse Project, is a popular repository option that allows for upload of most types of material, without any stringent requirements. The Dataverse organization also provides ample resources on how to organize, upload, and share data through Dataverse. These resources include very thorough, readable, and user guides and best practices. Screenshot of the main page of The Dataverse Project An easier way to think about Dataverse is to interpret it similar to a folder system on your computer. A Dataverse is just an online folder that contains files, data, or datasets that are all related to some topic, project, etc. Although Dataverse was started at Harvard and the base Dataverse lives there, there are many versions of Dataverse that are specific to and supported by various institutions. For example, these training modules are being developed primarily by faculty, staff, and students at the University of North Carolina at Chapel Hill. As such, the examples contained in this module will specifically connect with the UNC Dataverse; though many of the methods outlined here are applicable to other Dataverses and additional online repositories, in general. What is a Dataverse? Remember how we pointed out that a Dataverse is similar to a folder system on a computer? Well, here we are going to show you what that actually looks like. But first, something that can be confusing when starting to work with Dataverse is the fact that the term Dataverse is used for both the overarching repository as well as individual subsections (or folders) in which data are stored. For example, the UNC Dataverse is called a Dataverse, but to upload data, you need to upload it to a specific sub-Dataverse. So, what is the difference between the high level UNC Dataverse and smaller, sub-dataverses? Well, nothing, really. The UNC Dataverse is similar to a large folder that says, these are all the projects and research related to or contained within UNC. From there, we want to be more specific about where we store our research, so we are creating more sub-Dataverses (folders) within that higher, overarching UNC Dataverse. As an example, using the UNC Dataverse, here we can see various sub-Dataverses that have been created as repositories for specific projects or types of data. As another example looking within a specific Dataverse, here we can see the Dataverse that hosts datasets and publications for Dr. Julia Rager’s lab, the Ragerlab-Dataverse. Within this Datavere, we can see various datasets produced by her lab. It is worth noting that the datasets may not necessarily be directly related to each other in terms of exact topic, for example, the Ragerlab-Dataverse hosts data pertaining to wildfire smoke exposure as well as chemical exposures and breast cancer. But they are all pertaining to experiments and analyses run within her specific lab. Let’s now start talking more specifically about how to organize data and format files for Dataverse, create your own “Dataverse”, upload datasets, and what this all means! Dataset Structure Before uploading your data to any data repository, it is important to structure your data efficiently and effectively, making it easy for others to navigate, understand, and utilize. While we will cover this in various sections throughout these training modules, here are some basic tips for data structure and organization. Keep all data for one participant or subject within one column (or row) of your dataset Genomic data and other analytical assays tend to have subjects on columns and genes, expression, etc. as the rows Descriptive and demographic data often tend to have subjects or participants as the rows and each descriptor variable (including demographics and any other subject variables) as columns Create succinct, descriptive variable names For example, do not use something like “This Variable Contains Information Regarding Smoking Status”, and instead just using something like, “Smoking_Status” Be aware of using spacing, special characters, and capitalization within variable names Think about transforming data from wide to long format depending on your specific dataset and general conventions Be sure to follow specific guidelines of repository when appropriate TAME 2.0 Module 1.1 FAIR Data Management Practices and TAME 2.0 Module 1.4 Data Wrangling in Excel are also helpful resources to reference when thinking about organizing your data. A general example of an organized, long format dataset in Excel in provided below: Only .csv or .txt files can be uploaded to dataverse; therefore, the metadata and data tabs in an excel file will need to saved and uploaded as two separate .csv or .txt files. Answer to Environmental Health Question 1 With this, we can answer Environmental Health Question 1: How should I structure my data for upload into online repositories? Answer: It is ideal to have data clearly organized and filled, with succinct and descriptive variable names clearly labeled and values filled in. Most commonly, datasets should be saved as separate .csv or .txt files for upload into data repositories. Metadata There are many different definitions of what a metadata file is. Helpful explanations, for example, are provided by the UNC University Libraries: There are many definitions of metadata, but one of the simplest is data about data. More specifically… Metadata (in terms of data management) describe a dataset: how they were collected; when they were collected; what assumptions were made in their methodology; their geographic scope; if there are multiple files, how they relate to one another; the definitions of individual variables and, if applicable, what possible answers were (i.e., to survey questions); the calibration of any equipment used in data collection; the version of software used for analysis; etc. Very often, a dataset that has no metadata is incomprehensible. Metadata ARE data. They are pieces of information that have some meaning in relation to another piece of information. They can be created, managed, stored, and preserved like any other data. Metadata can be applied to anything. A computer file can be described in the same way that a book or piece of art can be described. For example, both can have a title, an author, and a year created. Metadata should be documented for research outputs of any kind. Metadata generally has little value on their own. Metadata adds value to other information, but are usually not valuable in themselves. There are exceptions to this rule, such as text transcription of an audio file. There are three kinds of metadata: Descriptive metadata consist of information about the content and context of your data. Examples: title, creator, subject keywords, and description (abstract) Structural metadata describe the physical structure of compound data. Examples: camera used, aperture, exposure, file format, and relation to other data or files Administrative metadata are information used to manage your data. Examples: when and how they were created, who can access them, software required to use them, and copyright permissions Therefore, after having organized your primary dataset for submission into online repositories, it is equally important to have a metadata file for easy comprehension and utilization of your data for future researchers or anyone downloading your data. While most repositories capture some metadata on the dataset page (e.g., descripton of data, upload date, contact information), there is generally little information about the specific data values and variables. In this section, we review some general guidelines and tips to better annotate your data. First, keep in mind, depending on the specific repository you are using, you may have to follow their metadata standards. But, if uploading to more generalist repository, this may be up to you to define. Generally, a metadata file consists of a set of descriptors for each variable in the data. If you are uploading data that contains many covariates or descriptive variables, it is essential that you provide a metadata file that describes these covariates. Both a description of the variable as well as any specific levels of any categorical or factor type variables. From the dataset presented previously, here we present an example of an associated metadata file: Answer to Environmental Health Question 2 With this, we can answer Environmental Health Question 2: What does the term ‘metadata’ mean and what does it look like? Answer: Metadata refers to the information that describes and explains data. It looks like an additional dataset that provides context with details such as the source, type, owner, and relationships to other datasets. This file can help users understand the relevance of a specific dataset and provide guidance on how to use it. Creating a Dataverse Now, let’s review how to actually create a Dataverse. First, navigate to the parent Dataverse that you would like to use as your primary host website. For example, our group uses the UNC Dataverse. If you do not already have one, create a username and login. Then, from the home Dataverse page, click “Add Data” and select “New Dataverse”. And fill in the information necessary. And that is it. After creating your Dataverse site, you will need to publish it; however, before it is accessible to the public, note that you can actually create a Dataverse within another Dataverse (similar to a folder within a folder on your computer). This makes sense even when you are creating a new Dataverse at the home, UNC Dataverse level, you are still technically creating a new Dataverse within an existing one (the large UNC Dataverse). Here are some tips as you create your Dataverse: Do not recreate a Dataverse that already exists Choose a name that is specific, but general enough that it doesn’t only pertain to one specific dataset You can add more than one contact email, if necessary Creating a Dataset Creating a dataset creates a page for your data containing information about that data, a citation for the data (something valuable and somewhat unique to Dataverse), as well the place from where you data can be directly accessed or downloaded. First, decide the specific Dataverse your data will live and navigate to that specific Dataverse site. Then carry out the following steps to create a dataset: Navigate to the Dataverse page under which your dataset will live Click “Add Data” and then select “New Dataset” Fill in the necessary information Upload your data and metadata file(s) structured as detailed above Now, you have a dataset within your Dataverse. Again, you will have to publish the dataset for someone to have access to it. The easy part of using a more generalist repository like Dataverse, is that you do not have to have a strict data structure adherence. However, this means it is up to you to make sure your data is readable and useable. Concluding Remarks In this training module, we set out to express the importance of uploading data to online repositories, demonstrate what the upload process may look like using a generalist repository (Dataverse), and give some examples and tips on structuring data for upload and creating metadata files. It is important to choose the appropriate repository for your data based on your field of study and specifications of your work. Test Your Knowledge Try creating your own Dataverse repository, format your files to be uploaded to Dataverse, and upload those files to your new repository! "],["file-management-using-github.html", "File Management using Github Introduction to Training Module Creating an Account Creating a Repository Uploading Code Adding Subfolders in a Repository Updating Code Updating Repository Titles and Structure to Support a Manuscript Tracking Code Changes using Github Branches Concluding Remarks", " File Management using Github This training module was developed by Alexis Payton, Lauren E. Koval, Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module Good data practices like file management and code tracking are imperative for data analysis initiatives, especially when working in research teams and/or shared project folders. Often times analyses and manuscripts are edited many times prior to being submitted for a grant or publication. Analysis methods are also shared between members of a research team and to external communities, as further detailed in TAME 2.0 Module 1.1 FAIR Data Management Practices. Therefore, Github has emerged as an effective way to manage, share, and track how code changes over time. Github is an open source or publicly accessible platform designed to facilitate version control and issue tracking of code. It is used by us and many of our colleagues to not only document versions of script written for data analysis and visualization, but to also make our code publicly available for open communication and dissemination of results. This training module serves a launch pad for getting acclimated with Github and includes… Creating an account Uploading code Creating a repository and making it legible for manuscript submission Creating an Account First, users must create their own accounts within github to start uploading/sharing code. To do this, navigate to github.com, click “Sign Up”, and follow the on screen instructions. Creating a Repository A repository, also known as a “repo”, is similar to a project folder that will contain all code pertaining to a specific project (which can be used for specific research programs, grants, or manuscripts, as examples). A repository can be set to public or private. If a repo is initially set to private to keep findings confidential prior to publication, it can always be updated to public once findings are ready for public dissemination. Multiple people can be allowed to work on a project together within a single repository. To access the repositories that are currently available to you through your user account, click the circle in top right-hand corner and click “Your repositories”. To create a new repository, click on the green button that says “New”. Then give your repository a descriptive name. We often edit the repo titles to match the title of specific manuscripts, though specific titling formats are up to the users/team’s preference. For more information, visit Github’s Create a repo documentation. Then click “Add a README file” to initiate the README file, which is important to continually edit to provide analysis-specific background information, and any additional information that would be helpful during and after code is drafted to better facilitate tracking information and project details. We provide further details surrounding specific information that can be included within the README file below. Uploading Code The simplest way to upload code is to first navigate to the repository that you would like to upload your code/associated files to. Note that this could represent a repo that you created or that someone granted you access to. Click “Add file” then click “Upload files”. Drag and drop your file containing your script into github and click “Commit changes”. A more advanced way to upload code is by using the command line, which allows a user to directly interact with the computer or software application. Further documentation can be found here. Adding Subfolders in a Repository To keep the repository organized, it might be necessary to create a new folder (like the folder labeled “1.1. Summary Statistics” in the above screenshot). Files can be grouped into these folders based on the type of analysis. To do so, click on the new file and then click on the pencil icon next to the “Blame” button. Click on the box that contains the title of the file. Write the title of your new folder and then end with a forward slash (/). In the screenshot below, we’re creating a new folder entitled “New Folder”. Click “Commit changes” and your file should now be in a new folder. Updating Code Saving iterations of code can save valuable time later as analyses are constantly being updated and edited. If your code undergoes substantial changes, (e.g., adding/ removing steps or if there’s code that is likely to be beneficial later on, but is no longer relevant to the current analysis), it is helpful to save that version in Github for future reference. To do so, create a subfolder named “Archive” and move the old file into it. If you have multiple versions of a file with the same name, add the current date to prevent the file from being overwritten later on as seen in the screenshot below. Once the old file version has been archived, now upload the most recent version of your code to the main folder. Based on the screenshot above, that would be under “3. ML Visualizations”. Note: If a file is uploaded with the same name it will be overwritten, which can’t be undone! Therefore, put the older file into the archive folder if you’d like it to be saved PRIOR to uploading the new version. Updating Repository Titles and Structure to Support a Manuscript If the code is for a manuscript, it’s helpful to include the table or figure name it pertains to in the manuscript in parentheses. For example, “Baseline Clusters (Figure 3)”. This allows viewers to find find the code for each table or figure faster. Using a README.md file A README.md file is used to describe the overall aims and purpose of the analyses in the repository or a folder within a repository. It is often the first file that someone will look at in a repo/folder, so it is important to include information that would be valuable to an outsider trying to make use of the work. To add a README.md file, click “Add file” and then “Create new file”. Name your file “README.md”. A README.md file uses R markdown syntax. This type of syntax is very helpful as you continue to develop R coding skills, as it provides a mechanism through which your code’s output can be visualized and saved as a rendered file version. There are many helpful resources for R markdown, including some that we find helpful: R Markdown Cheatsheet R Markdown Syntax Overview The final README.md file for the OVERALL repository for manuscript submission should look something like the screenshot below. Always include… The main goal of the project The final manuscript name, year it was published, Pub Med ID (if applicable) Graphical abstract (if needed for publication) Names and brief descriptions of each file Include both the goal of the analysis and the methodology used (ie. Using chi square tests to determine if there are statistically significant differences across demographic groups) If the code was written in the software Jupyter (ie. has the extension .ipynb not .R or .Rmd), NBViewer is a website that can render jupyter notebooks (files). This is helpful, because sometimes the files take too long to render, so link the repository from the NB viewer website. Go to nbviewer.org –&gt; type in the name of the repository –&gt; copy the url and add it to the README.md file The final README.md file for the a subfolder within a repository should look something like the screenshot below. Always include… The name of each file Brief description of each file Include both the goal of the analysis and the methodology used Table or Figure name in the corresponding manuscript (if applicable) Note: That the organization structure for the README.md files are simply recommendations and should be changed based on needs of the project. However, it is important to include information and organize the repository in a way that helps other readers and colleagues navigate it who aren’t familiar with the project. Example Repositories Below are links to repositories that contain code for analyses used in published manuscripts. These are examples of well organized Github repositories. Wildfires and Environmental Justice: Future Wildfire Events Predicted to Disproportionally Impact Socioeconomically Vulnerable Communities in North Carolina Plasma sterols and vitamin D are correlates and predictors of ozone-induced inflammation in the lung: A pilot study Cytokine signature clusters as a tool to compare changes associated with tobacco product use in upper and lower airway samples Tracking Code Changes using Github Branches Github is a useful platform for managing and facilitating code tracking performed by different collaborators through branches. When creating a repository on Github, it automatically creates a default branch entitled “main”. It’s possible to create a new branch which allows a programmer to make changes to files in a repository in isolation from the main branch. This is beneficial, because the same file can be compared across branches, potentially created by different scientists, and merged together to reflect those changes. Note: In order for this to work the file in main branch has to have the same name and the file in the newly created branch. Let’s start by creating a new branch. First, navigate to a repository, select “main” and then “View all branches”. Click “New branch”, give your branch a title, and click “Create new branch”. In the screenshot, you’ll see the new branch entitled “jr-changes”. As a new collaborator interested in comparing and merging code changes to a file, click on the new branch that was just created. Based on the screenshot, that means click “jr-changes”. After uploading the file(s) to this branch, you’ll see a notification that this branch is now a certain number of commits ahead of the main branch. A commit records the number of changes to files in a branch. Based on the screenshot, “jr-changes” is now 2 commits ahead of “main”. Click on “2 commits ahead” and scroll down to compare versions between the “main” and “jr-changes” branches. A pull request will need to be created. A pull request allows other collaborators to see changes made to a file within a branch. These proposed changes can be discussed and amended before merging them into the main branch. For more information, visit Github’s branches, pull requests and comparing branches in pull requests documentation. Go ahead and click on “Create pull request”. Click on “Create pull request” again on the next screen. Select “Merge pull request” and then “Confirm merge”. Concluding Remarks In summary, this training module serves as a basic tutorial for sharing code on Github in a way that is beneficial for scientific research. Concepts discussed include uploading and updating code, making a repository easily readable for manuscript submissions, and tracking code changes across collaborators. We encourage trainees and data scientists to implement code tracking and sharing through Github and to also keep up with current trends in data analysis documentation that continue to evolve over time. Test Your Knowledge Try creating your own Github profile, set up a practice repo with subfolders, and a detailed READ.md file paralleling the suggested formatting and content detailed above for your own data analyses! "],["data-wrangling-in-excel.html", "Data Wrangling in Excel Introduction to Training Module Save a Copy of the Soon-To-Be Organized and Cleaned Dataset as a New File Remove Extraneous White Space Replace Missing Data with “NA” Create a Metadata Tab Abbreviate and Capitalize Categorical Data Alphabetize (Sort) the Data by the Categorical Variable of Interest Create a New Subject Number Column Remove Special Symbols and Dashes Bold all Column Names and Center all Data Create a Subject Identifier Column Separate Subject Demographic Data from Experimental Measurements Convert Data from Wide to Long Format Pivoting Data from a Wide to Long Format Generating Summary-Level Statistics with Pivot Tables Excel vs. R: Which Should You Use? Concluding Remarks", " Data Wrangling in Excel This training module was developed by Alexis Payton, Elise Hickman, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module This module is intended to be a starting guide to cleaning and organizing an example toxicology dataset in Excel. Data wrangling involves cleaning, removing of erroneous data, and restructuring necessary for to preparing wet lab generated data for downstream analyses. These steps will ensure that: Data are amenable to downstream analyses in R, or your preferred programming language Data are clear and easily interpretable by collaborators, reviewers, and readers Click here for more information on data wrangling. In this training tutorial, we’ll make use of an example dataset that needs to be wrangled. The dataset contains concentration values for molecules that were measured using protein-based ELISA technologies. These molecules specifically span 17 sterols and cytokines, selected based upon their important roles in mediating biological responses. These measures were derived from human serum samples. Demographic information also exists for each subject. The following steps detailed in this training module are by no means exhaustive! Further resources are provided at the end. This module provides example steps that are helpful when wrangling your data in Excel. Datasets often come in many different formats from our wet bench colleagues, therefore some steps will likely need to be added, removed, or amended depending on your specific data. Save a Copy of the Soon-To-Be Organized and Cleaned Dataset as a New File Open Microsoft Excel and prior to ANY edits, click “File” –&gt; “Save As” to save a new version of the file that can serve as the cleaned version of the data. This is very important for file tracking purposes, and can help in the instance that the original version needs to be referred back to (e.g., if data are accidentally deleted or modified during downstream steps). The file needs to be named something indicative of the data it contains followed by the current date (e.g., “Allostatic Mediator Data_061622”). The title should be succinct and descriptive. It is okay to use dashes or underscores in the name of the title. Do not include special characters, such as $, #, @, !, %, &amp;, *, (, ), and +. Special characters tend to generate errors on local hard drives when syncing to cloud-based servers, and they are difficult to upload into programming software. Let’s first view what the dataset currently looks like: Helpful Excel Keyboard Shortcuts The following keyboard shortcuts can help you work more efficiently in Excel: Move to the last cell in use on the sheet Control + Fn + Right arrow key (Mac users) Control + End (PC users) Move to the beginning of the sheet Control + Fn + Left arrow key, then same Control + Fn + Up arrow key (Mac users) Control + Home (PC users) Highlight and grab all data Click on the first cell in the upper left hand corner then click and hold Shift + Command + Down arrow key + Right arrow key (Mac users) Shift + Command + Down arrow key + Right arrow key (PC users) Note: This only works if there are no cells with missing information or gaps in the columns/rows used to define the peripheral area. For more available shortcuts on various operating systems click here. Remove Extraneous White Space Before we can begin organizing the data, we need to remove the entirely blank rows of cells. This reduces the file size and allows for the use of the filter function in Excel, as well as other organizing functions, which will be used in the next few steps. This step also makes the data look more tidy and amenable to import for coding purposes. Excel Trick #1: Select all lines that need to be removed and press Control + minus key for Mac and PC users. (Note that there are other ways to do this for larger datasets, but this works fine for this small example.) Excel Trick #2: An easier way to remove blank rows and cells for larger datasets, includes clicking “Find &amp; Select”–&gt; “Special” –&gt; “Blanks” –&gt; click “OK” to select all blank rows and cells. Click “Delete” within the home tab –&gt; “Delete sheet rows”. After removing the blank rows, the file should look like the screenshot below. Replace Missing Data with “NA” There are many ways missing data can be encoded in datasets. This includes values like “blank”, “N/A”, “NA”, or leaving a cell blank. Replacing all missing values with “NA” values is done for 2 reasons: To confirm that the data is indeed missing R reads in “NA” values as missing values To check for missing values, the filter function can be used on each column and only select cells with missing values. You may need to scroll to the bottom of the filter pop up window for numerical data. Enter “NA” into the cell of the filtered column. Double click the bottom right corner of the cell to copy the “NA” down the rest of the column. There was no missing data in this dataset, so this step can be skipped. Create a Metadata Tab Metadata explains what each column represents in the dataset. Metadata is now a required component of data sharing, so it is best to initiate this process prior to data analysis. Ideally, this information is filled in by the scientist(s) who generated the data. Create a new tab (preferably as the first tab) and label it “XXXXX_METADATA” (ie., “Allostatic_METADATA”) Then relabel the original data tab as “XXXX_DATA” (ie., “Allostatic_DATA). Within the metadata tab, create three columns: the first, “Column Identifier”, contains each of the column names found in the data tab; the second, “Code”, contains the individual variable/ abbreviation for each column identifier; the third, “Description” contains additional information and definitions for abbreviations. Abbreviate and Capitalize Categorical Data Categorical data are easier to handle in programming languages when they are capitalized and abbreviated. It also helps reduce typos and potential typing mistakes within your script. For this dataset, the following variables were edited: Group “control” became “NS” for non-smoker “smoker” became “CS” for cigarette smoker Sex “f” became “F” for female “m” became “M” for male Race “AA” became “B” for Black “White” became “W” for White Excel Trick: To change cells that contain the same data simultaneously, navigate to “Edit”, click “Find”, and then “Replace”. Once the categorical data have been abbreviated, add those abbreviations to the metadata and describe what they symbolize. Alphabetize (Sort) the Data by the Categorical Variable of Interest For this dataset, we will sort by the column “Group”. This organizes the data and sets it up for the next step. Highlight all the column headers. Click on the “Sort &amp; Filter” button and click “Filter”. Click on the arrow on cell that contains the column name “Group” and click “Ascending”. Create a New Subject Number Column Analysis-specific subjects are created to give an ordinal subject number to each subject, which allows the scientist to easily identify the number of subjects. In addition, these new ordinal subject numbers will be used to create a subject identifier that combines both a subject’s group and subject number that is helpful for downstream visualization analyses. Relabel the subject number/identifier column as “Original_Subject_Number” and create an ordinal subject number column labeled “Subject_Number”. R reads in spaces between words as periods, therefore it’s common practice to replace spaces with underscores when doing data analysis in R. Avoid using dashes in column names or anywhere else in the dataset. Remove Special Symbols and Dashes Programming languages, in general, do not operate well with special symbols and dashes, particularly when included in column identifiers. For this reason, it is best to remove these while cleaning up your data, prior to importing it into R or your preferred programming software. In this case, this dataset contains dashes and Greek letters within some of the column header identifiers. Here, it is beneficial to remove these dashes (e.g., change IL-10 to IL10) and replace the Greek letters with first letter of the word in English (e.g., change TNF-\\(\\alpha\\) to TNFa). Bold all Column Names and Center all Data These data will likely be shared with collaborators, uploaded onto data deposition websites, and used as supporting information in published manuscripts. For these purposes, it is nice to format data in Excel such that it is visually appealing and easy to digest. For example, here, it is nice to bold column identifiers and center the data, as shown below: Create a Subject Identifier Column The subject identifier column labeled, “Group_Subject_No”, combines the subject number with the variable of interest (ie. Group for this dataset). This is useful for analyses to identify outliers by the subject number and the group. Insert 2 additional columns where the current “Sex” column is. To combine values from two different columns, type “=CONCAT(D1,” _ “,C1)” in the first cell in the first column inserted. Double click the right corner of the cell for the formula to be copied to last row in the dataset. Copy the entire column and paste only the values in the second column by navigating to the drop down arrow next to “Paste” and click “Paste Values”. Label the second column “Group_Subject_No” and delete the first column. Separate Subject Demographic Data from Experimental Measurements This example dataset is very small, so the demographic data (e.g., sex, race, age) was kept within the same file as the experimentally measured molecules. Though in larger datasets (e.g., genome-wide data, exposomic data, etc), it is often beneficial to separate the demographic data into one file that can be labeled according to the following format: “XXX_Subject_Info_061622” (ie. “Allostatic_Subject_Info_061622”). This step was not completed for this current data, since it had a smaller size and the downstream analyses were simple. Convert Data from Wide to Long Format A wide format contains values that DO NOT repeat the subject identifier column. For this dataset, each subject has one row containing all of its data, therefore the subject identifier occurs once in the dataset. Wide Format A long format contains values that DO repeat the subject identifier column. For this dataset, that means a new column was created entitled “Variable” containing all the mediator names and a column entitled “Value” containing all their corresponding values. In the screenshot, an additional column, “Category”, was added to help with the categorization of mediators in R analyses. Long Format The reason a long format is preferred is because it makes visualizations and statistical analyses more efficient in R. In the long format, we were able to add a column entitled “Category” to categorize the mediators into “AL Biomarker” or “Cytokine” allowing us to more easily subset the mediators in R. Read more about wide and long formats here. To convert the data from a wide to long format, follow the steps below: Pivoting Data from a Wide to Long Format To do this, a power query in Excel will be used. Note: If you are working on a Mac, you will need to have at least Excel 2016 installed to follow this tutorial, as Power Query is not avaialble for earlier versions. Add-ins are available for Windows users. See this link for more details. Start by copying all of the data, including the column titles. (Hint: Try using the keyboard shortcut mentioned above.) Click the tab at the top that says “Data”. Then click “Get Data (Power Query)” at the far left. It will ask you to choose a data source. Click “Blank table” in the bottom row. Paste the data into the table. (Hint: Use the shortcut Ctrl + “v”). At this point, your screen should look like the screenshot below. Click “Use first row as headers” and then click “Next” in the bottom right hand corner. Select all the columns with biomarker names. That should be the column “Cortisol” through the end. Click the “Transform” button in the upper left hand corner. Then click “Unpivot columns” in the middle of the pane. The final result should look like the sceenshot below with all the biomarkers now in one column entitled “Attribute” and their corresponding values in another column entitled “Value”. To save this, go back to the “Home” tab and click “Close &amp; load”. You should see something similar to the screenshot below. In the upper right with all the shaded tables (within the “Table” tab), click the arrow to the left of the green table until you see one with no shading. Then click the table with no colors. Click “Convert to Range” within the “Table” tab. This removes the power query capabilities, so that the data is a regular excel sheet. Now the “Category” column can be created to identify the types of biomarkers in the dataset. The allostatic load (AL) biomarkers denoted in the “Category” column include the variables Cortisol, CRP, Fibrinogen, Hba1c, HDL, and Noradrenaline. The rest of the variables were labeled as cytokines. Additionally, we can make this data more closely resemble the final long format screenshot by bolding the headers, centering all the data, etc. We have successfully wrangled our data and the final dataset now looks like this: Generating Summary-Level Statistics with Pivot Tables A PivotTable is a tool in Excel used to summarize numerical data. It’s called a pivot table, because it pivots or changes how the data is displayed to make statistical inferences. This can be useful for generating initial summary-level statistics to guage the distribution of data. To create a PivotTable, start by selecting all of the data. (Hint: Try using the keyboard shortcut mentioned above.) Click “Insert” tab on the upper left-hand side, click “PivotTable”, and click “OK”. The new PivotTable should be available in a new sheet as seen in the screenshot below. A PivotTable will be constructed based on the column headers that can be dragged into the PivotTable fields located on the right-hand side. For example, what if we were interested in determining if there were differences in average expression between non-smokers and cigarette smokers in each category of biomarkers? As seen below, drag the “Group” variable under the “Rows” field and drag the “Value” variable under the “Values” field. Notice that it automatically calculates the sum of the expression values for each group. To change the function to average, click the “i” icon and select “Average”. The output should mirror what’s below with non-smokers having an average expression that’s more than double that of cigarette smokers. Excel vs. R: Which Should You Use? For the most part, it’s better to perform final analyses in R (or another programming language) rather than Excel for the following reasons… R clearly shows the code (instructions), which makes editing, interpretability, and sharing easier. This makes analyses more reproducible and can save time. R has packages that makes more complex analyses possible (i.e., machine learning and heatmaps) that aren’t available in Excel. R can handle larger data sets. R can compute and process data faster. However, Excel is still a software that has many benefits for running analyses including… Excel is user-friendly and most people have experience in navigating the software at a basic level. Excel can be faster for rudimentary statistical analyses and visualizations. Depending on each scientist’s skill-level and the complexity of the analysis, Excel or R could be beneficial. Concluding Remarks In summary, this training module highlights the importance of data wrangling and how to do so in Microsoft Excel for downstream analyses. Concepts discussed include helpful Excel features like power queries and pivot tables and when to use Microsoft Excel vs. R. Additional Resources Data wrangling in Excel can be expedited with knowledge of useful features and functions to format data. Check out the resources below for additional information on Excel tricks. Data Analysis in Excel Excel Spreesheet Hacks Excel for Beginners Test Your Knowledge Try wrangling the “Module1_4_TYKInput.xlsx” to mimic the cleaned versions of the data found in “Module1_4_TYKSolution.xlsx”. This dataset includes sterol and cytokine concentration levels extracted from induced sputum samples collected after ozone exposure. After wrangling, you should end up with a sheet for subject information and a sheet for experimental data. Using the a PivotTable on the cleaned dataset, find the standard deviation of each cytokine variable stratified by the disease status. "],["downloading-and-programming-in-r.html", "Downloading and Programming in R Introduction to Training Module General Introduction and Installation of R and RStudio Introduction to R Packages Scripting Basics Code Troubleshooting Concluding Remarks", " Downloading and Programming in R This training module was developed by Kyle Roell, Elise Hickman, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module In this training module, we will provide a brief introduction of: R R Studio Packages in R Scripting basics Code troubleshooting General Introduction and Installation of R and RStudio What is R? R is a programming language. Computer script (lines of code) can be used to increase data analysis reproducibility, transparency, and methods sharing, and is becoming increasingly incorporated into exposure science, toxicology, and environmental health research. One of the most commonly used coding languages in the field of environmental health science is the R language. Some advantages of using R include the following: Free, open-source programming language that is licensed under the Free Software Foundation’s GNU General Public License Can be run across all major platforms and operating systems, including Unix, Windows, and MacOS Publicly available packages help you carry out analyses efficiently (without you having to code for everything yourself) Large, diverse collection of packages Comprehensive documentation When code is efficiently tracked during development/execution, it promotes reproducible analyses Because of these advantages, R has emerged as an avenue for world-wide collaboration in data science. Other commonly implemented scripting languages in the field of environmental health research include Python and SAS, among others; and these training tutorials focus on R as an important introductory-level example that also houses many relevant packages and example datasets as further described throughout TAME. Downloading and Installing R To download R, first navigate to https://cran.rstudio.com/ and download the .pkg file for your operating system. Install this file according to your computer’s typical program installation steps. What is RStudio? RStudio is an Integrated Development Environment (IDE) for R, which makes it more ‘user friendly’ when developing and using R script. It is a desktop application that can be downloaded for free, online. Downloading and Installing RStudio To download RStudio: Navigate to: https://posit.co/download/rstudio-desktop/ Scroll down and select “Download RStudio” Install according to your computer’s typical program installation steps RStudio Orientation Here is a screenshot demonstrating what the RStudio desktop app looks like: The default RStudio layout has four main panes (numbered above in the blue boxes): Source Editor: allows you to open and edit script files and view data. Console: where you can type code that will execute immediately when you press enter/return. This is also where code from script files will appear when you run the code. Environment: shows you the objects in your environment. Viewer: has a number of useful tabs, including: Files: a file manager that allows you to navigate similar to Finder or File Explorer Plots: where plots you generate by executing code will appear Packages: shows you packages that are loaded (checked) and those that can be loaded (unchecked) Help: where help pages will appear for packages and functions (see below for further instructions on the help option) Under “Tools” → “Global Options,” RStudio panes can be customized to appear in different configurations or with different color themes. A number of other options can also be changed. For example, you can choose to have colors highlighted the color they appear or rainbow colored parentheses that can help you visualize nested code. Introduction to R Packages One of the major benefits to coding in the R language is access to the continually expanding resource of thousands of user-developed packages. Packages represent compilations of code and functions fitted for a specialized focus or purpose. These are often written by R users and submitted to the CRAN, or another host such as BioConductor or Github. Packages aid in improved data analyses and methods sharing. Packages have varying utilities, spanning basic organization and manipulation of data, visualizing data, and more advanced approaches to parse and analyze data, with examples included in all of the proceeding training modules. Examples of some common packages that we’ll be using throughout these training modules include the following: tidyverse: A collection of open source R packages that share an underlying design philosophy, grammar, and data structures of tidy data. For more information on the tidyverse package, see its associated CRAN webpage, primary webpage, and peer-reviewed article released in 2018. ggplot2: A system for creating graphics. Users provide the data and tell R what type of graph to use, how to map variables to aesthetics (elements of the graph), and additional stylistic elements to include in the graph. For more information on the ggplot2 package, see its associated CRAN webpage and R Documentation. More information on these packages, as well as many others, is included throughout TAME training modules. Downloading/Installing R Packages R packages often do not need to be downloaded from a website. Instead, you can install packages and load them through running script in R. Note that you only need to install packages one time, but packages must be loaded each time you start a new R session. # Install the package install.packages(“tidyverse”) # Load the package for use library(tidyverse) Many packages also exist as part of the baseline configuration of an R working environment, and do not require manual loading each time you launch R. These include the following packages: datasets graphics methods stats utils You can learn more about a function by typing one question mark before the name of the function, which will bring up documentation in the Help tab of the Viewer window. Importantly, this documentation includes a description of the different arguments that can be passed to the function and examples for how to use the function. ?install.packages ## starting httpd help server ... done You can learn more about a package by typing two question marks before the name of the package. This will bring up vingettes and help pages associated with that package. ??tidyverse Scripting Basics Data Types Before writing any script, let’s first review different data types in R. Data types are what they imply – the type of data you are handling. It is important to understand data types because functions often require a specific data type as input. R has 5 basic data types: Logical (e.g., TRUE or FALSE) Integer (e.g., 1, 2, 3) Numeric (real or decimal) Character (e.g., ”apple”) Complex (e.g., 1 + 0i) Numeric variables are often stored as “double” values (sometimes shown as &lt; dbl &gt;), or a decimal type with at least two decimal places. Character variables can also be stored as factors, which are data structures that are implemented to store categorical data in a specific order (also known as levels). Data are stored in data structures. There are many different data structures in R. Some packages even implement unique data structures. The most common data structures are: Vectors: also known as an atomic vector, can contain characters, logical values, integers, or numeric values (but all elements must be the same data type). Matrices: a vector with multiple dimensions. Elements must still be all the same data type. Data frames: similar to a matrix but can contain different data types and additional attributes such as row names (and is one of the most common data structures in environmental health research). Tibbles are a stricter type of data frame implemented in the tidyverse package. Lists: a special type of vector that acts as a container – other data structures can be stored within the list, and lists can contain other lists. Lists can contain elements that are different data structures. Figure 1: Created with BioRender.com Writing Script R code is written line by line. It may take just one line or many lines of code for one step to be executed, depending on the number of arguments to the function you are using. R code is executed (run) by selecting the line(s) of code to run and pressing return/enter (or a keyboard shortcut), or by clicking “Run” in the upper right corner of the script. A very simple example of running code is as follows: 3 + 4 ## [1] 7 We can see that when we ran our code, the answer was returned. But what if we want to store that answer? We can assign that number to a variable named x using the assignment operator &lt;-: x &lt;- 3 + 4 Then, if we run a line of code with our variable, we will get that value: x ## [1] 7 The assignment operator can also be used to assign values to any of the data structures discussed above, such as vectors and data frames, as shown here: # Creating a vector of values called my_values my_values &lt;- c(7, 3, 8, 9) # Viewing the vector my_values ## [1] 7 3 8 9 # Creating a data frame of values corresponding to colors my_df &lt;- data.frame(values = my_values, color = c(&quot;Blue&quot;, &quot;Red&quot;, &quot;Yellow&quot;, &quot;Purple&quot;)) # Viewing the data frame my_df ## values color ## 1 7 Blue ## 2 3 Red ## 3 8 Yellow ## 4 9 Purple Comments You may have noticed in the code chunks above that there were # followed by phrases describing the code. R allows for scripts to contain non-code elements, called comments, that will not be run or interpreted. Comments are useful to help make code more interpretable for others or to add reminders of what and why parts of code may have been written. To make a comment, simply use a # followed by the comment. A # only comments out a single line of code. In other words, only that line will be commented and therefore not be run, but lines directly above/below it will still be run: # This is an R comment! For more on comments, see TAME 2.0 Module 2.2 Coding Best Practices. Autofilling RStudio will autofill function names and object names as you type, which can save a lot of time. When you are typing a variable or function name, you can press tab while typing. RStudio will look for variables or functions that match the first few letters you’ve typed. If multiple matches are found, RStudio will provide you with a drop down list to select from, which may be useful when searching through newly installed packages or trying to quickly type variable names in an R script. For example, let’s say we instead named our example data frame something much longer, and we had two data frames with similar names. If we start typing in my_ and pause our typing, all of the objects that start with that name will appear as options in a list. To select which one to autofill, navigate down the list and click return/enter. my_df_with_really_long_name &lt;- data.frame(values = my_values, color = c(&quot;Blue&quot;, &quot;Red&quot;, &quot;Yellow&quot;, &quot;Purple&quot;)) my_df_with_really_long_name_2 &lt;- data.frame(values = my_values, color = c(&quot;Green&quot;, &quot;Teal&quot;, &quot;Magenta&quot;, &quot;Orange&quot;)) Finding and Setting Your Working Directory Another step that is commonly done at the very beginning of your code is setting your working direction. This tells your computer where to look for files that you want to import and where to deposit output files produced during your scripted activities. To view your current working directory, run the following: getwd() To set or change the location of your working directory, run the following: setwd(&quot;/file path to where your input files are&quot;) Note that macOS file paths use / to separate folders, whereas PC file paths use \\. You can easily find the file path to your desired working directory by navigating to “Session”, then “Set Working Directory”, and “Choose Directory”: In the popup box, navigate to the folder you want to set as your working directory and click “Open.” Look in the R console, which will now contain a line of code with setwd() containing your file path. You can copy this line of code to the top of your script for future use. Alternatively, you can navigate to the folder you want in Finder or File Explorer and right click to see the file path. Within your working directory, you can make sub-folders to keep your analyses organized. Here is an example folder hierarchy: How you set up your folder hierarchy is highly dependent on your specific analysis and coding style. However, we recommend that you: Name your script something concise, but descriptive (no acronyms) Consider using dates when appropriate Separate your analysis into logical sections so that script doesn’t get too long or hard to follow Revisit and adapt your organization as the project evolves! Archive old code so you can revisit it A Quick Note About Projects Creating projects allows you to store your progress (open script, global environment) for one project in an R Project File. This facilitates quick transitions between multiple projects. Find detailed information about how to set up projects here. Importing Files After setting the working directory, you can import and export files using various functions based on the type of file being imported or exported. Often, it is easiest to import data into R that are in a comma separated value / comma delimited file (.csv) or tab / text delimited file (.txt). Other datatypes such as SAS data files or large .csv files may require different functions to be more efficiently read in, and some of these file formats will be discussed in future modules. Files can also be imported and exported from Excel using the openxlsx package. Below, we will demonstrate how to read in .csv and .txt files: # Read in the .csv data that&#39;s located in our working directory csv.dataset &lt;- read.csv(&quot;Module2_1_Input/Module2_1_InputData1.csv&quot;) # Read in the .txt data txt.dataset &lt;- read.table(&quot;Module2_1_Input/Module2_1_InputData1.txt&quot;) These datasets now appear as saved dataframes (“csv.dataset” and “txt.dataset”) in our working environment. Viewing Data After data have been loaded into R, or created within R, you will likely want to view what these datasets look like. Datasets can be viewed in their entirety, or datasets can be subsetted to quickly look at part of the data. Here’s some example script to view just the beginnings of a dataframe using the head() function: head(csv.dataset) ## Sample Var1 Var2 Var3 ## 1 sample1 1 2 1 ## 2 sample2 2 4 4 ## 3 sample3 3 6 9 ## 4 sample4 4 8 16 ## 5 sample5 5 10 25 Here, you can see that this automatically brings up a view of the first five rows of the dataframe. Another way to view the first five rows of a dataframe is to run the following: csv.dataset[1:5,] ## Sample Var1 Var2 Var3 ## 1 sample1 1 2 1 ## 2 sample2 2 4 4 ## 3 sample3 3 6 9 ## 4 sample4 4 8 16 ## 5 sample5 5 10 25 This brings us to an important concept - indexing! Brackets are used in R to index. Within the bracket, the first argument represents the row numbers, and the second argument represents the column numbers. A colon between two numbers means to select all of the columns in between the left and right numbers. The above line of code told R to select rows 1 to 5, and, by leaving the column argument blank, all of the columns. Expanding on this, to view the first 5 rows and 2 columns, we can run the following: csv.dataset[1:5, 1:2] ## Sample Var1 ## 1 sample1 1 ## 2 sample2 2 ## 3 sample3 3 ## 4 sample4 4 ## 5 sample5 5 For another example: What if we want to only view the first and third row, and first and fourth column? We can use a vector within the index to do this: csv.dataset[c(1, 3), c(1, 4)] ## Sample Var3 ## 1 sample1 1 ## 3 sample3 9 To view the entire dataset, use the View() function: View(csv.dataset) Another way to view a dataset is to just click on the name of the data in the environment pane. The view window will pop up in the same way that it did with the View() function. Determining Data Structures and Data Types As discussed above, there are a number of different data structures and types that can be used in R. Here, we will demonstrate functions that can be used to identify data structures and types within R objects. The glimpse() function, which is part of the tidyverse package, is helpful because it allows us to see an overview of our column names and the types of data contained within those columns. # Load tidyverse package library(tidyverse) glimpse(csv.dataset) ## Rows: 5 ## Columns: 4 ## $ Sample &lt;chr&gt; &quot;sample1&quot;, &quot;sample2&quot;, &quot;sample3&quot;, &quot;sample4&quot;, &quot;sample5&quot; ## $ Var1 &lt;int&gt; 1, 2, 3, 4, 5 ## $ Var2 &lt;int&gt; 2, 4, 6, 8, 10 ## $ Var3 &lt;int&gt; 1, 4, 9, 16, 25 Here, we see that our Sample column is a character column, while the rest are integers. The class() function is also helpful for understanding objects in our global environment: # What class (data structure) is our object? class(csv.dataset) ## [1] &quot;data.frame&quot; # What class (data type) is a specific column in our data? class(csv.dataset$Sample) ## [1] &quot;character&quot; These functions are particularly helpful when introducing new functions or troubleshooting code because functions often require input data to be a specific structure or data type. Exporting Data Now that we have these datasets saved as dataframes, we can use these as examples to export data files from the R environment back into our local directory. There are many ways to export data in R. Data can be written out into a .csv file, tab delimited .txt file, or RData file, for example. There are also many functions within packages that write out specific datasets generated by that package. To write out to a .csv file: write.csv(csv.dataset, &quot;Module2_1_SameCSVFileNowOut.csv&quot;) To write out a .txt tab delimited file: write.table(txt.dataset, &quot;Module2_1_SameTXTFileNowOut.txt&quot;) R also allows objects to be saved in RData files. These files can be read into R, as well, and will load the object into the current workspace. Entire workspaces are also able to be saved in RData files, such that when you open an RData file, your script and Global Environment will be just as you saved them. Below includes example code to carry out these tasks, and though these files are not provided, they are just example code for future reference. # Read in saved single R data object r.obj = readRDS(&quot;data.rds&quot;) # Write single R object to file saveRDS(object, &quot;single_object.rds&quot;) # Read in multiple saved R objects load(&quot;multiple_data.RData&quot;) # Save multiple R objects save(object1, object2, &quot;multiple_objects.RData&quot;) # Save entire workspace save.image(&quot;entire_workspace.RData&quot;) # Load entire workspace load(&quot;entire_workspace.RData&quot;) Code Troubleshooting Learning how to code is an iterative, exploratory process. The secret to coding is to… Make sure to include “R” and the package and/or function name in your search. Don’t be afraid to try out different solutions until you find one that works for you, but also know when it is time to ask for help. For example, when you have tried solutions available on forums, but they aren’t working for you, or you know a colleague has already spent a significant amount of time developing code for this specific task. Note that when reading question/answer forums, make sure to look at how recent a post is, as packages are updated frequently, and old answers may or may not work. Some common reasons that code doesn’t work and potential solutions to these problems include: Two packages are loaded that have functions with the same name, and the default function is not the one you are intending to run. Solutions: specify the package that you want the function to be called from each time you use it (e.g., dplyr::select()) or re-assign that function at the beginning of your script (e.g., select &lt;- dplyr::select) Your data object is the wrong input type (is a data frame and needs to be a matrix, is character but needs to be numeric) Solution: double check the documentation (?functionname) for the input/variable type needed You accidentally wrote over your data frame or variable with another section of code Solution: re-run your code from the beginning, checking that your input is in the correct format There is a bug in the function/package you are trying to use (this is most common after packages are updated or after you update your version of R) Solution: post an issue on GitHub for that package (or StackOverflow if there is not a GitHub) using a reproducible example There are a number of forums that can be extremely helpful when troubleshooting your code, such as: Stack Overflow: one of the most common forums to post questions related to coding and will often be the first few links in a Google search about any code troubleshooting. It is free to make an account, which allows you to post and answer questions. Cross Validated: a forum focused on statistics, including machine learning, data analysis, data mining, and data visualization, and is best for conceptual questions related to how statistical tests are carried out, when to use specific tests, and how to interpret tests (rather than code execution questions, which are more appropriate to post on Stack Overflow). BioConductor Forum: provides a platform for specific coding and conceptual questions about BioConductor packages. GitHub: can also be used to create posts about specific issues/bugs for functions within that package. Before you post a question, make sure you have thoroughly explored answers to existing similar questions and are able to explain in your question why those haven’t worked for you. You will also need to provide a reproducible example of your error or question, meaning that you provide all information (input data, packages, code) needed such that others can reproduce your exact issues. While demonstrating a reproducible example is beyond the scope of this module, see the below links and packages for help getting started: Detailed step-by-step guides for how to make reproducible examples: How to Reprex by Ariel Muldoon What’s a reproducible example (reprex) and how do I create one? Helpful packages: reprex: part of tidyverse, useful for preparing reproducible code for posting to forums. datapasta: useful for creating code you can copy and paste that creates a new data frame as a subset of your original data. Concluding Remarks Together, this training module provides introductory level information on installing and loading packages in R, scripting basics, importing and exporting data, and code troubleshooting. Additional Resources Coursera Stack Overflow How to Learn R R for Data Science Test Your Knowledge Install R and RStudio on your computer. Launch RStudio and explore installing packages (e.g., tidyverse) and understanding data types using the built-in datasets in R. Make a vector of the letters A-E. Make a data frame of the letters A-E in one column and their corresponding number in the alphabet order in the second column (e.g., A corresponds with 1). "],["coding-best-practices.html", "Coding “Best” Practices Introduction to Training Module Scripting File Types Script Headers and Annotation Coding Style Script Organization Concluding Remarks", " Coding “Best” Practices This training module was developed by Kyle Roell, Alexis Payton, Elise Hickman, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module In this training module, we will be going over coding “best” practices. The reason we put “best” in quotes is because these practices are what we currently consider best or better, though everyone has different coding styles, annotation styles, etc that also change over time. Here, we hope to give you a sense of what we do when coding, why we do it, and why we think it is important. We will also be pointing out other guides to style, annotations, and best practices that we suggest implementing into your own coding. Some of the questions we hope to answer in this module are: What type of scripting file should I use? What should I name my script? What should I put at the top of every script and why is it important? How should I annotate my code? Why are annotations important? How do I implement these coding practices into my own code? Where can I find other resources to help with coding best practices? In the following sections, we will be addressing these questions. Keep in mind that the advice and suggestions in this section are just that: advice and suggestions. So please take them into consideration and integrate them into your own coding style as appropriate. Scripting File Types Two of the most common scripting file types applicable to the R language are .R (normal R files) and .Rmd (R Markdown). Normal R files appear as plain text and can be used for running any normal R code. R Markdown files are used for more intensive documentation of code and allow for a combination of code, non-code text explaining the code, and viewing of code output, tables, and figures that are rendered together into an output file (typically .html, although other formats such as .pdf are also offered). For example, TAME is coded using R Markdown, which allows us to include blocks of non-code text, hyperlinks, annotated code, schematics, and output figures all in one place. We highly encourage the use of R Markdown as the default scripting file type for R-based projects because it produces a polished final document that is easy for others to follow, whereas .R files are more appropriate for short, one-off analyses and writing in-depth functions and packages. However, code executed in normal .R files and R Markdown will produce the same results, and ultimately, which file type to use is personal preference. See below for screenshots that demonstrate some of the stylistic differences between .R, .Rmd, and .Rmd knitted to HTML format: If you are interested in learning more about the basic features of R Markdown and how to use them, see the following resources: RStudio introduction to R Markdown R Markdown Cheat Sheet Bookdown R Markdown guide Including external images in R Markdown with knitr Interactive plots with plotly Interactive data tables with DT Naming the Script File The first thing we need to talk about, which is sometimes overlooked in the discussion of coding practices, is script file naming conventions and high level descriptive headers within a script. It is important to remember to name your code something concise, but descriptive. You want to be able to easily recognize what the script is for and does without a cumbersome, lengthy title. Some tips for naming conventions: Be concise, but descriptive Use dates when appropriate Avoid special characters Use full words if possible, avoiding non-standard acronyms Keep in mind that each script should have a clear purpose within a given project. And, it is sometimes necessary, and often common, to have multiple scripts within one project that all pertain to different parts of the analysis. For example, it may be appropriate to have one script for data cleaning and pre-processing and another script for analyzing data. When scripting an analysis with multiple sub-analyses, some prefer to keep code for each sub-analysis separate (e.g., one file for an ANOVA and one file for a k-means analysis on the same data input), while others prefer to have longer code files with more subsections. Whichever method you choose, we recommend maintaining clear documentation that indicates locations for input and output files for each sub-analysis (e.g., whether global environment objects or output files from a previous script are needed to run the current script). Script Headers and Annotation Script Header Once your script is created and named, it is generally recommended to include a header at the top of the script. The script header can be used for describing: Title of Script - This can be a longer or more readable name than script file name. Author(s) - Who wrote the script? Date - When was the script developed? Description - Provides a more detailed description of the purpose of the script and any notes or special considerations for this particular script. In R, it is common to include multiple #, the comment operator, or a # followed by another special character, to start and end a block of coding annotation or the script header. An example of this in an .R file is shown below: ######################################################################## ######################################################################## ### Script Longer Title ### ### Description of what this script does! ### Also can include special notes or anything else here. ### ### Created by: Kyle Roell and Julia Rager ### Last updated: 01 May 2023 ######################################################################## ######################################################################## This block of comment operators is common in .R but not .Rmd files because .Rmd files have their own specific type of header, known as the YAML, which contains the title, author, date, and formatting outputs for the .Rmd file: We will now review how annotations within the script itself can make a huge difference in understanding the code within. Annotations Before we review coding style considerations, it is important to address code annotating. So, what are annotations and why are they important? Annotations are notes embedded within your code as comments that will not be run. The beauty of annotating your code is that not only others, but future you, will be able to read through and better understand what a particular piece of code does. We suggest annotating your code while you write it and incorporate a lot of description. While not every single line needs an annotation, or a very detailed one, it is helpful to provide comments and annotation as much as you can while maintaining feasibility. General annotation style In general, annotations will be short sentences that describe what your code does or why you are executing that specific code. This can be helpful when you are defining a covariate a specific way, performing a specific analytical technique, or just generally explaining why you are doing what you’re doing. # Performing logistic regression to assess association between xyz and abc # Regression confounders: V1, V2, V3 ... xyz.regression.output = glm(xyz ~ abc + V1 + V2 + V3, family=binomial(), data=example.data) Mid-script headings Another common approach to annotations is to use mid-script type headings to separate out the script into various sections. For example, you might want to create distinct sections for “Loading Packages, Data, and Setup”, “Covariate Definition”, “Correlation Analysis”, “Regression Analysis”, etc. This can help you, and others, reading your script, to navigate the script more easily. It also can be more visually pleasing to see the script split up into multiple sections as opposed to one giant chunk of code interspersed with comments. Similar to above, the following example is specific to .R files. For .Rmd files, sub headers can be created by increasing the number of # before the header. ########################################################################### ########################################################################### ### ### Regression Analyses ### ### You can even add some descriptions or notes here about this section! ### ########################################################################### # Performing logistic regression to assess association between xyz and abc # Regression confounders: V1, V2, V3 ... xyz.regression.output = glm(xyz ~ abc + V1 + V2 + V3, family=binomial(), data=example.data) General tips for annotations: Make comments that are useful and meaningful You don’t need to comment every single line In general, you probably won’t over-comment your script, so more is generally better That being said, don’t write super long paragraphs every few lines Split up your script into various sections using mid-script headings when appropriate Quick, short comments and annotations While it is important to provide descriptive annotations, not every one needs to be a sentence or longer. As stated previously, it is not necessary to comment every single line. Here is an example of very brief commenting: # Loading necessary packages library(ggplot2) # Plotting package In the example above, we can see that these short comments clearly convey what the script does – load the necessary package and indicate what the package is needed for. Short, one line annotations can also be placed after lines to clarify that specific line or within the larger mid-script headings to split up these larger sections of code. Coding Style Coding style is often a contentious topic! There are MANY styles of coding, and no two coders have the same exact style, even if they are following the same reference. Here, we will provide some guides to coding style and go over some of the basic, general tips for making your code readable and efficient. Here is an example showing how you can use spacing to align variable assignment: # Example of using spacing for alignment of variable assignment Longer_variable_name_x = 1 Short_name_y = 2 Note that guides will suggest you use &lt;- as the assignment operator. However, for most situations, &lt;- and = will do the same thing. For spacing around certain symbols and operators: Include a space after if, before parenthesis Include a space on either side of symbols such as &lt; The first (opening) curly brace should not be on its own line, but the second (closing) should # Example of poor style if(Longer_variable_name_x &lt;Short_name_y) {Short_name_y = 0} # Example of better style if (Longer_variable_name_x &lt; Short_name_y) { Short_name_y = 0 } Summary of general tips for coding style: Variable names Make them intuitive, short, but descriptive Use the same convention throughout (ex: separating words with . or _) Data names Use _datatype (e.g., cytokines_df for a dataframe data type) following a short, concise name for your data so that readers know what type of input is being used AND/OR Clearly define important data frames or other input in your code comments (e.g., “cytokines” refers to a data frame containing all of the cytokine data from this experiment) Separate long lines onto two or more lines (typically for loops or functions) Use &lt;- for assignment operator Using spacing appropriately for readability Alignment of lines After certain keywords when appropriate Be consistent throughout Example: if you use Tidyverse conventions, continue to use it throughout your script Try to make your code as readable as possible Common style guides: Google’s Style Guide Tidyverse Style Guide R Blogger’s Best Coding Practices Script Organization Lastly, it is important to note that organizing your script efficiently can help with readability as well. In general, as stated before, the beginning of your script should contain some sort of header lines to describe the script. The basic ordering we suggest for most scripts is: Header section Loading libraries and data Function definitions (if any user defined functions exist) Data and variables manipulation Analyses While following this exact organization isn’t absolute, using this structure or something similar can greatly improve the clarity of your analyses and make them easier for others to follow. Concluding Remarks In this module, we demonstrate basic coding style and best practices. Please reference additional style guides (above) and create your own style from best practices that work for you. Aim to make your code understandable and readable, and consider what a future reader, including yourself, will need to understand why you wrote your code the way you did and how to apply your code to new analyses. Test Your Knowledge Using the input file provided (“Module2_2_TYKInput.R”): Convert the script and annotations into R Markdown format. Improve the organization, comments, and scripting to follow the coding best practices described in this module. List the changes you made at the bottom of the new R Markdown file. Notes on the starting code: This starting code uses dummy data to demonstrate how to make a graph in R that includes bars representing the mean, with standard deviation error bars overlaid. You don’t need to understand every step in the code to be able to improve the existing coding style! You can run each step of the code if needed to understand better what it does. "],["data-manipulation-and-reshaping.html", "Data Manipulation and Reshaping Introduction to Training Module Data Manipulation Using Base R Introduction to Tidyverse Concluding Remarks", " Data Manipulation and Reshaping This training module was developed by Kyle Roell, Alexis Payton, Elise Hickman, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module Data within the fields of exposure science, toxicology, and public health are very rarely prepared and ready for downstream statistical analyses and visualization code. The beginning of almost any scripted analysis includes important formatting steps that make the data easier to read and work with. This can be done in several ways, including: Base R operations and functions, or A collection of packages (and philosophy) known as The Tidyverse. In this training tutorial we will review some of the most common ways you can organize and manipulate data, including: Merging data Filtering and subsetting data Pivoting data wider and longer (also known as casting and melting) These approaches will first be demonstrated using the functions available in base R. Then, the exact same approaches will be demonstrated using the functions and syntax that are part of the Tidyverse package. We will demonstrate these data manipulation and organization methods using an environmentally relevant example data set from a human cohort. This dataset was generated by creating data distributions randomly pulled from our previously published cohorts, resulting in a unique data set for these training purposes. The dataset contains environmental exposure metrics from metal levels obtained using sources of drinking water and human urine samples and associated demographic data. Training Module’s Environmental Health Question This training module was specifically developed to answer the following environmental health question using data manipulation and reshaping approaches: What is the average urinary chromium concentration across different maternal education levels? We’ll use base R and Tidydverse to answer this question, but let’s start with Base R. Workspace Preparation and Data Import Set your working directory In preparation, first let’s set our working directory to the folder path that contains our input files: setwd(&quot;/file path to where your input files are&quot;) Note that macOS file paths use / as folder separators, and PC file paths use \\. Importing example datasets Next, let’s read in our example data sets: demographic_data &lt;- read.csv(&quot;Module2_3_Input/Module2_3_InputData1.csv&quot;) chemical_data &lt;- read.csv(&quot;Module2_3_Input/Module2_3_InputData2.csv&quot;) Viewing example datasets Let’s see what these datasets look like: dim(demographic_data) ## [1] 200 6 dim(chemical_data) ## [1] 200 7 The demographic data set includes 200 rows x 7 columns, while the chemical measurement data set includes 200 rows x 7 columns. We can preview the demographic data frame by using the head() function, which displays all the columns and the first 6 rows of a data frame: head(demographic_data) ## ID BMI MAge MEdu BW GA ## 1 1 27.7 22.99928 3 3180.058 34 ## 2 2 26.8 30.05142 3 3210.823 43 ## 3 3 33.2 28.04660 3 3311.551 40 ## 4 4 30.1 34.81796 3 3266.844 32 ## 5 5 37.4 42.68440 3 3664.088 35 ## 6 6 33.3 24.94960 3 3328.988 40 These demographic data are organized according to subject ID (first column) followed by the following subject information: ID: subject number BMI: body mass index MAge: maternal age in years MEdu: maternal education level; 1 = “less than high school”, 2 = “high school or some college”, 3 = “college or greater” BW: body weight in grams GA: gestational age in weeks We can also preview the chemical dataframe: head(chemical_data) ## ID DWAs DWCd DWCr UAs UCd UCr ## 1 1 6.426464 1.292941 51.67987 10.192695 0.7537104 42.60187 ## 2 2 7.832384 1.798535 50.10409 11.815088 0.9789506 41.30757 ## 3 3 7.516569 1.288461 48.74001 10.079057 0.1903262 36.47716 ## 4 4 5.906656 2.075259 50.92745 8.719123 0.9364825 42.47987 ## 5 5 7.181873 2.762643 55.16882 9.436559 1.4977829 47.78528 ## 6 6 9.723429 3.054057 51.14812 11.589403 1.6645837 38.26386 These chemical data are organized according to subject ID (first column), followed by measures of: DWAs: drinking water arsenic levels in µg/L DWCd: drinking water cadmium levels in µg/L DWCr: drinking water chromium levels in µg/L UAs: urinary arsenic levels in µg/L UCd: urinary cadmium levels in µg/L UCr: urinary chromium levels in µg/L Data Manipulation Using Base R Merging Data Using Base R Syntax Merging datasets represents the joining together of two or more datasets, using a common identifier (generally some sort of ID) to connect the rows. This is useful if you have multiple datasets describing different aspects of the study, different variables, or different measures across the same samples. Samples could correspond to the same study participants, animals, cell culture samples, environmental media samples, etc, depending on the study design. In the current example, we will be joining human demographic data and environmental metals exposure data collected from drinking water and human urine samples. Let’s start by merging the example demographic data with the chemical measurement data using the base R function merge(). To learn more about this function, you can type ?merge, which brings up helpful information in the R console. To merge these datasets with the merge function, use the following code. The by = argument specifies the column used to match the rows of data. full.data &lt;- merge(demographic_data, chemical_data, by = &quot;ID&quot;) dim(full.data) ## [1] 200 12 This merged dataframe contains 200 rows x 12 columns. Viewing this merged dataframe, we can see that the merge() function retained the first column in each original dataframe (ID), though did not replicate it since it was used as the identifier for merging. All other columns include their original data, just merged together by the IDs in the first column. head(full.data) ## ID BMI MAge MEdu BW GA DWAs DWCd DWCr UAs ## 1 1 27.7 22.99928 3 3180.058 34 6.426464 1.292941 51.67987 10.192695 ## 2 2 26.8 30.05142 3 3210.823 43 7.832384 1.798535 50.10409 11.815088 ## 3 3 33.2 28.04660 3 3311.551 40 7.516569 1.288461 48.74001 10.079057 ## 4 4 30.1 34.81796 3 3266.844 32 5.906656 2.075259 50.92745 8.719123 ## 5 5 37.4 42.68440 3 3664.088 35 7.181873 2.762643 55.16882 9.436559 ## 6 6 33.3 24.94960 3 3328.988 40 9.723429 3.054057 51.14812 11.589403 ## UCd UCr ## 1 0.7537104 42.60187 ## 2 0.9789506 41.30757 ## 3 0.1903262 36.47716 ## 4 0.9364825 42.47987 ## 5 1.4977829 47.78528 ## 6 1.6645837 38.26386 These datasets were actually quite easy to merge, since they had the same exact column identifier and number of rows. You can edit your script to include more specifics in instances when these may differ across datasets that you would like to merge. This option allows you to edit the name of the column that is used in each dataframe. Here, these are still the same “ID”, but you can see that adding the by.x and by.y arguments allows you to specify instances when different column names are used in the two datasets. full.data &lt;- merge(demographic_data, chemical_data, by.x = &quot;ID&quot;, by.y = &quot;ID&quot;) # Viewing data head(full.data) ## ID BMI MAge MEdu BW GA DWAs DWCd DWCr UAs ## 1 1 27.7 22.99928 3 3180.058 34 6.426464 1.292941 51.67987 10.192695 ## 2 2 26.8 30.05142 3 3210.823 43 7.832384 1.798535 50.10409 11.815088 ## 3 3 33.2 28.04660 3 3311.551 40 7.516569 1.288461 48.74001 10.079057 ## 4 4 30.1 34.81796 3 3266.844 32 5.906656 2.075259 50.92745 8.719123 ## 5 5 37.4 42.68440 3 3664.088 35 7.181873 2.762643 55.16882 9.436559 ## 6 6 33.3 24.94960 3 3328.988 40 9.723429 3.054057 51.14812 11.589403 ## UCd UCr ## 1 0.7537104 42.60187 ## 2 0.9789506 41.30757 ## 3 0.1903262 36.47716 ## 4 0.9364825 42.47987 ## 5 1.4977829 47.78528 ## 6 1.6645837 38.26386 Note that after merging datasets, it is always helpful to check that the merging was done properly before proceeding with your data analysis. Helpful checks could include viewing the merged dataset, checking the numbers of rows and columns to make sure chunks of data are not missing, and searching for values (or strings) that exist in one dataset but not the other, among other mechanisms of QA/QC. Filtering and Subsetting Data Using Base R Syntax Filtering and subsetting data are useful tools when you need to focus on specific parts of your dataset for downstream analyses. These could represent, for example, specific samples or participants that meet certain criteria that you are interested in evaluating. It is also useful for removing unneeded variables or samples from dataframes as you are working through your script. Note that in the examples that follow, we will create new dataframes that are distinguished from our original dataframe by adding sequential numbers to the end of the dataframe name (e.g., subset.data1, subset.data2, subset.data3). This style of dataframe naming is useful for the simple examples we are demonstrating, but in a full scripted analysis, we encourage the use of more descriptive dataframe names. For example, if you are subsetting your data to include only the first 100 rows, you could name that dataframe “data.first100.” For this example, let’s first define a vector of columns that we want to keep in our analysis, then subset the data by keeping only the columns specified in our vector: # Defining a vector of columns to keep in the analysis subset.columns &lt;- c(&quot;BMI&quot;, &quot;MAge&quot;, &quot;MEdu&quot;) # Subsetting the data by selecting the columns represented in the defined &#39;subset.columns&#39; vector subset.data1 &lt;- full.data[,subset.columns] # Viewing the top of this subsetted dataframe head(subset.data1) ## BMI MAge MEdu ## 1 27.7 22.99928 3 ## 2 26.8 30.05142 3 ## 3 33.2 28.04660 3 ## 4 30.1 34.81796 3 ## 5 37.4 42.68440 3 ## 6 33.3 24.94960 3 We can also easily subset data based on row numbers. For example, to keep only the first 100 rows: subset.data2 &lt;- full.data[1:100,] # Viewing the dimensions of this new dataframe dim(subset.data2) ## [1] 100 12 To remove the first 100 rows, we use the same code as above, but include a - sign before our vector to indicate that these rows should be removed: subset.data3 &lt;- full.data[-c(1:100),] # Viewing the dimensions of this new dataframe dim(subset.data3) ## [1] 100 12 Conditional statements are also written to filter and subset data. A conditional statement is written to execute one block of code if the statement is true and a different block of code if the statement is false. A conditional statement requires a Boolean or true/false statement that will be either TRUE or FALSE. A couple of the more commonly used functions used to create conditional statements include… if(){} or an if statement means “execute R code when the condition is met”. if(){} else{} or an if/else statement means “execute R code when condition 1 is met, if not execute R code for condition 2”. ifelse() is a function that executes the same logic as an if/else statement. The first argument specifies a condition to be met. If that condition is met, R code in the second argument is executed, and if that condition is not met, R code in the third argument is executed. There are six comparison operators that are used to created these Boolean values: == means “equals”. != means “not equal”. &lt; means “less than”. &gt; means “greater than”. &lt;= means “less than or equal to”. &gt;= mean “greater than or equal to”. There are also three logical operators that are used to create these Boolean values: &amp; means “and”. | means “or”. ! means “not”. We can filter data based on conditions using the subset() function. For example, the following code filters for subjects whose BMI is greater than 25 and who have a college education: subset.data4 &lt;- subset(full.data, BMI &gt; 25 &amp; MEdu == 3) Additionally, we can subset and select specific columns we would like to keep, using the select argument within the subset() function: # Filtering for subjects whose BMI is less than 22 or greater than 27 # Also selecting the BMI, maternal age, and maternal education columns subset.data5 &lt;- subset(full.data, BMI &lt; 22 | BMI &gt; 27, select = subset.columns) For more information on the subset() function, see its associated documentation. Melting and Casting Data using Base R Syntax Melting and casting refers to the conversion of data to “long” or “wide” form as discussed previously in TAME 2.0 Module 1.4 Data Wrangling in Excel. You will often see data within the environmental health field in wide format, though long format is necessary for some procedures, such as plotting with ggplot2 and performing certain analyses. Here, we’ll illustrate some example script to melt and cast data using the reshape2 package. Let’s first install and load the reshape2 package: if (!requireNamespace(&quot;reshape2&quot;)) install.packages(&quot;reshape2&quot;); library(reshape2) ## ## Attaching package: &#39;reshape2&#39; ## The following object is masked from &#39;package:tidyr&#39;: ## ## smiths Using the fully merged dataframe, let’s remind ourselves what these data look like in the current dataframe format: head(full.data) ## ID BMI MAge MEdu BW GA DWAs DWCd DWCr UAs ## 1 1 27.7 22.99928 3 3180.058 34 6.426464 1.292941 51.67987 10.192695 ## 2 2 26.8 30.05142 3 3210.823 43 7.832384 1.798535 50.10409 11.815088 ## 3 3 33.2 28.04660 3 3311.551 40 7.516569 1.288461 48.74001 10.079057 ## 4 4 30.1 34.81796 3 3266.844 32 5.906656 2.075259 50.92745 8.719123 ## 5 5 37.4 42.68440 3 3664.088 35 7.181873 2.762643 55.16882 9.436559 ## 6 6 33.3 24.94960 3 3328.988 40 9.723429 3.054057 51.14812 11.589403 ## UCd UCr ## 1 0.7537104 42.60187 ## 2 0.9789506 41.30757 ## 3 0.1903262 36.47716 ## 4 0.9364825 42.47987 ## 5 1.4977829 47.78528 ## 6 1.6645837 38.26386 These data are represented by single subject identifiers listed as unique IDs per row, with associated environmental measures and demographic data organized across the columns. Thus, this dataframe is currently in wide (also known as casted) format. Let’s convert this dataframe to long (also known as melted) format. Here, will will specify that we want a row for each unique sample ID + variable measure pair by using id = \"ID\": full.melted &lt;- melt(full.data, id = &quot;ID&quot;) # Viewing this new dataframe head(full.melted) ## ID variable value ## 1 1 BMI 27.7 ## 2 2 BMI 26.8 ## 3 3 BMI 33.2 ## 4 4 BMI 30.1 ## 5 5 BMI 37.4 ## 6 6 BMI 33.3 You can see here that each measure that was originally contained as a unique column has been reoriented, such that the original column header is now listed throughout the second column labeled variable. Then, the third column contains the value of this variable. Let’s see an example view of the middle of this new dataframe: full.melted[1100:1110,1:3] ## ID variable value ## 1100 100 DWAs 7.928885 ## 1101 101 DWAs 8.677403 ## 1102 102 DWAs 8.115183 ## 1103 103 DWAs 7.134189 ## 1104 104 DWAs 8.816142 ## 1105 105 DWAs 7.487227 ## 1106 106 DWAs 7.541973 ## 1107 107 DWAs 6.313516 ## 1108 108 DWAs 6.654474 ## 1109 109 DWAs 7.564429 ## 1110 110 DWAs 7.357122 Here, we can see a different variable (DWAs) now being listed. This continues throughout the entire dataframe, which has the following dimensions: dim(full.melted) ## [1] 2200 3 Let’s now re-cast this dataframe back into wide format using the dcast() function. Here, we are telling the dcast() function to give us a sample (ID) for every variable in the column labeled variable. The column names from the variable column and corresponding values from the value column are then used to fill in the dataset: full.cast &lt;- dcast(full.melted, ID ~ variable) head(full.cast) ## ID BMI MAge MEdu BW GA DWAs DWCd DWCr UAs ## 1 1 27.7 22.99928 3 3180.058 34 6.426464 1.292941 51.67987 10.192695 ## 2 2 26.8 30.05142 3 3210.823 43 7.832384 1.798535 50.10409 11.815088 ## 3 3 33.2 28.04660 3 3311.551 40 7.516569 1.288461 48.74001 10.079057 ## 4 4 30.1 34.81796 3 3266.844 32 5.906656 2.075259 50.92745 8.719123 ## 5 5 37.4 42.68440 3 3664.088 35 7.181873 2.762643 55.16882 9.436559 ## 6 6 33.3 24.94960 3 3328.988 40 9.723429 3.054057 51.14812 11.589403 ## UCd UCr ## 1 0.7537104 42.60187 ## 2 0.9789506 41.30757 ## 3 0.1903262 36.47716 ## 4 0.9364825 42.47987 ## 5 1.4977829 47.78528 ## 6 1.6645837 38.26386 Here, we can see that this dataframe is back in its original casted (or wide) format. Now that we’re familiar with some base R functions to reshape our data, let’s answer our original question: What is the average urinary chromium concentration for each maternal education level? Although it is not necessary to calculate the average, we could first subset our data frame to only include the two columns we are interested in (MEdu and UCr): subset.data6 &lt;- full.data[,c(&quot;MEdu&quot;, &quot;UCr&quot;)] head(subset.data6) ## MEdu UCr ## 1 3 42.60187 ## 2 3 41.30757 ## 3 3 36.47716 ## 4 3 42.47987 ## 5 3 47.78528 ## 6 3 38.26386 Next, we will make a new data frame for each maternal education level: # Creating new data frames based on maternal education category data.matedu.1 &lt;- subset(subset.data6, MEdu == 1) data.matedu.2 &lt;- subset(subset.data6, MEdu == 2) data.matedu.3 &lt;- subset(subset.data6, MEdu == 3) # Previewing the first data frame to make sure our function is working as specified head(data.matedu.1) ## MEdu UCr ## 14 1 38.59349 ## 18 1 47.77878 ## 37 1 35.33980 ## 63 1 34.72255 ## 66 1 34.13982 ## 76 1 31.38145 Last, we can calculate the average urinary chromium concentration using each of our data frames: mean(data.matedu.1$UCr) ## [1] 39.88055 mean(data.matedu.2$UCr) ## [1] 40.61807 mean(data.matedu.3$UCr) ## [1] 40.41556 With this, we can answer our Environmental Health Question: What is the average urinary chromium concentration across different maternal education levels? Answer: The average urinary Chromium concentrations are 39.9 µg/L for participants with less than high school education, 40.6 µg/L for participants with high school or some college education, and 40.4 µg/L for participants with college education or greater. Introduction to Tidyverse Tidyverse is a collection of packages that are commonly used to more efficiently organize and manipulate datasets in R. This collection of packages has its own specific type of syntax and formatting that differ slightly from base R functions. There are eight core tidyverse packages: For data visualization and exploration: ggplot2 For data wrangling and transformation: dplyr tidyr stringr forcats For data import and management: tibble readr For functional programming: purr Here, we will carry out all the of the same data organization exercises demonstrated above using packages that are part of The Tidyverse, specifically using functions that are part of the dplyr and tidyr packages. Downloading and Loading the Tidyverse Package If you don’t have tidyverse already installed, you will need to install it using: if(!require(tidyverse)) install.packages(&quot;tidyverse&quot;) And then load the tidyverse package using: library(tidyverse) Note that by loading the tidyverse package, you are also loading all of the packages included within The Tidyverse and do not need to separately load these packages. Merging Data Using Tidyverse Syntax To merge the same example dataframes using tidyverse, you can run the following script: full.data.tidy &lt;- inner_join(demographic_data, chemical_data, by = &quot;ID&quot;) head(full.data.tidy) ## ID BMI MAge MEdu BW GA DWAs DWCd DWCr UAs ## 1 1 27.7 22.99928 3 3180.058 34 6.426464 1.292941 51.67987 10.192695 ## 2 2 26.8 30.05142 3 3210.823 43 7.832384 1.798535 50.10409 11.815088 ## 3 3 33.2 28.04660 3 3311.551 40 7.516569 1.288461 48.74001 10.079057 ## 4 4 30.1 34.81796 3 3266.844 32 5.906656 2.075259 50.92745 8.719123 ## 5 5 37.4 42.68440 3 3664.088 35 7.181873 2.762643 55.16882 9.436559 ## 6 6 33.3 24.94960 3 3328.988 40 9.723429 3.054057 51.14812 11.589403 ## UCd UCr ## 1 0.7537104 42.60187 ## 2 0.9789506 41.30757 ## 3 0.1903262 36.47716 ## 4 0.9364825 42.47987 ## 5 1.4977829 47.78528 ## 6 1.6645837 38.26386 Note that you can still merge dataframes that have different ID column names with the argument by = c(\"ID.x\", \"ID.y\"). tidyverse also has other join, functions, shown in the graphic below (source): inner_join keeps only rows that have matching ID variables in both datasets full_join keeps the rows in both datasets left_join matches rows based on the ID variables in the first dataset (and omits any rows from the second dataset that do not have matching ID variables in the first dataset) right_join matches rows based on ID variables in the second dataset (and omits any rows from the first dataset that do not have matching ID variables in the second dataset) anti_join(x,y) keeps the rows that are unique to the first dataset anti_join(y,x) keeps the rows that are unique to the second dataset The Pipe Operator One of the most important elements of Tidyverse syntax is use of the pipe operator (%&gt;%). The pipe operator can be used to chain multiple functions together. It takes the object (typically a dataframe) to the left of the pipe operator and passes it to the function to the right of the pipe operator. Multiple pipes can be used in chain to execute multiple data cleaning steps without the need for intermediate dataframes. The pipe operator can be used to pass data to functions within all of the Tidyverse universe packages, not just the functions demonstrated here. Below, we can see the same code executed above, but this time with the pipe operator. The demographic_data dataframe is passed to inner_join() as the first argument to that function, with the following arguments remaining the same. full.data.tidy2 &lt;- demographic_data %&gt;% inner_join(chemical_data, by = &quot;ID&quot;) head(full.data.tidy2) ## ID BMI MAge MEdu BW GA DWAs DWCd DWCr UAs ## 1 1 27.7 22.99928 3 3180.058 34 6.426464 1.292941 51.67987 10.192695 ## 2 2 26.8 30.05142 3 3210.823 43 7.832384 1.798535 50.10409 11.815088 ## 3 3 33.2 28.04660 3 3311.551 40 7.516569 1.288461 48.74001 10.079057 ## 4 4 30.1 34.81796 3 3266.844 32 5.906656 2.075259 50.92745 8.719123 ## 5 5 37.4 42.68440 3 3664.088 35 7.181873 2.762643 55.16882 9.436559 ## 6 6 33.3 24.94960 3 3328.988 40 9.723429 3.054057 51.14812 11.589403 ## UCd UCr ## 1 0.7537104 42.60187 ## 2 0.9789506 41.30757 ## 3 0.1903262 36.47716 ## 4 0.9364825 42.47987 ## 5 1.4977829 47.78528 ## 6 1.6645837 38.26386 Because the pipe operator is often used in a chain, it is best practice is to start a new line after each pipe operator, with the new lines of code indented. This makes code with multiple piped steps easier to follow. However, if just one function is being executed, the pipe operator can be used on the same line as the input and function or omitted altogether (as shown in the previous two code chunks). Here is an example of placing the function to the right of the pipe operator on a new line, with placeholder functions shown as additional steps: full.data.tidy3 &lt;- demographic_data %&gt;% inner_join(chemical_data, by = &quot;ID&quot;) %&gt;% additional_function_1() %&gt;% additional_function_2() Filtering and Subsetting Data Using Tidyverse Syntax Column-wise functions The select() function is used to subset columns in Tidyverse. Here, we can use our previously defined vector subset.columns in the select() function to keep only the columns in our subset.columns vector. The all_of() function tells the select() to keep all of the columns that match elements of the subset.columns vector. subset.tidy1 &lt;- full.data.tidy %&gt;% select(all_of(subset.columns)) head(subset.tidy1) ## BMI MAge MEdu ## 1 27.7 22.99928 3 ## 2 26.8 30.05142 3 ## 3 33.2 28.04660 3 ## 4 30.1 34.81796 3 ## 5 37.4 42.68440 3 ## 6 33.3 24.94960 3 There are many different ways that select() can be used. See below for some examples using dummy variable names: # Select specific ranges in the dataframe data &lt;- data %&gt;% select(start_column_1:end_column_1) data &lt;- data %&gt;% select(c(start_column_1:end_column_1, start_column_2:end_column_2)) # Select columns that match the elements in a character vector an an additional range of columns data &lt;- data %&gt;% select(c(all_of(character_vector), start_column_1:end_column_1)) To select columns that have names that contain specific strings, you can use functions such as starts_with(), ends_with(), and contains(). These functions allow you to ignore the case of the strings with ignore.case = TRUE. These arguments can be combined with specific column names and other selection ranges. data &lt;- data %&gt;% select(starts_with(&quot;starting_string&quot;)) data &lt;- data %&gt;% select(other_column_to_keep, starts_with(&quot;starting_string&quot;)) To remove columns using tidyverse, you can use similar code, but include a - sign before the argument defining the columns. # Removing columns subset.tidy2 &lt;- full.data.tidy %&gt;% select(-all_of(subset.columns)) # Viewing this new dataframe head(subset.tidy2) ## ID BW GA DWAs DWCd DWCr UAs UCd UCr ## 1 1 3180.058 34 6.426464 1.292941 51.67987 10.192695 0.7537104 42.60187 ## 2 2 3210.823 43 7.832384 1.798535 50.10409 11.815088 0.9789506 41.30757 ## 3 3 3311.551 40 7.516569 1.288461 48.74001 10.079057 0.1903262 36.47716 ## 4 4 3266.844 32 5.906656 2.075259 50.92745 8.719123 0.9364825 42.47987 ## 5 5 3664.088 35 7.181873 2.762643 55.16882 9.436559 1.4977829 47.78528 ## 6 6 3328.988 40 9.723429 3.054057 51.14812 11.589403 1.6645837 38.26386 Row-wise functions The slice() function can be used to keep or remove a certain number of rows based on their position within the dataframe. For example, we can retain only the first 100 rows using the following code: subset.tidy3 &lt;- full.data.tidy %&gt;% slice(1:100) dim(subset.tidy3) ## [1] 100 12 Or, we can remove the first 100 rows: subset.tidy4 &lt;- full.data.tidy %&gt;% slice(-c(1:100)) dim(subset.tidy4) ## [1] 100 12 The related functions slice_min() and slice_max() can be used to select rows with the smallest or largest values of a variable. The filter() function can be used to keep or remove specific rows based on conditional statements. For example, we can keep only rows where BMI is greater than 25 and age is greater than 31: subset.tidy5 &lt;- full.data.tidy %&gt;% filter(BMI &gt; 25 &amp; MAge &gt; 31) dim(subset.tidy5) ## [1] 49 12 Combining column and row-wise functions Now, we can see how Tidyverse makes it easy to chain together multiple data manipulation steps. Here, we first filter rows based on values for BMI and age, then we select our columns of interest: subset.tidy6 &lt;- full.data.tidy %&gt;% filter(BMI &gt; 25 &amp; MAge &gt; 31) %&gt;% select(BMI, MAge, MEdu) head(subset.tidy6) ## BMI MAge MEdu ## 1 30.1 34.81796 3 ## 2 37.4 42.68440 3 ## 3 36.9 33.58589 3 ## 4 33.7 33.82961 3 ## 5 25.7 37.08028 3 ## 6 28.4 47.85761 3 Melting and Casting Data Using Tidyverse Syntax To melt and cast data in Tidyverse, you can use the pivot functions (i.e., pivot_longer() or pivot_wider()). The first argument in the pivot_longer() function specifies which columns should be pivoted. This can be specified with either positive or negative selection - i.e., naming columns to pivot with a vector or range or naming columns not to pivot with a - sign. Here, we are telling the function to pivot all of the columns except the ID column, which we need to keep to be able to trace back which values came from which subject. The names_to = argument allows you to set what you want to name the column that stores the variable names (the column names in wide format). The values_to = argument allows you to set what you want to name the column that stores the values. We almost always call these columns “var” and “value”, respectively, but you can name them anything that makes sense for your dataset. full.pivotlong &lt;- full.data.tidy %&gt;% pivot_longer(-ID, names_to = &quot;var&quot;, values_to = &quot;value&quot;) head(full.pivotlong, 15) ## # A tibble: 15 × 3 ## ID var value ## &lt;int&gt; &lt;chr&gt; &lt;dbl&gt; ## 1 1 BMI 27.7 ## 2 1 MAge 23.0 ## 3 1 MEdu 3 ## 4 1 BW 3180. ## 5 1 GA 34 ## 6 1 DWAs 6.43 ## 7 1 DWCd 1.29 ## 8 1 DWCr 51.7 ## 9 1 UAs 10.2 ## 10 1 UCd 0.754 ## 11 1 UCr 42.6 ## 12 2 BMI 26.8 ## 13 2 MAge 30.1 ## 14 2 MEdu 3 ## 15 2 BW 3211. To pivot our data back to wide format, we can use pivot_wider(), which will pull the column names from the column specified in the names_from = argument and the corresponding values from the column specified in the values_from = argument. full.pivotwide &lt;- full.pivotlong %&gt;% pivot_wider(names_from = &quot;var&quot;, values_from = &quot;value&quot;) head(full.pivotwide) ## # A tibble: 6 × 12 ## ID BMI MAge MEdu BW GA DWAs DWCd DWCr UAs UCd UCr ## &lt;int&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; ## 1 1 27.7 23.0 3 3180. 34 6.43 1.29 51.7 10.2 0.754 42.6 ## 2 2 26.8 30.1 3 3211. 43 7.83 1.80 50.1 11.8 0.979 41.3 ## 3 3 33.2 28.0 3 3312. 40 7.52 1.29 48.7 10.1 0.190 36.5 ## 4 4 30.1 34.8 3 3267. 32 5.91 2.08 50.9 8.72 0.936 42.5 ## 5 5 37.4 42.7 3 3664. 35 7.18 2.76 55.2 9.44 1.50 47.8 ## 6 6 33.3 24.9 3 3329. 40 9.72 3.05 51.1 11.6 1.66 38.3 Now that we’re familiar with some tidyverse functions to reshape our data, let’s answer our original question: What is the average urinary Chromium concentration for each maternal education level? We can use the group_by() function to group our dataset by education class, then the summarize function to calculate the mean of our variable of interest within each class. Note how much shorter and more efficient this code is than the code we used to calculate the same values using base R! full.data %&gt;% group_by(MEdu) %&gt;% summarize(Avg_UCr = mean(UCr)) ## # A tibble: 3 × 2 ## MEdu Avg_UCr ## &lt;int&gt; &lt;dbl&gt; ## 1 1 39.9 ## 2 2 40.6 ## 3 3 40.4 For more detailed and advanced examples of pivoting in Tidyverse, see the Tidyverse Pivoting Vignette. Concluding Remarks This training module provides an introductory level overview of data organization and manipulation basics in base R and Tidyverse, including merging, filtering, subsetting, melting, and casting, and demonstrates these methods with an environmentally relevant dataset. These methods are used regularly in scripted analyses and are important preparation steps for almost all downstream analyses and visualizations. Test Your Knowledge What subjects, arranged from highest to lowest drinking water cadmium levels, had babies at at least 35 weeks and had urinary cadmium levels of at least 1.5 µg/L? Hint: Try using the arrange() function from the tidyverse package. "],["improving-coding-efficiencies.html", "Improving Coding Efficiencies Introduction to Training Module Loops Functions List operations Concluding Remarks Additional Resources", " Improving Coding Efficiencies This training module was developed by Elise Hickman, Alexis Payton, Kyle Roell, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module In this module, we’ll explore how to improve coding efficiency. Coding efficiency involves performing a task in as few lines as possible and can… Shorten code by eliminating redundancies Reduce the number of typos Help other coders understand script better Specific approaches that we will discuss in this module include loops, functions, and list operations, which can all be used to make code more succinct. A loop is employed when we want to perform a repetitive task, while a function contains a block of code organized together to perform one specific task. List operations, in which the same function is applied to a list of dataframes, can also be used to code more efficiently. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: Are there statistically significant differences in drinking water arsenic, cadmium, and chromium between normal weight (BMI &lt; 25) and overweight (BMI \\(\\geq\\) 25) subjects? Are there statistically significant differences in drinking water arsenic, cadmium, and chromium between underweight (BMI &lt; 18.5) and non-underweight (BMI \\(\\geq\\) 18.5) subjects? Are there statistically significant difference in drinking water arsenic, cadmium, and chromium between non-obese (BMI &lt; 29.9) and obese (BMI \\(\\geq\\) 29.9) subjects? We will demonstrate how this analysis can be approached using for loops, functions, or list operations. We will introduce the syntax and structure of each approach first, followed by application of the approach to our data. First, let’s prepare the workspace and familiarize ourselves with the dataset we are going to use. Data Import and Workspace Preparation Installing required packages If you already have these packages installed, you can skip this step, or you can run the below code which checks installation status for you. We will be using the tidyverse package for data manipulation steps and the rstatix package for statistical tests, as it provides pipe friendly adaptations of the base R statistical tests and returns results in a dataframe rather than a list format, making results easier to access. This brings up an important aspect of coding efficiency - sometimes, there is already a package that has been designed with functions to help you execute your desired analysis in an efficient way, so you don’t need to write custom functions yourself! So, don’t forget to explore packages relevant to your analysis before spending a lot of time developing custom solutions (although, sometimes this is necessary). if (!requireNamespace(&quot;tidyverse&quot;)) install.packages(&quot;tidyverse&quot;) if (!requireNamespace(&quot;rstatix&quot;)) install.packages(&quot;rstatix&quot;) Loading required packages library(tidyverse) library(rstatix) Setting your working directory setwd(&quot;/file path to where your input files are&quot;) Importing example dataset The first example dataset contains subject demographic data, and the second dataset contains corresponding chemical data. Familiarize yourself with these data used previously in TAME 2.0 Module 2.3 Data Manipulation and Reshaping. # Load the demographic data demographic_data &lt;- read.csv(&quot;Module2_4_Input/Module2_4_InputData1.csv&quot;) # View the top of the demographic dataset head(demographic_data) ## ID BMI MAge MEdu BW GA ## 1 1 27.7 22.99928 3 3180.058 34 ## 2 2 26.8 30.05142 3 3210.823 43 ## 3 3 33.2 28.04660 3 3311.551 40 ## 4 4 30.1 34.81796 3 3266.844 32 ## 5 5 37.4 42.68440 3 3664.088 35 ## 6 6 33.3 24.94960 3 3328.988 40 # Load the chemical data chemical_data &lt;- read.csv(&quot;Module2_4_Input/Module2_4_InputData2.csv&quot;) # View the top of the chemical dataset head(chemical_data) ## ID DWAs DWCd DWCr UAs UCd UCr ## 1 1 6.426464 1.292941 51.67987 10.192695 0.7537104 42.60187 ## 2 2 7.832384 1.798535 50.10409 11.815088 0.9789506 41.30757 ## 3 3 7.516569 1.288461 48.74001 10.079057 0.1903262 36.47716 ## 4 4 5.906656 2.075259 50.92745 8.719123 0.9364825 42.47987 ## 5 5 7.181873 2.762643 55.16882 9.436559 1.4977829 47.78528 ## 6 6 9.723429 3.054057 51.14812 11.589403 1.6645837 38.26386 Preparing the example dataset For ease of analysis, we will merge these two datasets before proceeding. # Merging data full_data &lt;- inner_join(demographic_data, chemical_data, by = &quot;ID&quot;) # Previewing new data head(full_data) ## ID BMI MAge MEdu BW GA DWAs DWCd DWCr UAs ## 1 1 27.7 22.99928 3 3180.058 34 6.426464 1.292941 51.67987 10.192695 ## 2 2 26.8 30.05142 3 3210.823 43 7.832384 1.798535 50.10409 11.815088 ## 3 3 33.2 28.04660 3 3311.551 40 7.516569 1.288461 48.74001 10.079057 ## 4 4 30.1 34.81796 3 3266.844 32 5.906656 2.075259 50.92745 8.719123 ## 5 5 37.4 42.68440 3 3664.088 35 7.181873 2.762643 55.16882 9.436559 ## 6 6 33.3 24.94960 3 3328.988 40 9.723429 3.054057 51.14812 11.589403 ## UCd UCr ## 1 0.7537104 42.60187 ## 2 0.9789506 41.30757 ## 3 0.1903262 36.47716 ## 4 0.9364825 42.47987 ## 5 1.4977829 47.78528 ## 6 1.6645837 38.26386 Continuous demographic variables, like BMI, are often dichotomized (or converted to a categorical variable with two categories representing higher vs. lower values) to increase statistical power in analyses. This is particularly important for clinical data that tend to have smaller sample sizes. In our initial dataframe, BMI is a continuous or numeric variable; however, our questions require us to dichotomize BMI. We can use the following code, which relies on if/else logic (see TAME 2.0 Module 2.3 Data Manipulation and Reshaping for more information) to generate a new column representing our dichotomized BMI variable for our first environmental health question. # Adding dichotomized BMI column full_data &lt;- full_data %&gt;% mutate(Dichotomized_BMI = ifelse(BMI &lt; 25, &quot;Normal&quot;, &quot;Overweight&quot;)) # Previewing new data head(full_data) ## ID BMI MAge MEdu BW GA DWAs DWCd DWCr UAs ## 1 1 27.7 22.99928 3 3180.058 34 6.426464 1.292941 51.67987 10.192695 ## 2 2 26.8 30.05142 3 3210.823 43 7.832384 1.798535 50.10409 11.815088 ## 3 3 33.2 28.04660 3 3311.551 40 7.516569 1.288461 48.74001 10.079057 ## 4 4 30.1 34.81796 3 3266.844 32 5.906656 2.075259 50.92745 8.719123 ## 5 5 37.4 42.68440 3 3664.088 35 7.181873 2.762643 55.16882 9.436559 ## 6 6 33.3 24.94960 3 3328.988 40 9.723429 3.054057 51.14812 11.589403 ## UCd UCr Dichotomized_BMI ## 1 0.7537104 42.60187 Overweight ## 2 0.9789506 41.30757 Overweight ## 3 0.1903262 36.47716 Overweight ## 4 0.9364825 42.47987 Overweight ## 5 1.4977829 47.78528 Overweight ## 6 1.6645837 38.26386 Overweight We can see that we now have created a new column entitled Dichotomized_BMI that we can use to perform a statistical test to assess if there are differences between drinking water metals between normal and overweight subjects. Loops We will start with loops. There are three main types of loops in R: for, while, and repeat. We will focus on for loops in this module, but for more in-depth information on loops, including the additional types of loops, see here. Before applying loops to our data, let’s discuss how for loops work. The basic structure of a for loop is shown here: # Basic structure of a for loop for (i in 1:4){ print(i) } ## [1] 1 ## [1] 2 ## [1] 3 ## [1] 4 for loops always start with for followed by a statement in parentheses. The argument in the parentheses tells R how to iterate (or repeat) through the code in the curly brackets. Here, we are telling R to iterate through the code in curly brackets 4 times. Each time we told R to print the value of our iterator, or i, which has a value of 1, 2, 3, and then 4. Loops can also iterate through columns in a dataset. For example, we can use a for loop to print the ages of each subject: # Creating a smaller dataframe for our loop example full_data_subset &lt;- full_data[1:6, ] # Finding the total number of rows or subjects in the dataset number_of_rows &lt;- length(full_data_subset$MAge) # Creating a for loop to iterate from 1 to the last row for (i in 1:number_of_rows){ # Printing each subject age # Need to put `[i]` to index the correct value corresponding to the row we are evaluating print(full_data_subset$MAge[i]) } ## [1] 22.99928 ## [1] 30.05142 ## [1] 28.0466 ## [1] 34.81796 ## [1] 42.6844 ## [1] 24.9496 Now that we know how a for loop works, how can we apply this approach to determine whether there are statistically significant differences in drinking water arsenic, cadmium, and chromium between normal weight (BMI &lt; 25) and overweight (BMI \\(\\geq\\) 25) subjects. Because our data are normally distributed and there are two groups that we are comparing, we will use a t-test applied to each metal measured in drinking water. Testing for assumptions is outside the scope of this module, but see TAME 2.0 Module 3.3 Normality Tests and Data Transformation for more information on this topic. Running a t-test in R is very simple, which we can demonstrate by running a t-test on the drinking water arsenic data: # Running t-test and storing results in t_test_res t_test_res &lt;- full_data %&gt;% t_test(DWAs ~ Dichotomized_BMI) # Viewing results t_test_res ## # A tibble: 1 × 8 ## .y. group1 group2 n1 n2 statistic df p ## * &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;int&gt; &lt;int&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; ## 1 DWAs Normal Overweight 96 104 -0.728 192. 0.468 We can see that our p-value is 0.468. Because this is greater than 0.05, we cannot reject the null hypothesis that normal weight and overweight subjects are exposed to the same drinking water arsenic concentrations. Although this was a very simple line of code to run, what if we have many columns we want to run the same t-test on? We can use a for loop to iterate through these columns. Let’s break down the steps of our for loop before executing the code. First, we will define the variables (columns) we want to run our t-test on. This is different from our approach above, because in those code chunks, we were using numbers to indicate the number of iterations through the loop. Here, we are naming the specific variables instead, and R will iterate though each of these variables. Note that we could omit this step and instead use the numeric column index of our variables of interest [7:9]. However, naming the specific columns makes this approach more robust because if additional data are added to or removed from our dataframe, the numeric column index of our variables could change. Which approach you choose really depends on the purpose of your loop! Second, we will create an empty dataframe where we will store the results generated by our for loop. Third, we will actually run our for loop. This will tell R: for each variable in our vars_of_interest vector, run a t-test with that variable (and store the results in a temporary dataframe called “res”), then add those results to our final results dataframe. A row will be added to the results dataframe each time R iterates through a new variable, resulting in a dataframe that stores the results of all of our t-tests. # Defining variables (columns) we want to run a t-test on vars_of_interest &lt;- c(&quot;DWAs&quot;, &quot;DWCd&quot;, &quot;DWCr&quot;) # Creating an empty dataframe to store results t_test_res_DW &lt;- data.frame() # Running for loop for (i in vars_of_interest) { # Storing the results of each iteration of the loop in a temporary results dataframe res &lt;- full_data %&gt;% # Writing the formula needed for each iteration of the loop t_test(as.formula(paste(i, &quot;~ Dichotomized_BMI&quot;, sep = &quot;&quot;))) # Adding a row to the results dataframe each time the loop is iterated t_test_res_DW &lt;- bind_rows(t_test_res_DW, res) } # Viewing our results t_test_res_DW ## .y. group1 group2 n1 n2 statistic df p ## 1 DWAs Normal Overweight 96 104 -0.7279621 192.3363 0.468 ## 2 DWCd Normal Overweight 96 104 -0.5894360 196.1147 0.556 ## 3 DWCr Normal Overweight 96 104 0.1102933 197.9870 0.912 With this, we can answer Environmental Health Question #1: Are there statistically significant differences in drinking water arsenic, cadmium, and chromium between normal weight (BMI &lt; 25) and overweight (BMI \\(\\geq\\) 25) subjects? Answer: No, there are not any statistically significant differences in drinking water metals between normal weight and overweight subjects. Formulas and Pasting Note the use of the code as.formula(paste0(i, \"~ Dichotomized_BMI\")). Let’s take a quick detour to discuss the use of the as.formula() and paste() functions, as these are important functions often used in loops and user-defined functions. Many statistical test functions and regression functions require one argument to be a formula, which is typically formatted as y ~ x, where y is the dependent variable of interest and x is an independent variable. For some functions, additional variables can be included on the right side of the formula to represent covariates (additional variables of interest). The function as.formula() returns the argument in parentheses in formula format so that it can be correctly passed to other functions. We can demonstrate that here by assigning a dummy variable j the character string var1: # Assigning variable j &lt;- &quot;var1&quot; # Demonstrating output of as.formula() as.formula(paste(j, &quot; ~ Dichotomized_BMI&quot;, sep = &quot;&quot;)) ## var1 ~ Dichotomized_BMI We can use the paste() function to combine strings of characters. The paste function takes each argument (as many arguments as is needed) and pastes them together into one character string, with the separator between arguments set by the sep = argument. When our y variable is changing with each iteration of our for loop, we can use the paste() function to write our formula correctly by telling the function to paste the variable i, followed by the rest of our formula, which stays the same for each iteration of the loop. Let’s examine the output of just the paste() part of our code: paste(j, &quot; ~ Dichotomized_BMI&quot;, sep = &quot;&quot;) ## [1] &quot;var1 ~ Dichotomized_BMI&quot; The paste() function is very flexible and can be useful in many other settings when you need to create one character string from arguments from different sources! Notice that the output looks different from the output of as.formula(). There is a returned index ([1]), and there are quotes around the character string. The last function we will highlight here is the noquote() function, which can be helpful if you’d like a string without quotes: noquote(paste(j, &quot; ~ Dichotomized_BMI&quot;, sep = &quot;&quot;)) ## [1] var1 ~ Dichotomized_BMI However, this still returns an indexed number, so there are times when it will not allow code to execute properly (for example, when we need a formula format). Next, we will learn about functions and apply them to our dataset to answer our additional environmental health questions. Functions Functions are useful when you want to execute a block of code organized together to perform one specific task, and you want to be able to change parameters for that task easily rather than having to copy and paste code over and over that largely stays the same but might have small modifications in certain arguments. The basic structure of a function is as follows: function_name &lt;- function(parameter_1, parameter_2...){ # Function body (where the code goes) insert_code_here # What the function returns return() } A function requires you to name it as we did with function_name. In parentheses, the function requires you to specify the arguments or parameters. Parameters (i.e., parameter_1) act as placeholders in the body of the function. This allows us to change the values of the parameters each time a function is called, while the majority of the code remains the same. Lastly, we have a return() statement, which specifies what object (i.e., vector, dataframe, etc.) we want to retrieve from a function. Although a function can display the last expression from the function body in the absence of a return() statement, it’s a good habit to include it as the last expression. It is important to note that, although functions can take many input parameters and execute large code chunks, they can only return one item, whether that is a value, vector, dataframe, plot, code output, or list. When writing your own functions, it is important to describe the purpose of the function, its input, its parameters, and its output so that others can understand what your functions does and how to use it. This can be defined either in text above a code chunk if you are using R Markdown or as comments within the code itself. We’ll start with a simple function. Let’s say we want to convert temperatures from Fahrenheit to Celsius. We can write a function that takes the temperature in Fahrenheit and converts it to Celsius. Note that we have given our parameters descriptive names (fahrenheit_temperature, celsius_temperature), which makes our code more readable than if we assigned them dummy names such as x and y. # Function to convert temperatures in Fahrenheit to Celsius ## Parameters: temperature in Fahrenheit (input) ## Output: temperature in Celsius fahrenheit_to_celsius &lt;- function(fahrenheit_temperature){ celsius_temperature &lt;- (fahrenheit_temperature - 32) * (5/9) return(celsius_temperature) } Notice that the above code block was run, but there isn’t an output. Rather, running the code assigns the function code to that function. When you run code defining a function, that function will appear in your Global Environment under the “Functions” section. We can see the output of the function by providing an input value. Let’s start by converting 41 degrees Fahrenheit to Celsius: # Calling the function # Here, 41 is the `fahrenheit_temperature` in the function fahrenheit_to_celsius(41) ## [1] 5 41 degrees Fahrenheit is equivalent to 5 degrees Celsius. We can also have the function convert a vector of values. # Defining vector of temperatures vector_of_temperatures &lt;- c(81,74,23,65) # Calling the function fahrenheit_to_celsius(vector_of_temperatures) ## [1] 27.22222 23.33333 -5.00000 18.33333 Before getting back to answer our environmental health related questions, let’s look at one more example of a function. This time we’ll create a function that can calculate the circumference of a circle based on its radius in inches. Here you can also see a different style of commenting to describe the function’s purpose, inputs, and outputs. circle_circumference &lt;- function(radius){ # Calculating a circle&#39;s circumference based on the radius inches # :parameters: radius # :output: circumference and radius # Calculating diameter first diameter &lt;- 2 * radius # Calculating circumference circumference &lt;- pi * diameter return(circumference) } # Calling function circle_circumference(3) ## [1] 18.84956 So, if a circle had a radius of 3 inches, its circumference would be ~19 inches. What if we were interested in seeing the diameter to double check our code? diameter ## Error in eval(expr, envir, enclos): object &#39;diameter&#39; not found R throws an error, because the variable diameter was created inside the function and the function only returned the circumference variable. This is actually one of the ways that functions can improve coding efficiency - by not needing to store intermediate variables that aren’t of interest to the main goal of the code or analysis. However, there are two ways we can still see the diameter variable: Put print statements in the body of the function (print(diameter)). Have the function return a different variable or list of variables (c(circumference, diameter)). See the below section on List Operation for more on this topic. We can now move on to using a more complicated function to answer all three of our environmental health questions without repeating our earlier code three times. The main difference between each of our first three environmental health questions is the BMI cutoff used to dichotomize the BMI variable, so we can use that as one of the parameters for our function. We can also use arguments in our function to name our groups. We can adapt our previous for loop code into a function that will take different BMI cutoffs and return statistical results by including parameters to define the parts of the analysis that will change with each unique question. For example: Changing the BMI cutoff from a number (in our previous code) to our parameter name that specifies the cutoff Changing the group names for assigning category (in our previous code) to our parameter names # Function to dichotomize BMI into different categories and return results of t-test on drinking water metals between dichotomized groups ## Parameters: ### input_data: dataframe containing BMI and drinking water metals levels ### bmi_cutoff: numeric value specifying the cut point for dichotomizing BMI ### lower_group_name: name for the group of subjects with BMIs lower than the cutoff ### upper_group_name: name for the group of subjects with BMIs higher than the cutoff ### variables: vector of variable names that statistical test should be run on ## Output: dataframe with statistical results for each variable in the variables vector bmi_DW_ttest &lt;- function(input_data, bmi_cutoff, lower_group_name, upper_group_name, variables){ # Creating dichotomized variable dichotomized_data &lt;- input_data %&gt;% mutate(Dichotomized_BMI = ifelse(BMI &lt; bmi_cutoff, lower_group_name, upper_group_name)) # Creating an empty dataframe to store results t_test_res_DW &lt;- data.frame() # Running for loop for (i in variables) { # Storing the results of each iteration of the loop in a temporary results dataframe res &lt;- dichotomized_data %&gt;% # Writing the formula needed for each iteration of the loop t_test(as.formula(paste(i, &quot;~ Dichotomized_BMI&quot;, sep = &quot;&quot;))) # Adding a row to the results dataframe each time the loop is iterated t_test_res_DW &lt;- bind_rows(t_test_res_DW, res) } # Return results return(t_test_res_DW) } For the first example of using the function, we have included the name of each argument for clarity, but this isn’t necessary if you pass in the arguments in the order they were defined when writing the function. # Defining variables (columns) we want to run a t-test on vars_of_interest &lt;- c(&quot;DWAs&quot;, &quot;DWCd&quot;, &quot;DWCr&quot;) # Apply function for normal vs. overweight (bmi_cutoff = 25) bmi_DW_ttest(input_data = full_data, bmi_cutoff = 25, lower_group_name = &quot;Normal&quot;, upper_group_name = &quot;Overweight&quot;, variables = vars_of_interest) ## .y. group1 group2 n1 n2 statistic df p ## 1 DWAs Normal Overweight 96 104 -0.7279621 192.3363 0.468 ## 2 DWCd Normal Overweight 96 104 -0.5894360 196.1147 0.556 ## 3 DWCr Normal Overweight 96 104 0.1102933 197.9870 0.912 Here, we can see the same results as above in the Loops section. We can next apply the function to answer our additional environmental health questions: # Apply function for underweight vs. non-underweight (bmi_cutoff = 18.5) bmi_DW_ttest(full_data, 18.5, &quot;Underweight&quot;, &quot;Non-Underweight&quot;, vars_of_interest) ## .y. group1 group2 n1 n2 statistic df p ## 1 DWAs Non-Underweight Underweight 166 34 -0.86947835 53.57143 0.388 ## 2 DWCd Non-Underweight Underweight 166 34 -0.97359810 55.45450 0.334 ## 3 DWCr Non-Underweight Underweight 166 34 0.04305105 56.08814 0.966 # Apply function for non-obese vs. obese (bmi_cutoff = 29.9) bmi_DW_ttest(full_data, 29.9, &quot;Non-Obese&quot;, &quot;Obese&quot;, vars_of_interest) ## .y. group1 group2 n1 n2 statistic df p ## 1 DWAs Non-Obese Obese 144 56 -1.9312097 86.80253 0.0567 ## 2 DWCd Non-Obese Obese 144 56 0.3431076 94.52209 0.7320 ## 3 DWCr Non-Obese Obese 144 56 -0.6878311 89.61818 0.4930 With this, we can answer Environmental Health Questions #2 &amp; #3: Are there statistically significant differences in drinking water arsenic, cadmium, and chromium between underweight (BMI &lt; 18.5) and non-underweight (BMI \\(\\geq\\) 18.5) subjects or between non-obese (BMI &lt; 29.9) and obese (BMI \\(\\geq\\) 29.9) subjects? Answer: No, there are not any statistically significant differences in drinking water metals between underweight and non-underweight subjects or between non-obese and obese subjects. Here, we were able to answer all three of our environmental health questions within relatively few lines of code by using a function to efficiently assess different variations on our analysis. In the last section of this module, we will demonstrate how to use list operations to improve coding efficiency. List operations Lists are a data type in R that can store other data types (including lists, to make nested lists). This allows you to store multiple dataframes in one object and apply the same functions to each dataframe in the list. Lists can also be helpful for storing the results of a function if you would like to be able to access multiple outputs. For example, if we return to our example of a function that calculates the circumference of a circle, we can store both the diameter and circumference as list objects. The function will then return a list containing both of these values when called. # Adding list element to our function circle_circumference_4 &lt;- function(radius){ # Calculating a circle&#39;s circumference and diameter based on the radius in inches # :parameters: radius # :output: list that contains diameter [1] and circumference [2] # Calculating diameter first diameter &lt;- 2 * radius # Calculating circumference circumference &lt;- pi * diameter # Storing results in a named list results &lt;- list(&quot;diameter&quot; = diameter, &quot;circumference&quot; = circumference) # Return results results } # Calling function circle_circumference_4(10) ## $diameter ## [1] 20 ## ## $circumference ## [1] 62.83185 We can also call the results individually using the following code: # Storing results of function circle_10 &lt;- circle_circumference_4(10) # Viewing only diameter ## Method 1 circle_10$diameter ## [1] 20 ## Method 2 circle_10[1] ## $diameter ## [1] 20 # Viewing only circumference ## Method 1 circle_10$circumference ## [1] 62.83185 ## Method 2 circle_10[2] ## $circumference ## [1] 62.83185 In the context of our dataset, we can use list operations to clean up and combine our results from all three BMI stratification approaches. This is often necessary to prepare data to share with collaborators or for supplementary tables in a manuscript. Let’s revisit our code for producing our statistical results, this time assigning our results to a dataframe rather than viewing them. # Defining variables (columns) we want to run a t-test on vars_of_interest &lt;- c(&quot;DWAs&quot;, &quot;DWCd&quot;, &quot;DWCr&quot;) # Normal vs. overweight (bmi_cutoff = 25) norm_vs_overweight &lt;- bmi_DW_ttest(input_data = full_data, bmi_cutoff = 25, lower_group_name = &quot;Normal&quot;, upper_group_name = &quot;Overweight&quot;, variables = vars_of_interest) # Underweight vs. non-underweight (bmi_cutoff = 18.5) under_vs_nonunderweight &lt;- bmi_DW_ttest(full_data, 18.5, &quot;Underweight&quot;, &quot;Non-Underweight&quot;, vars_of_interest) # Non-obese vs. obese (bmi_cutoff = 29.9) nonobese_vs_obese &lt;- bmi_DW_ttest(full_data, 29.9, &quot;Non-Obese&quot;, &quot;Obese&quot;, vars_of_interest) # Viewing one results dataframe as an example norm_vs_overweight ## .y. group1 group2 n1 n2 statistic df p ## 1 DWAs Normal Overweight 96 104 -0.7279621 192.3363 0.468 ## 2 DWCd Normal Overweight 96 104 -0.5894360 196.1147 0.556 ## 3 DWCr Normal Overweight 96 104 0.1102933 197.9870 0.912 For publication purposes, let’s say we want to make the following formatting changes: Keep only the comparison of interest (for example Normal vs. Overweight) and the associated p-value, removing columns that are not as useful for interpreting or sharing the results Rename the .y. column so that its contents are clearer Collapse all of our data into one final dataframe We can first write a function to execute these cleaning steps: # Function to clean results dataframes ## Parameters: ### input_data: dataframe containing results of t-test ## Output: cleaned dataframe data_cleaning &lt;- function(input_data) { data &lt;- input_data %&gt;% # Rename .y. column rename(&quot;Variable&quot; = &quot;.y.&quot;) %&gt;% # Merge group1 and group2 unite(Comparison, group1, group2, sep = &quot; vs. &quot;) %&gt;% # Keep only columns of interest select(c(Variable, Comparison, p)) return(data) } Then, we can make a list of our dataframes to clean and apply: # Making list of dataframes t_test_res_list &lt;- list(norm_vs_overweight, under_vs_nonunderweight, nonobese_vs_obese) # Viewing list of dataframes head(t_test_res_list) ## [[1]] ## .y. group1 group2 n1 n2 statistic df p ## 1 DWAs Normal Overweight 96 104 -0.7279621 192.3363 0.468 ## 2 DWCd Normal Overweight 96 104 -0.5894360 196.1147 0.556 ## 3 DWCr Normal Overweight 96 104 0.1102933 197.9870 0.912 ## ## [[2]] ## .y. group1 group2 n1 n2 statistic df p ## 1 DWAs Non-Underweight Underweight 166 34 -0.86947835 53.57143 0.388 ## 2 DWCd Non-Underweight Underweight 166 34 -0.97359810 55.45450 0.334 ## 3 DWCr Non-Underweight Underweight 166 34 0.04305105 56.08814 0.966 ## ## [[3]] ## .y. group1 group2 n1 n2 statistic df p ## 1 DWAs Non-Obese Obese 144 56 -1.9312097 86.80253 0.0567 ## 2 DWCd Non-Obese Obese 144 56 0.3431076 94.52209 0.7320 ## 3 DWCr Non-Obese Obese 144 56 -0.6878311 89.61818 0.4930 And we can apply the cleaning function to each of the dataframes using the lapply() function, which takes a list as the first argument and the function to apply to each list element as the second argument: # Applying cleaning function t_test_res_list_cleaned &lt;- lapply(t_test_res_list, data_cleaning) # Vieweing cleaned dataframes head(t_test_res_list_cleaned) ## [[1]] ## Variable Comparison p ## 1 DWAs Normal vs. Overweight 0.468 ## 2 DWCd Normal vs. Overweight 0.556 ## 3 DWCr Normal vs. Overweight 0.912 ## ## [[2]] ## Variable Comparison p ## 1 DWAs Non-Underweight vs. Underweight 0.388 ## 2 DWCd Non-Underweight vs. Underweight 0.334 ## 3 DWCr Non-Underweight vs. Underweight 0.966 ## ## [[3]] ## Variable Comparison p ## 1 DWAs Non-Obese vs. Obese 0.0567 ## 2 DWCd Non-Obese vs. Obese 0.7320 ## 3 DWCr Non-Obese vs. Obese 0.4930 Last, we can collapse our list down into one dataframe using the do.call() and rbind.data.frame() functions, which together, take the elements of the list and collapse them into a dataframe by binding the rows together: t_test_res_cleaned &lt;- do.call(rbind.data.frame, t_test_res_list_cleaned) # Viewing final dataframe t_test_res_cleaned ## Variable Comparison p ## 1 DWAs Normal vs. Overweight 0.4680 ## 2 DWCd Normal vs. Overweight 0.5560 ## 3 DWCr Normal vs. Overweight 0.9120 ## 4 DWAs Non-Underweight vs. Underweight 0.3880 ## 5 DWCd Non-Underweight vs. Underweight 0.3340 ## 6 DWCr Non-Underweight vs. Underweight 0.9660 ## 7 DWAs Non-Obese vs. Obese 0.0567 ## 8 DWCd Non-Obese vs. Obese 0.7320 ## 9 DWCr Non-Obese vs. Obese 0.4930 The above example is just that - an example to demonstrate the mechanics of using list operations. However, there are actually a couple of even more efficient ways to execute the above cleaning steps: Build cleaning steps into the analysis function if you know you will not need to access the raw results dataframe. Bind all three dataframes together, then execute the cleaning steps. We will demonstrate #2 below: # Start by binding the rows of each of the results dataframes t_test_res_cleaned_2 &lt;- bind_rows(norm_vs_overweight, under_vs_nonunderweight, nonobese_vs_obese) %&gt;% # Rename .y. column rename(&quot;Variable&quot; = &quot;.y.&quot;) %&gt;% # Merge group1 and group2 unite(Comparison, group1, group2, sep = &quot; vs. &quot;) %&gt;% # Keep only columns of interest select(c(Variable, Comparison, p)) # Viewing results t_test_res_cleaned_2 ## Variable Comparison p ## 1 DWAs Normal vs. Overweight 0.4680 ## 2 DWCd Normal vs. Overweight 0.5560 ## 3 DWCr Normal vs. Overweight 0.9120 ## 4 DWAs Non-Underweight vs. Underweight 0.3880 ## 5 DWCd Non-Underweight vs. Underweight 0.3340 ## 6 DWCr Non-Underweight vs. Underweight 0.9660 ## 7 DWAs Non-Obese vs. Obese 0.0567 ## 8 DWCd Non-Obese vs. Obese 0.7320 ## 9 DWCr Non-Obese vs. Obese 0.4930 As you can see, this dataframe is the same as the one we produced using list operations. It was produced using fewer lines of code and without the need for a user-defined function! For our purposes, this was a more efficient approach. However, we felt it was important to demonstrate the mechanics of list operations because there may be times where you do need to keep dataframes separate during specific analyses. Concluding Remarks This module provided an introduction to loops, functions, and list operations and demonstrated how to use them to efficiently analyze an environmentally relevant dataset. When and how you implement these approaches depends on your coding style and the goals of your analysis. Although here we were focused on statistical tests and data cleaning, these flexible approaches can be used in a variety of data analysis steps. We encourage you to implement loops, functions, and list operations in your analyses when you find the need to iterate through statistical tests, visualizations, data cleaning, or other common workflow elements! Additional Resources Intro2r Loops Intro2r Functions in R Hadley Wickham Advanced R - Functionals Test Your Knowledge Use the same input data we used in this module to answer the following questions and produce a cleaned, publication-ready data table of results. Note that these data are normally distributed, so you can use a t-test. Are there statistically significant differences in urine metal concentrations (ie. arsenic levels, cadmium levels, etc.) between younger (MAge &lt; 40) and older (MAge \\(\\geq\\) 40) mothers? Are there statistically significant differences in urine metal concentrations (ie. arsenic levels, cadmium levels, etc.) between between normal weight (BMI &lt; 25) and overweight (BMI \\(\\geq\\) 25) subjects? "],["data-visualizations.html", "Data Visualizations Introduction to Data Visualizations Introduction to Training Module Density Plot Visualization Boxplot Visualization Correlation Visualizations Heatmap Visualization Concluding Remarks", " Data Visualizations This training module was developed by Alexis Payton, Kyle Roell, Lauren E. Koval, Elise Hickman, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Data Visualizations Selecting an approach to visualize data is an important consideration when presenting scientific research, given that figures have the capability to summarize large amounts of data efficiently and effectively. (At least that’s the goal!) This module will focus on basic data visualizations that we view to be most commonly used, both in and outside of the field of environmental health research, many of which you have likely seen before. This module is not meant to be an exhaustive representation of all figure types, rather it serves as an introduction to some types of figures and how to approach choosing the one that most optimally displays your data and primary findings. When selecting a data visualization approach, here are some helpful questions you should first ask yourself: What message am I trying to convey with this figure? How does this figure highlight major findings from the paper? Who is the audience? What type of data am I working with? A Guide To Getting Data Visualization Right is a great resource for determining which figure is best suited for various types of data. More complex methodology-specific charts are presented in succeeding TAME modules. These include visualizations for: Two Group Comparisons (e.g.,boxplots and logistic regression) in Module 3.4 Introduction to Statistical Tests and Module 4.4 Two Group Comparisons and Visualizations Multi-Group Comparisons (e.g.,boxplots) in Module 3.4 Introduction to Statistical Tests and Module 4.5 Mult-Group Comparisons and Visualizations Supervised Machine Learning (e.g.,decision boundary plots, variable importance plots) in Module 5.3 Supervised ML Model Interpretation Unsupervised Machine Learning Principal Component Analysis (PCA) plots and heatmaps in Module 5.4 Unsupervised Machine Learning I: K-Means Clustering &amp; PCA Dendrograms, clustering visualizations, heatmaps, and variable contribution plots in Module 5.5 Unsupervised Machine Learning II: Additional Clustering Applications -Omics Expression (e.g.,MA plots and volcano plots) in Module 6.2 -Omics and Systems Biology: Transcriptomic Applications Mixtures Methods Forest Plots in Module 6.3 Mixtures I: Overview and Quantile G-Computation Application Trace Plots in Module 6.4 Mixtures II: BKMR Application Sufficient Similarity (e.g.,heatmaps, clustering) in Module 6.5 Mixtures III: Sufficient Similarity Toxicokinetic Modeling (e.g.,line graph, dose response) in Module 6.6 Toxicokinetic Modeling Introduction to Training Module Visualizing data is an important step in any data analysis, including those carried out in environmental health research. Often, visualizations allow scientists to better understand trends and patterns within a particular dataset under evaluation. Even after statistical analysis of a dataset, it is important to then communicate these findings to a wide variety of target audiences. Visualizations are a vital part of communicating complex data and results to target audiences. In this module, we highlight some figures that can be used to visualize larger, more high-dimensional datasets using figures that are more simple (but still relevant!) than methods presented later on in TAME. This training module specifically reviews the formatting of data in preparation of generating visualizations, scaling datasets, and then guides users through the generation of the following example data visualizations: Density plots Boxplots Correlation plots Heatmaps These visualization approaches are demonstrated using a large environmental chemistry dataset. This example dataset was generated through chemical speciation analysis of smoke samples collected during lab-based simulations of wildfire events. Specifically, different biomass materials (eucalyptus, peat, pine, pine needles, and red oak) were burned under two combustion conditions of flaming and smoldering, resulting in the generation of 12 different smoke samples. These data have been previously published in the following environmental health research studies, with data made publicly available: Rager JE, Clark J, Eaves LA, Avula V, Niehoff NM, Kim YH, Jaspers I, Gilmour MI. Mixtures modeling identifies chemical inducers versus repressors of toxicity associated with wildfire smoke. Sci Total Environ. 2021 Jun 25;775:145759. doi: 10.1016/j.scitotenv.2021.145759. Epub 2021 Feb 10. PMID: 33611182. Kim YH, Warren SH, Krantz QT, King C, Jaskot R, Preston WT, George BJ, Hays MD, Landis MS, Higuchi M, DeMarini DM, Gilmour MI. Mutagenicity and Lung Toxicity of Smoldering vs. Flaming Emissions from Various Biomass Fuels: Implications for Health Effects from Wildland Fires. Environ Health Perspect. 2018 Jan 24;126(1):017011. doi: 10.1289/EHP2200. PMID: 29373863. GGplot2 ggplot2 is a powerful package used to create graphics in R. It was designed based on the philosophy that every figure can be built using a dataset, a coordinate system, and a geom that specifies the type of plot. As a result, it is fairly straightforward to create highly customizable figures and is typically preferred over using base R to generate graphics. We’ll generate all of the figures in this module using ggplot2. For additional resources on ggplot2 see ggplot2 Posit Documentation and Data Visualization with ggplot2. Script Preparations Cleaning the global environment rm(list=ls()) Installing required R packages If you already have these packages installed, you can skip this step, or you can run the below code which checks installation status for you if (!requireNamespace(&quot;GGally&quot;)) install.packages(&quot;GGally&quot;); if (!requireNamespace(&quot;corrplot&quot;)) install.packages(&quot;corrplot&quot;); if (!requireNamespace(&quot;pheatmap&quot;)) install.packages(&quot;pheatmap&quot;); Loading R packages required for this session library(tidyverse) library(GGally) library(corrplot) library(reshape2) library(pheatmap) Set your working directory setwd(&quot;/filepath to where your input files are&quot;) Importing example dataset Then let’s read in our example dataset. As mentioned in the introduction, this example dataset represents chemical measurements across 12 different biomass burn scenarios representing potential wildfire events. Let’s upload and view these data: # Load the data smoke_data &lt;- read.csv(&quot;Module3_1_Input/Module3_1_InputData.csv&quot;) # View the top of the dataset head(smoke_data) ## Chemical.Category Chemical CASRN Eucalyptus_Smoldering ## 1 n-Alkanes 2-Methylnonadecane 1560-86-7 0.06 ## 2 n-Alkanes 3-Methylnonadecane 6418-45-7 0.04 ## 3 n-Alkanes Docosane 629-97-0 0.21 ## 4 n-Alkanes Dodecylcyclohexane 1795-17-1 0.04 ## 5 n-Alkanes Eicosane 112-95-8 0.11 ## 6 n-Alkanes Heneicosane 629-94-7 0.13 ## Eucalyptus_Flaming Peat_Smoldering Peat_Flaming Pine_Smoldering Pine_Flaming ## 1 0.06 1.36 0.06 0.06 0.06 ## 2 0.04 1.13 0.90 0.47 0.04 ## 3 0.25 9.46 0.57 0.16 0.48 ## 4 0.04 0.25 0.04 0.04 0.04 ## 5 0.25 7.55 0.54 0.17 0.29 ## 6 0.28 6.77 0.34 0.13 0.42 ## Pine_Needles_Smoldering Pine_Needles_Flaming Red_Oak_Smoldering ## 1 0.06 0.06 0.06 ## 2 0.04 0.72 0.04 ## 3 0.32 0.18 0.16 ## 4 0.12 0.04 0.04 ## 5 0.28 0.16 0.15 ## 6 0.30 0.13 0.13 ## Red_Oak_Flaming Units ## 1 0.13 ng_per_uL ## 2 0.77 ng_per_uL ## 3 0.36 ng_per_uL ## 4 0.04 ng_per_uL ## 5 0.38 ng_per_uL ## 6 0.69 ng_per_uL Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: How do the distributions of the chemical concentration data differ based on each biomass burn scenario? Are there correlations between biomass burn conditions based on the chemical concentration data? Under which biomass burn conditions are concentrations of certain chemical categories the highest? We can create a density plot to answer the first question. Similar to a histogram, density plots are an effective way to show overall distributions of data and can be useful to compare across various test conditions or other stratifications of the data. In this example of a density plot, we’ll visualize the distributions of chemical concentration data on the x axis. A density plot automatically displays where values are concentrated on the y axis. Additionally, we’ll want to have multiple density plots within the same figure for each biomass burn condition. Before the data can be visualized, it needs to be converted from a wide to long format. This is because we need to have variable or column names entitled Chemical_Concentration and Biomass_Burn_Condition that can be placed into ggplot(). For review on converting between long and wide formats and using other tidyverse tools, see TAME 2.0 Module 2.3 Data Manipulation &amp; Reshaping. longer_smoke_data = pivot_longer(smoke_data, cols = 4:13, names_to = &quot;Biomass_Burn_Condition&quot;, values_to = &quot;Chemical_Concentration&quot;) head(longer_smoke_data) ## # A tibble: 6 × 6 ## Chemical.Category Chemical CASRN Units Biomass_Burn_Condition ## &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; ## 1 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus_Smoldering ## 2 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus_Flaming ## 3 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat_Smoldering ## 4 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat_Flaming ## 5 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine_Smoldering ## 6 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine_Flaming ## # ℹ 1 more variable: Chemical_Concentration &lt;dbl&gt; Scaling dataframes for downstream data visualizations A data preparation method that is commonly used to convert values into those that can be used to better illustrate overall data trends is data scaling. Scaling can be achieved through data transformations or normalization procedures, depending on the specific dataset and goal of analysis/visualization. Scaling is often carried out using data vectors or columns of a dataframe. For this example, we will normalize the chemical concentration dataset using a basic scaling and centering procedure using the base R function, scale(). This algorithm results in the normalization of a dataset using the mean value and standard deviation. This scaling step will convert chemical concentration values in our dataset into normalized values across samples, such that each chemical’s concentration distributions are more easily comparable between the different biomass burn conditions. For more information on the scale() function, see its associated RDocumentation and helpful tutorial on Implementing the scale() function in R. scaled_longer_smoke_data = longer_smoke_data %&gt;% # scaling within each chemical group_by(Chemical) %&gt;% mutate(Scaled_Chemical_Concentration = scale(Chemical_Concentration)) %&gt;% ungroup() head(scaled_longer_smoke_data) # see the new scaled values now in the last column (column 7) ## # A tibble: 6 × 7 ## Chemical.Category Chemical CASRN Units Biomass_Burn_Condition ## &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; ## 1 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus_Smoldering ## 2 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus_Flaming ## 3 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat_Smoldering ## 4 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat_Flaming ## 5 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine_Smoldering ## 6 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine_Flaming ## # ℹ 2 more variables: Chemical_Concentration &lt;dbl&gt;, ## # Scaled_Chemical_Concentration &lt;dbl[,1]&gt; We can see that in the Scaled_Chemical_Concentration column, each chemical is scaled based on a normal distribution centered around 0, with values now less than or greater than zero. Now that we have our dataset formatted, let’s plot it. Density Plot Visualization The following code can be used to generate a density plot: ggplot(scaled_longer_smoke_data, aes(x = Scaled_Chemical_Concentration, color = Biomass_Burn_Condition)) + geom_density() Answer to Environmental Health Question 1, Method I With this method, we can answer Environmental Health Question #1: How do the distributions of the chemical concentration data differ based on each biomass burn scenario? Answer: In general, there are a high number of chemicals that were measured at relatively lower abundances across all smoke samples (hence, the peak in occurrence density occurring towards the left, before 0). The three conditions of smoldering peat, flaming peat, and flaming pine contained the most chemicals at the highest relative concentrations (hence, these lines are the top three lines towards the right). Boxplot Visualization A boxplot can also be used to answer our first environmental health question: How do the distributions of the chemical concentration data differ based on each biomass burn scenario?. A boxplot also displays a data’s distribution, but it incorporates a visualization of a five number summary (i.e., minimum, first quartile, median, third quartile, and maximum). Any outliers are displayed as dots. For this example, let’s have Scaled_Chemical_Concentration on the x axis and Biomass_Burn_Condition on the y axis. The scaled_longer_smoke_data dataframe is the format we need, so we’ll use that for plotting. ggplot(scaled_longer_smoke_data, aes(x = Scaled_Chemical_Concentration, y = Biomass_Burn_Condition, color = Biomass_Burn_Condition)) + geom_boxplot() Answer to Environmental Health Question 1, Method II With this alternative method, we can answer, in a different way, Environmental Health Question #1: How do the distributions of the chemical concentration data differ based on each biomass burn scenario? Answer, Method II: The median chemical concentration is fairly low (less than 0) for all biomass burn conditions. Overall, there isn’t much variation in chemical concentrations with the exception of smoldering peat, flaming peat, and flaming eucalyptus. Correlation Visualizations Let’s turn our attention to the second environmental health question: Are there correlations between biomass burn conditions based on the chemical concentration data? We’ll use two different correlation visualizations to answer this question using the GGally package. GGally is a package that serves as an extension of ggplot2, the baseline R plotting system based on the grammar of graphics. GGally is very useful for creating plots that compare groups or features within a dataset, among many other utilities. Here we will demonstrate the ggpairs() function within GGally using the scaled chemistry dataset. This function will produce an image that shows correlation values between biomass burn sample pairs and also illustrates the overall distributions of values in the samples. For more information on GGally, see its associated RDocumentation and example helpful tutorial. GGally requires a wide dataframe with ids (i.e.,Chemical) as the rows and the variables that will be compared to each other (i.e.,Biomass_Burn_Condition) as the columns. Let’s create that dataframe. # first selecting the chemical, biomass burn condition, and # the scaled chemical concentration columns wide_scaled_data = scaled_longer_smoke_data %&gt;% pivot_wider(id_cols = Chemical, names_from = &quot;Biomass_Burn_Condition&quot;, values_from = &quot;Scaled_Chemical_Concentration&quot;) %&gt;% # converting the chemical names to row names column_to_rownames(var = &quot;Chemical&quot;) head(wide_scaled_data) ## Eucalyptus_Smoldering Eucalyptus_Flaming Peat_Smoldering ## 2-Methylnonadecane -0.3347765 -0.3347765 2.841935 ## 3-Methylnonadecane -0.8794448 -0.8794448 1.649829 ## Docosane -0.3465132 -0.3327216 2.842787 ## Dodecylcyclohexane -0.4240624 -0.4240624 2.646734 ## Eicosane -0.3802202 -0.3195928 2.841691 ## Heneicosane -0.3895328 -0.3166775 2.835527 ## Peat_Flaming Pine_Smoldering Pine_Flaming ## 2-Methylnonadecane -0.3347765 -0.3347765 -0.3347765 ## 3-Methylnonadecane 1.1161291 0.1183422 -0.8794448 ## Docosane -0.2223890 -0.3637526 -0.2534201 ## Dodecylcyclohexane -0.4240624 -0.4240624 -0.4240624 ## Eicosane -0.1940076 -0.3542370 -0.3022707 ## Heneicosane -0.2875354 -0.3895328 -0.2486793 ## Pine_Needles_Smoldering Pine_Needles_Flaming ## 2-Methylnonadecane -0.3347765 -0.3347765 ## 3-Methylnonadecane -0.8794448 0.6984509 ## Docosane -0.3085863 -0.3568568 ## Dodecylcyclohexane 0.7457649 -0.4240624 ## Eicosane -0.3066012 -0.3585675 ## Heneicosane -0.3069635 -0.3895328 ## Red_Oak_Smoldering Red_Oak_Flaming ## 2-Methylnonadecane -0.3347765 -0.1637228 ## 3-Methylnonadecane -0.8794448 0.8144726 ## Docosane -0.3637526 -0.2947948 ## Dodecylcyclohexane -0.4240624 -0.4240624 ## Eicosane -0.3628981 -0.2632960 ## Heneicosane -0.3895328 -0.1175398 By default, ggpairs() displays Pearson’s correlations. To show Spearman’s correlations takes more nuance, but can be done using the code that has been commented out below. # ggpairs with Pearson&#39;s correlations wide_scaled_data = data.frame(as.matrix(wide_scaled_data)) ggpairs(wide_scaled_data) # ggpairs with Spearman&#39;s correlations # pearson_correlations = cor(wide_scaled_data, method = &quot;spearman&quot;) # ggpairs(wide_scaled_data, upper = list(continuous = wrap(ggally_cor, method = &quot;spearman&quot;))) Many of these biomass burn conditions have significant correlations denoted by the asterisks. ’*’: p value &lt; 0.1 ’**’: p value &lt; 0.05 ’***’: p value &lt; 0.01 The upper right portion displays the correlation values, where a value less than 0 indicates negative correlation and a value greater than 0 signifies positive correlation. The diagonal shows the density plots for each variable. The lower left portion visualizes the values of the two variables compared using a scatterplot. Answer to Environmental Health Question 2 With this, we can answer Environmental Health Question #2: Are there correlations between biomass burn conditions based on the chemical concentration data? Answer: There is low correlation between many of the variables (-0.5 &lt; correlation value &lt; 0.5). Eucalyptus flaming and pine flaming are significantly positively correlated along with peat flaming and pine needles flaming (correlation value ~0.7 and p value &lt; 0.001). We can visualize correlations another way using the other function from GGally, ggcorr(), which visualizes each correlation as a square. Note that this function calculates Pearson’s correlations by default. However, this can be changed using the method parameter shown in the code commented out below. # Pearson&#39;s correlations ggcorr(wide_scaled_data) # Spearman&#39;s correlations # ggcorr(wide_scaled_data, method = &quot;spearman&quot;) We’ll visualize correlations between each of the groups using one more figure using the corrplot() function from the corrplot package. # Need to supply corrplot with a correlation matrix, here, using the &#39;cor&#39; function corrplot(cor(wide_scaled_data)) Each of these correlation figures displays the same information, but the one you choose to use is a matter of personal preference. Click on the following resources for additional information on ggpairs() and corrplot(). Heatmap Visualization Last, we’ll turn our attention to answering the final environmental health question: Under which biomass burn conditions are concentrations of certain chemical categories the highest? This can be addressed with the help of a heatmap. Heatmaps are a highly effective method of viewing an entire dataset at once. Heatmaps can appear similar to correlation plots, but typically illustrate other values (e.g., concentrations, expression levels, presence/absence, etc) besides correlation values. They are used to draw patterns between two variables of highest interest (that comprise the x and y axis, though additional bars can be added to display other layers of information). In this instance, we’ll use a heatmap to determine whether there are patterns apparent between chemical categories and biomass burn condition on chemical concentrations. For this example, we can plot Biomass_Burn_Condition and Chemical.Category on the axes and fill in the values with Scaled_Chemical_Concentration. When generating heatmaps, scaled values are often used to better distinguish patterns between groups/samples. In this example, we also plan to display the median scaled concentration value within the heatmap as an additional layer of helpful information to aid in interpretation. To do so, we’ll need to take the median chemical concentration for each biomass burn condition within each chemical category. However, since we want ggplot() to visualize the median scaled values with the color of the tiles this step was already necessary. # We&#39;ll find the median value and add that data to the dataframe as an additional column heatmap_df = scaled_longer_smoke_data %&gt;% group_by(Biomass_Burn_Condition, Chemical.Category) %&gt;% mutate(Median_Scaled_Concentration = median(Scaled_Chemical_Concentration)) head(heatmap_df) ## # A tibble: 6 × 8 ## # Groups: Biomass_Burn_Condition, Chemical.Category [6] ## Chemical.Category Chemical CASRN Units Biomass_Burn_Condition ## &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; ## 1 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus_Smoldering ## 2 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus_Flaming ## 3 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat_Smoldering ## 4 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat_Flaming ## 5 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine_Smoldering ## 6 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine_Flaming ## # ℹ 3 more variables: Chemical_Concentration &lt;dbl&gt;, ## # Scaled_Chemical_Concentration &lt;dbl[,1]&gt;, Median_Scaled_Concentration &lt;dbl&gt; Now we can plot the data and add the Median_Scaled_Concentration to the figure using geom_text(). Note that specifying the original Scaled_Chemical_Concentration in the fill parameter will NOT give you the same heatmap as specifying the median values in ggplot(). ggplot(data = heatmap_df, aes(x = Chemical.Category, y = Biomass_Burn_Condition, fill = Median_Scaled_Concentration)) + geom_tile() + # function used to specify a heatmap for ggplot geom_text(aes(label = round(Median_Scaled_Concentration, 2))) # adding concentration values as text, rounding to two values after the decimal Answer to Environmental Health Question 3 With this, we can answer Environmental Health Question #3: Under which biomass burn conditions are concentrations of certain chemical categories the highest? Answer: Peat flaming has the highest concentrations of inorganics and ions. Eucalyptus smoldering has the highest concentrations of levoglucosans. Pine smoldering has the highest concentrations of methoxyphenols. Peat smoldering has the highest concentrations of n-alkanes. Pine needles smoldering has highest concentrations of PAHs. This same heatmap can be achieved another way using the pheatmap() function from the pheatmap package. Using this function requires us to use a wide dataset, which we need to create. It will contain Chemical.Category, Biomass_Burn_Condition and Scaled_Chemical_Concentration. heatmap_df2 = scaled_longer_smoke_data %&gt;% group_by(Biomass_Burn_Condition, Chemical.Category) %&gt;% # using the summarize function instead of mutate function as was done previously since we only need the median values now summarize(Median_Scaled_Concentration = median(Scaled_Chemical_Concentration)) %&gt;% # transforming the data to a wide format pivot_wider(id_cols = Biomass_Burn_Condition, names_from = &quot;Chemical.Category&quot;, values_from = &quot;Median_Scaled_Concentration&quot;) %&gt;% # converting the chemical names to row names column_to_rownames(var = &quot;Biomass_Burn_Condition&quot;) head(heatmap_df2) ## Inorganics Ions Levoglucosan Methoxyphenols ## Eucalyptus_Flaming 0.05405359 0.05273246 0.4208870 -0.44781893 ## Eucalyptus_Smoldering -0.68595076 -0.80160192 1.7772753 -0.06449444 ## Peat_Flaming 2.24332901 1.77515899 -0.9383328 -0.51488738 ## Peat_Smoldering -0.51860591 -0.36146158 -0.8041211 0.05720971 ## Pine_Flaming -0.02063532 -0.05999543 -0.1992054 -0.50269422 ## Pine_Needles_Flaming 0.36405527 0.82229035 -0.8570130 -0.46331332 ## PAHs n-Alkanes ## Eucalyptus_Flaming 1.2885776 -0.3790357 ## Eucalyptus_Smoldering -0.4724635 -0.3465132 ## Peat_Flaming -0.5369746 -0.3093608 ## Peat_Smoldering -0.3162278 2.8238921 ## Pine_Flaming 1.7825403 -0.3347765 ## Pine_Needles_Flaming -0.4179505 -0.3850613 Now let’s generate the same heatmap this time using the pheatmap() function: pheatmap(heatmap_df2, # removing the clustering option from both rows and columns cluster_rows = FALSE, cluster_cols = FALSE, # adding the values for each cell, making those values black, and changing the font size display_numbers = TRUE, number_color = &quot;black&quot;, fontsize = 12) Notice that the pheatmap() function does not include axes or legend titles as with ggplot(), however those can be added to the figure after exporting from R in MS Powerpoint or Adobe. Additional parameters, including cluster_rows, for the pheatmap() function are discussed further in TAME 2.0 Module 5.4 Unsupervised Machine Learning. For basic heatmaps like the ones shown here, ggplot() or pheatmap() can both be used however, both have their pros and cons. For example, ggplot() figures tend to be more customizable and easily combined with other figures, while pheatmap() has additional parameters built into the function that can make plotting certain features advantageous like clustering. Concluding Remarks In conclusion, this training module provided example code to create highly customizable data visualizations using ggplot2 pertinent to environmental health research. Test Your Knowledge Replicate the figure below! The heatmap still visualizes the median chemical concentrations, but this time we’re separating the burn conditions, allowing us to determine if the concentrations of chemicals released are contingent upon the burn condition. For additional figures available and to view aspects of figures that can be changed in GGplot2, check out this GGPlot2 Cheat Sheet. You might need it to make this figure! Hint 1: Use the separate() function from tidyverse to split Biomass_Burn_Condition into Biomass and Burn_Condition. Hint 2: Use the function facet_wrap() within ggplot() to separate the heatmaps by Burn_Condition. "],["improving-data-visualizations.html", "Improving Data Visualizations Introduction to Data Visulization Conventions Introduction to Training Module Creating an Improved Boxplot Visualization Creating an Improved Heatmap Visualization Creating Multi-Plot Figures Concluding Remarks", " Improving Data Visualizations This training module was developed by Alexis Payton, Elise Hickman, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Data Visulization Conventions Data visualizations are used to convey key takeaways from a research study’s findings in a clear, succinct manner to highlight data trends, patterns, and/or relationships. In environmental health research, this is of particular importance for high-dimensional datasets that can typically be parsed using multiple methods, potentially resulting in many different approaches to visualize data. As a consequence, researchers are often faced with an overwhelming amount of options when deciding which visualization scheme(s) most optimally translate their results for effective dissemination. Effective data visualization approaches are vital to a researcher’s success for many reasons. For instance, manuscript readers or peer reviewers often scroll through a study’s text and focus on the quality and novelty of study figures before deciding whether to read/review the paper. Therefore, the importance of data visualizations cannot be understated in any research field. As a high-level introduction, it is important that we first communicate some traits that we think are imperative towards ensuring a successful data visualization approach as described in more detail below. Keys to successful data visualizations: Consider your audience, data type, and research question prior to selecting a figure to visualize your data For example, if more computationally complex methods are used in a manuscript that is intended for a journal with an audience that doesn’t have that same level of expertise, consider spending time focusing on how those results are presented in an approachable way for that audience. For a review of how to choose a rudimentary chart based on the data type, check out How to Choose the Right Data Visualization. Some of these basic charts will be presented in this module, while more complex analysis-specific visualizations, especially ones developed for high-dimensional data will be presented in later modules. Take the legibility of the figure into account This includes avoiding abbreviations when possible. (If they can’t be avoided explain them in the caption.) All titles should be capitalized, including titles for the legend(s) and axes. Underscores and periods between words should be replaced with spaces. Consider the legibility of the figure if printed in black and white. (However, that’s not as important these days.) Lastly, feel free to describe your plot in further detail in the caption to aid the reader in understanding the results presented. Minimize text Main titles aren’t necessary for single paneled figures (like the examples below), because in a publication the title of the figure is right underneath each figure. It’s good practice to remove this kind of extraneous text, which can make the figure seem more cluttered. Titles can be helpful in multi-panel figures, especially if there are multiple panels with the same figure type that present slightly different results. For example, in the Test Your Knowledge section, you’ll need to create two heatmaps, but one displays data under smoldering conditions and the other displays data under flaming conditions. In general, try to reduce the amount of extraneous text in a plot to keep a reader focused on the most important elements and takeaways in the plot. Use the minimal number of figures you need to support your narrative It is important to include an optimal number of figures within manuscripts and scientific reports. Too many figures might overwhelm the overall narrative, while too few might not provide enough substance to support your main findings. It can be helpful to also consider placing some figures in supplemental material to aid in the overall flow of your scientific writing. Select an appropriate color palette Packages have been developed to offer color palettes including MetBrewer and RColorBrewer. In addition, ggsci is a package that offers a collection of color palettes used in various scientific journals. For more information, check out MetBrewer, see its associated RDocumentation and example tutorial. For more information on RColorBrewer, see its associated RDocumentation and example tutorial. For more information on ggsci, see its associated RDocumentation. In general, it’s better to avoid bright and flashy colors that can be difficult to read. It’s advisable to use colors for manuscript figures that are color-blind friendly. Check out these Stack overflow answers about color blind-safe color palettes and packages. Popular packages for generating colorblind-friendly palettes include viridis and rcartocolor. Use color strategically Color can be used to visualize a variable. There are three ways to categorize color schemes - sequential, diverging, and qualitative. Below, definitions are provided for each along with example figures that we’ve previously published that illustrate each color scheme. In addition, figure titles and captions are also provided for context. Note that some of these figures have been simplified from what was originally published to show more streamlined examples for TAME. Sequential: intended for ordered categorical data (i.e., disease severity, likert scale, quintiles). The choropleth map below is from Winker, Payton et. al. Figure 2: Figure 1. Geospatial distribution of the risk of future wildfire events across North Carolina. Census tracts in North Carolina were binned into quintiles based on Wildfire Hazard Potential (WHP) with 1 (pale orange) having the lowest risk and 5 (dark red) having the highest risk. Figure regenerated here in alignment with its published Creative Commons Attribution 4.0 International License Diverging: intended to emphasize continuous data at extremes of the data range (typically using darker colors) and mid-range values (typically using lighter colors). This color scheme is ideal for charts like heatmaps. The heatmap below is from Payton, Perryman et. al. Figure 3: Figure 6. Individual cytokine expression levels across all subjects. Cytokine concentrations were derived from nasal lavage fluid samples. On the x axis, subjects were ordered first according to tobacco use status, starting with non-smokers then cigarette smokers and e-cigarette users. Within tobacco use groups, subjects are ordered from lowest to highest average cytokine concentration from left to right. Within each cluster shown on the y axis, cytokines are ordered from lowest to highest average cytokine concentration from bottom to top. Figure regenerated here in alignment with its published Creative Commons Attribution 4.0 International License Qualitative: intended for nominal categorical data to visualize clear differences between groups (i.e., soil types and exposure groups). The dendrogram below is from Koval et. al. Figure 4: Figure 2. Translating chemical use inventory data to inform human exposure patterning. Groups A-I illustrate the identified clusters of exposure source categories. Figure regenerated here in alignment with its published Creative Commons Attribution 4.0 International License Consider ordering axes to reveal patterns relevant to the research questions Ordering the axes can reveal potential patterns that may not be clear in the visualization otherwise. In the cytokine expression heatmap above, there are not clear differences in cytokine expression across the tobacco use groups. However, e-cigarette users seem to have slightly more muted responses compared to non-smokers and cigarette smokers in clusters B and C, which was corroborated in subsequent statistical analyses. It is also evident that Cluster A had the lowest cytokine concentrations, followed by Cluster B, and then Cluster C with the greatest concentrations. What makes these figures so compelling is how the aspects introduced above were thoughtfully incorporated. In the next section, we’ll put those principles into practice using data that were described and referenced previously in TAME 2.0 Module 3.1 Data Visualizations. Introduction to Training Module In this module, ggplot2, R’s data visualization package will be used to walk through ways to improve data visualizations. We’ll recreate two figures (i.e., the boxplot and heatmap) constructed previously in TAME 2.0 Module 3.1 Data Visualizations and improve them so they are publication-ready. Additionally, we’ll write figure titles and captions to contextualize the results presented for each visualization. When writing figure titles and captions, it is helpful to address the research question or overall concept that the figure seeks to capture rather than getting into the weeds of specific methods the plot is based on. This is especially important when visualizing more complex methods that your audience might not have as much knowledge on. Script Preparations Cleaning the global environment rm(list=ls()) Installing required R packages If you already have these packages installed, you can skip this step, or you can run the below code which checks installation status for you if (!requireNamespace(&quot;MetBrewer&quot;)) install.packages(&quot;MetBrewer&quot;); if (!requireNamespace(&quot;RColorBrewer&quot;)) install.packages(&quot;RColorBrewer&quot;); if (!requireNamespace(&quot;pheatmap&quot;)) install.packages(&quot;pheatmap&quot;); if (!requireNamespace(&quot;cowplot&quot;)) install.packages(&quot;cowplot&quot;); Loading required R packages library(tidyverse) library(MetBrewer) library(RColorBrewer) library(pheatmap) library(cowplot) Set your working directory setwd(&quot;/filepath to where your input files are&quot;) Importing example dataset Let’s now read in our example dataset. As mentioned in the introduction, this example dataset represents chemical measurements across 12 different biomass burn scenarios, representing chemicals emitted during potential wildfire events. Let’s upload and view these data: # Load the data smoke_data &lt;- read.csv(&quot;Module3_2_Input/Module3_2_InputData.csv&quot;) # View the top of the dataset head(smoke_data) ## Chemical.Category Chemical CASRN Eucalyptus_Smoldering ## 1 n-Alkanes 2-Methylnonadecane 1560-86-7 0.06 ## 2 n-Alkanes 3-Methylnonadecane 6418-45-7 0.04 ## 3 n-Alkanes Docosane 629-97-0 0.21 ## 4 n-Alkanes Dodecylcyclohexane 1795-17-1 0.04 ## 5 n-Alkanes Eicosane 112-95-8 0.11 ## 6 n-Alkanes Heneicosane 629-94-7 0.13 ## Eucalyptus_Flaming Peat_Smoldering Peat_Flaming Pine_Smoldering Pine_Flaming ## 1 0.06 1.36 0.06 0.06 0.06 ## 2 0.04 1.13 0.90 0.47 0.04 ## 3 0.25 9.46 0.57 0.16 0.48 ## 4 0.04 0.25 0.04 0.04 0.04 ## 5 0.25 7.55 0.54 0.17 0.29 ## 6 0.28 6.77 0.34 0.13 0.42 ## Pine_Needles_Smoldering Pine_Needles_Flaming Red_Oak_Smoldering ## 1 0.06 0.06 0.06 ## 2 0.04 0.72 0.04 ## 3 0.32 0.18 0.16 ## 4 0.12 0.04 0.04 ## 5 0.28 0.16 0.15 ## 6 0.30 0.13 0.13 ## Red_Oak_Flaming Units ## 1 0.13 ng_per_uL ## 2 0.77 ng_per_uL ## 3 0.36 ng_per_uL ## 4 0.04 ng_per_uL ## 5 0.38 ng_per_uL ## 6 0.69 ng_per_uL Now that we’ve been able to view the dataset, let’s come up with questions that can be answered with our boxplot and heatmap figure. This will inform how we format the dataframe for visualization. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: Boxplot: How do the distributions of the chemical concentration data differ based on each biomass burn scenario? Heatmap: Which classes of chemicals show the highest concentrations across the evaluated biomass burn conditions? How can these figures be combined into a single plot that can be then be exported from R? Formatting dataframes for downstream visualization code First, format the dataframe by changing it from a wide to long format and normalizing the chemical concentration data. For more details on this data reshaping visit TAME 2.0 Module 2.3 Data Manipulation &amp; Reshaping. scaled_longer_smoke_data = pivot_longer(smoke_data, cols = 4:13, names_to = &quot;Biomass_Burn_Condition&quot;, values_to = &quot;Chemical_Concentration&quot;) %&gt;% # scaling within each chemical group_by(Chemical) %&gt;% mutate(Scaled_Chemical_Concentration = scale(Chemical_Concentration)) %&gt;% ungroup() head(scaled_longer_smoke_data) ## # A tibble: 6 × 7 ## Chemical.Category Chemical CASRN Units Biomass_Burn_Condition ## &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; ## 1 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus_Smoldering ## 2 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus_Flaming ## 3 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat_Smoldering ## 4 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat_Flaming ## 5 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine_Smoldering ## 6 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine_Flaming ## # ℹ 2 more variables: Chemical_Concentration &lt;dbl&gt;, ## # Scaled_Chemical_Concentration &lt;dbl[,1]&gt; Creating an Improved Boxplot Visualization As we did in the previous module, a boxplot will be constructed to answer the first environmental heath question: How do the distributions of the chemical concentration data differ based on each biomass burn scenario?. Let’s remind ourselves of the original figure from the previous module. Based on the figure above, peat smoldering has the highest median scaled chemical concentration. However, this was difficult to determine given that the burn conditions aren’t labeled on the x axis and a sequential color palette was used, making it difficult to identify the correct boxplot with its burn condition in the legend. If you look closely, the colors in the legend are in a reverse order of the colors assigned to the boxplots. Let’s identify some elements of this graph that can be modified to make it easier to answer our research question. There are four main aspects we can adjust on this figure: 1. The legibility of the text in the legend and axes. Creating spaces between the text or exchanging the underscores for spaces improves the legibility of the figure. 2. The order of the boxplots. Ordering the biomass burn conditions from highest to lowest based on their median scaled chemical concentration allows the reader to easily determine the biomass burn condition that had the greatest or least chemical concentrations relative to each other. In R, this can be done by putting the Biomass_Burn_Condition variable into a factor. 3. Use of color. Variables can be visualized using color, text, size, etc. In this figure, it is redundant to have the biomass burn condition encoded in the legend and the color. Instead this variable can be put on the y axis and the legend will be removed to be more concise. The shades of the colors will also be changed, but to keep each burn condition distinct from each other, colors will be chosen that are distinct from one another. Therefore, we will choose a qualitative color scheme. 4. Show all data points when possible. Many journals now require that authors report every single value when making data visualizations, particularly for small n studies using bar graphs and boxplots to show results. Instead of just displaying the mean/median and surrounding data range, it is advised to show how every replicate landed in the study range when possible. Note that this requirement is not feasible for studies with larger sample sizes though should be considered for smaller in vitro and animal model studies. Let’s start with addressing #1: Legibility of Axis Text. The legend title and axis titles can easily be changed with ggplot(), so that will be done later. To remove the underscore from the Biomass_Burn_Condition column, we can use the function gsub(), which will replace all of the underscores with spaces, resulting in a cleaner-looking graph. # First adding spaces between the biomass burn conditions scaled_longer_smoke_data = scaled_longer_smoke_data %&gt;% mutate(Biomass_Burn_Condition = gsub(&quot;_&quot;, &quot; &quot;, Biomass_Burn_Condition)) # Viewing dataframe head(scaled_longer_smoke_data) ## # A tibble: 6 × 7 ## Chemical.Category Chemical CASRN Units Biomass_Burn_Condition ## &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; ## 1 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus Smoldering ## 2 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus Flaming ## 3 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat Smoldering ## 4 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat Flaming ## 5 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine Smoldering ## 6 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine Flaming ## # ℹ 2 more variables: Chemical_Concentration &lt;dbl&gt;, ## # Scaled_Chemical_Concentration &lt;dbl[,1]&gt; #2. Reordering the boxplots based on the median scaled chemical concentration. After calculating the median scaled chemical concentration for each biomass burn condition, the new dataframe will be arranged from lowest to highest median scaled concentration from the top of the dataframe to the bottom. This order will be saved in a vector, median_biomass_order. Although the biomass burn conditions are saved from lowest to highest concentration, ggplot() will plot them in reverse order with the highest concentration at the top and the lowest at the bottom of the y axis. Axis reordering can also be accomplished using reorder within the ggplot() function as described here and here. median_biomass = scaled_longer_smoke_data %&gt;% group_by(Biomass_Burn_Condition) %&gt;% summarize(Median_Concentration = median(Scaled_Chemical_Concentration)) %&gt;% # arranges dataframe from lowest to highest from top to bottom arrange(Median_Concentration) head(median_biomass) ## # A tibble: 6 × 2 ## Biomass_Burn_Condition Median_Concentration ## &lt;chr&gt; &lt;dbl&gt; ## 1 Red Oak Smoldering -0.459 ## 2 Eucalyptus Smoldering -0.451 ## 3 Pine Smoldering -0.424 ## 4 Pine Needles Smoldering -0.370 ## 5 Pine Needles Flaming -0.350 ## 6 Red Oak Flaming -0.337 # Saving that order median_biomass_order = median_biomass$Biomass_Burn_Condition # Putting into factor to organize the burn conditions scaled_longer_smoke_data$Biomass_Burn_Condition = factor(scaled_longer_smoke_data$Biomass_Burn_Condition, levels = median_biomass_order) # Final dataframe to be used for plotting head(scaled_longer_smoke_data) ## # A tibble: 6 × 7 ## Chemical.Category Chemical CASRN Units Biomass_Burn_Condition ## &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;fct&gt; ## 1 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus Smoldering ## 2 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus Flaming ## 3 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat Smoldering ## 4 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat Flaming ## 5 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine Smoldering ## 6 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine Flaming ## # ℹ 2 more variables: Chemical_Concentration &lt;dbl&gt;, ## # Scaled_Chemical_Concentration &lt;dbl[,1]&gt; Now that the dataframe has been finalized, we can plot the new boxplot. The final revision, #3: Making Use of Color, will be addressed with ggplot(). However, a palette can be chosen from the MetBrewer package. # Choosing the &quot;Jurarez&quot; palette from the `MetBrewer` package # `n = 12`, since there are 12 biomass burn conditions juarez_colors = met.brewer(name = &quot;Juarez&quot;, n = 12)[1:12] #4. Show all data points when possible will also be addressed with ggplot() by simply using geom_point(). FigureX1 = ggplot(scaled_longer_smoke_data, aes(x = Scaled_Chemical_Concentration, y = Biomass_Burn_Condition, color = Biomass_Burn_Condition)) + geom_boxplot() + # jittering the points, so they&#39;re not all on top of each other and adding transparency geom_point(position = position_jitter(h = 0.1), alpha = 0.7) + theme_light() + # changing the theme theme(axis.text = element_text(size = 9), # changing size of axis labels axis.title = element_text(face = &quot;bold&quot;, size = rel(1.5)), # changes axis titles legend.position = &quot;none&quot;) + # removes legend xlab(&#39;Scaled Chemical Concentration (pg/uL)&#39;) + ylab(&#39;Biomass Burn Condition&#39;) + # changing axis labels scale_color_manual(values = c(juarez_colors)) # changing the colors FigureX1 An appropriate title for this figure could be: “Figure X. Chemical concentration distributions of biomass burn conditions. The boxplots are based on the scaled chemical concentration values, which used the raw chemical concentrations values scaled within each chemical. The individual dots represent the concentrations of each chemical. The biomass burn conditions on the y axis are ordered from greatest (top) to least (bottom) based on median scaled chemical concentration.” Answer to Environmental Health Question 1 With this, we can answer Environmental Health Question #1: Which biomass burn condition has the highest total chemical concentration? Answer: Smoldering peat has the highest median chemical concentration, however the median concentrations are comparable across all biomass burn conditions. All the flaming conditions have the highest median chemical concentrations and more overall variation than their respective smoldering conditions with the exception of smoldering peat. You may notice that the scaled chemical concentration was put on the x axis and burn condition was put on the y axis and not vice versa. When names are longer in length, they are more legible if placed on the y axis. Other aspects of the figure were changed in the latest version, but those are minor compared to changing the order of the boxplots, revamping the text, and changing the usage of color. For example, the background was changed from gray to white. Figure backgrounds are generally white, because the figure is easier to read if the paper is printed in black and white. A plot’s background can easily be changed to white in R using theme_light(), theme_minimal(), or theme_bw(). Posit provides a very helpful GGplot2 cheat sheet for changing a figure’s parameters. Creating an Improved Heatmap Visualization We’ll use a heatmap to answer the second environmental health question: Which classes of chemicals show the highest concentrations across the evaluated biomass burn conditions? Let’s view the original heatmap from the previous module and find aspects of it that can be improved. # Changing the biomass condition variable back to a character from a factor scaled_longer_smoke_data$Biomass_Burn_Condition = as.character(scaled_longer_smoke_data$Biomass_Burn_Condition) # Calculating the median value within each biomass burn condition and category scaled_longer_smoke_data = scaled_longer_smoke_data %&gt;% group_by(Biomass_Burn_Condition, Chemical.Category) %&gt;% mutate(Median_Scaled_Concentration = median(Scaled_Chemical_Concentration)) # Plotting ggplot(data = scaled_longer_smoke_data, aes(x = Chemical.Category, y = Biomass_Burn_Condition, fill = Median_Scaled_Concentration)) + geom_tile() + geom_text(aes(label = round(Median_Scaled_Concentration, 2))) # adding concentration values as text, rounding to two values after the decimal From the figure above, it’s clear that certain biomass burn conditions are associated with higher chemical concentrations for some of the chemical categories. For example, peat flaming exposure was associated with higher levels of inorganics and ions, while pine smoldering exposure was associated with higher levels of methoxyphenols. Although these are important findings, it is still difficult to determine if there are greater similarities in chemical profiles based on the biomass or the incineration temperature. Therefore, let’s identify some elements of this chart that can be modified to make it easier to answer our research question. There are three main aspects we can adjust on this figure: 1. The legibility of the text in the legend and axes. Similar to what we did previously, we’ll replace underscores and periods with spaces in the axis labels and titles. 2. The order of the axis labels. Ordering the biomass burn condition and chemical category from highest to lowest based on their median scaled chemical concentration allows the reader to easily determine the biomass burn condition that had the greatest or least total chemical concentrations relative to each other. From the previous boxplot figure, biomass burn condition is already in this order, however we need to order the chemical category by putting the variable into a factor. 3. Use of color. Notice that in the boxplot we used a qualitative palette, which is best for creating visual differences between different classes or groups. In this heatmap, we’ll choose a diverging color palette that uses two or more contrasting colors. A diverging color palette is able to highlight mid range with a lighter color and values at either extreme with a darker color or vice versa. #1: Legibility of Text can be addressed in ggplot() and so can #2: Reordering the heatmap. Biomass_Burn_Condition has already been reordered and put into a factor, but we need to do the same with Chemical.Category. Similar to before, median scaled chemical concentration for each chemical category will be calculated. However, this time the new dataframe will be arranged from highest to lowest median scaled concentration from the top of the dataframe to the bottom. ggplot() will plot them in the SAME order with the highest concentration on the left side and the lowest on the right side of the figure. # Order the chemical category by the median scaled chemical concentration median_chemical = scaled_longer_smoke_data %&gt;% group_by(Chemical.Category) %&gt;% summarize(Median_Concentration = median(Scaled_Chemical_Concentration)) %&gt;% arrange(-Median_Concentration) head(median_chemical) ## # A tibble: 6 × 2 ## Chemical.Category Median_Concentration ## &lt;chr&gt; &lt;dbl&gt; ## 1 Inorganics -0.265 ## 2 n-Alkanes -0.335 ## 3 Ions -0.359 ## 4 Levoglucosan -0.417 ## 5 Methoxyphenols -0.434 ## 6 PAHs -0.459 # Saving that order median_chemical_order = median_chemical$Chemical.Category # Putting into factor to organize the chemical categories scaled_longer_smoke_data$Chemical.Category = factor(scaled_longer_smoke_data$Chemical.Category, levels = median_chemical_order) # Putting burn conditons back into a factor to organize them scaled_longer_smoke_data$Biomass_Burn_Condition = factor(scaled_longer_smoke_data$Biomass_Burn_Condition, levels = median_biomass_order) # Viewing the dataframe to be plotted head(scaled_longer_smoke_data) ## # A tibble: 6 × 8 ## # Groups: Biomass_Burn_Condition, Chemical.Category [6] ## Chemical.Category Chemical CASRN Units Biomass_Burn_Condition ## &lt;fct&gt; &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;fct&gt; ## 1 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus Smoldering ## 2 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Eucalyptus Flaming ## 3 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat Smoldering ## 4 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Peat Flaming ## 5 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine Smoldering ## 6 n-Alkanes 2-Methylnonadecane 1560-86-7 ng_per_… Pine Flaming ## # ℹ 3 more variables: Chemical_Concentration &lt;dbl&gt;, ## # Scaled_Chemical_Concentration &lt;dbl[,1]&gt;, Median_Scaled_Concentration &lt;dbl&gt; Now that the dataframe has been finalized, we can plot the new boxplot. The final revision, #3: Making Use of Color, will be addressed with ggplot(). Here a palette is chosen from the RColorBrewer package. # Only needed to choose 2 colors for &#39;low&#39; and &#39;high&#39; in the heatmap # `n = 8` in the code to generate more colors that can be chosen from rcolorbrewer_colors = brewer.pal(n = 8, name = &#39;Accent&#39;) FigureX2 = ggplot(data = scaled_longer_smoke_data, aes(x = Chemical.Category, y = Biomass_Burn_Condition, fill = Median_Scaled_Concentration)) + geom_tile(color = &#39;white&#39;) + # adds white space between the tiles geom_text(aes(label = round(Median_Scaled_Concentration, 2))) + # adding concentration values as text theme_minimal() + # changing the theme theme(axis.text = element_text(size = 9), # changing size of axis labels axis.title = element_text(face = &quot;bold&quot;, size = rel(1.5)), # changes axis titles legend.title = element_text(face = &#39;bold&#39;, size = 10), # changes legend title legend.text = element_text(size = 9)) + # changes legend text labs(x = &#39;Chemical Category&#39;, y = &#39;Biomass Burn Condition&#39;, fill = &quot;Scaled Chemical\\nConcentration (pg/mL)&quot;) + # changing axis labels scale_fill_gradient(low = rcolorbrewer_colors[5], high = rcolorbrewer_colors[6]) # changing the colors FigureX2 An appropriate title for this figure could be: “Figure X. Chemical category concentrations across biomass burn conditions. Scaled chemical concentration values are based on the raw chemical concentration values scaled within each chemical. Chemical category on the x axis is ordered from highest to lowest median concentration from left to right. Biomass burn condition on the y axis is ordered from the highest to lowest median concentration from top to bottom. The values in each tile represent the median scaled chemical concentration.” Answer to Environmental Health Question 2 With this, we can answer Environmental Health Question #2: Which classes of chemicals show the highest concentrations across the evaluated biomass burn conditions? Answer: Ordering the axes from highest to lowest concentration didn’t help organize the data as much as we would’ve liked given some of the variance of chemical concentrations across the chemical categories. Nevertheless, it’s still clear that peat flaming produces the highest concentration of inorganics and ions, peat smoldering with n-Alkanes, eucalyptus smoldering with Levoglucosan, pine smoldering with methoxyphenols, and pine flaming with PAHs. In addition, flaming conditions seem to have higher levels of inorganics and ions while smoldering conditions seem to have higher levels of levoglucosan and PAHs. It would be helpful if there was a way to group these chemical profiles based on similarity and that’s where the pheatmap() function can be helpful when it can be difficult to spot those patterns using visual inspection alone. Just for fun, let’s briefly visualize a hierarchical clustering heatmap, which will be used to group both the biomass burn conditions and chemical categories based on their chemical concentrations. In this module, we’ll focus only on the pheatmap() visualization, but more information on hierarchical clustering can be found in Module 5.5 Unsupervised Machine Learning II: Additional Clustering Applications. As we showed in the previous module, this function requires a wide dataframe which we’ll need to create. It will contain Chemical.Category, Biomass_Burn_Condition and Scaled_Chemical_Concentration. heatmap_df2 = scaled_longer_smoke_data %&gt;% group_by(Biomass_Burn_Condition, Chemical.Category) %&gt;% # using the summarize function instead of mutate function as was done previously since we only need the median values now summarize(Median_Scaled_Concentration = median(Scaled_Chemical_Concentration)) %&gt;% # transforming the data to a wide format pivot_wider(id_cols = Biomass_Burn_Condition, names_from = &quot;Chemical.Category&quot;, values_from = &quot;Median_Scaled_Concentration&quot;) %&gt;% # converting the chemical names to row names column_to_rownames(var = &quot;Biomass_Burn_Condition&quot;) head(heatmap_df2) ## Inorganics n-Alkanes Ions Levoglucosan ## Red Oak Smoldering -0.6236516 -0.3895328 -0.8011155 1.1042911 ## Eucalyptus Smoldering -0.6859508 -0.3465132 -0.8016019 1.7772753 ## Pine Smoldering -0.8012749 -0.3637526 -0.8016019 0.9961855 ## Pine Needles Smoldering -0.8436790 -0.3076779 -0.4611222 -0.6357051 ## Pine Needles Flaming 0.3640553 -0.3850613 0.8222903 -0.8570130 ## Red Oak Flaming 0.4858796 -0.3016948 0.6793662 -0.8642615 ## Methoxyphenols PAHs ## Red Oak Smoldering 0.04674277 -0.5179641 ## Eucalyptus Smoldering -0.06449444 -0.4724635 ## Pine Smoldering 2.74452541 -0.4580922 ## Pine Needles Smoldering -0.22835290 -0.3162278 ## Pine Needles Flaming -0.46331332 -0.4179505 ## Red Oak Flaming -0.51488738 -0.5318127 Now let’s generate the same heatmap this time using the pheatmap() function: # creating a color palette blue_pink_palette = colorRampPalette(c(rcolorbrewer_colors[5], rcolorbrewer_colors[6])) pheatmap(heatmap_df2, # changing the color scheme color = blue_pink_palette(40), # hierarchical clustering of the biomass burn conditions cluster_rows = TRUE, # creating white space between the two largest clusters cutree_row = 2, # adding the values for each cell and making those values black display_numbers = TRUE, number_color = &quot;black&quot;, # changing the font size and the angle of the column names fontsize = 12, angle_col = 45) By using incorporating the dendrogram into the visualization, it’s easier to see that the chemical profiles have greater similarities within incineration temperatures rather than biomasses (with the exception of pine needles smoldering). Creating Multi-Plot Figures We can combine figures using the plot_grid() function from the cowplot package. For additional information on the plot_grid() function and parameters that can be changed see Arranging Plots in a Grid. Other packages that have figure combining capabilities include the patchwork package and the grid_arrange() function from the gridExtra package. Figures can also be combined after they’re exported from R using other applications like MS powerpoint and Adobe pdf. FigureX = plot_grid(FigureX1, FigureX2, # Adding labels, changing size their size and position labels = &quot;AUTO&quot;, label_size = 15, label_x = 0.04, rel_widths = c(1, 1.5)) FigureX An appropriate title for this figure could be: “Figure X. Chemical concentration distributions across biomass burn conditions. (A) The boxplots are based on the scaled chemical concentration values, which used the raw chemical concentrations values scaled within each chemical. The individual dots represent the concentrations of each chemical. The biomass burn conditions on the y axis are ordered from greatest (top) to least (bottom) based on median scaled chemical concentration. (B) The heatmap visualizes concentrations across chemical categories. Chemical category on the x axis is ordered from highest to lowest median concentration from left to right. Biomass burn condition on the y axis is ordered from the highest to lowest median concentration from top to bottom. The values in each tile represent the median scaled chemical concentration. By putting these two figures side by side, it’s now easier to compare the distributions of each biomass burn condition in figure A alongside the median chemical category concentrations in figure B that are responsible for the variation seen on the left. Concluding Remarks In conclusion, this training module provided information and example code for improving, streamlining, and making ggplot2 figures publication ready. Keep in mind that concepts and ideas presented in this module can be subjective and might need to be amended given the situation, dataset, and visualization. Additional Resources Beginner’s Guide to Data Visualizations and Improving Data Visualizations in R Generating Colors for Visualizations Additional Hands on Training Brewer, Cynthia A. 1994. Color use guidelines for mapping and visualization. Chapter 7 (pp. 123-147) in Visualization in Modern Cartography Hattab, G., Rhyne, T.-M., &amp; Heider, D. (2020). Ten simple rules to colorize biological data visualization. PLOS Computational Biology, 16(10), e1008259. PMID: 33057327 Lastly, for researchers who are newer to R programming, ggpubr is a package specifically designed to create publication-ready graphs similar to ggplot2 with more concise syntax. This package is particularly useful for statistically relevant visualizations, which are further explored in later modules including, TAME 2.0 Module 3.4 Introduction to Statistical Tests, TAME 2.0 Module 4.4 Two Group Comparisons and Visualizations, and TAME 2.0 Module 4.5 Multigroup Comparisons and Visualizations. Test Your Knowledge Replicate the figure below! The heatmap is the same as the “Test Your Knowledge” figure from TAME 2.0 Module 3.1 Data Visualizations. This time we’ll focus on making the figure look more publication ready by cleaning up the titles, cleaning up the labels, and changing the colors. The heatmap still visualizes the median chemical concentrations, but this time we’re separating the burn conditions, allowing us to determine if the concentrations of chemicals released are contingent upon the burn condition. Hint: To view additional aspects of figures that can be changed in ggplot2 check out this GGPlot2 Cheat Sheet. It might come in handy! "],["normality-tests-and-data-transformations.html", "Normality Tests and Data Transformations Introduction to Training Module What is a Normal Distribution? Qualitative Assessment of Normality Quantitative Normality Assessment Data Transformation Additional Considerations Regarding Normality Concluding Remarks", " Normality Tests and Data Transformations This training module was developed by Elise Hickman, Alexis Payton, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module When selecting the appropriate statistical tests to evaluate potential trends in your data, selection often relies upon whether or not underlying data are normally distributed. Many statistical tests and methods that are commonly implemented in exposure science, toxicology, and environmental health research rely on assumptions of normality. Applying a statistical test intended for data with a specific distribution when your data do not fit within that distribution can generate unreliable results, with the potential for false positive and false negative findings. Thus, one of the most common statistical tests to perform at the beginning of an analysis is a test for normality. In this training module, we will: Review the normal distribution and why it is important Demonstrate how to test whether your variable distributions are normal… Qualitatively, with histograms and Q-Q plots Quantitatively, with the Shapiro-Wilk test Discuss data transformation approaches Demonstrate log2 data transformation for non-normal data Discuss additional considerations related to normality We will demonstrate normality assessment using example data derived from a study in which chemical exposure profiles were collected across study participants through silicone wristbands. This exposure monitoring technique has been described through previous publications, including the following examples: O’Connell SG, Kincl LD, Anderson KA. Silicone wristbands as personal passive samplers. Environ Sci Technol. 2014 Mar 18;48(6):3327-35. doi: 10.1021/es405022f. Epub 2014 Feb 26. PMID: 24548134; PMCID: PMC3962070. Kile ML, Scott RP, O’Connell SG, Lipscomb S, MacDonald M, McClelland M, Anderson KA. Using silicone wristbands to evaluate preschool children’s exposure to flame retardants. Environ Res. 2016 May;147:365-72. doi: 10.1016/j.envres.2016.02.034. Epub 2016 Mar 3. PMID: 26945619; PMCID: PMC4821754. Hammel SC, Hoffman K, Phillips AL, Levasseur JL, Lorenzo AM, Webster TF, Stapleton HM. Comparing the Use of Silicone Wristbands, Hand Wipes, And Dust to Evaluate Children’s Exposure to Flame Retardants and Plasticizers. Environ Sci Technol. 2020 Apr 7;54(7):4484-4494. doi: 10.1021/acs.est.9b07909. Epub 2020 Mar 11. PMID: 32122123; PMCID: PMC7430043. Levasseur JL, Hammel SC, Hoffman K, Phillips AL, Zhang S, Ye X, Calafat AM, Webster TF, Stapleton HM. Young children’s exposure to phenols in the home: Associations between house dust, hand wipes, silicone wristbands, and urinary biomarkers. Environ Int. 2021 Feb;147:106317. doi: 10.1016/j.envint.2020.106317. Epub 2020 Dec 17. PMID: 33341585; PMCID: PMC7856225. In the current example dataset, chemical exposure profiles were obtained from the analysis of silicone wristbands worn by 97 participants for one week. Chemical concentrations on the wristbands were measured with gas chromatography mass spectrometry. The subset of chemical data used in this training module are all phthalates, a group of chemicals used primarily in plastic products to increase flexibility and durability. Script Preparations Cleaning the global environment rm(list=ls()) Installing required R packages If you already have these packages installed, you can skip this step, or you can run the below code which checks installation status for you if (!requireNamespace(&quot;openxlsx&quot;)) install.packages(&quot;openxlsx&quot;); if (!requireNamespace(&quot;tidyverse&quot;)) install.packages(&quot;tidyverse&quot;); if (!requireNamespace(&quot;ggpubr&quot;)) install.packages(&quot;ggpubr&quot;); Loading R packages required for this session library(openxlsx) # for importing data library(tidyverse) # for manipulating and plotting data library(ggpubr) # for making Q-Q plots with ggplot Set your working directory setwd(&quot;/filepath to where your input files are&quot;) Importing example dataset # Import data wrist_data &lt;- read.xlsx(&quot;Module3_3_Input/Module3_3_InputData.xlsx&quot;) # Viewing the data head(wrist_data) ## S_ID Age DEP DBP BBP DEHA DEHP DEHT ## 1 1 24.76986 335.58857 574.5443 40.67286 755.8157 10621.7029 30420.68 ## 2 2 25.39452 56.38286 1075.7114 243.48857 2716.7314 3036.5757 23991.82 ## 3 3 34.55068 515.65429 121.1657 205.86857 3286.5886 3056.2743 46188.06 ## 4 4 23.83562 1009.00714 373.4957 66.97571 3966.5371 729.7971 17900.74 ## 5 5 39.29315 33.74143 104.0629 77.17286 1654.3317 2599.7129 13597.44 ## 6 6 36.15616 168.79714 503.8300 61.98429 398.6314 1492.6143 29875.76 ## DINP TOTM ## 1 26534.290 1447.86000 ## 2 10073.704 39.46143 ## 3 1842.949 112.67714 ## 4 78779.567 92.31000 ## 5 3682.956 161.84571 ## 6 23845.493 182.56429 Our example dataset contains subject IDs (S_ID), subject ages, and measurements of 8 different phthalates from silicone wristbands: DEP: Diethyl phthalate DBP : Dibutyl phthalate BBP : Butyl benzyl phthalate DEHA : Di(2-ethylhexyl) adipate DEHP : Di(2-ethylhexyl) phthalate DEHT: Di(2-ethylhexyl) terephthalate DINP : Diisononyl phthalate TOTM : Trioctyltrimellitate The units for the chemical data are nanogram of chemical per gram of silicone wristband (ng/g) per day the participant wore the wristband. One of the primary questions in this study was whether there were significant differences in chemical exposure between subjects with different levels of social stress or between subjects with differing demographic characteristics. However, before we can analyze the data for significant differences between groups, we first need to assess whether our numeric variables are normally distributed. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: Are these data normally distributed? How does the distribution of data influence the statistical tests performed on the data? Before answering these questions, let’s define normality and how to test for it in R. What is a Normal Distribution? A normal distribution is a distribution of data in which values are distributed roughly symmetrically out from the mean such that 68.3% of values fall within one standard deviation of the mean, 95.4% of values fall within 2 standard deviations of the mean, and 99.7% of values fall within three standard deviations of the mean. Figure 5: Figure Credit: D Wells, CC BY-SA 4.0 https://creativecommons.org/licenses/by-sa/4.0, via Wikimedia Commons Common parametric statistical tests, such as t-tests, one-way ANOVAs, and Pearson correlations, rely on the assumption that data fall within the normal distribution for calculation of z-scores and p-values. Non-parametric tests, such as the Wilcoxon Rank Sum test, Kruskal-Wallis test, and Spearman Rank correlation, do not rely on assumptions about data distribution. Some of the aforementioned between-group comparisons were introduced in TAME 2.0 Module 3.4 Introduction to Statistical Tests. They, along with non-parametric tests, are explored further in later modules including TAME 2.0 Module 4.4 Two-Group Comparisons &amp; Visualizations and TAME 2.0 Module 4.5 Multi-group Comparisons &amp; Visualizations. Qualitative Assessment of Normality We can begin by assessing the normality of our data through plots. For example, plotting data using histograms, densities, or Q-Q plots can graphically help inform if a variable’s values appear to be normally distributed or not. We will start with visualizing our data distributions with histograms. Histograms Let’s start with visualizing the distribution of the participant’s ages using the hist() function that is part of base R. hist(wrist_data$Age) We can edit some of the parameters to improve this basic histogram visualization. For example, we can decrease the size of each bin using the breaks parameter: hist(wrist_data$Age, breaks = 10) The hist() function is useful for plotting single distributions, but what if we have many variables that need normality assessment? We can leverage ggplot2’s powerful and flexible graphics functions such as geom_histogram() and facet_wrap() to inspect histograms of all of our variables in one figure panel. For more information on data manipulation in general, see TAME 2.0 Module 2.3 Data Manipulation &amp; Reshaping and for more on ggplot2 including the use of facet_wrap(), see TAME 2.0 Module 3.2 Improving Data Visualizations. First, we’ll pivot our data to longer to prepare for plotting. Then, we’ll make our plot. We can use the theme_set() function to set a default graphing theme for the rest of the script. A graphing theme represents a set of default formatting parameters (mostly colors) that ggplot will use to make your graphs. theme_bw() is a basic theme that includes a white background for the plot and dark grey axis text and minor axis lines. The theme that you use is a matter of personal preference. For more on the different themes available through ggplot2, see here. # Pivot data longer to prepare for plotting wrist_data_long &lt;- wrist_data %&gt;% pivot_longer(!S_ID, names_to = &quot;variable&quot;, values_to = &quot;value&quot;) # Set theme for graphing theme_set(theme_bw()) # Make figure panel of histograms ggplot(wrist_data_long, aes(value)) + geom_histogram(fill = &quot;gray40&quot;, color = &quot;black&quot;, binwidth = function(x) {(max(x) - min(x))/25}) + facet_wrap(~ variable, scales = &quot;free&quot;) + labs(y = &quot;# of Observations&quot;, x = &quot;Value&quot;) From these histograms, we can see that our chemical variables do not appear to be normally distributed. Q-Q Plots Q-Q (quantile-quantile) plots are another way to visually assess normality. Similar to the histogram above, we can create a single Q-Q plot for the age variable using base R functions. Normal Q-Q plots (Q-Q plots where the theoretical quantiles are based on a normal distribution) have theoretical quantiles on the x-axis and sample quantiles, representing the distribution of the variable of interest from the dataset, on the y-axis. If the variable of interest is normally distributed, the points on the graph will fall along the reference line. # Plot points qqnorm(wrist_data$Age) # Add a reference line for theoretically normally distributed data qqline(wrist_data$Age) Small variations from the reference line, as seen above, are to be expected for the most extreme values. Overall, we can see that the age data are relatively normally distributed, as the points fall along the reference line. To make a figure panel with Q-Q plots for all of our variables of interest, we can use the ggqqplot() function within the ggpubr package. This function generates Q-Q plots and has arguments that are similar to ggplot2. ggqqplot(wrist_data_long, x = &quot;value&quot;, facet.by = &quot;variable&quot;, ggtheme = theme_bw(), scales = &quot;free&quot;) With this figure panel, we can see that the chemical data have very noticeable deviations from the reference, suggesting non-normal distributions. To answer our first environmental health question, age is the only variable that appears to be normally distributed in our dataset. This is based on our histograms and Q-Q plots with data centered in the middle and spreading with a distribution on both the lower and upper sides that follow typical normal data distributions. However, chemical concentrations appear to be non-normally distributed. Next, we will implement a quantitative approach to assessing normality, based on a statistical test for normality. Quantitative Normality Assessment Single Variable Normality Assessment We will use the Shapiro-Wilk test to quantitatively assess whether our data distribution is normal, again looking at the age data. This test can be carried out simply using the shapiro.test() function from the base R stats package. When using this test and interpreting its results, it is important to remember that the null hypothesis is that the sample distribution is normal, and a significant p-value means the distribution is non-normal. shapiro.test(wrist_data$Age) ## ## Shapiro-Wilk normality test ## ## data: wrist_data$Age ## W = 0.9917, p-value = 0.8143 This test resulted in a p-value of 0.8143, so we cannot reject the null hypothesis (that data are normally distributed). This means that we can assume that age is normally distributed, which is consistent with our visualizations above. Multiple Variable Normality Assessment With a large dataset containing many variables of interest (e.g., our example data with multiple chemicals), it is more efficient to test each column for normality and then store those results in a dataframe. We can use the base R function apply() to apply the Shapiro Wilk test over all of the numeric columns of our dataframe. This function generates a list of results, with a list element for each variable tested. There are also other ways that you could iterate through each of your columns, such as a for loop or a function as discussed in TAME 2.0 Module 2.4 Improving Coding Efficiencies. # Apply Shapiro Wilk test shapiro_res &lt;- apply(wrist_data %&gt;% select(-S_ID), 2, shapiro.test) # View first three list elements glimpse(shapiro_res[1:3]) ## List of 3 ## $ Age:List of 4 ## ..$ statistic: Named num 0.992 ## .. ..- attr(*, &quot;names&quot;)= chr &quot;W&quot; ## ..$ p.value : num 0.814 ## ..$ method : chr &quot;Shapiro-Wilk normality test&quot; ## ..$ data.name: chr &quot;newX[, i]&quot; ## ..- attr(*, &quot;class&quot;)= chr &quot;htest&quot; ## $ DEP:List of 4 ## ..$ statistic: Named num 0.225 ## .. ..- attr(*, &quot;names&quot;)= chr &quot;W&quot; ## ..$ p.value : num 2.74e-20 ## ..$ method : chr &quot;Shapiro-Wilk normality test&quot; ## ..$ data.name: chr &quot;newX[, i]&quot; ## ..- attr(*, &quot;class&quot;)= chr &quot;htest&quot; ## $ DBP:List of 4 ## ..$ statistic: Named num 0.658 ## .. ..- attr(*, &quot;names&quot;)= chr &quot;W&quot; ## ..$ p.value : num 1.08e-13 ## ..$ method : chr &quot;Shapiro-Wilk normality test&quot; ## ..$ data.name: chr &quot;newX[, i]&quot; ## ..- attr(*, &quot;class&quot;)= chr &quot;htest&quot; We can then convert those list results into a dataframe. Each variable is now in a row, with columns describing outputs of the statistical test. # Create results dataframe shapiro_res &lt;- do.call(rbind.data.frame, shapiro_res) # View results dataframe shapiro_res ## statistic p.value method data.name ## Age 0.9917029 8.143367e-01 Shapiro-Wilk normality test newX[, i] ## DEP 0.2248611 2.736536e-20 Shapiro-Wilk normality test newX[, i] ## DBP 0.6584967 1.076529e-13 Shapiro-Wilk normality test newX[, i] ## BBP 0.2367689 3.757059e-20 Shapiro-Wilk normality test newX[, i] ## DEHA 0.6646692 1.454576e-13 Shapiro-Wilk normality test newX[, i] ## DEHP 0.6163531 1.519572e-14 Shapiro-Wilk normality test newX[, i] ## DEHT 0.8072684 6.315917e-10 Shapiro-Wilk normality test newX[, i] ## DINP 0.5741864 2.486638e-15 Shapiro-Wilk normality test newX[, i] ## TOTM 0.3397424 6.901903e-19 Shapiro-Wilk normality test newX[, i] Finally, we can clean up our results dataframe and add a column that will quickly tell us whether our variables are normally or non-normally distributed based on the Shapiro-Wilk normality test results. # Clean dataframe shapiro_res &lt;- shapiro_res %&gt;% # Add normality conclusion mutate(normal = ifelse(p.value &lt; 0.05, F, T)) %&gt;% # Remove columns that do not contain informative data select(c(p.value, normal)) # View cleaned up dataframe shapiro_res ## p.value normal ## Age 8.143367e-01 TRUE ## DEP 2.736536e-20 FALSE ## DBP 1.076529e-13 FALSE ## BBP 3.757059e-20 FALSE ## DEHA 1.454576e-13 FALSE ## DEHP 1.519572e-14 FALSE ## DEHT 6.315917e-10 FALSE ## DINP 2.486638e-15 FALSE ## TOTM 6.901903e-19 FALSE The results from the Shapiro-Wilk test demonstrate that age data are normally distributed, while the chemical concentration data are non-normally distributed. These results support the conclusions we made based on our qualitative assessment above with histograms and Q-Q plots. Answer to Environmental Health Question 1 With this, we can now answer Environmental Health Question #1: Are these data normally distributed? Answer: Age is normally distributed, while chemical concentrates are non-normally distributed. Answer to Environmental Health Question 2 We can also answer Environmental Health Question #2: How does the distribution of data influence the statistical tests performed on the data? Answer: Parametric statistical tests should be used when analyzing the age data, and non-parametric tests should be used when analyzing the chemical concentration data Data Transformation There are a number of approaches that can be used to change the range and/or distribution of values within each variable. Typically, the purpose for applying these changes is to reduce bias in a dataset, remove known sources of variation, or prepare data for specific downstream analyses. The following are general definitions for common terms used when discussing these changes: Transformation refers to any process used to change data into other, related values. Normalization and standardization are types of data transformation. Transformation can also refer to performing the same mathematical operation on every value in your dataframe. For example, taking the log2 or log10 of every value is referred to as log transformation. Normalization is the process of transforming variables so that they are on a similar scale and therefore are comparable. This can be important when variables in a dataset contain a mixture of data types that are represented by vastly different numeric magnitudes or when there are known sources of variability across samples. Normalization methods are highly dependent on the type of input data. One example of normalization is min-max scaling, which results in a range for each variable of 0 to 1. Although normalization in computational methodologies typically refers to min-max scaling or other similar methods where the variable’s range is bounded by specific values, wet-bench approaches also employ normalization - for example, using a reference gene for RT-qPCR assays or dividing a total protein amount for each sample by the volume of each sample to obtain a concentration. Standardization, also known as Z-score normalization, is a specific type of normalization that involves subtracting each value from the mean of that variable and dividing by that variable’s standard deviation. The standardized values for each variable will have a mean of 0 and a standard deviation of 1. The scale() function in R performs standardization by default when the data are centered (argument center = TRUE is included within the scale function). Transformation of example data When data are non-normally distributed, such as with the chemical concentrations in our example dataset, it may be desirable to transform the data so that the distribution becomes closer to a normal distribution, particularly if there are only parametric tests available to test your hypothesis. A common transformation used in environmental health research is log2 transformation, in which data are transformed by taking the log2 of each value in the dataframe. Let’s log2 transform our chemical data and examine the resulting histograms and Q-Q plots to qualitatively assess whether data appear more normal following transformation. We will apply a pseudo-log2 transformation, where we will add 1 to each value before log2 transforming so that all resulting values are positive and any zeroes in the dataframe do not return -Inf. # Apply psuedo log2 (pslog2) transformation to chemical data wrist_data_pslog2 &lt;- wrist_data %&gt;% mutate(across(DEP:TOTM, ~ log2(.x + 1))) # Pivot data longer wrist_data_pslog2_long &lt;- wrist_data_pslog2 %&gt;% pivot_longer(!S_ID, names_to = &quot;variable&quot;, values_to = &quot;value&quot;) # Make figure panel of histograms ggplot(wrist_data_pslog2_long, aes(value)) + geom_histogram(fill = &quot;gray40&quot;, color = &quot;black&quot;, binwidth = function(x) {(max(x) - min(x))/25}) + facet_wrap(~ variable, scales = &quot;free&quot;) + labs(y = &quot;# of Observations&quot;, x = &quot;Value&quot;) # Make a figure panel of Q-Q plots ggqqplot(wrist_data_pslog2_long, x = &quot;value&quot;, facet.by = &quot;variable&quot;, ggtheme = theme_bw(), scales = &quot;free&quot;) Both the histograms and the Q-Q plots demonstrate that our log2 transformed data are more normally distributed than the raw data graphed above. Let’s apply the Shapiro-Wilk test to our log2 transformed data to determine if the chemical distributions are normally distributed. # Apply Shapiro Wilk test shapiro_res_pslog2 &lt;- apply(wrist_data_pslog2 %&gt;% select(-S_ID), 2, shapiro.test) # Create results dataframe shapiro_res_pslog2 &lt;- do.call(rbind.data.frame, shapiro_res_pslog2) # Clean dataframe shapiro_res_pslog2 &lt;- shapiro_res_pslog2 %&gt;% ## Add normality conclusion mutate(normal = ifelse(p.value &lt; 0.05, F, T)) %&gt;% ## Remove columns that do not contain informative data select(c(p.value, normal)) # View cleaned up dataframe shapiro_res_pslog2 ## p.value normal ## Age 0.814336705 TRUE ## DEP 0.001335217 FALSE ## DBP 0.368954224 TRUE ## BBP 0.052805523 TRUE ## DEHA 0.979072298 TRUE ## DEHP 0.304963678 TRUE ## DEHT 0.770066136 TRUE ## DINP 0.883662530 TRUE ## TOTM 0.004399442 FALSE The results from the Shapiro-Wilk test demonstrate that the the log2 chemical concentration data are more normally distributed than the raw data. Overall, the p-values, even for the chemicals that are still non-normally distributed, are much higher, and only 2 out of the 8 chemicals are non-normally distributed by the Shapiro-Wilk test. We can also calculate average p-values across all variables for our raw and log2 transformed data to further demonstrate this point. # Calculate the mean Shapiro-Wilk p-value for the raw chemical data mean(shapiro_res$p.value) ## [1] 0.09048186 # Calculate the mean Shapiro-Wilk p-value for the pslog2 transformed chemical data mean(shapiro_res_pslog2$p.value) ## [1] 0.4643995 Therefore, the log2 chemical data would be most appropriate to use if researchers are wanting to perform parametric statistical testing (and particularly if there is not a non-parametric statistical test for a given experimental design). It is important to note that if you proceed to statistical testing using log2 or other transformed data, graphs you make of significant results should use the transformed values on the y-axis, and findings should be interpreted in the context of the transformed values. Additional Considerations Regarding Normality The following sections detail additional considerations regarding normality. Similar to other advice in TAME, appropriate methods for handling normality assessment and normal versus non-normal data can be dependent on your field, lab, endpoints of interest, and downstream analyses. We encourage you to take those elements of your study into account, alongside the guidance provided here, when assessing normality. Regardless of the specific steps you take, be sure to report normality assessment steps and the data transformation or statistical test decisions you make based on them in your final report or manuscript. Determining which data should go through normality testing: Values for all samples (rows) that will be going into statistical testing should be tested for normality. If you are only going to be statistically testing a subset of your data, perform the normality test on that subset. Another way to think of this is that data points that are on the same graph together and/or that have been used as input for a statistical test should be tested for normality together. Analyzing datasets with a mixture of normally and non-normally distributed variables: There are a couple of different routes you can pursue if you have a mixture of normally and non-normally distributed variables in your dataframe: Perform parametric statistical tests on the normally distributed variables and non-parametric tests on the non-normally distributed variable. Perform the statistical test across all variables that fits with the majority of the variable distributions in your dataset. Our preference is to perform one test across all variables of the same data type/endpoint (e.g., all chemical concentrations, all cytokine concentrations). Aim to choose an approach that fits best rather than perfectly. Improving efficiency for normality assessment: If you find yourself frequently performing the same normality assessment workflow, consider writing a function that will execute each normality testing step (making a histogram, making a Q-Q plot, determining Shapiro-Wilk normality variable by variable, and determining the average Shapiro-Wilk p-value across all variables) and store the results in a list for easy inspection. Concluding Remarks In conclusion, this training module serves as an introduction to and step by step tutorial for normality assessment and data transformations. Approaches described in this training module include visualizations to qualitatively assess normality, statistical tests to quantitatively assess normality, data transformation, and other distribution considerations relating to normality. These methods are an important step in data characterization and exploration prior to downstream analyses and statistical testing, and they can be applied to nearly all studies carried out in environmental health research. Additional Resources Descriptive Statistics and Normality Tests for Statistical Data STHDA Normality Test in R Normalization vs. Standardization Test Your Knowledge Use the input file provided (“Module3_3_TYKInput.xlsx”), which represents a similar dataset to the one used in the module, to answer the following questions: Are any variables normally distributed in the raw data? Does psuedo log2 transforming the values make the distributions overall more or less normally distributed? What are the average Shapiro-Wilk p-values for the raw and psuedo log2 transformed data? "],["intoduction-to-statistical-tests.html", "Intoduction to Statistical Tests Introduction to Training Module Assessing Normality &amp; Homogeneity of Variance Two-Group Visualizations and Statistical Comparisons using the T-Test Three-Group Visualizations and Statistical Comparisons using an ANOVA Regression Modeling and Visualization: Linear and Logistic Regressions Statistical Evaluations of Categorical Data using the Chi-Squared Test and Fisher’s Exact Test Concluding Remarks", " Intoduction to Statistical Tests This training module was developed by Alexis Payton, Kyle Roell, Elise Hickman, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module This training module provides a brief introduction to some of the most commonly implemented statistics and associated visualizations used in exposure science, toxicology, and environmental health studies. This module first uploads an example dataset that is similar to the data used in TAME 2.0 Module 2.3 Data Manipulation &amp; Reshaping], though it includes some expanded subject information data to allow for more example statistical tests. Then, methods to evaluate data normality are presented, including visualization-based and statistical-based approaches. Basic statistical tests discussed in this module include: T test Analysis of Variance (ANOVA) with a Tukey’s Post-Hoc test Regression Modeling (Linear and Logistic) Chi-squared test Fisher’s exact test These statistical tests are very simple, with more extensive examples and associated descriptions of statistical models in the proceeding applications-based training modules in: TAME 2.0 Module 4.4 Two-Group Comparisons &amp; Visualizations TAME 2.0 Module 4.5 Multi-Group Comparisons &amp; Visualizations TAME 2.0 Module 4.6 Advanced Multi-Group Comparisons &amp; Visualizations Script Preparations Cleaning the global environment rm(list=ls()) Installing required R packages If you already have these packages installed, you can skip this step, or you can run the below code which checks installation status for you if (!requireNamespace(&quot;tidyverse&quot;)) install.packages(&quot;tidyverse&quot;); if (!requireNamespace(&quot;car&quot;)) install.packages(&quot;car&quot;); if (!requireNamespace(&quot;ggpubr&quot;)) install.packages(&quot;ggpubr&quot;); if(!requireNamespace(&quot;effects&quot;)) install.packages(&quot;effects&quot;); Loading R packages required for this session library(tidyverse) # all tidyverse packages, including dplyr and ggplot2 library(car) # package for statistical tests library(ggpubr) # ggplot2 based plots library(effects) # for linear modeling Set your working directory setwd(&quot;/filepath to where your input files are&quot;) Importing example datasets Let’s read in our example dataset. Note that these data are similar to those used previously, except that demographic and chemical measurement data were previously merged, and a few additional columns of subject information/demographics were added to serve as more thorough examples of data for use in this training module. # Loading data full.data &lt;- read.csv(&quot;Module3_4_Input/Module3_4_InputData.csv&quot;) Let’s view the top of the first 9 columns of data in this dataframe: full.data[1:10,1:9] ## ID BMI BMIcat MAge MEdu BW GA Smoker Smoker3 ## 1 1 27.7 Overweight 22.99928 College_Degree 3180.058 34 NS Never ## 2 2 26.8 Overweight 30.05142 College_Degree 3210.823 43 S Never ## 3 3 33.2 Overweight 28.04660 College_Degree 3311.551 40 NS Never ## 4 4 30.1 Overweight 34.81796 College_Degree 3266.844 32 S Never ## 5 5 37.4 Overweight 42.68440 College_Degree 3664.088 35 NS Never ## 6 6 33.3 Overweight 24.94960 College_Degree 3328.988 40 NS Never ## 7 7 24.8 Overweight 29.54798 College_Degree 3061.949 30 NS Never ## 8 8 16.9 Underweight 24.94954 College_Degree 3332.539 38 NS Current ## 9 9 36.9 Overweight 33.58589 College_Degree 3260.482 39 NS Never ## 10 10 21.7 Normal 39.29018 College_Degree 3141.723 35 NS Current These represent the subject information/demographic data, which include the following columns: ID: subject number BMI: body mass index BMIcat: BMI &lt;= 18.5 binned as “Underweight”, 18.5 &lt; BMI &lt;= 24.5 binned as “Normal”, BMI &gt; 24.5 binned as “Overweight” MAge: maternal age in years MEdu: maternal education level; “No_HS_Degree” = “less than high school”, “No_College_Degree” = “high school or some college”, “College_Degree” = “college or greater” BW: body weight in grams GA: gestational age in weeks Smoker: “NS” = non-smoker, “S” = smoker Smoker3: “Never”, “Former”, or “Current” smoking status Let’s now view the remaining columns (columns 10-15) in this dataframe: full.data[1:10,10:15] ## DWAs DWCd DWCr UAs UCd UCr ## 1 6.426464 1.292941 51.67987 10.192695 0.7537104 42.60187 ## 2 7.832384 1.798535 50.10409 11.815088 0.9789506 41.30757 ## 3 7.516569 1.288461 48.74001 10.079057 0.1903262 36.47716 ## 4 5.906656 2.075259 50.92745 8.719123 0.9364825 42.47987 ## 5 7.181873 2.762643 55.16882 9.436559 1.4977829 47.78528 ## 6 9.723429 3.054057 51.14812 11.589403 1.6645837 38.26386 ## 7 6.268547 1.218410 52.08578 8.887948 0.6347667 39.45535 ## 8 6.718448 1.414975 54.96740 9.304968 0.6658849 45.09987 ## 9 9.074928 2.727755 55.72826 10.818153 1.6585757 42.58577 ## 10 5.771691 2.410993 47.06552 8.747217 1.7354305 34.80661 These columns represent the environmental exposure measures, including: DWAs: drinking water arsenic levels in µg/L DWCd: drinking water cadmium levels in µg/L DWCr: drinking water chromium levels in µg/L UAs: urinary arsenic levels in µg/L UCd: urinary cadmium levels in µg/L UCr: urinary chromium levels in µg/L Now that the script is prepared and the data are uploaded, we can start by asking some initial questions about the data that can be answered by running some basic statistical tests and visualizations. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: Are there statistically significant differences in BMI between non-smokers and smokers? Are there statistically significant differences in BMI between current, former, and never smokers? Is there a relationship between maternal BMI and birth weight? Are maternal age and gestational age considered potential covariates in the relationship between maternal BMI and birth weight? Are there statistically significant differences in gestational age based on whether a subject is a non-smoker or a smoker? Is there a relationship between smoking status and BMI? Assessing Normality &amp; Homogeneity of Variance Statistical test selection often relies upon whether or not the underlying data are normally distributed and that variance across the groups is the same (homogeneity of variances). Many statistical tests and methods that are commonly implemented in exposure science, toxicology, and environmental health research rely on assumptions of normality. Thus, one of the most common statistical tests to perform at the beginning of an analysis is a test for normality. As discussed in the previous module, there are a few ways to evaluate the normality of a dataset: First, you can visually gauge whether a dataset appears to be normally distributed through plots. For example, plotting data using histograms, densities, or Q-Q plots can graphically help inform if a variable’s values appear to be normally distributed or not. Second, you can evaluate normality using statistical tests, such as the Kolmogorov-Smirnov (K-S) test and Shapiro-Wilk test. When using these tests and interpreting their results, it is important to remember that the null hypothesis is that the sample distribution is normal, and a significant p-value means the distribution is non-normal. Let’s start with the first approach based on data visualizations. In this module, we’ll primarily be generating figures using the ggubr package which is specifically designed to generate ggplot2-based figures using more streamlined coding syntax. In addition, this package has statistical parameters for plotting that are useful for basic statistical analysis, especially for people with introductory experience to plotting in R. For further documentation on ggubr, click here. Let’s begin with a histogram to view the distribution of BMI data using the gghistogram() function from the ggubr package: gghistogram(data = full.data, x = &quot;BMI&quot;, bins = 20) Let’s also view the Q–Q (quantile-quantile) plot using the ggqqplot() function also from the ggubr package: ggqqplot(full.data$BMI, ylab = &quot;BMI&quot;) From these visualizations, the BMI variable appears to be normally distributed, with data centered in the middle and spreading with a distribution on both the lower and upper sides that follow typical normal data distributions. Let’s now implement the second approach based on statistical tests for normality. Here, let’s use the Shapiro-Wilk test as an example, again looking at the BMI data. shapiro.test(full.data$BMI) ## ## Shapiro-Wilk normality test ## ## data: full.data$BMI ## W = 0.99232, p-value = 0.3773 This test resulted in a p-value of 0.3773, so we cannot reject the null hypothesis (that the BMI data are normally distributed). These findings support the assumption that these data are normally distributed. Next, we’ll assess homogeneity of variance using the Levene’s test. This will be done using the leveneTest()function from the car package: # First convert the smoker variable to a factor full.data$Smoker = factor(full.data$Smoker, levels = c(&quot;NS&quot;, &quot;S&quot;)) leveneTest(BMI ~ Smoker, data = full.data) ## Levene&#39;s Test for Homogeneity of Variance (center = median) ## Df F value Pr(&gt;F) ## group 1 0.2631 0.6086 ## 198 The p value, (Pr&gt;F), is 0.6086 indicating that variance in BMI across the smoking groups is the same. Therefore, the assumptions of a t-test, including normality and homogeneity of variance, have been met. Two-Group Visualizations and Statistical Comparisons using the T-Test T-tests are commonly used to test for a significant difference between the means of two groups in normally distributed data. In this example, we will be answering Environmental Health Question 1: Are there statistically significant differences in BMI between non-smokers and smokers? We will specifically implement a two sample t-test (or independent samples t-test). Let’s first visualize the BMI data across these two groups using boxplots: ggboxplot(data = full.data, x = &quot;Smoker&quot;, y = &quot;BMI&quot;) From this plot, it looks like non-smokers (labeled “NS”) may have significantly higher BMI than smokers (labeled “S”), though we need statistical evaluation of these data to more thoroughly evaluate this potential data trend. It is easy to perform a t-test on these data using the t.test() function from the base R stats package: t.test(data = full.data, BMI ~ Smoker) ## ## Welch Two Sample t-test ## ## data: BMI by Smoker ## t = 2.5372, df = 80.362, p-value = 0.01311 ## alternative hypothesis: true difference in means between group NS and group S is not equal to 0 ## 95 percent confidence interval: ## 0.583061 4.823447 ## sample estimates: ## mean in group NS mean in group S ## 26.11176 23.40851 Answer to Environmental Health Question 1 With this, we can answer Environmental Health Question #1: Are there statistically significant differences in BMI between non-smokers and smokers? Answer: From this statistical output, we can see that the overall mean BMI in non-smokers (group “NS”) is 26.1, and the overall mean BMI in smokers (group “S”) is 23.4. We can also see that the resulting p-value comparison between the means of these two groups is, indeed, significant (p-value = 0.013), meaning that the means between these groups are significantly different (i.e., are not equal). It’s also helpful to save these results into a variable within the R global environment, which then allows us to access specific output values and extract them more easily for our records. For example, we can run the following to specifically extract the resulting p-value from this test: ttest.res &lt;- t.test(data = full.data, BMI ~ Smoker) # making a list in the R global environment with the statistical results signif(ttest.res$p.value, 2) # pulling the p-value and using the `signif` function to round to 2 significant figures ## [1] 0.013 Three-Group Visualizations and Statistical Comparisons using an ANOVA Analysis of Variance (ANOVA) is a statistical method that can be used to compare means across three or more groups in normally distributed data. To demonstrate an ANOVA test on this dataset, let’s answer Environmental Health Question 2: Are there statistically significant differences in BMI between current, former, and never smokers? To do this we’ll use the Smoker3 variable from our dataset. Let’s again start by viewing these data distributions using a boxplot: ggboxplot(data = full.data, x = &quot;Smoker3&quot;, y = &quot;BMI&quot;) From this cursory review of the data, it looks like the current smokers likely demonstrate significantly different BMI measures than the former and never smokers, though we need statistical tests to verify this potential trend. We also require statistical tests to evaluate potential differences (or lack of differences) between former and never smokers. Let’s now run the ANOVA to compare BMI between smoking groups, using the aov() function to fit an ANOVA model: smoker_anova = aov(data = full.data, BMI ~ Smoker3) smoker_anova ## Call: ## aov(formula = BMI ~ Smoker3, data = full.data) ## ## Terms: ## Smoker3 Residuals ## Sum of Squares 2046.713 6817.786 ## Deg. of Freedom 2 197 ## ## Residual standard error: 5.882861 ## Estimated effects may be unbalanced We need to extract the typical ANOVA results table using either the summary() or anova() function on the resulting fitted object: anova(smoker_anova) ## Analysis of Variance Table ## ## Response: BMI ## Df Sum Sq Mean Sq F value Pr(&gt;F) ## Smoker3 2 2046.7 1023.36 29.57 5.888e-12 *** ## Residuals 197 6817.8 34.61 ## --- ## Signif. codes: 0 &#39;***&#39; 0.001 &#39;**&#39; 0.01 &#39;*&#39; 0.05 &#39;.&#39; 0.1 &#39; &#39; 1 This table outputs a lot of information, including the F value referring to the resulting F-statistic, Pr(&gt;F) referring to the p-value of the F-statistic, and other values that are described in detail through other available resources including this helpful video through PennState’s statistics online resources. Answer to Environmental Health Question 2 With this, we can answer Environmental Health Question #2: Are there statistically significant differences in BMI between current, former, never smokers? Answer: From this ANOVA output table, we can conclude that the group means across all three groups are not equal given that the p value, written as Pr(&gt;F) is significant (p value = 5.88 x 10-12). However, it doesn’t tell us which groups differ from each other and that’s where post hoc tests like Tukey’s are useful. Let’s run a Tukey’s post hoc test using the TukeyHSD() function in base R to determine which of the current, former, and never smokers have significant differences in BMI: smoker_tukey = TukeyHSD(smoker_anova) smoker_tukey ## Tukey multiple comparisons of means ## 95% family-wise confidence level ## ## Fit: aov(formula = BMI ~ Smoker3, data = full.data) ## ## $Smoker3 ## diff lwr upr p adj ## Former-Current 7.4436765 4.203006 10.684347 0.0000005 ## Never-Current 8.1167857 5.595470 10.638102 0.0000000 ## Never-Former 0.6731092 -2.011764 3.357982 0.8245448 Although the above Tukey object contains a column p adj, those are the raw unadjusted p values. It is common practice to adjust p values from multiple comparisons to prevent the reporting of false positives or reporting of a significant difference that doesn’t actually exist (Feise, 2002). There are a couple of different methods that are used to adjust p values including the Bonferroni and the Benjamini &amp; Hochberg approaches. For this example, we’ll use the p.adjust() function to obtain the Benjamini &amp; Hochberg adjusted p values. Check out the associated RDocumentation to discover other methods that can be used to adjust p values using the p.adjust() function: # First converting the Tukey object into a dataframe smoker_tukey_df = data.frame(smoker_tukey$Smoker3) %&gt;% # renaming the `p adj` to `P Value` for clarity rename(`P Value` = p.adj) # Adding a column with the adjusted p values smoker_tukey_df$`P Adj` = p.adjust(smoker_tukey_df$`P Value`, method = &quot;fdr&quot;) smoker_tukey_df ## diff lwr upr P Value P Adj ## Former-Current 7.4436765 4.203006 10.684347 5.064863e-07 7.597295e-07 ## Never-Current 8.1167857 5.595470 10.638102 3.432921e-12 1.029876e-11 ## Never-Former 0.6731092 -2.011764 3.357982 8.245448e-01 8.245448e-01 Answer to Environmental Health Question 2 We can use this additional information to further answer Environmental Health Question #2: Are there statistically significant differences in BMI between current, former, and never smokers? Answer: Current smokers have significantly lower BMIs than people who have never smoked and people who have formerly smoked. This is made evident by the 95% confidence intervals (lwr and upr) that don’t cross 0 and the p values that are less than 0.05 even after adjusting. Regression Modeling and Visualization: Linear and Logistic Regressions Regression modeling aims to find a relationship between a dependent variable (or outcome, response, y) and an independent variable (or predictor, explanatory variable, x). There are many forms of regression analysis, but here we will focus on two: linear regression and logistic regression. In brief, linear regression is generally used when you have a continuous dependent variable and there is assumed to be some sort of linear relationship between the dependent and independent variables. Conversely, logistic regression is often used when the dependent variable is dichotomous. Let’s first run through an example linear regression model to answer Environmental Health Question 3: Is there a relationship between maternal BMI and birth weight? Linear Regression We will first visualize the data and a run simple correlation analysis to evaluate whether these data are generally correlated. Then, we will run a linear regression to evaluate the relationship between these variables in more detail. Plotting the variables against one another and adding a linear regression line using the function ggscatter() from the ggubr package: ggscatter(full.data, x = &quot;BMI&quot;, y = &quot;BW&quot;, # Adding a linear line with 95% condfidence intervals as the shaded region add = &quot;reg.line&quot;, conf.int = TRUE, # Customize reg. line add.params = list(color = &quot;blue&quot;, fill = &quot;lightgray&quot;), # Adding Pearson&#39;s correlation coefficient cor.coef = TRUE, cor.method = &quot;pearson&quot;, cor.coeff.args = list(label.sep = &quot;\\n&quot;)) We can also run a basic correlation analysis between these two variables using the cor.test() function. This function uses the Pearson’s correlation test as default, which we can implement here due to the previously discussed assumption of normality for this dataset. Note that other tests are needed in instances when data are not normally distributed (e.g., Spearman Rank). This function is used here to extract the Pearson’s correlation coefficient and p-value (which also appear above in the upper left corner of the graph): cor.res &lt;- cor.test(full.data$BW, full.data$BMI) signif(cor.res$estimate, 2) ## cor ## 0.25 signif(cor.res$p.value, 2) ## [1] 0.00039 Together, it looks like there may be an association between BW and BMI, based on these correlation results, demonstrating a significant p-value of 0.0004. To test this further, let’s run a linear regression analysis using the lm() function, using BMI (X) as the independent variable and BW as the dependent variable (Y): crude_lm &lt;- lm(data = full.data, BW ~ BMI) summary(crude_lm) # viewing the results summary ## ## Call: ## lm(formula = BW ~ BMI, data = full.data) ## ## Residuals: ## Min 1Q Median 3Q Max ## -598.39 -116.72 8.11 136.54 490.11 ## ## Coefficients: ## Estimate Std. Error t value Pr(&gt;|t|) ## (Intercept) 3069.201 52.576 58.38 &lt; 2e-16 *** ## BMI 7.208 1.997 3.61 0.000388 *** ## --- ## Signif. codes: 0 &#39;***&#39; 0.001 &#39;**&#39; 0.01 &#39;*&#39; 0.05 &#39;.&#39; 0.1 &#39; &#39; 1 ## ## Residual standard error: 188 on 198 degrees of freedom ## Multiple R-squared: 0.06176, Adjusted R-squared: 0.05702 ## F-statistic: 13.03 on 1 and 198 DF, p-value: 0.0003876 Answer to Environmental Health Question 3 With this, we can answer Environmental Health Question #3: Is there a relationship between maternal BMI and birth weight? Answer: Not only is there a slight positive correlation between maternal BMI and BW as indicated by ~0.25 correlation coefficient, this linear relationship is significant due to the p-value being ~0.0004. Additionally, we can derive confidence intervals for the BMI estimate using: confint(crude_lm)[&quot;BMI&quot;,] ## 2.5 % 97.5 % ## 3.270873 11.145740 Notice that the r-squared (R2) value in regression output is the squared value of the previously calculated correlation coefficient (R). signif(sqrt(summary(crude_lm)$r.squared), 2) ## [1] 0.25 In epidemiological studies, the potential influence of confounders is considered by including important covariates within the final regression model. Let’s go ahead and investigate Environmental Health Question 4: Are maternal age and gestational age considered potential covariates in the relationship between maternal BMI and birth weight? We can do that by adding those variables to the linear model. adjusted_lm = lm(data = full.data, BW ~ BMI + MAge + GA) summary(adjusted_lm) ## ## Call: ## lm(formula = BW ~ BMI + MAge + GA, data = full.data) ## ## Residuals: ## Min 1Q Median 3Q Max ## -454.04 -111.24 5.79 116.46 488.41 ## ## Coefficients: ## Estimate Std. Error t value Pr(&gt;|t|) ## (Intercept) 2247.995 117.946 19.060 &lt; 2e-16 *** ## BMI 6.237 1.774 3.515 0.000547 *** ## MAge 4.269 1.887 2.263 0.024752 * ## GA 19.612 2.656 7.385 4.28e-12 *** ## --- ## Signif. codes: 0 &#39;***&#39; 0.001 &#39;**&#39; 0.01 &#39;*&#39; 0.05 &#39;.&#39; 0.1 &#39; &#39; 1 ## ## Residual standard error: 165.5 on 196 degrees of freedom ## Multiple R-squared: 0.2805, Adjusted R-squared: 0.2695 ## F-statistic: 25.47 on 3 and 196 DF, p-value: 5.884e-14 Let’s further visualize these regression modeling results by adding a regression line to the original scatterplot. Before doing so, we’ll use the effect() function from the effects package to make estimated predictions of birth weight values for the crude and adjusted linear models. The crude model only has BMI as the dependent variable, while the adjusted model includes BMI, maternal age, and gestational age as dependent variables. This function creates a table that contains 5 columns: fitted values for BMI (BMI), predictor values (fit), standard errors of the predictions (se), lower confidence limits (lower), and upper confidence limits (upper). An additional column, Model, was added to specify whether the values correspond to the crude or adjusted model. For additional information on visualizing adjusted linear models, see Plotting Adjusted Associations in R. crude_lm_predtable = data.frame(effect(term = &quot;BMI&quot;, mod = crude_lm), Model = &quot;Crude&quot;) adjusted_lm_predtable = data.frame(effect(term = &quot;BMI&quot;, mod = adjusted_lm), Model = &quot;Adjusted&quot;) # Viewing one of the tables crude_lm_predtable ## BMI fit se lower upper Model ## 1 10 3141.284 33.63898 3074.948 3207.621 Crude ## 2 19 3206.159 18.54497 3169.588 3242.730 Crude ## 3 28 3271.034 14.21563 3243.000 3299.067 Crude ## 4 36 3328.700 24.86346 3279.669 3377.732 Crude ## 5 45 3393.575 41.18575 3312.356 3474.794 Crude Now we can plot each linear model and their corresponding 95% confidence intervals (CI). It’s easier to visualize this using ggplot2 instead of ggubr so that’s what we’ll use: options(repr.plot.width=9, repr.plot.height=6) # changing dimensions of the entire figure ggplot(full.data, aes(x = BMI, y = BW)) + geom_point() + # Crude line geom_line(data = crude_lm_predtable, mapping = aes(x = BMI, y = fit, color = Model)) + # Adjusted line geom_line(data = adjusted_lm_predtable, mapping = aes(x = BMI, y = fit, color = Model)) + # Crude 95% CI geom_ribbon(data = crude_lm_predtable, mapping = aes(x = BMI, y = fit, ymin = lower, ymax = upper, fill = Model), alpha = 0.25) + # Adjusted 95% CI geom_ribbon(data = adjusted_lm_predtable, mapping = aes(x = BMI, y = fit, ymin = lower, ymax = upper, fill = Model), alpha = 0.25) Answer to Environmental Health Question 4 With this, we can answer Environmental Health Question #4: Are maternal age and gestational age considered potential covariates in the relationship between maternal BMI and birth weight? Answer: BMI is still significantly associated with BW and the included covariates are also shown to be significantly related to birth weight in this model. However, the addition of gestational age and maternal age did not have much of an impact on modifying the relationship between BMI and birth weight. Logistic Regression To carry out a logistic regression, we need to evaluate one continuous variable (here, we select gestational age, using the GA variable) and one dichotomous variable (here, we select smoking status, using the Smoker variable) to evaluate Environmental Health Question 5: Are there statistically significant differences in gestational age based on whether a subject is a non-smoker or a smoker? Because smoking status is a dichotomous variable, we will use logistic regression to look at this relationship. Let’s first visualize these data using a stacked bar plot for the dichotomous smoker dataset: ggboxplot(data = full.data, x = &quot;Smoker&quot;, y = &quot;GA&quot;) With this visualization, it’s difficult to tell whether or not there are significant differences in maternal education based on smoking status. Let’s now run the statistical analysis, using logistic regression modeling: # Before running the model, &quot;Smoker&quot;, needs to be binarized to 0&#39;s or 1&#39;s for the glm function glm_data = full.data %&gt;% mutate(Smoker = ifelse(Smoker == &quot;NS&quot;, 0,1)) # Use GLM (generalized linear model) and specify the family as binomial # This tells GLM to run a logistic regression log.res = glm(Smoker ~ GA, family = &quot;binomial&quot;, data = glm_data) summary(log.res) # viewing the results ## ## Call: ## glm(formula = Smoker ~ GA, family = &quot;binomial&quot;, data = glm_data) ## ## Coefficients: ## Estimate Std. Error z value Pr(&gt;|z|) ## (Intercept) 1.26669 1.37042 0.924 0.3553 ## GA -0.06764 0.03796 -1.782 0.0747 . ## --- ## Signif. codes: 0 &#39;***&#39; 0.001 &#39;**&#39; 0.01 &#39;*&#39; 0.05 &#39;.&#39; 0.1 &#39; &#39; 1 ## ## (Dispersion parameter for binomial family taken to be 1) ## ## Null deviance: 218.10 on 199 degrees of freedom ## Residual deviance: 214.89 on 198 degrees of freedom ## AIC: 218.89 ## ## Number of Fisher Scoring iterations: 4 Similar to the regression modeling analysis, we can also derive confidence intervals: confint(log.res)[&quot;GA&quot;,] ## Waiting for profiling to be done... ## 2.5 % 97.5 % ## -0.14301103 0.00640128 Answer to Environmental Health Question 5 With this, we can answer Environmental Health Question #5: Are there statistically significant differences in maternal education level based on whether they are a non-smoker or a smoker? Answer: Collectively, these results show a non-significant p-value relating gestational age to smoking status. The confidence intervals also overlap across zero. Therefore, these data do not demonstrate a significant association between gestational age and smoking status. Statistical Evaluations of Categorical Data using the Chi-Squared Test and Fisher’s Exact Test Chi-squared test and Fisher’s exact tests are used primarily when evaluating data distributions between two categorical variables. The difference between a Chi-squared test and the Fisher’s exact test surrounds the specific procedure being run. The Chi-squared test is an approximation and is run with larger sample sizes to determine whether there is a statistically significant difference between the expected vs. observed frequencies in one or more categories of a contingency table. The Fisher’s exact test is similar, though is an exact measure that can be run on any sample size, including smaller sample sizes. The number of samples or subjects (n) considered to be sufficiently large enough is subjective, contingent upon the research question being asked, and the experimental design. However, smaller sample sizes can be more permissible if the sample is normally distributed, but generally speaking having n &gt; 30 is a common convention in statistics (Alexander, 2022). For this example, we are interested in evaluating the potential relationship between two categorical variables: smoking status (using the Smoker variable) and categorical BMI group (using the BMIcat variable) to address Environmental Health Question 6: Is there a relationship between smoking status and BMI? To run these categorical statistical tests, let’s first create and view a 2-way contingency table describing the frequencies of observations across the categorical BMI and smoking groups: ContingencyTable &lt;- with(full.data, table(BMIcat, Smoker)) ContingencyTable ## Smoker ## BMIcat NS S ## Normal 43 14 ## Overweight 87 22 ## Underweight 23 11 Now let’s run the Chi-squared test on this table: chisq.test(ContingencyTable) ## ## Pearson&#39;s Chi-squared test ## ## data: ContingencyTable ## X-squared = 2.1849, df = 2, p-value = 0.3354 Note that we can also run the Chi-squared test using the following code, without having to generate the contingency table: chisq.test(full.data$BMIcat, full.data$Smoker) ## ## Pearson&#39;s Chi-squared test ## ## data: full.data$BMIcat and full.data$Smoker ## X-squared = 2.1849, df = 2, p-value = 0.3354 Or: with(full.data, chisq.test(BMIcat, Smoker)) ## ## Pearson&#39;s Chi-squared test ## ## data: BMIcat and Smoker ## X-squared = 2.1849, df = 2, p-value = 0.3354 Answer to Environmental Health Question 6 Note that these all produce the same results. With this, we can answer Environmental Health Question #6: Is there a relationship between smoking status and BMI? Answer: This results in a p-value = 0.34, demonstrating that there is no significant relationship between BMI categories and smoking status. We can also run a Fisher’s Exact Test when considering sample sizes. We won’t run this here due to computing time, but here is some example code for your records: #With small sample sizes, can use Fisher&#39;s Exact Test #fisher.test(full.data$BMI, full.data$Smoker) Concluding Remarks In conclusion, this training module serves as a high-level introduction to basic statistics and visualization methods. Statistical approaches described in this training module include tests for normality, t-test, analysis of variance, regression modeling, chi-squared test, and Fisher’s exact test. Visualization approaches include boxplots, histograms, scatterplots, and regression lines. These methods serve as an important foundation for nearly all studies carried out in environmental health research. Test Your Knowledge If we’re interested in investigating if there are significant differences in birth weight based on maternal education level, which statistical test should you use? Is that relationship considered to be statistically significant and how can we visualize the distributions of these groups? "],["overview-of-experimental-design-and-example-data.html", "Overview of Experimental Design and Example Data Introduction to Training Module Replicates Orientation to Example Data for Chapter 4 Concluding Remarks", " Overview of Experimental Design and Example Data This training module was developed by Elise Hickman, Sarah Miller, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module Converting wet lab experimentation data into dry lab analyses facilitates reproducibility and transparency in data analysis. This is helpful for consistency across members of the same research group, review of analyses by collaborators or reviewers, and implementation of similar future analyses. In comparison with analysis workflows that use subscription- or license-based applications, such as Prism or SAS, analysis workflows that leverage open-source programming languages such as R also increase accessibility of analyses. Additionally, scripted analyses minimize the risk for copy-paste error, which can occur when cleaning experimental data, transferring it to an analysis application, and exporting and formatting analysis results. Some of the barriers in converting wet lab experimentation into dry lab analyses include data cleaning, selection and implementation of appropriate statistical tests, and reporting results. This chapter will provide introductory material guiding wet-bench scientists in R analyses, bridging the gap between commonly available R tutorials (which, while helpful, may not provide sufficient level of detail or relevant examples) and intensive data science workflows (which may be too detailed). In this module, we will provide an overview of key experimental design features and terms that will be used throughout this chapter, and we will provide a detailed overview of the example data. In the subsequent modules, we will dive into analyzing the example data. Replicates One of the most important components of selecting an appropriate analysis is first understanding how data should be compared between samples, which often means addressing experimental replicates. There are two main types of replicates that are used in environmental health research: biological replicates and technical replicates. Biological Replicates Biological replicates are the preferred unit of statistical comparison because they represent biologically distinct samples, demonstrating biological variation in the system. What is considered to be a biological replicate can depend on what model system is being used. For example, in studies with human clinical samples or cells from different human donors, the different humans are considered the biological replicates. In studies using animals as model organisms, individual animals are typically considered biological replicates, although this can vary depending on the experimental design. In studies that use cell lines, which are derived from one human or animal and are modified to continuously grow in culture, a biological replicate could be either cells from different passages (different thawed aliquots) grown in completely separate flasks, all experimented with on the same day, or repeating an experiment on the same set of cells (one thawed aliquot) but on separate experimental days, so the cells have grown/replicated between experiments. The final “N” that you report should reflect your biological replicates, or independent experiments. What constitutes an independent experiment or biological replicate is highly field-, lab-, organism-, and endpoint-dependent, so make sure to discuss this within your research group in the experiment planning phase and again before your analysis begins. No matter what you choose, ensure that when you report your results, you are transparent about what your biological replicates are. For example, the below diagram (adapted from BitesizeBio) illustrates different ways of defining replicates in experiments with cell lines: Figure 6: Created with BioRender.com N = 3 cells could be considered technical replicates if the endpoint of interest is very low throughput, such as single cell imaging or analyses. N = 3 cell culture wells is a more common approach to technical replicates and is typically used when one sample is collected from each well, such as in the case of media or cell lysate collection. Note that each well within the Week 1 biological replicate would be considered a technical replicate for Week 1’s experiment. Similarly, each well within the Week 2 biological replicate would be considered a technical replicate for Week 2’s experiment. For more on technical replicates, see the next section. Although N = 3 cell lines is a less common approach to biological replicates, some argue for this approach because each cell line is typically derived from one biological source. In this scenario, each of the cell lines would be unique but would represent the same cell type or lineage (e.g., for respiratory epithelium, A549, 16HBE, and BEAS-2B cell lines). Also note that to perform statistical analyses, an N of at least 3 biological replicates is needed, and an even higher N may be needed for a sufficiently powered study. Although power calculations are outside the scope of this module, we encourage you to use power calculation resources, such as G*Power to assist in selecting an appropriate N for your study. Technical Replicates Technical replicates are repeated measurements on the same sample or biological source, demonstrating the variation underlying protocols, equipment, and sample handling. In environmental health research, there can be technical replicates separately related to either the experimental design or the downstream analyses. Technical replicates related to experimental design refer to the chemical exposure for cell-based (in vitro) experiments, where there may be multiple wells of cells from the same passage or human/mouse exposed to the same treatment. Technical replicates related to downstream analyses refer to the endpoints that are measured after chemical exposure in each sample. To illustrate this, consider an experiment where cells from four unique human donors (D1-D4) are grown in cell culture plates, and then three wells of cells from each donor are exposed to a chemical treatment (Tx) or a vehicle control (Ctrl). The plate layout might look something like this, with technical replicates related to experimental design, i.e. chemical exposure, in the same color: Figure 7: Created with BioRender.com For this experiment, we have four biological replicates (the four donors) and three technical exposure replicates per dose (because three wells from each donor were exposed to each condition). The technical replicates here capture potential unintended variation between wells in cell growth and chemical exposure. Following the exposure of the cells to a chemical of interest, the media is collected from each well and assayed using a plate reader assay for concentrations of a marker of inflammation. For each sample collected (from each well), there are three technical replicates used to measure the concentration of the inflammatory marker. The purpose of these technical replicates is to capture potential unintended well-to-well variation in the plate reader assay. The plate layout might look something like this, with the letter and number in each well of the plate layout representing the well in the exposure plate layout that the media sample being assayed came from: Figure 8: Created with BioRender.com Technical replicates should typically be averaged before performing any statistical analysis. For the experiment described above, we would: Average the technical replicates for the plate reader assay to obtain one value per original cell culture well for inflammatory marker concentration. Then, average the technical replicates for the chemical exposure to obtain one value per biological replicate (donor). This would result in a dataset with eight values (four control and four treatment) for statistical analysis. Number and inclusion of technical replicates The above example is just one approach to experimental design. As mentioned above in the biological replicates section, selection of appropriate biological and technical replicates can vary greatly depending on your model organism, experimental design, assay, and standards in the field. For example, there may be cases where well-to-well variation for certain assays is minimal compared with variation between biological replicates, or when including technical replicates for each donor is experimentally or financially unfeasible, resulting in a lack of technical replicates. Matched Experimental Design Matching (also known as paired or repeated measures) in an experimental design is also a very important concept when selecting the appropriate statistical analysis. In experiments with matched design, multiple measurements are collected from the same biological replicate. This typically provides increased statistical power because changes are observed within each biological replicate relative to its starting point. In environmental health research, this can include study designs such as: Samples were collected from the same individuals, animals, or cell culture wells pre- and post-exposure. Cells from the same biological replicate were exposed to different doses of a chemical. The experimental design described above represents a matched design because cells from the same donor are exposed to both the treatment and the vehicle control. Orientation to Example Data for Chapter 4 In this chapter, we will be using an example dataset derived from an in vitro, or cell culture, experiment. Before diving into analysis of these data in the subsequent modules, we will provide an overview of where these data came from and preview what the input data frames look like. Experimental Design In this experiment, primary human bronchial epithelial cells (HBECs) from sixteen different donors were exposed to the gas acrolein, which is emitted from the combustion of fossil fuels, tobacco, wood, and plastic. Inhalation exposure to acrolein is associated with airway inhalation, and this study aimed to understand how exposure to acrolein changes secretion of markers of inflammation. Prior to experimentation, the HBECs were grown on a permeable membrane support for 24 days with air on one side and liquid media on the other side, allowing them to differentiate into a form that is very similar to what is found in the human body. The cells were then exposed for 2 hours to 0 (filtered air), 0.6, 1, 2, or 4 ppm acrolein, with two technical replicate wells from each donor per dose. Twenty-four hours later, the media was collected, and concentrations of inflammatory markers were measured using an enzyme-linked immunosorbent assay (ELISA). Figure 9: Created with BioRender.com Note that this is a matched experimental design because cells from every donor were exposed to every concentration of acrolein, rather than cells from different donors being exposed to each of the different doses. Starting Data Next, let’s familiarize ourselves with the data that resulted from this experiment. There are two input data files, one that contains cytokine concentration data and one that contains demographic information about the donors: The cytokine data contains information about the cytokine measurements for each of the six proteins measured in the basolateral media for each sample (units = pg/mL), which can be identified by the donor, dose, and replicate columns. The demographic data contains information about the age and sex of each donor. In the subsequent modules, we’ll be using these data to assess whether exposure to acrolein significantly changes secretion of inflammatory markers and whether donor characteristics, such as sex and age, modify these responses. Concluding Remarks This module reviewed important components of experimental design, such as replicates and matching, which are critical for data pre-processing and selecting appropriate statistical tests. Test Your Knowledge Read the following experimental design descriptions. For each description, determine the number of biological replicates (per group), the number of technical replicates, and whether the experimental design is matched. One hundred participants are recruited to a study aiming to determine whether people who use e-cigarettes have different concentrations of inflammatory markers in their airways. Fifty participants are non e-cigarette users and 50 participants are e-cigarette users. After the airway samples are collected, each sample is analyzed with an ELISA, with three measurements taken per sample. Twenty mice are used in a study aiming to understand the effects of particulate matter on cardiovascular health. The mice are randomized such that half of the mice are exposed to filtered air and half are exposed to particulate matter. During the exposures, the mice are continuously monitored for endpoints such as heart rate and heart function. One month later, the mice that were exposed to particulate matter are exposed to filtered air, and the mice that were exposed to filtered air are exposed to particulate matter, with the same cardiovascular endpoints collected. "],["data-import-processing-and-summary-statistics.html", "Data Import, Processing, and Summary Statistics Introduction to Training Module Data Import Handling Missing Values Averaging Replicates Descriptive Statistics Normality Assessment and Data Transformation Concluding Remarks", " Data Import, Processing, and Summary Statistics This training module was developed by Elise Hickman, Alexis Payton, Sarah Miller, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module The first steps in any scripted analysis of wet-bench data include importing the data, cleaning the data to prepare for analyses, and conducting preliminary data exploration steps, such as addressing missing values, calculating summary statistics, and assessing normality. Although less exciting than diving right into the statistical analysis, these steps are crucial in guiding downstream analyses and ensuring accurate results. In this module, we will discuss each of these steps and work through them using an example dataset (introduced in TAME 2.0 Module 4.1 Overview of Experimental Design and Example Data of inflammatory markers secreted by airway epithelial cells after exposure to different concentrations of acrolein. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: What is the mean concentration of each inflammatory biomarker by acrolein concentration? Are our data normally distributed? Data Import First, we need to import our data. Data can be imported into R from many different file formats, including .csv (as demonstrated in previous chapters), .txt, .xlsx, and .pdf. Often, data are formatted in Excel prior to import, and the openxlsx package provides helpful functions that allow the user to import data from Excel, create workbooks for storing results generated in R, and export data from R to Excel workbooks. Below, we will use the read.xlsx() function to import our data directly from Excel. Other useful packages include pdftools (PDF import), tm (text mining of PDFs), and plater (plate reader formatted data import). Workspace Preparation and Data Import Set working directory In preparation, first let’s set our working directory to the folder path that contains our input files: setwd(&quot;/filepath to where your input files are&quot;) Load required packages And load required packages: library(openxlsx) # for importing Excel files library(DT) # for easier viewing of data tables library(tidyverse) # for data cleaning and graphing library(imputeLCMD) # for data imputation with QRILC library(table1) # for summary table library(vtable) # for summary table library(ggpubr) # for making Q-Q plots with ggplot Import example datasets Next, let’s read in our example datasets: biomarker_data &lt;- read.xlsx(&quot;Module4_2_Input/Module4_2_InputData1.xlsx&quot;) demographic_data &lt;- read.xlsx(&quot;Module4_2_Input/Module4_2_InputData2.xlsx&quot;) View example datasets First, let’s preview our example data. Using the datatable() function from the DT package allows us to interactively scroll through our biomarker data. datatable(biomarker_data) We can see that our biomarker data are arranged with samples in rows and sample information and biomarker measurements in the columns. datatable(demographic_data) Our demographic data provide information about the donors that our cells came from, matching to the Donor column in our biomarker data. Handling Missing Values Next, we will investigate whether we have missing values and which variables and donors have missing values. # Calculate the total number of NAs per variable biomarker_data %&gt;% summarise(across(IL1B:VEGF, ~sum(is.na(.)))) ## IL1B IL6 IL8 IL10 TNFa VEGF ## 1 0 0 6 5 0 0 # Calculate the number of missing values per subject biomarker_data %&gt;% group_by(Donor) %&gt;% summarise(across(IL1B:VEGF, ~sum(is.na(.)))) ## # A tibble: 16 × 7 ## Donor IL1B IL6 IL8 IL10 TNFa VEGF ## &lt;chr&gt; &lt;int&gt; &lt;int&gt; &lt;int&gt; &lt;int&gt; &lt;int&gt; &lt;int&gt; ## 1 D1 0 0 0 0 0 0 ## 2 D10 0 0 0 2 0 0 ## 3 D11 0 0 0 0 0 0 ## 4 D12 0 0 2 0 0 0 ## 5 D13 0 0 0 3 0 0 ## 6 D14 0 0 0 0 0 0 ## 7 D15 0 0 0 0 0 0 ## 8 D16 0 0 0 0 0 0 ## 9 D2 0 0 2 0 0 0 ## 10 D3 0 0 0 0 0 0 ## 11 D4 0 0 0 0 0 0 ## 12 D5 0 0 0 0 0 0 ## 13 D6 0 0 0 0 0 0 ## 14 D7 0 0 0 0 0 0 ## 15 D8 0 0 0 0 0 0 ## 16 D9 0 0 2 0 0 0 Here, we can see that we do have a few missing values. What should we do with these values? Missing Values and Data Imputation Missing values Before deciding what to do about our missing values, it’s important to understand why they are missing. There are a few different types of missing values that could be present in a dataset: Missing completely at random (MCAR): has nothing to do with the experimental unit being studied (e.g., a sample is damaged or lost in the lab) Missing at random (MAR): there may be a systematic difference between missing and measured values, but they can be explained by observed differences in the data or experimental unit Missing not at random (MNAR): data are missing due to factors that are not observed/measured (e.g., measurement for a specific endpoint is below the limit of detection (LOD) of an assay) We know from the researchers who generated this dataset that the values are missing because these specific proteins were below the limit of detection for the assay for certain samples; therefore, our data are missing not at random. This can help us with our choice of imputation method, described below. Imputation Imputation is the assignment of a value to a missing data point by inferring that value from other properties of the dataset or externally defined limits. Whether or not you should impute your data is not a one-size-fits-all approach and may vary depending on your field, experimental design, the type of data, and the type of missing values in your dataset. Two questions you can ask yourself when deciding whether or not to impute data are: Is imputation needed for downstream analyses? Some analyses are not permissive to including NAs or 0s; others are. Will imputing values bias my analyses unnecessarily? If so, consider analyzing subsets of the data that are complete separately. There are many different imputation methods (too many to cover them all in this module); here, we will introduce a few that we use most often. We encourage you to explore these in more depth and to understand typical imputation workflows for your lab, data type, and/or discipline. For variables where imputed values are expected to be generally bound by the existing range of data (e.g., MCAR): missForest For variables with samples below the limit of detection for the assay, such as for mass spectrometry or ELISAs (e.g., MNAR) Replace non-detects with the limit of detection divided by the square root of 2 Quantile Regression Imputation of Left-Censored Data (QRILC) GSimp (can also be used to impute values above a specific threshold) If you do impute missing values, make sure to include both your raw and imputed data, along with detailed information about the imputation method, within your manuscript, supplemental information, and/or GitHub. You can even present summary statistics for both raw and imputed data for additional transparency. Imputation of Our Data Before imputing our data, it is a good idea to implement a background filter that checks to see if a certain percentage of values for each variable are missing. For variables with a very high percentage of missing values, imputation can be unreliable because there is not enough information for the imputation algorithm to reference. The threshold for what this percentage should be can vary by study design and the extent to which your data are subset into groups that may have differing biomarker profiles; however, a common threshold we frequently use is to remove variables with missing data for 25% or more of samples. We can use the following code to calculate the percentage values missing for each endpoint: biomarker_data %&gt;% summarise(across(IL1B:VEGF, ~sum(is.na(.))/nrow(biomarker_data)*100)) ## IL1B IL6 IL8 IL10 TNFa VEGF ## 1 0 0 3.75 3.125 0 0 Here, we can see that only about 3-4% of values are missing for our variables with missing data, so we will proceed to imputation with our dataset as-is. We will impute values using QRILC, which pulls from the left side of the data distribution (the lower values) to impute missing values. We will write a function that will apply QRILC imputation to our dataframe. This function takes a data frame with missing values as input and returns a dataframe with QRILC imputed values in place of NAs as output. QRILC_imputation = function(df){ # Normalize data before applying QRILC per QRILC documentation ## Select only numeric columns, log2 transform, and convert to a matrix ### 4 comes from there being 3 metadata columns before the numeric data starts QRILC_prep = df[,4:dim(df)[2]] %&gt;% mutate_all(., function(x) log2(x + 1)) %&gt;% as.matrix() imputed_QRILC_object = impute.QRILC(QRILC_prep, tune.sigma = 0.1) QRILC_log2_df = data.frame(imputed_QRILC_object[1]) # Converting back the original scale QRILC_df = QRILC_log2_df %&gt;% mutate_all(., function(x) 2^x - 1) # Adding back in metadata columns QRILC_df = cbind(Donor = df$Donor, Dose = df$Dose, Replicate = df$Replicate, QRILC_df) return(QRILC_df) } Now we can apply the QRILC function to our dataframe. We use the function set.seed() to ensure that the QRILC function generates the same numbers each time we run the script. For more on setting seeds, see here. # Set random seed to ensure reproducibility in results set.seed(1104) # Apply function biomarker_data_imp &lt;- QRILC_imputation(biomarker_data) Averaging Replicates The last step we need to take before our data are ready for analysis is averaging the two technical replicates for each donor and dose. We will do this by creating an ID column that represents the donor and dose together and using that column to group and average the data. This results in a dataframe where our rows contain data representing each biological replicate exposed to each of the five concentrations of acrolein. biomarker_data_imp_avg &lt;- biomarker_data_imp %&gt;% # Create an ID column that represents the donor and dose unite(Donor_Dose, Donor, Dose, sep = &quot;_&quot;) %&gt;% # Average replicates with each unique Donor_Dose group_by(Donor_Dose) %&gt;% summarize(across(IL1B:VEGF, mean)) %&gt;% # Round results to the same number of significant figures as the original data mutate(across(IL1B:VEGF, \\(x) round(x, 2))) %&gt;% # Separate back out the Donor_Dose column separate(Donor_Dose, into = c(&quot;Donor&quot;, &quot;Dose&quot;), sep = &quot;_&quot;) # View new data frame datatable(biomarker_data_imp_avg) Descriptive Statistics Generating descriptive statistics (e.g., mean, median, mode, range, standard deviation) can be helpful for understanding the general distribution of your data and for reporting results either in the main body of a manuscript/report (for small datasets) or in the supplementary material (for larger datasets). There are a number of different approaches that can be used to calculate summary statistics, including functions that are part of base R and that are part of packages. Here, we will demonstrate a few different ways to efficiently calculate descriptive statistics across our dataset. Method #1 - Tidyverse and Basic Functions The mean, or average of data points, is one of the most commonly reported summary statistics and is often reported as mean ± standard deviation to demonstrate the spread in the data. Here, we will make a table of mean ± standard deviation for each of our biomarkers across each of the dose groups using tidyverse functions. # Calculate means biomarker_group_means &lt;- biomarker_data_imp_avg %&gt;% group_by(Dose) %&gt;% summarise(across(IL1B:VEGF, \\(x) mean(x))) # View data datatable(biomarker_group_means) You’ll notice that there are a lot of decimal places in our calculated means, while in our original data, there are only two decimal places. We can add a step to round the data to our above code chunk to produce cleaner results. # Calculate means biomarker_group_means &lt;- biomarker_data_imp_avg %&gt;% group_by(Dose) %&gt;% summarise(across(IL1B:VEGF, \\(x) mean(x))) %&gt;% mutate(across(IL1B:VEGF, \\(x) round(x, 2))) # View data datatable(biomarker_group_means) With this, we can answer Environmental Health Question 1: What is the mean concentration of each inflammatory biomarker by acrolein concentration? Answer: With the above table, we can see the mean concentrations for each of our inflammatory biomarkers by acrolein dose. IL-8 overall has the highest concentrations, followed by VEGF and IL-6. For IL1B, IL8, TNFa, and VEGF, it appears that the concentration of the biomarker goes up with increasing dose. We can use very similar code to calculate our standard deviations: # Calculate means biomarker_group_sds &lt;- biomarker_data_imp_avg %&gt;% group_by(Dose) %&gt;% summarise(across(IL1B:VEGF, \\(x) sd(x))) %&gt;% mutate(across(IL1B:VEGF, \\(x) round(x, 1))) # View data datatable(biomarker_group_sds) Now we’ve calculated both the means and standard deviations! However, these are typically presented as mean ± standard deviation. We can merge these dataframes by executing the following steps: Pivot each dataframe to a long format, with each row containing the value for one biomarker at one dose. Create a variable that represents each unique row (combination of Dose and variable). Join the data frames by row. Unite the two columns with mean and standard deviation, with ± in between them. Pivot the dataframe wider so that the dataframe resembles what we started with for the means and standard deviations. First, we’ll pivot each dataframe to a long format and create a variable that represents each unique row. # Pivot dataframes longer and create variable column for each row biomarker_group_means_long &lt;- pivot_longer(biomarker_group_means, !Dose, names_to = &quot;variable&quot;, values_to = &quot;mean&quot;) %&gt;% unite(Dose_variable, Dose, variable, remove = FALSE) biomarker_group_sds_long &lt;- pivot_longer(biomarker_group_means, !Dose, names_to = &quot;variable&quot;, values_to = &quot;sd&quot;) %&gt;% unite(Dose_variable, Dose, variable, remove = FALSE) # Preview what dataframe looks like datatable(biomarker_group_means_long) Next, we will join the mean and standard deviation datasets. Notice that we are only joining the Dose_variable and sd columns from the standard deviation data frame to prevent duplicate columns (Dose, variable) from being included. # Merge the dataframes by row biomarker_group_summstats &lt;- left_join(biomarker_group_means_long, biomarker_group_sds_long %&gt;% select(c(Dose_variable, sd)), by = &quot;Dose_variable&quot;) # Preview the new dataframe datatable(biomarker_group_summstats) Then, we can unite the mean and standard deviation columns and add the ± symbol between them by storing that character as a variable and pasting that variable in our paste() function. # Store plus/minus character plusminus &lt;-&quot;\\u00b1&quot; Encoding(plusminus)&lt;-&quot;UTF-8&quot; # Create new column with mean +/- standard deviation biomarker_group_summstats &lt;- biomarker_group_summstats %&gt;% mutate(mean_sd = paste(mean, plusminus, sd, sep = &quot; &quot;)) # Preview the new dataframe datatable(biomarker_group_summstats) Last, we can pivot the dataframe wider to revert it to its original layout, which is more digestible to read. # Pivot dataframe wider biomarker_group_summstats &lt;- biomarker_group_summstats %&gt;% # Remove columns we don&#39;t need any more select(-c(Dose_variable, mean, sd)) %&gt;% # Pivot wider pivot_wider(id_cols = Dose, names_from = &quot;variable&quot;, values_from = &quot;mean_sd&quot;) # View final dataframe datatable(biomarker_group_summstats) These data are now in a publication-ready format that can be exported to a .txt, .csv., or .xlsx file for easy sharing. Method #2 - Applying a List of Functions Calculating our mean and standard deviation separately using tidyverse wasn’t too difficult, but what if we want to calculate other descriptive statistics, such as minimum, median, and maximum? We could use the above approach, but we would need to make a separate dataframe for each and then merge them all together. Instead, we can use the map_dfr() function from the purrr package, which is also part of tidyverse. This function takes a list of functions you want to apply to your data and applies these functions over specified columns in the data. Let’s see how it works: # Define summary functions summary_functs &lt;- lst(min, median, mean, max, sd) # Apply functions to data, grouping by dose # .id = &quot;statistic&quot; tells the function to create a column describing which statistic that row is reporting biomarker_descriptive_stats_all &lt;- map_dfr(summary_functs, ~ summarize(biomarker_data_imp_avg %&gt;% group_by(Dose), across(IL1B:VEGF, .x)), .id = &quot;statistic&quot;) # View data datatable(biomarker_descriptive_stats_all) Depending on your final goal, descriptive statistics data can then be extracted from this dataframe and cleaned up or reformatted as needed to create a publication-ready table! Other Methods There are also packages that have been developed for specifically making summary tables, such as table1 and vtable. These packages can create summary tables in HTML format, which appear nicely in R Markdown and can be copied and pasted into Word. Here, we will briefly demonstrate how these packages work, and we encourage you to explore more using the package vignettes! Table1 The table1 package makes summary tables using the function table1(), which takes the columns that you want in the rows of the table on the left side of the first argument, followed by | and then the grouping variable. The output table can be customized in a number of ways, including what summary statistics are output and whether or not statistical comparisons are run between groups (see package vignette for more details). # Get names of all of the columns to include in the table paste(names(biomarker_data_imp_avg %&gt;% select(IL1B:VEGF)), collapse=&quot; + &quot;) ## [1] &quot;IL1B + IL6 + IL8 + IL10 + TNFa + VEGF&quot; # Make the table table1(~ IL1B + IL6 + IL8 + IL10 + TNFa + VEGF | Dose, data = biomarker_data_imp_avg) Vtable The vtable package includes the function st(), which can also be used to make HTML tables (and other output formats; see out argument). For example: # HTML output st(biomarker_data_imp_avg, group = &#39;Dose&#39;) Table 1: Summary Statistics Dose 0 0.6 1 2 4 Variable N Mean SD N Mean SD N Mean SD N Mean SD N Mean SD IL1B 16 9.8 1.9 16 10 1.9 16 11 1.5 16 11 1.6 16 12 2 IL6 16 597 658 16 670 810 16 565 652 16 536 474 16 288 297 IL8 16 21987 7570 16 24703 10768 16 22607 10648 16 41687 17939 16 94439 33535 IL10 16 1.4 1.3 16 0.44 0.56 16 1.2 1.2 16 0.88 0.43 16 1.2 0.78 TNFa 16 3.2 1.8 16 3.1 1.7 16 4.1 2.4 16 5.2 2 16 6.7 2.7 VEGF 16 971 417 16 1023 486 16 1196 621 16 1754 749 16 1617 836 # Dataframe output st(biomarker_data_imp_avg, group = &#39;Dose&#39;, out = &#39;return&#39;) ## Variable N Mean SD N Mean SD N Mean SD N Mean SD N Mean ## 1 Dose 0 0.6 1 2 4 ## 2 IL1B 16 9.8 1.9 16 10 1.9 16 11 1.5 16 11 1.6 16 12 ## 3 IL6 16 597 658 16 670 810 16 565 652 16 536 474 16 288 ## 4 IL8 16 21987 7570 16 24703 10768 16 22607 10648 16 41687 17939 16 94439 ## 5 IL10 16 1.4 1.3 16 0.44 0.56 16 1.2 1.2 16 0.88 0.43 16 1.2 ## 6 TNFa 16 3.2 1.8 16 3.1 1.7 16 4.1 2.4 16 5.2 2 16 6.7 ## 7 VEGF 16 971 417 16 1023 486 16 1196 621 16 1754 749 16 1617 ## SD ## 1 ## 2 2 ## 3 297 ## 4 33535 ## 5 0.78 ## 6 2.7 ## 7 836 Similar to table1, see the package vignette for detailed information about how to customize tables using this package. Normality Assessment and Data Transformation The last step we will take before beginning to test our data for statistical differences between groups (in the next module) is to understand our data’s distribution through normality assessment. This will inform which statistical tests we will perform on our data. For more detail on normality testing, including detailed explanations of each type of normality assessment and explanations of the code underlying the following graphs and tables, see TAME Module 3.3, Normality Tests and Data Transformations. We’ll start by looking at histograms of our data for qualitative normality assessment: # Set theme theme_set(theme_bw()) # Pivot data longer to prepare for plotting biomarker_data_imp_avg_long &lt;- biomarker_data_imp_avg %&gt;% pivot_longer(-c(Donor, Dose), names_to = &quot;variable&quot;, values_to = &quot;value&quot;) # Make figure panel of histograms ggplot(biomarker_data_imp_avg_long, aes(value)) + geom_histogram(fill = &quot;gray40&quot;, color = &quot;black&quot;, binwidth = function(x) {(max(x) - min(x))/25}) + facet_wrap(~ variable, scales = &quot;free&quot;, nrow = 2) + labs(y = &quot;# of Observations&quot;, x = &quot;Value&quot;) From these histograms, we can see that IL-1B appears to be normally distributed, while the other endpoints do not appear to be normally distributed. We can also use Q-Q plots to assess normality qualitatively: ggqqplot(biomarker_data_imp_avg_long, x = &quot;value&quot;, facet.by = &quot;variable&quot;, ggtheme = theme_bw(), scales = &quot;free&quot;) With this figure panel, we can see that most of the variables have very noticeable deviations from the reference, suggesting non-normal distributions. To assess normality quantitatively, we can use the Shapiro-Wilk test. Note that the null hypothesis is that the sample distribution is normal, and a significant p-value means the distribution is non-normal. # Apply Shapiro Wilk test to dataframe shapiro_res &lt;- apply(biomarker_data_imp_avg %&gt;% select(IL1B:VEGF), 2, shapiro.test) # Create results data frame shapiro_res &lt;- do.call(rbind.data.frame, shapiro_res) # Clean data frame shapiro_res &lt;- shapiro_res %&gt;% ## Add normality conclusion mutate(normal = ifelse(p.value &lt; 0.05, F, T)) %&gt;% ## Remove columns that do not contain informative data select(c(p.value, normal)) # View cleaned up data frame datatable(shapiro_res) With this, we can answer Environmental Health Question 2: Are our data normally distributed? Answer: The results from the Shapiro-Wilk test demonstrate that the IL1B data are normally distributed, while the other variables are non-normally distributed. These results support the conclusions we made based on our qualitative assessment above with histograms and Q-Q plots. Log2 Transforming and Re-Assessing Normality Log2 transformation is a common transformation used in environmental health research and can move data closer to a normal distribution. For more on data transformation, see TAME Module 3.3, Normality Tests and Data Transformations. We will pseudo-log2 transform our data, which adds a 1 to each value before log2 transformation and ensures that resulting values are positive real numbers. Let’s see if the log2 data are more normally distributed than the raw data. # Apply log2 transformation to data biomarker_data_imp_avg_log2 &lt;- biomarker_data_imp_avg %&gt;% mutate(across(IL1B:VEGF, ~ log2(.x + 1))) Make histogram panel: # Pivot data longer and make figure panel of histograms biomarker_data_imp_avg_log2_long &lt;- biomarker_data_imp_avg_log2 %&gt;% pivot_longer(-c(Donor, Dose), names_to = &quot;variable&quot;, values_to = &quot;value&quot;) # Make histogram panel ggplot(biomarker_data_imp_avg_log2_long, aes(value)) + geom_histogram(fill = &quot;gray40&quot;, color = &quot;black&quot;, binwidth = function(x) {(max(x) - min(x))/25}) + facet_wrap(~ variable, scales = &quot;free&quot;) + labs(y = &quot;# of Observations&quot;, x = &quot;Value&quot;) Make Q-Q plot panel: ggqqplot(biomarker_data_imp_avg_log2_long, x = &quot;value&quot;, facet.by = &quot;variable&quot;, ggtheme = theme_bw(), scales = &quot;free&quot;) Run Shapiro-Wilk test: # Apply Shapiro Wilk test shapiro_res_log2 &lt;- apply(biomarker_data_imp_avg_log2 %&gt;% select(IL1B:VEGF), 2, shapiro.test) # Create results data frame shapiro_res_log2 &lt;- do.call(rbind.data.frame, shapiro_res_log2) # Clean data frame shapiro_res_log2 &lt;- shapiro_res_log2 %&gt;% ## Add normality conclusion mutate(normal = ifelse(p.value &lt; 0.05, F, T)) %&gt;% ## Remove columns that do not contain informative data select(c(p.value, normal)) # View cleaned up data frame shapiro_res_log2 ## p.value normal ## IL1B 0.250821677 TRUE ## IL6 0.001445717 FALSE ## IL8 0.017386740 FALSE ## IL10 0.018935719 FALSE ## TNFa 0.194774045 TRUE ## VEGF 0.047231367 FALSE The histograms and Q-Q plots demonstrate that the log2 data are more normally distributed than the raw data. The results from the Shapiro-Wilk test also demonstrate that the the log2 data are more normally distributed as a whole than the raw data. Overall, the p-values, even for the variables that are still non-normally distributed, are much higher. So, should we proceed with the raw data or the log2 data? This depends on what analyses we plan to do. In general, it is best to keep the data in as close to its raw format as possible, so if all of our analyses are available with a non-parametric test, we could use our raw data. However, some statistical tests do not have a non-parametric equivalent, in which case it would likely be best to use the log2-transformed data. For subsequent modules, we will proceed with the log2 data for consistency; however, choices regarding normality assessment can vary, so be sure to discuss these choices within your research group before proceeding with your analysis. For more on decisions regarding normality, see TAME Module 3.3 Normality Tests and Data Transformations. For more on parametric vs. non-parametric tests, see TAME Module 4.4 Two Group Comparisons and Visualizations and TAME Module 4.5 Multi-Group Comparisons and Visualizations. Concluding Remarks Taken together, this module demonstrates important data processing steps necessary before proceeding with between-group statistical testing, including data import, handling missing values, averaging replicates, generating descriptive statistics tables, and assessing normality. Careful consideration and description of these steps in the methods section of a manuscript or report increases reproducibility of analyses and helps to improve the accuracy and statistical validity of subsequent statistical results. Test Your Knowledge Functional endpoints from these cultures were also measured. These endpoints were: 1) Membrane Permeability (MemPerm), 2) Trans-Epithelial Electrical Resistance (TEER), 3) Ciliary Beat Frequency (CBF), and 4) Expression of Mucin (MUC5AC). Work through the same processes demonstrated in this module using the provided data (“Module4_2_TYKInput.xlsx”) to answer the following questions: How many technical replicates are there for each dose? Are there any missing values? What are the average values for each endpoint by dose? Are the raw data normally distributed? "],["data-import-from-pdf-sources.html", "Data Import from PDF Sources Introduction to Training Module Importing Data from Many Single PDFs with the Same Formatting Importing Data Stored in PDF Tables Concluding Remarks", " Data Import from PDF Sources This training module was developed by Elise Hickman, Alexis Payton, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module Most tutorials for R rely on importing .csv, .xlsx, or .txt files, but there are numerous other file formats that can store data, and these file formats can be more difficult to import into R. PDFs can be particularly difficult to interface with in R because they are not formatted with defined rows/columns/cells as is done in Excel or .csv/.txt formatting. In this module, we will demonstrate how to import data from from PDFs into R and format it such that it is amenable for downstream analyses or export as a table. Familiarity with tidyverse, for loops, and functions will make this module much more approachable, so be sure to review TAME 2.0 Modules 2.3 Data Manipulation and Reshaping and 2.4 Improving Coding Efficiencies if you need a refresher. Overview of Example Data To demonstrate import of data from PDFs, we will be leveraging two example datasets, described in more detail in their respective sections later on in the module. PDFs generated by Nanoparticle Tracking Analysis (NTA), a technique used to quantify the size and distribution of particles (such as extracellular vesicles) in a sample. We will be extracting data from an experiment in which epithelial cells were exposed to four different environmental chemicals or a vehicle control, and secreted particles were isolated and characterized using NTA. A PDF containing information about variables collected as part of a study whose samples are part of NIH’s BioLINCC Repository. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: Which chemical(s) increase and decrease the concentration of particles secreted by epithelial cells? How many variables total are available to us to request from the study whose data are store in the repository, and what are these variables? Importing Data from Many Single PDFs with the Same Formatting Getting Familiar with the Example Dataset The following example is based on extracting data from PDFs generated by Nanoparticle Tracking Analysis (NTA), a technique used to quantify the size and distribution of particles in a sample. Each PDF file is associated with one sample, and each PDF contains multiple values that we want to extract. Although this is a very specific type of data, keep in mind that this general approach can be applied to any data stored in PDF format - you will just need to make modifications based on the layout of your PDF file! For this example, we will be extracting data from 5 PDFs that are identically formatted but contain information unique to each sample. The samples represent particles isolated from epithelial cell media following an experiment where cells were exposed to four different environmental chemicals (labeled “A”, “B”, “C”, and “D”) or a vehicle control (labeled “Ctrl”). Here is what a full view of one of the PDFs looks like, with values we want to extract highlighted in yellow: Our goal is to extract these values and end up with a dataframe that looks like this, with each sample in a row and each variable in a column: If your files are not already named in a way that reflects unique sample information, such as the date of the experiment or sample ID, update your file names to contain this information before proceeding with the script. Here are the names for the example PDF files: Workspace Preparation and Data Import Installing and loading required R packages If you already have these packages installed, you can skip this step, or you can run the below code which checks installation status for you. We will be using the pdftools and tm packages to extract text from the PDF. And instead of using head() to preview dataframes, we will be using the function datatable() from the DT package. This function produces interactive tables and generates better formatting for viewing dataframes that have long character strings (like the ones we will be viewing in this section). if (!requireNamespace(&quot;pdftools&quot;)) install.packages(&quot;pdftools&quot;) if (!requireNamespace(&quot;tm&quot;)) install.packages(&quot;tm&quot;) if (!requireNamespace(&quot;DT&quot;)) install.packages(&quot;DT&quot;) if (!requireNamespace(&quot;janitor&quot;)) install.packages(&quot;janitor&quot;) Next, load the packages. library(tidyverse) library(pdftools) library(tm) library(DT) library(janitor) Initial data import from PDF files The following code stores the file names of all of the files in your directory that end in .pdf. To ensure that only PDFs of interest are imported, consider making a subfolder within your directory containing only the PDF extraction script file and the PDFs you want to extract data from. pdf_list &lt;- list.files(path = &quot;./Module4_3_Input&quot;, pattern = &quot;488.pdf$&quot;) We can see that each of our file names are now contained in the list. head(pdf_list) ## [1] &quot;20230214_0002_Expt1_A_size_488.pdf&quot; ## [2] &quot;20230214_0006_Expt1_Ctrl_size_488.pdf&quot; ## [3] &quot;20230214_0014_Expt1_C_size_488.pdf&quot; ## [4] &quot;20230214_0023_Expt1_D_size_488.pdf&quot; ## [5] &quot;20230214_0024_Expt1_B_size_488.pdf&quot; Next, we need to make a dataframe to store the extracted data. The PDF Identifier column will store the file name, and the Text column will store extracted text from the PDF. pdf_raw &lt;- data.frame(&quot;PDF Identifier&quot; = c(), &quot;Text&quot; = c()) The following code uses a for loop to loop through each file (as stored in the pdf_list vector) and extract the text from the PDF. Sometimes this code generates duplicates, so we will also remove the duplicates with distinct(). for (i in 1:length(pdf_list)){ # Iterating through each pdf file and separating each line of text document_text = pdf_text(paste(&quot;./Module4_3_Input/&quot;, pdf_list[i], sep = &quot;&quot;)) %&gt;% strsplit(&quot;\\n&quot;) # Saving the name of each PDF file and its text document = data.frame(&quot;PDF Identifier&quot; = gsub(x = pdf_list[i], pattern = &quot;.pdf&quot;, replacement = &quot;&quot;), &quot;Text&quot; = document_text, stringsAsFactors = FALSE) colnames(document) &lt;- c(&quot;PDF Identifier&quot;, &quot;Text&quot;) # Appending the new text data to the dataframe pdf_raw &lt;- rbind(pdf_raw, document) } pdf_raw &lt;- pdf_raw %&gt;% distinct() The new dataframe contains the data from all of the PDFs, with the PDF Identifier column containing the name of the input PDF file that corresponds to the text in the column next to it. datatable(pdf_raw) Extracting Variables of Interest Specific variables of interest can be extracted from the pdf_raw dataframe by filtering the dataframe for rows that contain a specific character string. This character string could be the variable of interest (if that word or set of words is unique and only occurs in that one place in the document) or a character string that occurs in the same line of the PDF as your variable of interest. Examples of both of these approaches are shown below. It is important to note that there can be different numbers of spaces in each row and after each semicolon, which will change the sep argument for each variable. For example, there are a different number of spaces after the semicolon for “Dilution Factor” than there are for “Concentration” (see above PDF screen shot for reference). We will work through an example for the first variable of interest, dilution factor, in detail. First, we can see what the dataframe looks like when we just filter rows based on keeping only rows that contain the string “Dilution Factor” in the text column using the grepl() function. dilution_factor_df &lt;- pdf_raw %&gt;% filter(grepl(&quot;Dilution Factor&quot;, Text)) datatable(dilution_factor_df) The value we are trying to extract is at the end of a long character string. We will want to use the tidyverse function separate() to isolate those values, but we need to know what part of the character string will separate the dilution factor values from the rest of the text. To determine this, we can call just one of the data cells and copy the semicolon and following spaces for use in the separate() function. # Return the value in the first row and second column. dilution_factor_df[1,2] ## [1] &quot; Temperature: 24.64 °C sensed Dilution Factor: 200&quot; Building on top of the previous code, we can now separate the dilution factor value from the rest of the text in the string. The separate() function takes an input data column and separates it into two or more columns based on the character passed to the separation argument. Here, everything before the separation string is discarded by setting the first new column to NA. Everything after the separation string will be stored in a new column called Dilution Factor, The starting Text column is removed by default. dilution_factor_df &lt;- pdf_raw %&gt;% filter(grepl(&quot;Dilution Factor&quot;, Text)) %&gt;% separate(Text, into = c(NA, &quot;Dilution Factor&quot;), sep = &quot;: &quot;) datatable(dilution_factor_df) For the “Original Concentration” variable, we filter rows by the string “pH” because the word concentration is found in multiple locations in the document. concentration_df = pdf_raw %&gt;% filter(grepl(&quot;pH&quot;, Text)) %&gt;% separate(Text, c(NA, &quot;Concentration&quot;), sep = &quot;: &quot;) datatable(concentration_df) With the dilution factor variable, there were no additional characters after the value of interest, but here, “Particles / mL” remains and needs to be removed so that the data can be used in downstream analyses. We can add an additional cleaning step to remove “Particles / mL” from the data and add the units to the column title. sep = \" P\" refers to the space before and first letter of the string to be removed. concentration_df = pdf_raw %&gt;% filter(grepl(&quot;pH&quot;, Text)) %&gt;% separate(Text, c(NA, &quot;Concentration&quot;), sep = &quot;: &quot;) %&gt;% separate(Concentration, c(&quot;Concentration (Particles/ mL)&quot;, NA), sep = &quot; P&quot;) datatable(concentration_df) Next, we want to extract size distribution data from the lower table. Note that the space in the first separate() function comes from the space between the “Number” and “Concentration” column in the string, and the space in the second separate() function comes from the space between the variable name and the number of interest. We can also convert values to numeric since they are currently stored as characters. size_distribution_df = pdf_raw %&gt;% filter(grepl(&quot;X10&quot;, Text)| grepl(&quot;X50 &quot;, Text)| grepl(&quot;X90&quot;, Text) | grepl(&quot;Mean&quot;, Text)| grepl(&quot;StdDev&quot;, Text)) %&gt;% separate(Text, c(&quot;Text&quot;, NA), sep = &quot; &quot;) %&gt;% separate(Text, c(&quot;Text&quot;, &quot;Size&quot;), sep = &quot; &quot;) %&gt;% mutate(Size = as.numeric(Size)) %&gt;% pivot_wider(names_from = Text, values_from = Size) datatable(size_distribution_df) Creating the final dataframe Now that we have created dataframes for all of the variables that we are interested in, we can join them together into one final dataframe. # Make list of all dataframes to include all_variables &lt;- list(dilution_factor_df, concentration_df, size_distribution_df) # Combine dataframes using reduce function. Sometimes, duplicate rows are generated by full_join. full_df = all_variables %&gt;% reduce(full_join, by = &quot;PDF Identifier&quot;) %&gt;% distinct() # View new dataframe datatable(full_df) For easier downstream analysis, the last step is to separate the PDF Identifier column into an informative sample ID that matches up with other experimental data. final_df &lt;- full_df %&gt;% separate(&#39;PDF Identifier&#39;, # Split sample identifier column into new columns, retaining the original column into = c(&quot;Date&quot;, &quot;FileNumber&quot;, &quot;Experiment Number&quot;, &quot;Sample_ID&quot;, &quot;Size&quot;, &quot;Wavelength&quot;), sep = &quot;_&quot;, remove = FALSE) %&gt;% select(-c(FileNumber, Size)) %&gt;% # Remove uninformative columns mutate(across(&#39;Dilution Factor&#39;:&#39;StdDev&#39;, as.numeric)) # Change variables to numeric where appropriate datatable(final_df) Let’s make a graph to help us answer Environmental Health Question 1. theme_set(theme_bw()) data_for_graphing &lt;- final_df %&gt;% clean_names() data_for_graphing$sample_id &lt;- factor(data_for_graphing$sample_id, levels = c(&quot;Ctrl&quot;, &quot;A&quot;, &quot;B&quot;, &quot;C&quot;, &quot;D&quot;)) ggplot(data_for_graphing, aes(x = sample_id, y = concentration_particles_m_l)) + geom_bar(stat = &quot;identity&quot;, fill = &quot;gray70&quot;, color = &quot;black&quot;) + ylab(&quot;Particle Concentration (Particles/mL)&quot;) + xlab(&quot;Exposure&quot;) With this, we can answer Environmental Health Question #1: Which chemical(s) increase and decrease the concentration of particles secreted by epithelial cells? Answer: Chemicals B and C appear to increase the concentration of secreted particles. However, additional replicates of this experiment are needed to assess statistical significance. Importing Data Stored in PDF Tables The above workflow is useful if you just want to extract a few specific values from PDFs, but isn’t as useful if data are already in a table format in a PDF. The tabulapdf package provides helpful functions for extracting dataframes from tables in PDF format. Getting Familiar with the Example Dataset The following example is based on extracting dataframes from a long PDF containing many individual data tables. This particular PDF came from the NIH’s BioLINCC Repository and details variables that researchers can request from the repository. Variables are part of larger datasets that contain many variables, with each dataset in a separate table. All of the tables are stored in one PDF file, and some of the tables are longer than one page (this will become relevant later on!). Similar to the first PDF workflow, remember that this is a specific example intended to demonstrate how to work through extracting data from PDFs. Modifications will need to be made for differently formatted PDFs. Here is what the first three pages of our 75-page starting PDF look like: If we zoom in a bit more on the first page, we can see that the dataset name is defined in bold above each table. This formatting is consistent throughout the PDF. The zoomed in view also allows us to see the columns and their contents more clearly. Some are more informative than others. The columns we are most interested in are listed below along with a description to guide you through the contents. Num: The number assigned to each variable in the dataset. This numbering restarts with 1 for each table. Variable: The variable name. Type: The type (or class) of the variable, either numeric or character. Label: A description of the variable and values associated with the variable. After extracting the data, we want to end up with a dataframe that contains all of the variables, their corresponding columns, and a column that indicates which dataset the variable is associated with: Workspace Preparation and Data Import Installing and loading required R packages Similar to previous sections, we need to install and load a few packages before proceeding. The tabulapdf package needs to be installed in a specific way as shown below and can sometimes be difficult to install on Macs. If errors are produced, follow the troubleshooting tips outlined in this Stack Overflow solution. # To install all of the packages except for tabulapdf if (!requireNamespace(&quot;stringr&quot;)) install.packages(&quot;stringr&quot;) if (!requireNamespace(&quot;pdftools&quot;)) install.packages(&quot;pdftools&quot;) if (!requireNamespace(&quot;rJava&quot;)) install.packages(&quot;rJava&quot;) # To install tabulapdf if (!require(&quot;remotes&quot;)) { install.packages(&quot;remotes&quot;) } library(remotes) remotes::install_github(c(&quot;ropensci/tabulizerjars&quot;, &quot;ropensci/tabulapdf&quot;), force=TRUE, INSTALL_opts = &quot;--no-multiarch&quot;) ## ## * checking for file &#39;C:\\Users\\Jessie PC\\AppData\\Local\\Temp\\Rtmpm0tPsD\\remotes62a857c530a8\\ropensci-tabulizerjars-d1924e0/DESCRIPTION&#39; ... OK ## * preparing &#39;tabulizerjars&#39;: ## * checking DESCRIPTION meta-information ... OK ## * checking for LF line-endings in source and make files and shell scripts ## * checking for empty or unneeded directories ## * building &#39;tabulizerjars_1.0.1.tar.gz&#39; ## glue (1.7.0 -&gt; 1.8.0) [CRAN] ## ## There is a binary version available but the source version is later: ## binary source needs_compilation ## glue 1.7.0 1.8.0 TRUE ## ## Binaries will be installed ## package &#39;glue&#39; successfully unpacked and MD5 sums checked ## Warning: cannot remove prior installation of package &#39;glue&#39; ## Warning in file.copy(savedcopy, lib, recursive = TRUE): problem copying ## C:\\Users\\Jessie ## PC\\AppData\\Local\\R\\win-library\\4.4\\00LOCK\\glue\\libs\\x64\\glue.dll to ## C:\\Users\\Jessie PC\\AppData\\Local\\R\\win-library\\4.4\\glue\\libs\\x64\\glue.dll: ## Permission denied ## Warning: restored &#39;glue&#39; ## ## The downloaded binary packages are in ## C:\\Users\\Jessie PC\\AppData\\Local\\Temp\\Rtmpm0tPsD\\downloaded_packages ## * checking for file &#39;C:\\Users\\Jessie PC\\AppData\\Local\\Temp\\Rtmpm0tPsD\\remotes62a83035ee4\\ropensci-tabulapdf-7325a8d/DESCRIPTION&#39; ... OK ## * preparing &#39;tabulapdf&#39;: ## * checking DESCRIPTION meta-information ... OK ## * checking for LF line-endings in source and make files and shell scripts ## * checking for empty or unneeded directories ## Removed empty directory &#39;tabulapdf/docs&#39; ## * building &#39;tabulapdf_1.0.5-4.tar.gz&#39; Load packages: library(tabulapdf) library(tidyverse) library(janitor) library(pdftools) library(stringr) Initial data import from PDF file The extract_tables() function automatically extracts tables from PDFs and stores them as tibbles (a specific tidyverse data structure similar to a dataframe) within a list. One table is extracted per page, even if the table spans multiple pages. This line of code can take a few seconds to run depending on the length of your PDF. tables &lt;- extract_tables(&quot;Module4_3_Input/Module4_3_InputData4.pdf&quot;, output = &quot;tibble&quot;) Glimpsing the first three elements in the tables list, we can see that each list element is a dataframe containing the columns from the PDF tables. glimpse(tables[1:3]) ## List of 3 ## $ : spc_tbl_ [30 × 7] (S3: spec_tbl_df/tbl_df/tbl/data.frame) ## ..$ Num : num [1:30] 1 NA NA NA 2 NA NA 3 NA NA ... ## ..$ Variable: chr [1:30] &quot;AAQ_1000&quot; NA NA NA ... ## ..$ Type : chr [1:30] &quot;Num&quot; NA NA NA ... ## ..$ Len : num [1:30] 8 NA NA NA 8 NA NA 8 NA NA ... ## ..$ Format : chr [1:30] &quot;2.&quot; NA NA NA ... ## ..$ Informat: chr [1:30] &quot;2.&quot; NA NA NA ... ## ..$ Label : chr [1:30] &quot;In the past 3 days, how much of the time did your asthma keep you from&quot; &quot;doing your usual activities at work, school, or at home? 0=None of the&quot; &quot;time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All&quot; &quot;of the time&quot; ... ## ..- attr(*, &quot;spec&quot;)= ## .. .. cols( ## .. .. Num = col_double(), ## .. .. Variable = col_character(), ## .. .. Type = col_character(), ## .. .. Len = col_double(), ## .. .. Format = col_character(), ## .. .. Informat = col_character(), ## .. .. Label = col_character() ## .. .. ) ## ..- attr(*, &quot;problems&quot;)=&lt;externalptr&gt; ## $ : spc_tbl_ [46 × 7] (S3: spec_tbl_df/tbl_df/tbl/data.frame) ## ..$ Num : num [1:46] 1 2 3 4 5 6 NA NA 7 NA ... ## ..$ Variable: chr [1:46] &quot;ABP_1000&quot; &quot;ABP_1010&quot; &quot;ABP_1020&quot; &quot;ABP_1030&quot; ... ## ..$ Type : chr [1:46] &quot;Num&quot; &quot;Num&quot; &quot;Num&quot; &quot;Num&quot; ... ## ..$ Len : num [1:46] 8 8 8 8 8 8 NA NA 8 NA ... ## ..$ Format : num [1:46] 2 2 2 2 2 2 NA NA 2 NA ... ## ..$ Informat: num [1:46] 2 2 2 2 2 2 NA NA 2 NA ... ## ..$ Label : chr [1:46] &quot;Are you currently retired? 1=Yes,0=No&quot; &quot;Are you retired because of asthma? 1=Yes,0=No&quot; &quot;Are you currently unemployed? 1=Yes,0=No&quot; &quot;Are you unemployed because of asthma? 1=Yes,0=No&quot; ... ## ..- attr(*, &quot;spec&quot;)= ## .. .. cols( ## .. .. Num = col_double(), ## .. .. Variable = col_character(), ## .. .. Type = col_character(), ## .. .. Len = col_double(), ## .. .. Format = col_double(), ## .. .. Informat = col_double(), ## .. .. Label = col_character() ## .. .. ) ## ..- attr(*, &quot;problems&quot;)=&lt;externalptr&gt; ## $ : spc_tbl_ [21 × 7] (S3: spec_tbl_df/tbl_df/tbl/data.frame) ## ..$ Num : num [1:21] 20 NA NA 21 NA NA 22 NA NA 23 ... ## ..$ Variable: chr [1:21] &quot;ABP_1190&quot; NA NA &quot;ABP_1200&quot; ... ## ..$ Type : chr [1:21] &quot;Num&quot; NA NA &quot;Num&quot; ... ## ..$ Len : num [1:21] 8 NA NA 8 NA NA 8 NA NA 8 ... ## ..$ Format : chr [1:21] &quot;2.&quot; NA NA &quot;2.&quot; ... ## ..$ Informat: chr [1:21] &quot;2.&quot; NA NA &quot;2.&quot; ... ## ..$ Label : chr [1:21] &quot;How much bother is the worry that you will have an asthma attack when&quot; &quot;visiting a new place? 0=No bother at all, 1=Minor irritation, 2=Slight&quot; &quot;bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery&quot; &quot;How much bother is the worry that you will catch a cold? 0=No bother at&quot; ... ## ..- attr(*, &quot;spec&quot;)= ## .. .. cols( ## .. .. Num = col_double(), ## .. .. Variable = col_character(), ## .. .. Type = col_character(), ## .. .. Len = col_double(), ## .. .. Format = col_character(), ## .. .. Informat = col_character(), ## .. .. Label = col_character() ## .. .. ) ## ..- attr(*, &quot;problems&quot;)=&lt;externalptr&gt; Exploring further, here is how each dataframe is formatted: datatable(tables[[1]]) Notice that, although the dataframe format mirrors the PDF table format, the label column is stored across multiple rows with NAs in the other columns of that row because the text was across multiple lines. In our final dataframe, we will want the entire block of text in one cell. We can also remove the “Len”, “Format”, and “Informat” columns because they are not informative and they are not found in every table. Next, we will walk through how to clean up this table using a series of steps in tidyverse. Cleaning dataframes First, we will select the columns we are interested in and use the fill() function to change the NAs in the “Num” column so that each line of text in the “Label” column has the correct “Num” value in the same row. cleaned_table1 &lt;- data.frame(tables[[1]]) %&gt;% # Extract the first table in the list # Select only the columns of interest select(c(Num, Variable, Type, Label)) %&gt;% # Change the &quot;Num&quot; column to numeric, which is required for the fill function mutate(Num = as.numeric(Num)) %&gt;% # Fill in the NAs in the &quot;Num&quot; column down the column fill(Num, .direction = &quot;down&quot;) datatable(cleaned_table1) We still need to move all of the Label text for each variable into one cell in one row instead of across multiple rows. For this, we can use the unlist() function. Here is a demonstration of how the unlist() function works using just the first variable: cleaned_table1_var1 &lt;- cleaned_table1 %&gt;% # Filter dataframe to just contain rows associated with the first variable filter(Num == 1) %&gt;% # Paste all character strings in the Label column with a space in between them into a new column called &quot;new_label&quot; mutate(new_label = paste(unlist(Label), collapse = &quot; &quot;)) datatable(cleaned_table1_var1) We now have all of the text we want in one cell, but we have duplicate rows that we don’t need. We can get rid of these rows by assigning blank values “NA” and then omitting rows that contain NAs. cleaned_table1_var1 &lt;- cleaned_table1_var1 %&gt;% mutate(across(Variable, na_if, &quot;&quot;)) %&gt;% na.omit() datatable(cleaned_table1_var1) We need to apply this code to the whole dataframe and not just one variable, so we can add group_by(Num) to our cleaning workflow, followed by the code we just applied to our filtered dataframe. cleaned_table1 &lt;- data.frame(tables[[1]]) %&gt;% # Extract the first table in the list # Select only the columns of interest select(c(Num, Variable, Type, Label)) %&gt;% # Change the &quot;Num&quot; column to numeric, which is required for the fill function mutate(Num = as.numeric(Num)) %&gt;% # Fill in the NAs in the &quot;Num&quot; column down the column fill(Num, .direction = &quot;down&quot;) %&gt;% # Group by variable number group_by(Num) %&gt;% # Unlist the text replace the text in the &quot;Label&quot; column with the unlisted text mutate(Label = paste(unlist(Label), collapse =&quot; &quot;)) %&gt;% # Make blanks in the &quot;Variable&quot; column into NAs mutate(across(Variable, na_if, &quot;&quot;)) %&gt;% # Remove rows with NAs na.omit() datatable(cleaned_table1) Ultimately, we need to clean up each dataframe in the list the same way, and we need all of the dataframes to be in one dataframe, instead of in a list. There are a couple of different ways to do this. Both rely on the code shown above for cleaning up each dataframe. Option #1 uses a for loop, while Option #2 uses application of a function on the list of dataframes. Both result in the same ending dataframe! Option #1 # Create a dataframe for storing variables variables &lt;- data.frame() # Make a for loop to format each dataframe and add it to the variables for (i in 1:length(tables)) { table &lt;- data.frame(tables[[i]]) %&gt;% select(c(Num, Variable, Type, Label)) %&gt;% mutate(Num = as.numeric(Num)) %&gt;% fill(Num, .direction = &quot;down&quot;) %&gt;% group_by(Num) %&gt;% mutate(Label = paste(unlist(Label), collapse =&quot; &quot;)) %&gt;% mutate(across(Variable, na_if, &quot;&quot;)) %&gt;% na.omit() variables &lt;- bind_rows(variables, table) } # View resulting dataframe datatable(variables) Option #2 # Write a function that applies all of the cleaning steps to an dataframe (output = cleaned dataframe) clean_tables &lt;- function(data) { data &lt;- data %&gt;% select(c(Num, Variable, Type, Label)) %&gt;% mutate(Num = as.numeric(Num)) %&gt;% fill(Num, .direction = &quot;down&quot;) %&gt;% group_by(Num) %&gt;% mutate(Label = paste(unlist(Label), collapse =&quot; &quot;)) %&gt;% mutate(across(Variable, na_if, &quot;&quot;)) %&gt;% na.omit() return(data) } # Apply the function over each table in the list of tables tables_clean &lt;- lapply(X = tables, FUN = clean_tables) # Unlist the dataframes and combine them into one dataframe tables_clean_unlisted &lt;- do.call(rbind, tables_clean) # View resulting dataframe datatable(tables_clean_unlisted) Adding Dataset Names We now have a dataframe with all of the information from the PDFs contained in one long table. However, now we need to add back in the label on top of each table. We can’t do this with the tabulapdf package because the name isn’t stored in the table. But we can use the pdftools package for this! First, we will read in the pdf using the PDF tools package. This results in a vector containing a long character string for each page of the PDF. Notice a few features of these character strings: Each line is separated by \\n Elements [1] and [2] of the vector contain the text “dataset Name:”, while element [3] does not because the third page was a continuation of the table from the second page and therefore did not have a table title. table_names &lt;- pdf_text(&quot;Module4_3_Input/Module4_3_InputData4.pdf&quot;) head(table_names[1:3]) ## [1] &quot; 10:13 Tuesday, February 11, 2020 1\\n\\n\\nData Set Name: aaaq.sas7bdat\\n\\nNum Variable Type Len Format Informat Label\\n 1 AAQ_1000 Num 8 2. 2. In the past 3 days, how much of the time did your asthma keep you from\\n doing your usual activities at work, school, or at home? 0=None of the\\n time, 1=A little of the time, 2=Some of the time, 3=Most of the time, 4=All\\n of the time\\n 2 AAQ_1010 Num 8 2. 2. During the past 3 days, how often have you had asthma symptoms? 0=Not\\n at all, 1=Once per day, 2=2-3 times per day, 3=4-5 times per day, 4=6 or\\n more times per day\\n 3 AAQ_1020 Num 8 2. 2. During the past 3 days, how often have you used your rescue inhaler or\\n nebulizer medication ? 0=Not at all, 1=Once per day, 2=2-3 times per day,\\n 3=4-5 times per day, 4=6 or more times per day\\n 4 AAQ_1030 Num 8 2. 2. During the past 3 days, how many total times did your asthma symptoms\\n wake you up from sleep? 0=Not at all, 1=1 time in the last 3 days, 2=2-3\\n times in the last 3 days, 3=4-5 times in the last 3 days, 4=&gt;=6 times in the\\n last 3 days\\n 5 AAQ_1040 Num 8 2. 2. How would you rate the amount of impairment you have experienced due\\n to your asthma in the past 3 days? 0=No impairment, 1=Mild impairment,\\n 2=Moderate impairment, 3=Severe impairment, 4=Very severe impairment\\n 6 AAQ_1050 Num 8 2. 2. How stressed or frightened were you by your asthma symptoms in the past\\n 3 days? 0=Not at all, 1=Mildly, 2=Moderately, 3=Severely, 4=Very\\n severely\\n 7 AAQ_1060 Num 8 2. 2. Why do you think your asthma was worse in the past 3 days compared to\\n what is normal for you? 0=I have not been worse over the past 3 days. My\\n asthma symptoms have been usual., 1=Common cold, 2=Allergies,\\n 3=Pollution or chemical irritant, 4=Too little asth\\n 8 VNUM_C Char 3 $3. $3. Visit Number (character)\\n 9 VNUM Num 8 Visit Number (numeric)\\n 10 VDATE Num 8 Number of days from Visit 1 to this visit\\n 11 RAND_ID Char 6 Randomized Master ID\\n 12 ENROLL_TYPE Char 15 Enrollment Type (Screen Fail, Randomized, Healthy Control)\\n 13 ENROLL_ORDER Num 8 Enrollment Order Number\\n&quot; ## [2] &quot; 10:13 Tuesday, February 11, 2020 2\\n\\n\\n\\nData Set Name: abp.sas7bdat\\n\\nNum Variable Type Len Format Informat Label\\n 1 ABP_1000 Num 8 2. 2. Are you currently retired? 1=Yes,0=No\\n 2 ABP_1010 Num 8 2. 2. Are you retired because of asthma? 1=Yes,0=No\\n 3 ABP_1020 Num 8 2. 2. Are you currently unemployed? 1=Yes,0=No\\n 4 ABP_1030 Num 8 2. 2. Are you unemployed because of asthma? 1=Yes,0=No\\n 5 ABP_1040 Num 8 2. 2. Do you get paid to do work? 1=Yes,0=No\\n 6 ABP_1050 Num 8 2. 2. How much does your asthma bother you at your paid work? 0=No bother\\n at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of\\n bother, 5=Makes my life a misery\\n 7 ABP_1060 Num 8 2. 2. Overall, how much does your asthma bother you when you do jobs\\n around the house? 0=No bother at all, 1=Minor irritation, 2=Slight bother,\\n 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery,\\n 0=None of these really apply to me\\n 8 ABP_1070 Num 8 2. 2. Overall, how much does your asthma bother your social life? 0=No bother\\n at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of\\n bother, 5=Makes my life a misery\\n 9 ABP_1080 Num 8 2. 2. Overall, how much does your asthma bother your personal life? 0=No\\n bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother,\\n 4=A lot of bother, 5=Makes my life a misery, 0=None of these really\\n apply to me\\n 10 ABP_1090 Num 8 2. 2. Are you involved in leisure activities, such as: walking for pleasure,\\n sports, exercise, travelling, taking vacations? 1=Yes,0=No\\n 11 ABP_1100 Num 8 2. 2. When involved in leisure activities, how much does your asthma bother\\n you? 0=No bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate\\n bother, 4=A lot of bother, 5=Makes my life a misery\\n 12 ABP_1110 Num 8 2. 2. Would you say that you can&#39;t do some of these sorts of things because of\\n asthma? 1=Yes,0=No\\n 13 ABP_1120 Num 8 2. 2. How much does your asthma bother you when you sleep? 0=No bother at\\n all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of\\n bother, 5=Makes my life a misery\\n 14 ABP_1130 Num 8 2. 2. How much does the cost of your asthma medicines bother you? 0=No\\n bother at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother,\\n 4=A lot of bother, 5=Makes my life a misery\\n 15 ABP_1140 Num 8 2. 2. Do you get free prescriptions? 1=Yes,0=No\\n 16 ABP_1150 Num 8 2. 2. How much does the inconvenience or embarrassment of taking your\\n asthma medicines bother you? 0=No bother at all, 1=Minor irritation,\\n 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life\\n a misery\\n 17 ABP_1160 Num 8 2. 2. How much do coughs and colds bother you? 0=No bother at all, 1=Minor\\n irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of bother,\\n 5=Makes my life a misery, 0=None of these really apply to me\\n 18 ABP_1170 Num 8 2. 2. Feeling upset is also a bother. Does your asthma make you feel anxious,\\n depressed, tired, or helpless? 1=Yes,0=No\\n 19 ABP_1180 Num 8 2. 2. Feeling upset is also a bother. Does your asthma make you feel anxious,\\n depressed, tired, or helpless? 0=No bother at all, 1=Minor irritation,\\n 2=Slight bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life\\n a misery\\n&quot; ## [3] &quot; 10:13 Tuesday, February 11, 2020 3\\n\\n\\nNum Variable Type Len Format Informat Label\\n 20 ABP_1190 Num 8 2. 2. How much bother is the worry that you will have an asthma attack when\\n visiting a new place? 0=No bother at all, 1=Minor irritation, 2=Slight\\n bother, 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery\\n 21 ABP_1200 Num 8 2. 2. How much bother is the worry that you will catch a cold? 0=No bother at\\n all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of\\n bother, 5=Makes my life a misery\\n 22 ABP_1210 Num 8 2. 2. How much bother is the worry that you will let others down? 0=No bother\\n at all, 1=Minor irritation, 2=Slight bother, 3=Moderate bother, 4=A lot of\\n bother, 5=Makes my life a misery\\n 23 ABP_1220 Num 8 2. 2. How much bother is the worry that your health may get worse in the\\n future? 0=No bother at all, 1=Minor irritation, 2=Slight bother,\\n 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery\\n 24 ABP_1230 Num 8 2. 2. How much bother is the worry that you won&#39;t be able to cope with an\\n asthma attack? 0=No bother at all, 1=Minor irritation, 2=Slight bother,\\n 3=Moderate bother, 4=A lot of bother, 5=Makes my life a misery\\n 25 VNUM_C Char 3 $3. $3. Visit Number (character)\\n 26 VNUM Num 8 Visit Number (numeric)\\n 27 VDATE Num 8 Number of days from Visit 1 to this visit\\n 28 RAND_ID Char 6 Randomized Master ID\\n 29 ENROLL_TYPE Char 15 Enrollment Type (Screen Fail, Randomized, Healthy Control)\\n 30 ENROLL_ORDER Num 8 Enrollment Order Number\\n&quot; Similar to the table cleaning section, we will work through an example of extracting the text of interest from one of these character vectors, then apply the same code to all of the character vectors. First, we will select just the first element in the vector and make it into a dataframe. # Create dataframe dataset_name_df_var1 &lt;- data.frame(strsplit(table_names[1], &quot;\\n&quot;)) # Clean column name colnames(dataset_name_df_var1) &lt;- c(&quot;Text&quot;) # View dataframe datatable(dataset_name_df_var1) Next, we will extract the dataset name using the same approach used in extracting values from the nanoparticle tracking example above and assign the name to a variable. We filter by the string “Data Set Name” because this is the start of the text string in the row where our dataset name is stored and is the same across all of our datasets. # Create dataframe dataset_name_df_var1 &lt;- dataset_name_df_var1 %&gt;% filter(grepl(&quot;Data Set Name&quot;, dataset_name_df_var1$Text)) %&gt;% separate(Text, into = c(NA, &quot;dataset&quot;), sep = &quot;Data Set Name: &quot;) # Assign variable dataset_name_var1 &lt;- dataset_name_df_var1[1,1] # View variable name dataset_name_var1 ## [1] &quot;aaaq.sas7bdat&quot; Now that we have the dataset name stored as a variable, we can create a dataframe that will correspond to the rows in our variables dataframe. The challenge is that each dataset contains a different number of variables! We can determine how many rows each dataset contains by returning to our variables dataframe and calculating the number of rows associated with each dataset. The following code splits the variables dataframe into a list of dataframes by each occurrence of 1 in the “Num” column (when the numbering restarts for a new dataset). # Calculate the number of rows associated with each dataset for reference dataset_list &lt;- split(variables, cumsum(variables$Num == 1)) glimpse(dataset_list[1:3]) ## List of 3 ## $ 1:&#39;data.frame&#39;: 13 obs. of 4 variables: ## ..$ Num : num [1:13] 1 2 3 4 5 6 7 8 9 10 ... ## ..$ Variable: chr [1:13] &quot;AAQ_1000&quot; &quot;AAQ_1010&quot; &quot;AAQ_1020&quot; &quot;AAQ_1030&quot; ... ## ..$ Type : chr [1:13] &quot;Num&quot; &quot;Num&quot; &quot;Num&quot; &quot;Num&quot; ... ## ..$ Label : chr [1:13] &quot;In the past 3 days, how much of the time did your asthma keep you from doing your usual activities at work, sch&quot;| __truncated__ &quot;During the past 3 days, how often have you had asthma symptoms? 0=Not at all, 1=Once per day, 2=2-3 times per d&quot;| __truncated__ &quot;During the past 3 days, how often have you used your rescue inhaler or nebulizer medication ? 0=Not at all, 1=O&quot;| __truncated__ &quot;During the past 3 days, how many total times did your asthma symptoms wake you up from sleep? 0=Not at all, 1=1&quot;| __truncated__ ... ## $ 2:&#39;data.frame&#39;: 30 obs. of 4 variables: ## ..$ Num : num [1:30] 1 2 3 4 5 6 7 8 9 10 ... ## ..$ Variable: chr [1:30] &quot;ABP_1000&quot; &quot;ABP_1010&quot; &quot;ABP_1020&quot; &quot;ABP_1030&quot; ... ## ..$ Type : chr [1:30] &quot;Num&quot; &quot;Num&quot; &quot;Num&quot; &quot;Num&quot; ... ## ..$ Label : chr [1:30] &quot;Are you currently retired? 1=Yes,0=No&quot; &quot;Are you retired because of asthma? 1=Yes,0=No&quot; &quot;Are you currently unemployed? 1=Yes,0=No&quot; &quot;Are you unemployed because of asthma? 1=Yes,0=No&quot; ... ## $ 3:&#39;data.frame&#39;: 11 obs. of 4 variables: ## ..$ Num : num [1:11] 1 2 3 4 5 6 7 8 9 10 ... ## ..$ Variable: chr [1:11] &quot;ACT_1&quot; &quot;ACT_2&quot; &quot;ACT_3&quot; &quot;ACT_4&quot; ... ## ..$ Type : chr [1:11] &quot;Num&quot; &quot;Num&quot; &quot;Num&quot; &quot;Num&quot; ... ## ..$ Label : chr [1:11] &quot;In the past 4 weeks, how much of the time did your asthma keep you from getting as much done at work, school or&quot;| __truncated__ &quot;During the past 4 weeks, how often have your had shortness of breath? 1=All of the time, 2=Most of the time, 3=&quot;| __truncated__ &quot;During the past 4 weeks, how often did your asthma symptoms wake you up at night or earlier than usual in the m&quot;| __truncated__ &quot;During the past 4 weeks, how often have you used your rescue inhaler or nebulizer medication? 1=All of the time&quot;| __truncated__ ... The number of rows in each list is the number of variables in that dataset. We can use this value in creating our dataframe of dataset names. # Store the number of rows in a variable n_rows = nrow(data.frame(dataset_list[1])) # Repeat the dataset name for the number of variables there are dataset_name_var1 = data.frame(&quot;dataset_name&quot; = rep(dataset_name_var1, times = n_rows)) # View data farme datatable(dataset_name_var1) We now have a dataframe that can be joined with our variables dataframe for the first table. We can apply this approach to each table in our original PDF using a for loop. # Make dataframe to store dataset names dataset_names &lt;- data.frame() # Create list of datasets dataset_list &lt;- split(variables, cumsum(variables$Num == 1)) # Remove elements from the table_names vector that do not contain the string &quot;Data Set Name&quot; table_names_filtered &lt;- stringr::str_subset(table_names, &#39;Data Set Name&#39;) # Populate dataset_names dataframe for (i in 1:length(table_names_filtered)) { # Get dataset name dataset_name_df &lt;- data.frame(strsplit(table_names_filtered[i], &quot;\\n&quot;)) base::colnames(dataset_name_df) &lt;- c(&quot;Text&quot;) dataset_name_df &lt;- dataset_name_df %&gt;% filter(grepl(&quot;Data Set Name&quot;, dataset_name_df$Text)) %&gt;% separate(Text, into = c(NA, &quot;dataset&quot;), sep = &quot;Data Set Name: &quot;) dataset_name &lt;- dataset_name_df[1,1] # Determine number of variables in that dataset data_set &lt;- data.frame(dataset_list[i]) n_rows = nrow(data_set) # Repeat the dataset name for the number of variables there are dataset_name = data.frame(&quot;Data Set Name&quot; = rep(dataset_name, times = n_rows)) # Bind to dataframe dataset_names &lt;- bind_rows(dataset_names, dataset_name) } # Rename column colnames(dataset_names) &lt;- c(&quot;Data Set Name&quot;) # View datatable(dataset_names) Combining Dataset Names and Variable Information Last, we will merge together the dataframe containing dataset names and variable information. # Merge together final_variable_df &lt;- cbind(dataset_names, variables) %&gt;% rename(&quot;Variable Description&quot; = &quot;Label&quot;, &quot;Variable Number Within Dataset&quot; = &quot;Num&quot;) %&gt;% clean_names() datatable(final_variable_df) We can also determine how many total variables we have, all of which are accessible via the table we just generated. # Total number of variables nrow(final_variable_df) ## [1] 1190 # Total number of variables With this, we can answer Environmental Health Question #2: How many variables total are available to us to request from the study whose data are stored in the repository, and what are these variables? Answer: There are 1190 variable available to us. We can browse through the variables, including the sub-table they were from, the type of variable they are, and how they were derived using the table we generated. Concluding Remarks This training module provides example case studies demonstrating how to import PDF data into R and clean it so that it is more useful and accessible for analyses. The approaches demonstrated in this module, though specific to our specific example data, can be adapted to many different types of PDF data. Test Your Knowledge Using the same input files that we used in part 1, “Importing Data from Many Single PDFs with the Same Formatting”, found in the Module4_3_TYKInput folder, extract the remaining variables of interest (Original Concentration and Positions Removed) from the PDFs and summarize them in one dataframe. "],["two-group-comparisons-and-visualizations.html", "Two-Group Comparisons and Visualizations Introduction to Training Module Overview of Two Group Statistical Tests Statistical vs Biological Significance Unpaired Test Example Paired Test Example Visualizing Results Concluding Remarks", " Two-Group Comparisons and Visualizations This training module was developed by Elise Hickman, Alexis Payton, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module Two group statistical comparisons, in which we want to know whether the means between two different groups are significantly different, are some of the most common statistical tests in environmental health research and even biomedical research as a field. In this training module, we will demonstrate how to run two group statistical comparisons and how to present publication-quality figures and tables of these results. We will continue to use the same example dataset as used in this chapter’s previous modules, which represents concentrations of inflammatory biomarkers secreted by airway epithelial cells after exposure to different concentrations of acrolein. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: Are there significant differences in inflammatory biomarker concentrations between cells from male and female donors at baseline? Are there significant differences in inflammatory biomarker concentrations between cells exposed to 0 and 4 ppm acrolein? Workspace Preparation and Data Import Here, we will import the processed data that we generated at the end of TAME 2.0 Module 4.2 Data Import, Processing, and Summary Statistics. These data, along with the associated demographic data, were introduced in TAME 2.0 Module 4.1 Overview of Experimental Design and Example Data. These data represent log2 concentrations of inflammatory biomarkers secreted by airway epithelial cells after exposure to four different concentrations of acrolein (plus filtered air as a control). We will also load packages that will be needed for the analysis, including previously introduced packages such as openxlsx, tidyverse, DT, and ggpubr, and additional packages relevant to statistical analysis and graphing that will be discussed in greater detail below. # Load packages library(openxlsx) library(tidyverse) library(DT) library(rstatix) library(ggpubr) # Import data biomarker_data &lt;- read.xlsx(&quot;Module4_4_Input/Module4_4_InputData1.xlsx&quot;) demographic_data &lt;- read.xlsx(&quot;Module4_4_Input/Module4_4_InputData2.xlsx&quot;) # View data datatable(biomarker_data) datatable(demographic_data) Overview of Two Group Statistical Tests Before applying statistical tests to our data, let’s first review common two-group statistical tests, their underlying assumptions, and variations on these tests. Common Tests The two most common two-group statistical tests are the… t-test (also known as the student’s t-test) and the Wilcoxon test (also known as the Wilcox test, Wilcoxon test, or Mann Whitney test) Both of these tests are testing the null hypothesis that the means of the two populations (groups) are the same; the alternative hypothesis is that they are not the same. A significant p-value means that we can reject the null hypothesis that the means of the two groups are the same. Whether or not a p-value meets criteria for significance is experiment-specific, though commonly implemented p-value filters for significance include p&lt;0.05 and p&lt;0.01. P-values can also be called alpha values, and they indicate the probability of a type I error, or false positive, where the null hypothesis is rejected despite it actually being true. On the other hand, a type II error, or false negative, occurs when the null hypothesis is not rejected when it actually should have been. Assumptions The main difference between these two tests is in the assumption about the underlying distribution of the data. T-tests assume that the data are pulled from a normal distribution, while Wilcoxon tests do not assume that the data are pulled from a normal distribution. Therefore, it is most appropriate to use a t-test when data are, in general, normally distributed and a Wilcoxon test when data are not normally distributed. Additional assumptions underlying t-tests and Wilcoxon test are: The dependent variable is continuous or ordinal (discrete, ordered values). The data is collected from a representative, random sample. T-tests also assume that: The standard deviations of the two groups are approximately equal (also called homogeneity of variance). When to Use a Parametric vs Non-Parametric Test? Deciding whether to use a parametric or non-parametric test isn’t a one size fits all approach, and the decision should be made holistically for each dataset. Typically, parametric tests should be used when the data are normally distributed, continuous, random sampled, without extreme outliers, and representative of independent samples or participants. A non-parametric test can be used when the sample size (n) is small, outliers are present in the dataset, and/or the data are not normally distributed. This decision matters more when dealing with smaller sample sizes (n&lt;10) as smaller sample sizes are more prone to being skewed, and parametric tests are more sensitive to outliers. Therefore, when dealing with a smaller n, it might be best to perform a data transformation as discussed in TAME 2.0 Module 3.3 Normality Testing &amp; Data Transformations and then perform a parametric test if more parametric assumptions are able to be met, or to use non-parametric tests. For larger sample sizes (n&gt;50), outliers can potentially be removed and the dataset can be retested for assumptions. Lastly, what’s considered “small” or “large” in regards to sample size can be subjective and should be taken into consideration within the context of the experiment. Variations Unequal Variance: When the assumption of homogeneity of variance is not met, a Welch’s t-test is generally preferred over a student’s t-test. This can be implemented easily by setting var.equal = FALSE as an argument to the function executing the t-test (e.g., t.test(), t_test()). For more on testing homogeneity of variance in R, see here. Paired vs Unpaired: Variations on the t-test and Wilcoxon test are used when the experimental design is paired (also called repeated measures or matching). This occurs when there are different treatments, exposures, or time points collected from the same biological/experimental unit. For example, cells from the same donor or passage number exposed to different concentrations of a chemical represents a paired design. Matched/paired experiments have increased power to detect significant differences because samples can be compared back to their own controls. One vs Two-Sided: A one-sided test evaluates the hypothesis that the mean of the treatment group significantly differs in a specific direction from the control. A two-sided test evaluates the hypothesis that the mean of the treatment group significantly differs from the control but does not specify a direction for that change. A two-sided test is the preferred approach and the default in R because, typically, either direction of change is possible and represents an informative finding. However, one-sided tests may be appropriate if an effect can only possibly occur in one direction. This can be implemented by setting alternative = \"one.sided\" within the statistical testing function. Which test should I choose? We provide the following flowchart to help guide your choice of statistical test to compare two groups: Figure 10: Created with BioRender.com Statistical vs Biological Significance Another important topic to discuss before proceeding to statistical testing is the true meaning of statistical significance. Statistical significance simply means that it is unlikely that the patterns being observed are due to random chance. However, just because an effect is statistically significant does not mean that it is biologically significant (i.e., has notable biological consequences). Often, there also needs to be a sufficient magnitude of effect (also called effect size) for the effects on a system to be meaningful. Although a p-value &lt; 0.05 is often considered the threshold for significance, this is just a standard threshold set to a generally “acceptable” amount of error (5%). What about a p-value of 0.058 with a very large biological effect? Accounting for effect size is also why filters such as log2 fold change are often applied alongside p-value filters in -omics based analysis. In discussions of effect size, the population size is also a consideration - a small percentage increase in a very large population can represent tens of thousands of individuals (or more). Another consideration is that we frequently do not know what magnitude of biological effect should be considered “significant.” These discussions can get complicated very quickly, and here we do not propose to have a solution to these thought experiments; rather, we recommend considering both statistical and biological significance when interpreting data. And, as stated in other sections of TAME, transparent reporting of statistical results will aid the audience in interpreting the data through their preferred perspectives. Unpaired Test Example We will start by performing a statistical test to determine whether there are significant differences in biomarker concentrations between male and female donors at baseline (0 ppm exposure). Previously we determined that the majority of our data was non-normally distributed (see TAME 2.0 Module 4.2 Data Import, Processing, and Summary Statistics), so we’ll skip testing for that assumption in this module. Based on those results, we will use the Wilcoxon test to determine if there are significant differences between groups. The Wilcoxon test does not assume homogeneity of variance, so we do not need to test for that prior to applying the test. This is an unpaired analysis because samples collected from the cells derived from male and female donor cells are different sets of cells (i.e., independent from each other). Thus, the specific statistical test applied will be the Wilcoxon Rank Sum test. First, we will filter our dataframe to only data representing the control (0 ppm) exposure: biomarker_data_malevsfemale &lt;- biomarker_data %&gt;% filter(Dose == &quot;0&quot;) Next, we need to add the demographic data to our dataframe: biomarker_data_malevsfemale &lt;- biomarker_data_malevsfemale %&gt;% left_join(demographic_data %&gt;% select(Donor, Sex), by = &quot;Donor&quot;) Here is what our data look like now: datatable(biomarker_data_malevsfemale) We can demonstrate the basic anatomy of the Wilcoxon test function wilcox.test() by running the function on just one variable. wilcox.test(IL1B ~ Sex, data = biomarker_data_malevsfemale) ## ## Wilcoxon rank sum exact test ## ## data: IL1B by Sex ## W = 29, p-value = 0.8371 ## alternative hypothesis: true location shift is not equal to 0 The p-value of 0.8371 indicates that males and females do not have significantly different concentrations of IL-1\\(\\beta\\). The wilcox.test() function is part of the pre-loaded package stats. The package rstatix provides identical statistical tests to stats but in a pipe-friendly (tidyverse-friendly) format, and these functions output results as dataframes rather than the text displayed above. biomarker_data_malevsfemale %&gt;% wilcox_test(IL1B ~ Sex) ## # A tibble: 1 × 7 ## .y. group1 group2 n1 n2 statistic p ## * &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;int&gt; &lt;int&gt; &lt;dbl&gt; &lt;dbl&gt; ## 1 IL1B F M 7 9 29 0.837 Here, we can see the exact same results as with the wilcox.test() function. For the rest of this module, we’ll proceed with using the rstatix version of statistical testing functions. Although it is simple to run the Wilcoxon test with the code above, it’s impractical for a large number of endpoints and doesn’t store the results in an organized way. Instead, we can run the Wilcoxon test over every variable of interest using a for loop. There are also other ways you could approach this, such as a function applied over a list. This for loop runs the Wilcoxon test on each endpoint, stores the results in a dataframe, and then binds together the results dataframes for each variable of interest. Note that you could easily change wilcox_test() to t_test() and add additional arguments to modify the way the statistical test is run. # Create a vector with the names of the variables you want to run the test on endpoints &lt;- colnames(biomarker_data_malevsfemale %&gt;% select(IL1B:VEGF)) # Create dataframe to store results sex_wilcoxres &lt;- data.frame() # Run for loop for (i in 1:length(endpoints)) { # Assign a name to the endpoint variable. endpoint &lt;- endpoints[i] # Run wilcox test and store in results dataframe. res_df &lt;- biomarker_data_malevsfemale %&gt;% wilcox_test(as.formula(paste0(endpoint, &quot;~ Sex&quot;, sep = &quot;&quot;))) # Bind results from this test with other tests in this loop sex_wilcoxres &lt;- rbind(sex_wilcoxres, res_df) } # View results sex_wilcoxres ## # A tibble: 6 × 7 ## .y. group1 group2 n1 n2 statistic p ## &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;int&gt; &lt;int&gt; &lt;dbl&gt; &lt;dbl&gt; ## 1 IL1B F M 7 9 29 0.837 ## 2 IL6 F M 7 9 28.5 0.791 ## 3 IL8 F M 7 9 34 0.832 ## 4 IL10 F M 7 9 43 0.252 ## 5 TNFa F M 7 9 39 0.458 ## 6 VEGF F M 7 9 30 0.916 With this, we can answer Environmental Health Question #1: Are there significant differences in inflammatory biomarker concentrations between cells from male and female donors at baseline? Answer: There are not any significant differences in concentrations of any of our biomarkers between male and female donors at baseline. Adjusting for Multiple Hypothesis Testing Above, we compared concentrations between males and females for six different endpoints or variables. Each time we run a comparison (with a p-value threshold of &lt; 0.05), we are accepting that there is a 5% chance that a significant result will actually be due to random chance and that we are rejecting the null hypothesis when it is actually true (type I error). Since we are testing six different hypotheses simultaneously, what is the probability then of observing at least one significant result due just to chance? \\[\\mathbb{P}({\\rm At Least One Significant Result}) = 1 - \\mathbb{P}({\\rm NoSignificantResults}) = 1 - (1 - 0.05)^{6} = 0.26\\] Here, we can see that we have a 26% chance of observing at least one significant result, even if all the tests are actually not significant. This chance increases as our number of endpoints increases; therefore, adjusting for multiple hypothesis testing becomes even more important with larger datasets. Many methods exist for adjusting for multiple hypothesis testing, with some of the most popular including Bonferroni, False Discovery Rate (FDR), and Benjamini-Hochberg (BH). However, opinions about when and how to adjust for multiple hypothesis testing can vary and also depend on the question you are trying to answer. For example, when there are a low number of variables (e.g., &lt; 10), it’s often not necessary to adjust for multiple hypothesis testing, and when there are many variables (e.g., 100s to 1000s), it is necessary, but what about for an intermediate number of comparisons? Whether or not to apply multiple hypothesis test correction also depends on whether each endpoint is of interest on its own or whether the analysis seeks to make general statements about all of the endpoints together and on whether reducing type I or type II error is most important in the analysis. For this analysis, we will not adjust for multiple hypothesis testing due to our relatively low number of variables. For more on multiple hypothesis testing, check out the following publications: Mohieddin J; Naser AP. “Why, When and How to Adjust Your P Values?”. Cell Journal (Yakhteh), 20, 4, 2018, 604-607. doi: 10.22074/cellj.2019.5992 PUBMID: 30124010 Feise, R.J. Do multiple outcome measures require p-value adjustment?. BMC Med Res Methodol 2, 8 (2002). https://doi.org/10.1186/1471-2288-2-8 PUBMID: 12069695 Paired Test Example To demonstrate an example of a paired two-group test, we can also determine whether exposure to 4 ppm acrolein significantly changes biomarker concentrations. This is now a paired design because each donor’s cells were exposed to both 0 and 4 ppm acrolein. To prepare the data, we will filter the dataframe to only include 0 and 4 ppm: biomarker_data_0vs4 &lt;- biomarker_data %&gt;% filter(Dose == &quot;0&quot; | Dose == &quot;4&quot;) Let’s view the dataframe. Note how the measurements for each donor are next to each other - this an important element of the default handling of the paired analysis in R. The dataframe should have the donors in the same order for the 0 and 4 ppm data. datatable(biomarker_data_0vs4) We can now run the same type of loop that we ran before, changing the independent variable in the formula to ~ Dose and adding paired = TRUE to the wilcox_test() function. # Create a vector with the names of the variables you want to run the test on endpoints &lt;- colnames(biomarker_data_0vs4 %&gt;% select(IL1B:VEGF)) # Create dataframe to store results dose_wilcoxres &lt;- data.frame() # Run for loop for (i in 1:length(endpoints)) { # Assign a name to the endpoint variable. endpoint &lt;- endpoints[i] # Run wilcox test and store in results dataframe. res_df &lt;- biomarker_data_0vs4 %&gt;% wilcox_test(as.formula(paste0(endpoint, &quot;~ Dose&quot;, sep = &quot;&quot;)), paired = TRUE) # Bind results from this test with other tests in this loop dose_wilcoxres &lt;- rbind(dose_wilcoxres, res_df) } # View results dose_wilcoxres ## # A tibble: 6 × 7 ## .y. group1 group2 n1 n2 statistic p ## &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;int&gt; &lt;int&gt; &lt;dbl&gt; &lt;dbl&gt; ## 1 IL1B 0 4 16 16 18 0.00763 ## 2 IL6 0 4 16 16 114 0.0155 ## 3 IL8 0 4 16 16 0 0.0000305 ## 4 IL10 0 4 16 16 65 0.9 ## 5 TNFa 0 4 16 16 5 0.000305 ## 6 VEGF 0 4 16 16 28 0.0386 Although this dataframe contains useful information about our statistical test, such as the groups being compared, the sample size (n) of each group, and the test statistic, what we really want (and what would likely be shared in supplemental material), is a more simplified version of these results in table format and more detailed information (n, specific statistical test, groups being compared) in the table legend. We can clean up the results using the following code to make clearer column names and ensure that the p-values are formatted consistently. dose_wilcoxres &lt;- dose_wilcoxres %&gt;% select(c(.y., p)) %&gt;% mutate(p = format(p, digits = 3, scientific = TRUE)) %&gt;% rename(&quot;Variable&quot; = &quot;.y.&quot;, &quot;P-Value&quot; = &quot;p&quot;) datatable(dose_wilcoxres) With this, we can answer Environmental Health Question #2: Are there significant differences in inflammatory biomarker concentrations between cells exposed to 0 and 4 ppm acrolein? Answer: Yes, there are significant differences in IL-1\\(\\beta\\), IL-6, IL-8, TNF-\\(\\alpha\\), and VEGF concentrations between cells exposed to 0 and 4 ppm acrolein. Visualizing Results Now, let’s visualize our results using ggplot2. For an introduction to ggplot2 visualizations, see TAME 2.0 Modules 3.1 Data Visualizations and 3.2 Improving Data Visualizations, as well as the extensive online documentation available for ggplot2. Single Plots We will start by making a very basic box and whisker plot of the IL-1\\(\\beta\\) data with individual data points overlaid. It is best practice to show all data points, allowing the reader to view the whole spread of the data, which can be obscured by plots such as bar plots with mean and standard error. # Setting theme for plot theme_set(theme_bw()) # Making plot ggplot(biomarker_data_0vs4, aes(x = Dose, y = IL1B)) + geom_boxplot() + geom_jitter(position = position_jitter(0.15)) We could add statistical markings to denote significance to this graph manually in PowerPoint or Adobe Illustrator, but there are actually R packages that act as extensions to ggplot2 and will do this for you! Two of our favorites are ggpubr and ggsignif. Here is an example using ggpubr: ggplot(biomarker_data_0vs4, aes(x = Dose, y = IL1B)) + geom_boxplot() + geom_jitter(position = position_jitter(0.15)) + # Adding a p value from a paired Wilcoxon test stat_compare_means(method = &quot;wilcox.test&quot;, paired = TRUE) We can further clean up our figure by modifying elements of the plot’s theme, including the font sizes, axis range, colors, and the way that the statistical results are presented. Perfecting figures can be time consuming but ultimately worth it, because clear figures aid greatly in presenting a coherent story that is understandable to readers/listeners. ggplot(biomarker_data_0vs4, aes(x = Dose, y = IL1B)) + # outlier.shape = NA removes outliers geom_boxplot(aes(fill = Dose), outlier.shape = NA) + # Changing box plot colors scale_fill_manual(values = c(&quot;#BFBFBF&quot;, &quot;#EE2B2B&quot;)) + geom_jitter(size = 3, position = position_jitter(0.15)) + # Adding a p value from a paired Wilcoxon test stat_compare_means(method = &quot;wilcox.test&quot;, paired = TRUE, # Changing the value to asterisks and moving to the middle of the plot label = &quot;p.signif&quot;, label.x = 1.5, label.y = 4.5, size = 12) + ylim(2.5, 5) + # Changing y axis label labs(y = &quot;Log2(IL-1\\u03B2 (pg/mL))&quot;) + # Removing legend theme(legend.position = &quot;none&quot;, axis.title = element_text(color = &quot;black&quot;, size = 15), axis.title.x = element_text(vjust = -0.75), axis.title.y = element_text(vjust = 2), axis.text = element_text(color = &quot;black&quot;, size = 12)) Multiple plots Making one plot was relatively straightforward, but to graph all of our endpoints, we would either need to repeat that code chunk for each individual biomarker or write a function to create similar plots given a specific biomarker as input. Then, we would need to stitch together the individual plots in external software or using a package such as patchwork (which is a great package if you need to combine individual figures from different sources or different size ratios!). While these are workable solutions and would get us to the same place, ggplot2 actually contains a function - facet_wrap() - that can be used to graph multiple endpoints from the same groups in one figure panel, which takes care of a lot of the work for us! To prepare our data for facet plotting, first we will pivot it longer: biomarker_data_0vs4_long &lt;- biomarker_data_0vs4 %&gt;% pivot_longer(-c(Donor, Dose), names_to = &quot;variable&quot;, values_to = &quot;value&quot;) datatable(biomarker_data_0vs4_long) Then, we can use similar code to what we used to make our single graph, with a few modifications to plot multiple panels simultaneously and adjust the style of the plot. Although it is beyond the scope of this module to explain the mechanics of each line of code, here are a few specific things to note about the code below that may be helpful when constructing similar plots: To create the plot with all six endpoints instead of just one, we: Changed input dataframe from wide to long format Changed y = from one specific endpoint to value Added the facet_wrap() argument ~ variable tells the function to make an individual plot for each variable nrow = 2 tells the function to put the plots into two rows scales = \"free_y\" tells the function to allow each individual graph to have a unique y-scale that best shows all of the data on that graph labeller feeds the edited (more stylistically correct) names for each panel to the function To ensure that the statistical results appear cleanly, within stat_compare_means(), we: Added hide.ns = TRUE so that only significant results are shown Added label.x.npc = \"center\" and hjust = 0.5 to ensure that asterisks are centered on the plot and that the text is center justified To add padding along the y axis, allowing space for significance asterisks, we added scale_y_continuous(expand = expansion(mult = c(0.1, 0.4))) # Create clean labels for the graph titles new_labels &lt;- c(&quot;IL10&quot; = &quot;IL-10&quot;, &quot;IL1B&quot; = &quot;IL-1\\u03B2 &quot;, &quot;IL6&quot; = &quot;IL-6&quot;, &quot;IL8&quot; = &quot;IL-8&quot;, &quot;TNFa&quot; = &quot;TNF-\\u03b1&quot;, &quot;VEGF&quot; = &quot;VEGF&quot;) # Make graph ggplot(biomarker_data_0vs4_long, aes(x = Dose, y = value)) + # outlier.shape = NA removes outliers geom_boxplot(aes(fill = Dose), outlier.shape = NA) + # Changing box plot colors scale_fill_manual(values = c(&quot;#BFBFBF&quot;, &quot;#EE2B2B&quot;)) + geom_jitter(size = 1.5, position = position_jitter(0.15)) + # Adding a p value from a paired Wilcoxon test stat_compare_means(method = &quot;wilcox.test&quot;, paired = TRUE, # Changing the value to asterisks and moving to the middle of the plot label = &quot;p.signif&quot;, size = 10, hide.ns = TRUE, label.x.npc = &quot;center&quot;, hjust = 0.5) + # Adding padding y axis scale_y_continuous(expand = expansion(mult = c(0.1, 0.4))) + # Changing y axis label ylab(expression(Log[2]*&quot;(Concentration (pg/ml))&quot;)) + # Faceting by each biomarker facet_wrap(~ variable, nrow = 2, scales = &quot;free_y&quot;, labeller = labeller(variable = new_labels)) + # Removing legend theme(legend.position = &quot;none&quot;, axis.title = element_text(color = &quot;black&quot;, size = 12), axis.title.x = element_text(vjust = -0.75), axis.title.y = element_text(vjust = 2), axis.text = element_text(color = &quot;black&quot;, size = 10), strip.text = element_text(size = 12, face = &quot;bold&quot;)) An appropriate title for this figure could be: “Figure X. Exposure to 4 ppm acrolein increases inflammatory biomarker secretion in primary human bronchial epithelial cells. Groups were compared using the Wilcoxon signed rank test. * p &lt; 0.05, ** p &lt; 0.01, *** p &lt; 0.001, **** p &lt; 0.0001, n = 16 per group (paired).” Concluding Remarks In this module, we introduced two-group statistical tests, which are some of the most common statistical tests applied in biomedical research. We applied these tests to our example dataset and demonstrated how to produce publication-quality tables and figures of our results. Implementing a workflow such as this enables efficient analysis of wet-bench generated data and customization of output figures and tables suited to your personal preferences. Test Your Knowledge Functional endpoints from these cultures were also measured. These endpoints were: 1) Membrane Permeability (MemPerm), 2) Trans-Epithelial Electrical Resistance (TEER), 3) Ciliary Beat Frequency (CBF), and 4) Expression of Mucin (MUC5AC). These data were already processed and tested for normality (see Test Your Knowledge for TAME 2.0 Module 4.2 Data Import, Processing, and Summary Statistics), with results indicating that two of the endpoints are normally distributed and two non-normally distributed. Due to the relatively low n of this dataset, we therefore recommend using non-parametric statistical tests. Use the same processes demonstrated in this module and the provided data (“Module4_4_TYKInput1.xlsx” (functional data) and “Module4_4_TYKInput2.xlsx” (demographic data)), run analyses and make publication-quality figures and tables to answer the following questions to determine: Are there significant differences in functional endpoints between cells from male and female donors at baseline? Are there significant differences in functional endpoints between cells exposed to 0 and 4 ppm acrolein? Go ahead and use non-parametric tests for these analyses. "],["multi-group-and-multi-variable-comparisons-and-visualizations.html", "Multi-Group and Multi-Variable Comparisons and Visualizations Introduction to Training Module Overview of Multi-Group Statistical Tests Multi-Group Analysis Example Visualization of Multi-Group Statistical Results Concluding Remarks", " Multi-Group and Multi-Variable Comparisons and Visualizations This training module was developed by Elise Hickman, Alexis Payton, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module In the previous module, we covered how to apply two-group statistical testing, one of the most basic types of statistical tests. In this module, we will build on the concepts introduced previously to apply statistical testing to datasets with more than two groups, which are also very common in environmental health research. We will review common multi-group overall effects tests and post-hoc tests, and we will demonstrate how to apply these tests and how to graph the results using the same example dataset as in previous modules in this chapter, which represents concentrations of inflammatory biomarkers secreted by airway epithelial cells after exposure to different concentrations of acrolein. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: Are there significant differences in inflammatory biomarker concentrations between different doses of acrolein? Do TNF-\\(\\alpha\\) concentrations significantly increase with increasing dose of acrolein? Workspace Preparation and Data Import Here, we will import the processed data that we generated at the end of TAME 2.0 Module 4.2, introduced in TAME 2.0 Module 4.1 Overview of Experimental Design and Example Data and the associated demographic data. These data represent log2 concentrations of inflammatory biomarkers secreted by airway epithelial cells after exposure to four different concentrations of acrolein (plus filtered air as a control). We will also load packages that will be needed for the analysis, including previously introduced packages such as openxlsx, tidyverse, DT, ggpubr, and rstatix. # Load packages library(openxlsx) library(tidyverse) library(DT) library(rstatix) library(ggpubr) # Import data biomarker_data &lt;- read.xlsx(&quot;Module4_5_Input/Module4_5_InputData1.xlsx&quot;) demographic_data &lt;- read.xlsx(&quot;Module4_5_Input/Module4_5_InputData2.xlsx&quot;) # View data datatable(biomarker_data) datatable(demographic_data) Overview of Multi-Group Statistical Tests Before applying statistical tests to our data, let’s first review the mechanics of multi-group statistical tests, including overall effects tests and post-hoc tests. Figure 11: Created with BioRender.com Overall Effects Tests The first step for multi-group statistical testing is to run an overall effects test. The null hypothesis for the overall effects test is that there are no differences among group means. A significant p-value rejects the null hypothesis that the groups are drawn from populations with the same mean and indicates that at least one group differs significantly from the overall mean. Similar to two-group statistical testing, choice of the specific overall statistical test to run depends on whether the data are normally or non-normally distributed and whether the experimental design is paired: Figure 12: Created with BioRender.com Importantly, overall effects tests return one p-value regardless of the number of groups being compared. To determine which pairwise comparisons are significant, post-hoc testing is needed. Post-Hoc Testing If significance is obtained with an overall effects test, we can use post-hoc testing to determine which specific pairs of groups are significantly different from each other. Just as with two group statistical tests and overall effects multi-group statistical tests, choosing the appropriate post-hoc test depends on the data’s normality and whether the experimental design is paired: Figure 13: Created with BioRender.com Note that the above diagram represents commonly selected post-hoc tests; others may also be appropriate depending on your specific experimental design. As with other aspects of the analysis, be sure to report which post-hoc test(s) you performed! Correcting for Multiple Hypothesis Testing Correcting for multiple hypothesis testing is important for both the overall effects test (if you are running it over many endpoints) and post-hoc tests; however, it is particularly important for post-hoc tests. This is because even an analysis of a relatively small number of experimental groups results in quite a few pairwise comparisons. Comparing each of our five dose groups to each other in our example data, there are 10 separate statistical tests being performed! Therefore, it is generally advisable to adjust pairwise post-hoc testing p-values. The Tukey’s HSD function within rstatix does this automatically, while pairwise t-tests, pairwise Wilcoxon tests, and Dunn’s test do not. P-value adjustment can be added to their respective rstatix functions using the p.adjust.method = argument. When applying a post-hoc test, you may choose to compare every group to every other group, or you may only be interested in significant differences between specific groups (e.g., treatment groups vs. a control). This choice will be governed by your hypothesis. Statistical testing functions will typically default to comparing all groups to each other, but the comparisons can be defined using the comparisons = argument if you want to restrict the test to specific comparisons. It is important to decide at the beginning of your analysis which comparisons are relevant to your hypothesis because the number of pairwise tests performed in the post-hoc analysis will influence how much the resulting p-values will be adjusted for multiple hypothesis testing. Which test should I choose? Use the following flowchart to help guide your choice of statistical test to compare multiple groups: Figure 14: Created with BioRender.com Multi-Group Analysis Example To determine whether there are significant differences across all of our doses, the Friedman test is the most appropriate due to our matched experimental design and non-normally distributed data. The friedman_test() function is part of the rstatix package. This package also has many other helpful functions for statistical tests that are pipe/tidyverse friendly. To demonstrate how this test works, we will first perform the test on one variable: biomarker_data %&gt;% friedman_test(IL1B ~ Dose | Donor) ## # A tibble: 1 × 6 ## .y. n statistic df p method ## * &lt;chr&gt; &lt;int&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;chr&gt; ## 1 IL1B 16 12.5 4 0.0140 Friedman test A p-value of 0.01 indicates that we can reject the null hypothesis that all of our data are drawn from groups that have equivalent means. Now, we can run a for loop similar to our two-group comparisons in TAME 2.0 Module 4.4, Two Group Comparisons and Visualizations to determine the overall p-value for each endpoint: # Create a vector with the names of the variables you want to run the test on endpoints &lt;- colnames(biomarker_data %&gt;% select(IL1B:VEGF)) # Create data frame to store results dose_friedmanres &lt;- data.frame() # Run for loop for (i in 1:length(endpoints)) { # Assign a name to the endpoint variable. endpoint &lt;- endpoints[i] # Run wilcox test and store in results data frame. res &lt;- biomarker_data %&gt;% friedman_test(as.formula(paste0(endpoint, &quot;~ Dose | Donor&quot;, sep = &quot;&quot;))) %&gt;% select(c(.y., p)) dose_friedmanres &lt;- rbind(dose_friedmanres, res) } # View results datatable(dose_friedmanres) These results demonstrate that all of our endpoints have significant overall differences across doses (p &lt; 0.05). To determine which pairwise comparisons are significant, we next need to apply a post-hoc test. We will apply a pairwise, paired Wilcoxon test due to our experimental design and data distribution, with the Benjamini-Hochberg (BH) correction for multiple testing: dose_wilcox_posthoc_IL1B &lt;- biomarker_data %&gt;% pairwise_wilcox_test(IL1B ~ Dose, paired = TRUE, p.adjust.method = &quot;BH&quot;) dose_wilcox_posthoc_IL1B ## # A tibble: 10 × 9 ## .y. group1 group2 n1 n2 statistic p p.adj p.adj.signif ## * &lt;chr&gt; &lt;chr&gt; &lt;chr&gt; &lt;int&gt; &lt;int&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;dbl&gt; &lt;chr&gt; ## 1 IL1B 0 0.6 16 16 17 0.006 0.025 * ## 2 IL1B 0 1 16 16 11 0.002 0.017 * ## 3 IL1B 0 2 16 16 21 0.013 0.033 * ## 4 IL1B 0 4 16 16 18 0.008 0.025 * ## 5 IL1B 0.6 1 16 16 38 0.13 0.186 ns ## 6 IL1B 0.6 2 16 16 31 0.058 0.096 ns ## 7 IL1B 0.6 4 16 16 25 0.025 0.05 * ## 8 IL1B 1 2 16 16 43 0.211 0.234 ns ## 9 IL1B 1 4 16 16 40 0.159 0.199 ns ## 10 IL1B 2 4 16 16 62 0.782 0.782 ns Here, we can now see whether there are statistically significant differences in IL-1\\(\\beta\\) secretion between each of our doses. To generate pairwise comparison results for each of our inflammatory biomarkers, we can run a for loop similar to the one we ran for our overall test: # Create a vector with the names of the variables you want to run the test on endpoints &lt;- colnames(biomarker_data %&gt;% select(IL1B:VEGF)) # Create data frame to store results dose_wilcox_posthoc &lt;- data.frame() # Run for loop for (i in 1:length(endpoints)) { # Assign a name to the endpoint variable. endpoint &lt;- endpoints[i] # Run wilcox test and store in results data frame. res &lt;- biomarker_data %&gt;% pairwise_wilcox_test(as.formula(paste0(endpoint, &quot;~ Dose&quot;, sep = &quot;&quot;)), paired = TRUE, p.adjust.method = &quot;BH&quot;) dose_wilcox_posthoc &lt;- rbind(dose_wilcox_posthoc, res) } # View results datatable(dose_wilcox_posthoc) We now have a dataframe storing all of our pairwise comparison results. However, this is a lot to scroll through, making it hard to interpret. We can generate a publication-quality table by manipulating the table and joining it with the overall test data. dose_results_cleaned &lt;- dose_wilcox_posthoc %&gt;% unite(comparison, group1, group2, sep = &quot; vs. &quot;) %&gt;% select(c(.y., comparison, p.adj)) %&gt;% pivot_wider(id_cols = &quot;.y.&quot;, names_from = &quot;comparison&quot;, values_from = &quot;p.adj&quot;) %&gt;% left_join(dose_friedmanres, by = &quot;.y.&quot;) %&gt;% relocate(p, .after = &quot;.y.&quot;) %&gt;% rename(&quot;Variable&quot; = &quot;.y.&quot;, &quot;Overall&quot; = &quot;p&quot;) %&gt;% mutate(across(&#39;Overall&#39;:&#39;2 vs. 4&#39;, \\(x) format(x, scientific = TRUE, digits = 3))) datatable(dose_results_cleaned) To more easily see overall significance patterns, we could also make the same table but with significance stars instead of p-values by keeping the p.adjust.signif column instead of the p.adj column in our post-hoc test results dataframe: dose_results_cleaned_2 &lt;- dose_wilcox_posthoc %&gt;% unite(comparison, group1, group2, sep = &quot; vs. &quot;) %&gt;% select(c(.y., comparison, p.adj.signif)) %&gt;% pivot_wider(id_cols = &quot;.y.&quot;, names_from = &quot;comparison&quot;, values_from = &quot;p.adj.signif&quot;) %&gt;% left_join(dose_friedmanres, by = &quot;.y.&quot;) %&gt;% relocate(p, .after = &quot;.y.&quot;) %&gt;% rename(&quot;Variable&quot; = &quot;.y.&quot;, &quot;Overall&quot; = &quot;p&quot;) %&gt;% mutate(across(&#39;Overall&#39;:&#39;2 vs. 4&#39;, \\(x) format(x, scientific = TRUE, digits = 3))) datatable(dose_results_cleaned_2) Answer to Environmental Health Question 1 With this, we can answer Environmental Health Question #1 : Are there significant differences in inflammatory biomarker concentrations between different doses of acrolein? Answer: Yes, there are significant differences in inflammatory biomarker concentrations between different doses of acrolein. The overall p-values for all biomarkers are significant. Within each biomarker, at least one pairwise comparison was significant between doses, with a majority of these significant comparisons being with the highest dose (4 ppm). Visualization of Multi-Group Statistical Results The statistical results we generated are a lot to digest in table format, so it can be helpful to graph the results. As our statistical testing becomes more complicated, so does the code used to generate results. The ggpubr package can perform statistical testing and overlay the results onto graphs for a specific set of tests, such as overall effects tests and unpaired t-tests or Wilcoxon tests. However, for tests that aren’t available by default, the package also contains the helpful stat_pvalue_manual() function that can be added to plots. This is what we will need to use to add the results of our pairwise, paired Wilcoxon test with BH correction, as there is no option for BH correction within the default function we might otherwise use (stat_compare_means()). We will first work through an example of this using one of our endpoints, and then we will demonstrate how to apply it to facet plotting. Single Plot We first need to format our existing statistical results so that they match the format that the function needs as input. Specifically, the dataframe needs to contain the following columns: group1 and group2: the groups being compared A column containing the results you want displayed (p, p.adj, or p.adj.signif typically) y.position, which tells the function where to plot the significance markers Our results dataframe for IL-1\\(\\beta\\) already contains our groups and p-values: datatable(dose_wilcox_posthoc_IL1B) We can add the position columns using the function add_xy_position(): dose_wilcox_posthoc_IL1B &lt;- dose_wilcox_posthoc_IL1B %&gt;% add_xy_position(x = &quot;Dose&quot;, step.increase = 2) datatable(dose_wilcox_posthoc_IL1B) Now, we are ready to make a graph of our results. We will use stat_friedman_test() to add our overall p-value and stat_pvalue_manual() to add our pairwise values. # Set graphing theme theme_set(theme_bw()) # Make plot ggplot(biomarker_data, aes(x = Dose, y = IL1B)) + geom_boxplot(aes(fill = Dose), outlier.shape = NA) + scale_fill_manual(values = c(&quot;#BFBFBF&quot;, &quot;#D5A298&quot;, &quot;#E38273&quot;, &quot;#EB5F4E&quot;, &quot;#EE2B2B&quot;)) + geom_jitter(size = 3, position = position_jitter(0.15)) + stat_friedman_test(wid = &quot;Donor&quot;, p.adjust.method = &quot;none&quot;, label = &quot;p = {p.format}&quot;, label.x.npc = &quot;left&quot;, label.y = 9.5, hjust = 0.5, size = 6) + stat_pvalue_manual(dose_wilcox_posthoc_IL1B, label = &quot;p.adj.signif&quot;, size = 12, hide.ns = TRUE) + ylim(2.5, 10) + labs(y = &quot;Log2(IL-1\\u03B2 (pg/mL))&quot;, x = &quot;Acrolein (ppm)&quot;) + theme(legend.position = &quot;none&quot;, axis.title = element_text(color = &quot;black&quot;, size = 15), axis.title.x = element_text(vjust = -0.75), axis.title.y = element_text(vjust = 2), axis.text = element_text(color = &quot;black&quot;, size = 12)) However, to make room for all of our annotations, our data become compressed, and it makes it difficult to see our data. Although presentation of statistical results is largely a matter of personal preference, we could clean up this plot by making our annotations appear on top of the bars, with indication in the figure legend that the comparison is with a specific dose. We will do this by: Filtering our results to those that are significant. Changing the symbol for comparisons that are not to the 0 dose. Layering this text onto the plot with geom_text() rather than stat_pvalue_manual(). First, let’s filter our results to significant results and change the symbol for comparisons that are not to the 0 dose to a caret (^) instead of stars. We can do this by creating a new column called label that keeps the existing label if group1 is 0, and if not, changes the label to a caret of the same length. We then use the summarize function to paste the labels for each of the groups together, resulting in a final dataframe containing our annotations for our plot. dose_wilcox_posthoc_IL1B_2 &lt;- dose_wilcox_posthoc_IL1B %&gt;% # Filter results to those that are significant filter(p.adj &lt;= 0.05) %&gt;% # Make new symbol mutate(label = ifelse(group1 == &quot;0&quot;, p.adj.signif, strrep(&quot;^&quot;, nchar(p.adj.signif)))) %&gt;% # Select only the columns we need select(c(group1, group2, label)) %&gt;% # Combine symbols for the same group group_by(group2) %&gt;% summarise(label = paste(label, collapse=&quot; &quot;)) %&gt;% # Remove duplicate row distinct(group2, .keep_all = TRUE) %&gt;% # Rename group2 to dose rename(&quot;Dose&quot; = &quot;group2&quot;) dose_wilcox_posthoc_IL1B_2 ## # A tibble: 4 × 2 ## Dose label ## &lt;chr&gt; &lt;chr&gt; ## 1 0.6 * ## 2 1 * ## 3 2 * ## 4 4 * ^ Then, we can use the same code as for our previous plot, but instead of using stat_pvalue_manual(), we will use geom_text() in combination with the dataframe we just created. ggplot(biomarker_data, aes(x = Dose, y = IL1B)) + geom_boxplot(aes(fill = Dose), outlier.shape = NA) + scale_fill_manual(values = c(&quot;#BFBFBF&quot;, &quot;#D5A298&quot;, &quot;#E38273&quot;, &quot;#EB5F4E&quot;, &quot;#EE2B2B&quot;)) + geom_jitter(size = 3, position = position_jitter(0.15)) + stat_friedman_test(wid = &quot;Donor&quot;, p.adjust.method = &quot;none&quot;, label = &quot;p = {p.format}&quot;, label.x.npc = &quot;left&quot;, label.y = 4.85, hjust = 0.5, size = 6) + geom_text(data = dose_wilcox_posthoc_IL1B_2, aes(x = Dose, y = 4.5, label = paste0(label)), size = 10, hjust = 0.5) + ylim(2.5, 5) + labs(y = &quot;Log2(IL-1\\u03B2 (pg/mL))&quot;, x = &quot;Acrolein (ppm)&quot;) + theme(legend.position = &quot;none&quot;, axis.title = element_text(color = &quot;black&quot;, size = 15), axis.title.x = element_text(vjust = -0.75), axis.title.y = element_text(vjust = 2), axis.text = element_text(color = &quot;black&quot;, size = 12)) An appropriate title for this figure could be: “Figure X. Exposure to 0.6-4 ppm acrolein increases IL-1\\(\\beta\\) secretion in primary human bronchial epithelial cells. Groups were compared using the Friedman test to obtain overall p-value and Wilcoxon signed rank test for post-hoc testing. * p &lt; 0.05 in comparison with 0 ppm, ^ p &lt; 0.05 in comparison with 0.6 ppm, n = 16 per group (paired).” Faceted Plot Ideally, we would extend this sort of graphical approach to our faceted plot showing all of our endpoints. However, there are quite a few statistically significant comparisons to graph, including comparisons that are significant between different pairs of doses (not just back to the control). While we could attempt to graph all of them, ultimately, this will lead to a cluttered figure panel. When thinking about how to simplify our plots, some options are: Instead of using the number of symbols to represent p-values, we could use a single symbol to represent any comparison with a p-value with at least p &lt; 0.05, and that symbol could be different depending on which group the significance is in comparison to. Symbols can be difficult to parse in R, so we could use letters or even the group names above the column of interest. For example, if the concentration of an endpoint at 2 ppm was significant in comparison with both 0 and 0.6 ppm, we could annotate “0, 0.6” above the 2 ppm column, or we could choose a letter (“a, b”) or symbol (“*, ^“) to convey these results. If the pattern is the same across many of the endpoints measured, we could graph a subset of the endpoints with the most notable data trends or the most biological meaning for the main body of the manuscript, with data for additional endpoints referred to in the text and shown in the supplemental figures or tables. If most of the significant comparisons are back to the control group, we could choose to only show comparisons with the control group, with textual description of the other significant comparisons and indication that those specific p-values can be viewed in the supplemental table of results. Which approach you decide to take (or maybe another approach altogether) is a matter of both personal preference and your specific study goals. You may also decide that it is important to you to show all significant comparisons, which will require more careful formatting of the plots to ensure that all text and annotations are legible. For this module, we will proceed with option #3 because many of our comparisons to the control dose (0) are significant, and we have enough groups that there likely will not be space to annotate all of them above our data. We will take similar steps here that we did when constructing our single endpoint graph, with a couple of small differences. Specifically, we need to: Create a dataframe of labels/annotations as we did above, but now filtered to only significant comparisons with the 0 group. Add to the label/annotation dataframe what we want the y position for each of the labels to be, which will be different for each endpoint. First, let’s create our annotations dataframe. We will start with the results dataframe from our posthoc testing: datatable(dose_wilcox_posthoc) dose_wilcox_posthoc_forgraph &lt;- dose_wilcox_posthoc %&gt;% filter(p.adj &lt;= 0.05) %&gt;% # Filter for only comparisons to 0 filter(group1 == &quot;0&quot;) %&gt;% # Rename columns rename(&quot;variable&quot; = &quot;.y.&quot;, &quot;Dose&quot; = &quot;group2&quot;) datatable(dose_wilcox_posthoc_forgraph) The Dose column will be used to tell ggplot2 where to place the annotations on the x axis, but we need to also specify where to add the annotations on the y axis. This will be different for each variable because each variable is on a different scale. We can approach this by computing the maximum value of each variable, then increasing that by 20% to add some space on top of the points. sig_labs_y &lt;- biomarker_data %&gt;% summarise(across(IL1B:VEGF, \\(x) max(x))) %&gt;% t() %&gt;% as.data.frame() %&gt;% rownames_to_column(&quot;variable&quot;) %&gt;% rename(&quot;y_pos&quot; = &quot;V1&quot;) %&gt;% mutate(y_pos = y_pos*1.2) sig_labs_y ## variable y_pos ## 1 IL1B 4.967953 ## 2 IL6 14.074658 ## 3 IL8 20.752105 ## 4 IL10 3.224473 ## 5 TNFa 4.478723 ## 6 VEGF 14.148027 Then, we can join these data to our labeling dataframe to complete what we need to make the annotations. dose_wilcox_posthoc_forgraph &lt;- dose_wilcox_posthoc_forgraph %&gt;% left_join(sig_labs_y, by = &quot;variable&quot;) Now, it’s time to graph! Keep in mind that although the plotting script can get long and unweildy, each line is just a new instruction to ggplot about a formatting element or an additional layer to add to the graph. # Pivot data longer biomarker_data_long &lt;- biomarker_data %&gt;% pivot_longer(-c(Donor, Dose), names_to = &quot;variable&quot;, values_to = &quot;value&quot;) # Create clean labels for the graph titles new_labels &lt;- c(&quot;IL10&quot; = &quot;IL-10&quot;, &quot;IL1B&quot; = &quot;IL-1\\u03B2 &quot;, &quot;IL6&quot; = &quot;IL-6&quot;, &quot;IL8&quot; = &quot;IL-8&quot;, &quot;TNFa&quot; = &quot;TNF-\\u03b1&quot;, &quot;VEGF&quot; = &quot;VEGF&quot;) # Make graph ggplot(biomarker_data_long, aes(x = Dose, y = value)) + # outlier.shape = NA removes outliers geom_boxplot(aes(fill = Dose), outlier.shape = NA) + # Changing box plot colors scale_fill_manual(values = c(&quot;#BFBFBF&quot;, &quot;#D5A298&quot;, &quot;#E38273&quot;, &quot;#EB5F4E&quot;, &quot;#EE2B2B&quot;)) + geom_jitter(size = 1.5, position = position_jitter(0.15)) + # Adding a p value from Friedman test stat_friedman_test(wid = &quot;Donor&quot;, p.adjust.method = &quot;none&quot;, label = &quot;p = {p.format}&quot;, label.x.npc = &quot;left&quot;, vjust = -3.5, hjust = 0.1, size = 3.5) + # Add label geom_text(data = dose_wilcox_posthoc_forgraph, aes(x = Dose, y = y_pos, label = p.adj.signif, size = 5, hjust = 0.5)) + # Adding padding y axis scale_y_continuous(expand = expansion(mult = c(0.1, 0.6))) + # Changing y axis label ylab(expression(Log[2]*&quot;(Concentration (pg/ml))&quot;)) + # Changing x axis label xlab(&quot;Acrolein (ppm)&quot;) + # Faceting by each biomarker facet_wrap(~ variable, nrow = 2, scales = &quot;free_y&quot;, labeller = labeller(variable = new_labels)) + # Removing legend theme(legend.position = &quot;none&quot;, axis.title = element_text(color = &quot;black&quot;, size = 12), axis.title.x = element_text(vjust = -0.75), axis.title.y = element_text(vjust = 2), axis.text = element_text(color = &quot;black&quot;, size = 10), strip.text = element_text(size = 12, face = &quot;bold&quot;)) An appropriate title for this figure could be: “Figure X. Exposure to acrolein increases secretion of proinflammatory biomarkers in primary human bronchial epithelial cells. Groups were compared using the Friedman test to obtain overall p-value and Wilcoxon signed rank test for post-hoc testing. * p &lt; 0.05, ** p &lt; 0.01, *** p &lt; 0.001, **** p &lt; 0.0001 for comparison with control. For additional significant comparisons, see Supplemental Table X. n = 16 per group (paired).” Answer to Environmental Health Question 2 With this, we can answer Environmental Health Question #2 : Do TNF-\\(\\alpha\\) concentrations significantly increase with increasing dose of acrolein? Answer: Yes, TNF-\\(\\alpha\\) concentrations significantly increase with increasing dose of acrolein, which we were able to visualize, along with other mediators, in our facet plot. Concluding Remarks In this module, we introduced common multi-group statistical tests, including both overall effects tests and post-hoc testing. We applied these tests to our example dataset and demonstrated how to produce publication-quality tables and figures of our results. Implementing a workflow such as this enables efficient analysis of wet-bench generated data and customization of output figures and tables suited to your personal preferences. Additional Resources STHDA: How to Add P-Values and Significance Levels to ggplots using ggpubr Adding p-values with ggprism Overview of ggsignif Test Your Knowledge Functional endpoints from these cultures were also measured. These endpoints were: 1) Membrane Permeability (MemPerm), 2) Trans-Epithelial Electrical Resistance (TEER), 3) Ciliary Beat Frequency (CBF), and 4) Expression of Mucin (MUC5AC). These data were already processed and tested for normality (see Test Your Knowledge for TAME 2.0 Module 4.2 Data Import, Processing, and Summary Statistics), with results indicating that two of the endpoints are normally distributed and two non-normally distributed. Use the same processes demonstrated in this module and the provided data (“Module4_5_TYKInput.xlsx” (functional data)) to run analyses and make a publication-quality figure panel and table to answer the following question: Are there significant differences in functional endpoints between cells treated with different concentrations of acrolein? For an extra challenge, try also making your faceted plot in the style of option #1 above, with different symbols, letters, or group names above columns to indicate which group that column in significant in comparison with. "],["introduction-to-artificial-intelligence-machine-learning-and-predictive-modeling-for-environmental-health.html", "Introduction to Artificial Intelligence, Machine Learning, and Predictive Modeling for Environmental Health Introduction to Training Module General Historical Context and Taxonomy of Modern AI/ML Application of Machine Learning in Environmental Health Science Concluding Remarks", " Introduction to Artificial Intelligence, Machine Learning, and Predictive Modeling for Environmental Health This training module was developed by David M. Reif, with contributions from Elise Hickman, Alexis Payton, and Julia E. Rager All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module Artificial intelligence (AI), machine learning (ML), and predictive modeling are becoming increasingly popular buzzwords both in the public domain and within research fields, including environmental health. Within environmental health, these computational techniques are implemented to integrate large, high dimensional datasets (e.g., chemical, biological, clinical/medical, model estimates, etc) to better understand links between environmental exposures and biological responses. In this training module, we will: Provide general historical context and taxonomy of modern AI/ML Provide an overview of the intersection between environmental health science ML through discussing… Why there is a need for ML in environmental health science The differences between ML and traditional statistical methods Predictive modeling in the context of environmental health science Additional applications of ML in environmental health science Training Module’s Environmental Health Question This training module was specifically developed to answer the following environmental health question: How and why are artificial intelligence, machine learning, and predictive modeling used in environmental health research? General Historical Context and Taxonomy of Modern AI/ML Before diving in to the applications of AI and ML in environmental health, let’s first establish what these term mean and how they are related. Note that the definitions surrounding AI and ML can be subjective, however the purpose of this module is not to get caught up in semantics, but to broadly understand how AI and ML can be applied to environmental health research. Artificial Intelligence (AI) encompasses computer systems that perform tasks typically associated with human cognition and intelligence. AI is found in our everyday lives, for instance, within face recognition, internet search queries, email spam detection, smart home devices, auto-navigation, and digital assistants. Machine Learning (ML) can be thought of as a subset of AI and describes a computer system that iteratively learns and improves from that experience autonomously. Below is a high level taxonomy of AI. It’s not meant to be an exhaustive depiction of all AI techniques but a simple visualization of how some of these methodologies are nested within each other. Note: AI can be categorized in different ways and may deviate from what is illustrated below. Advantages of AI and ML include the automation of repetitive tasks, complex problem solving, and reducing human error. However, disadvantages include learning from biased datasets or patterns that are reflected in the decisions of AI/ML and the potential limited interpretability of algorithms created by AI/ML. Check out the following resources for… Further explanation on differences in Artificial Intelligence vs. Machine Learning Other subsets of AI that fall outside of the scope of these modules in Types of Artificial Intelligence Additional discussion on the utility of ML approaches for high-dimensional data common in environmental health research in Payton et. al It is important to understand the methodological “roots” of current methods. Otherwise, it seems like every approach is novel! AI and ML methods have been around since the mid- to late- 1900s and continue to evolve in the present day. The earliest conceptual roots for these approaches can be traced from antiquity; however, it is generally thought that the field was named “artificial intelligence” at the “Dartmouth Workshop” in 1956, led by John McCarthy and others. The following schematic demonstrates the general taxonomy (categories, sub-fields, and specific methods) of modern AI and ML: A Brief Detour to Discuss ChatGPT ChatGPT (Chat Generative Pre-trained Transformer) is a publicly available chatbot developed by OpenAI. It was released in November of 2022 and quickly gained popularity due to its accessibility and ability to have human-like conversations with the user across almost any imaginable topic. Language Models (LLMs), including large language models like GPT-3 (a predecessor to ChatGPT), generally fall under the “Connectionist AI” category, which use deep learning techniques and are considered a subset of artificial neural networks. They fall under the deep learning subset due to their use of deep neural networks with many layers, allowing them to learn from large amounts of data and find intricate patterns. LLMs are trained to predict the probability of a word given its context in a dataset (a form of next-word prediction), which is a machine learning methodology. It’s notable that they use architectures like Transformer Networks, which are known for their efficiency in handling sequential data, making them a go-to choice for natural language processing (NLP) tasks. The use of attention mechanisms in these architectures allows the model to focus on different parts of the input sequence when producing an output sequence, offering a substantial improvement in performance for many natural language processing tasks. The role of ChatGPT and similar tools in the environmental health research space is still being explored. Although ChatGPT has the potential to streamline certain parts of the research process, such as text and language polishing, synthesizing existing information, and suggesting custom coding solutions, it is not an intellectual replacement for the expertise and diverse viewpoints of scientists and must be used transparently and with caution. Application of Machine Learning in Environmental Health Science For the rest of this module and chapter, we will focus on machine learning (ML). Generally speaking, ML is considered to encompass the study of computer algorithms that can improve automatically through experience and by the use of data. It is seen as a part of artificial intelligence (AI), discussed broadly above. Why do we need machine learning in environmental health science? There are many avenues to incorporate ML into environmental health research, all aimed at better identifying patterns amongst large datasets spanning medical health records, clinical data, exposure monitoring data, chemistry profiles, and the rapidly expanding realm of biological response data including multiple -omics endpoints. One well-known problem that can be better addressed by incorporating ML is the ‘too many chemicals, too little data’ problem. To detail, there are thousands of chemicals in commerce today. Testing these chemicals one by one for toxicity using comprehensive animal screening experiments would take decades and is not feasible financially. Current efforts to address this problem include using cell-based high throughput screening to efficiently determine biological responses to a variety of chemical exposures and treatment conditions. These screening efforts result in increasing amounts of data, which can be gathered to start building big databases. When many of these datasets and databases are combined, including diversity across different types of screening platforms, technologies, cell types, species, and other experimental variables, the associated dimensionality of the data gets “big.” This presents a problem because these data are diverse and high dimensional (the number of features or endpoints exceeds the number of observations/chemicals). To appropriately analyze and model these data, new approaches beyond traditional statistical methods are needed. Machine Learning vs. Traditional Statistical Methods There is plenty of debate as to where the line(s) between ML and traditional statistics should be drawn. In our opinion, a perfect delineation is not necessary for our purposes. Rather, we will focus on the usual goals/intent of each to help us understand the distinction for environmental health research. Traditional statistics may be able to handle 1:1 or 1:many comparisons of singular quantities (e.g., activity concentrations for two chemicals). However, once the modeling becomes more complex or exploratory, assumptions of most traditional methods will be violated. Furthermore, statistics draws population inferences from a sample, while AI/ML finds generalizable predictive patterns (Bzdok et al 2018). This is particularly helpful in predictive toxicology, in which we leverage high dimensional data to obtain generalizable forecasts for the effects of chemicals on biological systems. This image shows graphical abstractions of how a “problem” is solved using: Traditional statistics ((A) logistic regression and (B) linear regression), OR Machine learning ((C) support vector machines, (D) artificial neural networks, and (E) decision trees) Predictive Modeling in the Context of Environmental Health Science In the previous section, we briefly mentioned predictive toxicology. We often think of predictions as having a forward-time component (i.e. What will happen next?) … what about “prediction” in a different sense as applied to toxicology? Our working definition is that predictive toxicology describes a multidisciplinary approach to chemical toxicity evaluation that more efficiently uses animal test results, when needed, and leverages expanding non-animal test methods to forecast the effects of a chemical on biological systems. Examples of the questions we can answer using predictive toxicology include: Can we more efficiently design animal studies and analyze data from shorter assays using fewer animals to predict long-term health outcomes? Can this suite of in vitro assays predict what would happen in an organism? Can we use diverse, high dimensional data to cluster chemicals into predicted activity classes? Similar logic applies to the field of exposure science. What about “prediction” applied to exposure science? Our working definition is that predictive exposure science describes a multidisciplinary approach to chemical exposure evaluations that more efficiently uses biomonitoring, chemical inventory, and other exposure science-relevant databases to forecast exposure rates in target populations. For example: Can we use existing biomonitoring data from NHANES to predict exposure rates for chemicals that have yet to be measured in target populations? (see ExpoCast program, e.g., Wambagh et al 2014) Can I use chemical product use inventory data to predict the likelihood of a chemical being present in a certain consumer product? (e.g., Phillips et al 2018) There are many different types of ML methods that we can employ in predictive toxicology and exposure science, depending on the data type / purpose of data analysis. A recent review written together with Erin Baker’s lab provides a high-level overview on some of the types of ML methods and challenges to address when analyzing multi-omic data (including chemical signature data). Answer to Environmental Health Question With this, we can now answer our Environmental Health Question: How and why are machine learning, predictive modeling, and artificial intelligence used in environmental health research? Answer: Machine learning, a subcategory of artificial intelligence, can be used in environmental health science to better understand patterns between chemical exposure and biological response in complex, high dimensional datasets. These datasets are often generated as part of efforts to screen many chemicals efficiently. Predictive modeling, which can include machine learning approaches, leverages these data to forecast the effects of a chemical on biological systems. Additional Applications of Machine Learning in Environmental Health Science In addition to the predictive toxicology questions above, ML can also be applied in the analysis of complex, high dimensional data in observational clinical (human subjects) studies in environmental health, such as: Do subjects cluster by chemical exposure? Are there similarities between subjects that cluster together for chemical exposure, suggesting underlying factors relevant to chemical exposure? Are biological signatures in different exposure groups different enough overall that ML can predict which group a subject belongs to based on their signature? Concluding Remarks In conclusion, this training module provides an overview of the field of AI and ML and discusses applications of these tools in environmental health science through predictive modeling. These methods represent common tools that are used in high dimensional data analyses within the field of environmental health sciences. In the following modules, we will provide specific examples detailing how to apply both supervised and unsupervised machine learning methods to environmental health questions and how to interpret the results of these analyses. For a review article on ML, see: Odenkirk MT, Reif DM, Baker ES. Multiomic Big Data Analysis Challenges: Increasing Confidence in the Interpretation of Artificial Intelligence Assessments. Anal Chem. 2021 Jun 8;93(22):7763-7773. PMID: 34029068 For additional case studies that leverage more advanced ML techniques, see the following recent publications that also address environmental health questions from our research groups, with bracketed tags at the end of each citation denoting ML methods used in that study: Clark J, Avula V, Ring C, Eaves LA, Howard T, Santos HP, Smeester L, Bangma JT, O’Shea TM, Fry RC, Rager JE. Comparing the Predictivity of Human Placental Gene, microRNA, and CpG Methylation Signatures in Relation to Perinatal Outcomes. Toxicol Sci. 2021 Sep 28;183(2):269-284. PMID: 34255065 [hierarchical clustering, principal component analysis, random forest] Green AJ, Mohlenkamp MJ, Das J, Chaudhari M, Truong L, Tanguay RL, Reif DM. Leveraging high-throughput screening data, deep neural networks, and conditional generative adversarial networks to advance predictive toxicology. PLoS Comput Biol. 2021 Jul 2;17(7):e1009135. PMID: 3421407 [conditional generative adversarial network, deep neural network, support vector machine, random forest, multilayer perceptron] To KT, Truong L, Edwards S, Tanguay RL, Reif DM. Multivariate modeling of engineered nanomaterial features associated with developmental toxicity. NanoImpact. 2019 Apr;16:10.1016. PMID: 32133425 [random forest] Ring C, Sipes NS, Hsieh JH, Carberry C, Koval LE, Klaren WD, Harris MA, Auerbach SS, Rager JE. Predictive modeling of biological responses in the rat liver using in vitro Tox21 bioactivity: Benefits from high-throughput toxicokinetics. Comput Toxicol. 2021 May;18:100166. PMID: 34013136 [random forest] Hickman E, Payton A, Duffney P, Wells H, Ceppe AS, Brocke S, Bailey A, Rebuli ME, Robinette C, Ring B, Rager JE, Alexis NE, Jaspers I. Biomarkers of Airway Immune Homeostasis Differ Significantly with Generation of E-Cigarettes. Am J Respir Crit Care Med. 2022 Nov 15; 206(10):1248-1258. PMID: 35731626 [hierarchical clustering, quadratic discriminant analysis, multinomial logistic regression] Perryman AN, Kim H-YH, Payton A, Rager JE, McNell EE, Rebuli ME, et al. (2023) Plasma sterols and vitamin D are correlates and predictors of ozone-induced inflammation in the lung: A pilot study. PLoS ONE 18(5): e0285721. PMID: 37186612 [random forest, support vector machine, k nearest neighbor] Payton AD, Perryman AN, Hoffman JR, Avula V, Wells H, Robinette C, Alexis NE, Jaspers I, Rager JE, Rebuli ME. Cytokine signature clusters as a tool to compare changes associated with tobacco product use in upper and lower airway samples. American Journal of Physiology-Lung Cellular and Molecular Physiology 2022 322:5, L722-L736. PMID: 35318855 [k-means clustering, principal component analysis] "],["supervised-machine-learning.html", "Supervised Machine Learning Introduction to Training Module Types of Machine Learning Types of Supervised Machine Learning Algorithms Training Supervised Machine Learning Models Assessing Classification-Based Model Performance Introduction to Activity and Example Dataset Testing for Differences in Predictor Variables across the Outcome Classes Predicting iAs Detection with a Random Forest (RF) Model Class Imbalance Concluding Remarks", " Supervised Machine Learning This training module was developed by Alexis Payton, Oyemwenosa N. Avenbuan, Lauren E. Koval, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module Machine learning is a field that has been around for decades but has exploded in popularity and utility in recent years due to the proliferation of big and/or high dimensional data. Machine learning has the ability to sift through and learn from large volumes of data and use that knowledge to solve problems. The challenges of high dimensional data as they pertain to environmental health and the applications of machine learning to mitigate some of those challenges are discussed further in Payton et. al. In this module, we will introduce different types of machine learning and then focus in on supervised machine learning, including how to train and assess supervised machine learning models. We will then analyze an example dataset with supervised machine learning highlighting an example with random forest modeling. Types of Machine Learning Within the field of machine learning, there are many different types of algorithms that can be leveraged to address environmental health research questions. The two broad categories of machine learning frequently applied to environmental health research are: (1) supervised machine learning and (2) unsupervised machine learning. Supervised machine learning involves training a model using a labeled dataset, where each independent or predictor variable is associated with a dependent variable with a known outcome. This allows the model to learn how to predict the labeled outcome on data it hasn’t “seen” before based on the patterns and relationships it previously identified in the data. For example, supervised machine learning has been used for cancer prediction and prognosis based on variables like tumor size, stage, and age (Lynch et. al, Asadi et. al). Supervised machine learning includes: Classification: Using algorithms to classify a categorical outcome (ie. plant species, disease status, etc.) Regression: Using algorithms to predict a continuous outcome (ie. gene expression, chemical concentration, etc.) Soni, D. (2018, March 22). Supervised vs. Unsupervised Learning. Towards Data Science; Towards Data Science. https://towardsdatascience.com/supervised-vs-unsupervised-learning-14f68e32ea8d Unsupervised machine learning, on the other hand, involves using models to find patterns or associations between variables in a dataset that lacks a known or labeled outcome. For example, unsupervised machine learning has been used to identify new patterns across genes that are co-expressed, informing potential biological pathways mediating human disease (Botía et. al, Pagnuco et. al). Langs, G., Röhrich, S., Hofmanninger, J., Prayer, F., Pan, J., Herold, C., &amp; Prosch, H. (2018). Machine learning: from radiomics to discovery and routine. Der Radiologe, 58(S1), 1–6. PMID: 34013136. Figure regenerated here in alignment with its published Creative Commons Attribution 4.0 International License. Overall, the distinction between supervised and unsupervised learning is an important concept in machine learning, as it can inform the choice of algorithms and techniques used to analyze and make predictions from data. It is worth noting that there are also other types of machine learning, such as semi-supervised learning, reinforcement learning, and deep learning, though we will not further discuss these topics in this module. Types of Supervised Machine Learning Algorithms Although this module’s example will focus on a random forest model in the coding example below, other commonly used algorithms for supervised machine learning include: K-Nearest Neighbors (KNN): Uses distance to classify a data point in the test set based upon the most common class of neighboring data points from the training set. For more information on KNN, see K-Nearest Neighbor. Support Vector Machine (SVM): Creates a decision boundary line (hyperplane) in n-dimensional space to separate the data into each class so that when new data is presented, they can be easily categorized. For more information on SVM, see Support Vector Machine. Random Forest (RF): Uses a multitude of decision trees trained on a subset of different samples from the training set and the resulting classification of a data point in the test set is aggregated from all the decision trees. A decision tree is a hierarchical model that depicts decisions from predictors and their resulting outcomes. It starts with a root node, which represents an initial test from a single predictor. The root node splits into subsequent decision nodes that test another feature. These decision nodes can either feed into more decision nodes or leaf nodes that represent the predicted class label. A branch or a sub-tree refers to a subsection of an entire decision tree. Here is an example decision tree with potential variables and decisions informing a college basketball player’s likelihood of being drafted to the NBA: While decision trees are highly interpretable, they are prone to overfitting, thus they may not always generalize well to data outside of the training set. To address this, random forests are comprised of many different decision trees. Each tree is trained on a subset of the samples in the training data, selected with replacement, and a randomly selected set of predictor variables. For a dataset with p predictors, it is common to test \\(\\sqrt{p}\\), \\(\\frac{p}{2}\\), and p predictors to see which gives the best results. This process decorrelates the trees. For a classification problem, majority vote of the decision trees determines the final class for a prediction. This process loses interpretability inherent to individual trees, but reduces the risk of overfitting. For more information on RF and decision trees, check out Random Forest and Decision Trees. Note: One algorithm is not inherently better than the others with each having their respective advantages and disadvantages. Each algorithm’s predictive ability will be largely dependent on the size of the dataset, the distribution of the data points, and the scenario. Training Supervised Machine Learning Models In supervised machine learning, algorithms need to be trained before they can be used to predict on new data. This involves selecting a smaller portion of the dataset to train the model so it will learn how to predict the outcome as accurately as possible. The process of training an algorithm is essential for enabling the model to learn and improve over time, allowing it to make more accurate predictions and better adapt to new and changing circumstances. Ultimately, the quality and relevance of the training data will have a significant impact on the effectiveness of a machine learning model. Common partitions of the full dataset used to train and test a supervised machine learning model are the following: Training Set: a subset of the data that the algorithm “sees” and uses to identify patterns. Validation Set: a subset of the training set that is used to evaluate the model’s fit in an unbiased way allowing us to fine-tune its parameters and optimize performance. Test Set: a subset of data that is used to evaluate the final model’s fit based on the training and validation sets. This provides an objective assessment of the model’s ability to generalize new data. It is common to split the dataset into a training set that contains 60% of the data and the test set that contains 40% of the data, though other common splits include 70% training / 30% test and 80% training / 20% test. It is important to note that the test set should only be examined after the algorithm has been trained using the training/validation sets. Using the test set during the development process can lead to overfitting, where the model performs well on the test data but poorly on new data. The ideal algorithm is generalizable or flexible enough to accurately predict unseen data. This is known as the bias-variance tradeoff. For further information on the bias-variance tradeoff, see Understanding the Bias-Variance Tradeoff. Cross Validation Finally, we will discuss cross validation, which is an approach used during training to expose the model to more patterns in the data and aid in model evaluation. For example, if a model is trained and tested on a 60:40 split, our model’s accuracy will likely be influenced by where this 60:40 split occurs in the dataset. This will likely bias the data and reduce the algorithm’s ability to predict accurately for data not in the training set. Overall, cross validation (CV) is implemented to fine tune a model’s parameters and improve prediction accuracy and ability to generalize. Although there are a number of cross validation approaches, we will specifically highlight k-fold cross validation. k-fold cross validation works by splitting the samples in the training dataset into k equally sized folds or groups. For example, if we implement 5-fold CV, we start by… Splitting the training data into 5 groups, or “folds”. Five iterations of training/testing are then run where each of the 5 folds serves as the test data once and as part of the training set four times, as seen in the figure below. To measure predictive ability of each of the parameters tested, like the number of features to include, values like accuracy and specificity are calculated for each iteration. The parameters that optimize performance are selected for the final model which will be evaluated against the test set not used in training. Check out these resources for additional information on Cross Validation in Machine Learning and Cross Validation Pros &amp; Cons. Assessing Classification-Based Model Performance Evaluation metrics from a confusion matrix are often used to determine the best model during training and measure model performance during testing for classification-based supervised machine learning models. A confusion matrix consists of a table that displays the numbers of how often the algorithm correctly and incorrectly predicted the outcome. Let’s imagine you’re interested in predicting whether or not a player will be drafted to the National Basketball Association (NBA) based on a dataset that contains variables regarding a player’s assists, points, height etc. Let’s say that this dataset contains information on 253 players with 114 that were actually drafted and 139 that weren’t drafted. The confusion matrix below shows a model’s results where a player that is drafted is the “positive” class and a player that is not drafted is the “negative” class. Helpful confusion matrix terminology: True positive (TP): the number of correctly classified “positive” data points (i.e., the number of correctly classified players to be drafted) True negative (TN): the number of correctly classified “negative” data points (i.e., the number of correctly classified players to be not drafted) False positive (FP): the number of incorrectly classified “positive” data points (i.e., the number of players not drafted incorrectly classified as draft picks) False negative (FN): the number of incorrectly classified “negative” data points (i.e., the number of draft picks incorrectly classified as players not drafted) Some of the metrics that can be obtained from a confusion matrix are listed below: Overall Accuracy: indicates how often the model makes a correct prediction relative to the total number of predictions made and is typically used to assess overall model performance (\\(\\frac{TP+TN}{TP+TN+FP+FN}\\)). Sensitivity or Recall: evaluates how well the model was able to predict the “positive” class. It is calculated as the ratio of correctly classified true positives to the total number of positive cases (\\(\\frac{TP}{TP+FN}\\)). Specificity: evaluates how well the model was able to predict the “negative” class. It is calculated as the ratio of correctly classified true negatives to total number of negatives cases (\\(\\frac{TN}{TN+FP}\\)). Balanced Accuracy: is the mean of sensitivity and specificity and is often used in the case of a class imbalance to gauge how well the model can correctly predict values for both classes (\\(\\frac{sensitivity+specificity}{2}\\)). Positive Predictive Value (PPV) or Precision: evaluates how accurate predictions of the “positive” class are. It is calculated as the ratio of correctly classified true positives to total number of predicted positives (\\(\\frac{TP}{TP+FN}\\)). Negative Predictive Value (NPV): evaluates how accurate predictions of the “negative” class are. It is calculated as the ratio of correctly classified true negatives to total number of predicted negatives (\\(\\frac{TN}{TN+FP}\\)). For the above metrics, values fall between 0 and 1. Instances of 0 indicate that the model was not able to classify any data points correctly, and instances of 1 indicate that the model was able to classify all test data correctly. Although subjective, an overall accuracy of at least 0.7 is considered respectable (Barkved, 2022). Furthermore, a variety of additional metrics exist for evaluating model performance for classification problems (24 Evaluation Metrics for Binary Classification (And When to Use Them)). Selecting a metric for evaluating model performance varies by situation and is dependent not only on the individual dataset, but also the question being answered. Note: For multi-class classification (more than two labeled outcomes to be predicted), the same metrics are often used, but are obtained in a slightly different way. Regression based supervised machine learning models use loss functions to evaluate model performance. For more information regarding confusion matrices and loss functions for regression-based models, see: Additional Confusion Matrix Metrics Precision vs. Recall or Specificity vs. Sensitivity Loss Functions for Machine Learning Regression Introduction to Activity and Example Dataset In this activity, we will analyze an example dataset to see whether we can use environmental monitoring data to predict areas of contamination using random forest (RF). This example model will leverage a dataset of well water variables that span geospatial location, sampling date, and well water attributes, with the goal of predicting whether detectable levels of inorganic arsenic (iAs) are present. This dataset was obtained through the sampling of 713 private wells across North Carolina through the University of North Carolina Superfund Research Program (UNC-SRP) using an analytical method that was capable of detecting levels of iAs greater than 5ppm. As demonstrated through the script below, the algorithm will first be trained and tested, and then resulting model performance will be assessed using the previously detailed confusion matrix and related performance metrics. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: Which well water variables, spanning various geospatial locations, sampling dates, and well water attributes, significantly differ between samples containing detectable levels of iAs vs samples that are not contaminated/ non-detectable? How can we train a random forest (RF) model to predict whether a well might be contaminated with iAs? With this RF model, can we predict if iAs will be detected based on well water information? How could this RF model be improved upon, acknowledging that there is class imbalance? Script Preparations Cleaning the global environment rm(list=ls()) Installing required R packages If you already have these packages installed, you can skip this step, or you can run the below code which checks installation status for you if (!requireNamespace(&quot;readxl&quot;)) install.packages(&quot;readxl&quot;); if (!requireNamespace(&quot;lubridate&quot;)) install.packages(&quot;lubridate&quot;); if (!requireNamespace(&quot;tidyverse&quot;)) install.packages(&quot;tidyverse&quot;); if (!requireNamespace(&quot;gtsummary&quot;)) install.packages(&quot;gtsummary&quot;); if (!requireNamespace(&quot;flextable&quot;)) install.packages(&quot;flextable&quot;); if (!requireNamespace(&quot;caret&quot;)) install.packages(&quot;caret&quot;); if (!requireNamespace(&quot;randomForest&quot;)) install.packages(&quot;randomForest&quot;); Loading R packages required for this session library(readxl); library(lubridate); library(tidyverse); library(gtsummary); library(flextable); library(caret); library(randomForest); Set your working directory setwd(&quot;/filepath to where your input files are&quot;) Importing example dataset # Load the data arsenic_data &lt;- data.frame(read_xlsx(&quot;Module5_2_Input/Module5_2_InputData.xlsx&quot;)) # View the top of the dataset head(arsenic_data) ## Well_ID Water_Sample_Date Casing_Depth Well_Depth Static_Water_Depth ## 1 W_1 9/24/12 52 165 41 ## 2 W_2 12/17/15 40 445 42 ## 3 W_3 2/2/15 45 160 40 ## 4 W_4 10/22/12 42 440 57 ## 5 W_5 1/3/11 48 120 42 ## 6 W_6 12/15/15 60 280 32 ## Flow_Rate pH Detect_Concentration ## 1 60.0 7.7 ND ## 2 2.0 7.3 ND ## 3 40.0 7.4 ND ## 4 1.5 8.0 D ## 5 25.0 7.1 ND ## 6 10.0 8.2 D The columns in this dataset are described below: Well_ID: Unique id for each well (This is the sample identifier and not a predictive feature) Water_Sample_Date: Date that the well was sampled Casing_Depth: Depth of the casing of the well (ft) Well_Depth: Depth of the well (ft) Static_Water_Depth: Static water depth in the well (ft) Flow_Rate: Well flow rate (gallons per minute) pH: pH of water sample Detect_Concentration: Binary identifier (either non-detect “ND” or detect “D”) if iAs concentration detected in water sample Changing Data Types First, Detect_Concentration needs to be converted from a character to a factor so that Random Forest knows that the non-detect class is the baseline or “negative” class, while the detect class will be the “positive” class. Water_Sample_Date will be converted from a character to a date type using the mdy() function from the lubridate package. This is done so that the model understands this column contains dates. arsenic_data &lt;- arsenic_data %&gt;% # Converting `Detect_Concentration` from a character to a factor mutate(Detect_Concentration = relevel(factor(Detect_Concentration), ref = &quot;ND&quot;), # Converting water sample date from a character to a date type Water_Sample_Date = mdy(Water_Sample_Date)) %&gt;% # Removing tax id and only keeping the predictor and outcome variables in the dataset # This allows us to put the entire dataframe as is into RF select(-Well_ID) # Look at the top of the revised dataset head(arsenic_data) ## Water_Sample_Date Casing_Depth Well_Depth Static_Water_Depth Flow_Rate pH ## 1 2012-09-24 52 165 41 60.0 7.7 ## 2 2015-12-17 40 445 42 2.0 7.3 ## 3 2015-02-02 45 160 40 40.0 7.4 ## 4 2012-10-22 42 440 57 1.5 8.0 ## 5 2011-01-03 48 120 42 25.0 7.1 ## 6 2015-12-15 60 280 32 10.0 8.2 ## Detect_Concentration ## 1 ND ## 2 ND ## 3 ND ## 4 D ## 5 ND ## 6 D Testing for Differences in Predictor Variables across the Outcome Classes It is useful to run summary statistics on the variables that will be used as predictors in the algorithm to see if there are differences in distributions between the outcomes classes (either non-detect or detect in this case). Typically, greater significance often leads to better predictivity for a certain variable, since the model is better able to separate the classes. We’ll use the tbl_summary() function from the gtsummary package. Note, this may only be practical with smaller datasets or for a subset of predictors if there are many. For more information on the tbl_summary() function, check out this helpful Tutorial. arsenic_data %&gt;% # Displaying the mean and standard deviation in parentheses for all continuous variables tbl_summary(by = Detect_Concentration, statistic = list(all_continuous() ~ &quot;{mean} ({sd})&quot;)) %&gt;% # Adding a column that displays the total number of samples for each variable. This will be 713 for all variables since we have no missing data add_n() %&gt;% # Adding a column that displays the p-value from an anova test add_p(test = list(all_continuous() ~ &quot;aov&quot;)) %&gt;% as_flex_table() %&gt;% bold(bold = TRUE, part = &quot;header&quot;) ## The following errors were returned during `as_flex_table()`: ## ✖ For variable `Casing_Depth` (`Detect_Concentration`) and &quot;p.value&quot; ## statistic: The package &quot;cardx&quot; (&gt;= 0.2.1) is required. ## ✖ For variable `Flow_Rate` (`Detect_Concentration`) and &quot;p.value&quot; ## statistic: The package &quot;cardx&quot; (&gt;= 0.2.1) is required. ## ✖ For variable `Static_Water_Depth` (`Detect_Concentration`) and ## &quot;p.value&quot; statistic: The package &quot;cardx&quot; (&gt;= 0.2.1) is required. ## ✖ For variable `Water_Sample_Date` (`Detect_Concentration`) and ## &quot;p.value&quot; statistic: The package &quot;cardx&quot; (&gt;= 0.2.1) is required. ## ✖ For variable `Well_Depth` (`Detect_Concentration`) and &quot;p.value&quot; ## statistic: The package &quot;cardx&quot; (&gt;= 0.2.1) is required. ## ✖ For variable `pH` (`Detect_Concentration`) and &quot;p.value&quot; statistic: ## The package &quot;cardx&quot; (&gt;= 0.2.1) is required. .cl-74794e7e{}.cl-7473d14c{font-family:'Arial';font-size:11pt;font-weight:bold;font-style:normal;text-decoration:none;color:rgba(0, 0, 0, 1.00);background-color:transparent;}.cl-7473d156{font-family:'Arial';font-size:6.6pt;font-weight:bold;font-style:normal;text-decoration:none;color:rgba(0, 0, 0, 1.00);background-color:transparent;position: relative;bottom:3.3pt;}.cl-7473d160{font-family:'Arial';font-size:11pt;font-weight:normal;font-style:normal;text-decoration:none;color:rgba(0, 0, 0, 1.00);background-color:transparent;}.cl-7473d161{font-family:'Arial';font-size:6.6pt;font-weight:normal;font-style:normal;text-decoration:none;color:rgba(0, 0, 0, 1.00);background-color:transparent;position: relative;bottom:3.3pt;}.cl-7475e59a{margin:0;text-align:left;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);padding-bottom:2pt;padding-top:2pt;padding-left:5pt;padding-right:5pt;line-height: 1;background-color:transparent;}.cl-7475e59b{margin:0;text-align:center;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);padding-bottom:2pt;padding-top:2pt;padding-left:5pt;padding-right:5pt;line-height: 1;background-color:transparent;}.cl-7475e5a4{margin:0;text-align:left;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);padding-bottom:5pt;padding-top:5pt;padding-left:5pt;padding-right:5pt;line-height: 1;background-color:transparent;}.cl-7475e5a5{margin:0;text-align:center;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);padding-bottom:5pt;padding-top:5pt;padding-left:5pt;padding-right:5pt;line-height: 1;background-color:transparent;}.cl-7475e5a6{margin:0;text-align:left;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);padding-bottom:5pt;padding-top:5pt;padding-left:5pt;padding-right:5pt;line-height: 1;background-color:transparent;}.cl-7475fa58{width:1.703in;background-color:transparent;vertical-align: middle;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 1pt solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa62{width:0.54in;background-color:transparent;vertical-align: middle;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 1pt solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa63{width:2.528in;background-color:transparent;vertical-align: middle;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 1pt solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa6c{width:0.82in;background-color:transparent;vertical-align: middle;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 1pt solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa6d{width:1.703in;background-color:transparent;vertical-align: top;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa76{width:0.54in;background-color:transparent;vertical-align: top;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa77{width:2.528in;background-color:transparent;vertical-align: top;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa78{width:0.82in;background-color:transparent;vertical-align: top;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa80{width:1.703in;background-color:transparent;vertical-align: top;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa81{width:0.54in;background-color:transparent;vertical-align: top;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa82{width:2.528in;background-color:transparent;vertical-align: top;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa83{width:0.82in;background-color:transparent;vertical-align: top;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa8a{width:1.703in;background-color:transparent;vertical-align: middle;border-bottom: 0 solid rgba(255, 255, 255, 0.00);border-top: 0 solid rgba(255, 255, 255, 0.00);border-left: 0 solid rgba(255, 255, 255, 0.00);border-right: 0 solid rgba(255, 255, 255, 0.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa8b{width:0.54in;background-color:transparent;vertical-align: middle;border-bottom: 0 solid rgba(255, 255, 255, 0.00);border-top: 0 solid rgba(255, 255, 255, 0.00);border-left: 0 solid rgba(255, 255, 255, 0.00);border-right: 0 solid rgba(255, 255, 255, 0.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa8c{width:2.528in;background-color:transparent;vertical-align: middle;border-bottom: 0 solid rgba(255, 255, 255, 0.00);border-top: 0 solid rgba(255, 255, 255, 0.00);border-left: 0 solid rgba(255, 255, 255, 0.00);border-right: 0 solid rgba(255, 255, 255, 0.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa8d{width:0.82in;background-color:transparent;vertical-align: middle;border-bottom: 0 solid rgba(255, 255, 255, 0.00);border-top: 0 solid rgba(255, 255, 255, 0.00);border-left: 0 solid rgba(255, 255, 255, 0.00);border-right: 0 solid rgba(255, 255, 255, 0.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}CharacteristicNND N = 5151D N = 1981p-valueWater_Sample_Date7132013-06-05 (979.174260670888)2013-03-05 (957.843005291701)Casing_Depth71374 (33)55 (23)Well_Depth713301 (144)334 (128)Static_Water_Depth71335 (12)36 (13)Flow_Rate71325 (33)14 (16)pH7137.45 (0.55)7.82 (0.40)1Mean (SD) Note that N refers to the total sample number; ND refers to the samples that contained non-detectable levels of iAs; and D refers to the samples that contained detectable levels of iAs. Answer to Environmental Health Question 1 With this, we can answer Environmental Health Question #1: Which well water variables, spanning various geospatial locations, sampling dates, and well water attributes, significantly differ between samples containing detectable levels of iAs vs samples that are not contaminated/ non-detect? Answer: All of the evaluated descriptor variables are significantly different, with p&lt;0.05 between detect and non-detect iAs samples, with the exception of the sample date and the static water depth. With these findings, we feel comfortable moving forward with these well water descriptive variables as predictors in our model. Setting up Cross Validation At this point, we can move forward with training and testing a RF model aimed at predicting whether or not detectable levels of iAs are present in well water samples. We’ll take a glance at the distribution of Detect_Concentration between the two classes. # Set seed for reproducibility set.seed(17) # Establish a list of indices that will used to identify our training and testing data with a 60-40 split tt_indices &lt;- createDataPartition(y = arsenic_data$Detect_Concentration, p = 0.6, list = FALSE) # Use indices to make our training and testing datasets and view the number of Ds and NDs iAs_train &lt;- arsenic_data[tt_indices,] table(iAs_train$Detect_Concentration) ## ## ND D ## 309 119 iAs_test &lt;- arsenic_data[-tt_indices,] table(iAs_test$Detect_Concentration) ## ## ND D ## 206 79 We can see that there are notably more non-detects (ND) than detects (D) in both our training and testing sets. This is something important to consider when evaluating our model’s performance. Now we can set up our cross validation and train our model. We will be using the trainControl() function from the caret package for this task. It is one of the most commonly used libraries for supervised machine learning in R and can be leveraged for a variety algorithms including RF, SVM, KNN, and others. This model will be trained with 5-fold cross validation. Additionally, we will test 2, 3, and 6 predictors through the mtry parameter. See the caret documentation here. # Establish the parameters for our cross validation with 5 folds control &lt;- trainControl(method = &#39;cv&#39;, number = 5, search = &#39;grid&#39;, classProbs = TRUE) # Establish grid of predictors to test in our model as part of hyperparameter tuning p &lt;- ncol(arsenic_data) - 1 # p is the total number of predictors in the dataset tunegrid_rf &lt;- expand.grid(mtry = c(floor(sqrt(p)), p/2, p)) # We will test sqrt(p), p/2, and p predictors (2,3,&amp; 6 predictors, respectively) to see which performs best Predicting iAs Detection with a Random Forest (RF) Model # Look at the column names in training dataset colnames(iAs_train) ## [1] &quot;Water_Sample_Date&quot; &quot;Casing_Depth&quot; &quot;Well_Depth&quot; ## [4] &quot;Static_Water_Depth&quot; &quot;Flow_Rate&quot; &quot;pH&quot; ## [7] &quot;Detect_Concentration&quot; # Train model rf_train &lt;- train(x = iAs_train[,1:6], # Our predictor variables are in columns 1-6 of the dataframe y = iAs_train[,7], # Our outcome variable is in column 7 of the dataframe trControl = control, # Specify the cross-validation parameters we defined above method = &#39;rf&#39;, # Specify we want to train a Random Forest importance = TRUE, # This parameter calculates the variable importance for RF models specifically which can help with downstream analyses tuneGrid = tunegrid_rf, # Specify the number of predictors we want to test as defined above metric = &quot;Accuracy&quot;, ) # Specify what evaluation metric we want to use to decide which model is the best # Look at the results of training rf_train ## Random Forest ## ## 428 samples ## 6 predictor ## 2 classes: &#39;ND&#39;, &#39;D&#39; ## ## No pre-processing ## Resampling: Cross-Validated (5 fold) ## Summary of sample sizes: 344, 342, 342, 342, 342 ## Resampling results across tuning parameters: ## ## mtry Accuracy Kappa ## 2 0.7548726 0.3329518 ## 3 0.7361019 0.2891243 ## 6 0.7501661 0.3257679 ## ## Accuracy was used to select the optimal model using the largest value. ## The final value used for the model was mtry = 2. # Save the best model from our training. The best performing model is determined by the number of predictor variables we tested that resulted in the highest accuracy during the cross validation step. rf_final &lt;- rf_train$finalModel # View confusion matrix for best model rf_final ## ## Call: ## randomForest(x = x, y = y, mtry = param$mtry, importance = TRUE) ## Type of random forest: classification ## Number of trees: 500 ## No. of variables tried at each split: 2 ## ## OOB estimate of error rate: 24.77% ## Confusion matrix: ## ND D class.error ## ND 275 34 0.1100324 ## D 72 47 0.6050420 Answer to Environmental Health Question 2 With this, we can answer Environmental Health Question #2: How can we train a random forest (RF) model to predict whether a well might be contaminated with iAs? Answer: As is standard practice with supervised ML, we split our full dataset into a training dataset and a test dataset using a 60-40 split. Using the caret package, we implemented 5-fold cross validation to train a RF while also testing different numbers of predictors to see which optimized performance. The model that resulted in the greatest accuracy was selected as the final model. Now we can see how well our model does on data it hasn’t seen before by applying it to our testing data. # Use our best model to predict the classes for our test data. We need to make sure we remove the column of Ds/NDs from our test data. rf_res &lt;- predict(rf_final, iAs_test %&gt;% select(!Detect_Concentration)) # View a confusion matrix of the results and gauge model performance # Be sure to include the &#39;positive&#39; parameter to specify the correct positive class confusionMatrix(rf_res, iAs_test$Detect_Concentration, positive = &quot;D&quot;) ## Confusion Matrix and Statistics ## ## Reference ## Prediction ND D ## ND 178 47 ## D 28 32 ## ## Accuracy : 0.7368 ## 95% CI : (0.6817, 0.787) ## No Information Rate : 0.7228 ## P-Value [Acc &gt; NIR] : 0.32445 ## ## Kappa : 0.2907 ## ## Mcnemar&#39;s Test P-Value : 0.03767 ## ## Sensitivity : 0.4051 ## Specificity : 0.8641 ## Pos Pred Value : 0.5333 ## Neg Pred Value : 0.7911 ## Prevalence : 0.2772 ## Detection Rate : 0.1123 ## Detection Prevalence : 0.2105 ## Balanced Accuracy : 0.6346 ## ## &#39;Positive&#39; Class : D ## Answer to Environmental Health Question 3 With this, we can answer Environmental Health Question #3: With this RF model, can we predict if iAs will be detected based on well water information? Answer: We can use this model to predict if iAs can be detected in well water given that an overall accuracy of ~0.72 is decent, however we should consider other metrics that may influence how good we feel about this model depending on what is important to the question we are trying to answer. For example, the model did a good job at predicting non-detect data based on a sensitivity of ~0.85 and a NPV ~0.78, but struggled at predicting detect data based on a specificity of ~0.39 and a PPV of ~0.50. Additionally, the balanced accuracy of ~0.62 further emphasizes the difference in predictive ability of the model for non-detects and detects. If it is highly important to us that detects are classified correctly, we may want to improve this model before implementing it. Class Imbalance It is worth noting this discrepancy in predictive capabilities for detects vs. non-detects makes sense due to the observed class imbalance in our training data. There were notably more non-detects than detects in the training set, so the model was exposed to more of these data points and struggles to distinguish unique characteristics of detects when compared to non-detects. Additionally, we told the training algorithm to prioritize selecting a final model based on its overall accuracy. In the instances of a heavy class imbalance, it is common for a high accuracy to be achieved as the more prevalent class is predicted more often, though this doesn’t give the full picture of the model’s predictive capabilities. For example, if you consider a dog/cat case with a set of 90 dogs and 10 cats, a model could achieve 90% accuracy by predicting dog every time, which isn’t at all helpful in predicting cats. This is particularly important, because for toxicology related datasets, the “positive” class often represents the class with greater public health risk/ interest but can have less data. For example, when you classify subjects based upon whether or not they have asthma based on gene expression data. Asthmatics would likely be the “positive” class, but given that asthmatics are less prevalent than non-asthmatics in the general population, they would likely represent the minority class too. To address this issue, a few methods can be considered. Full implementation of these approaches is beyond the scope of this module, but relevant resources for further exploration are given. Synthetic Minority Oversampling Technique (SMOTE)- increases the number of minority classes in the training data, thereby reducing the class imbalance by synthetically generating additional samples derived from the existing minority class samples. SMOTE Oversampling &amp; Tutorial On How To Implement In Python And R How to Use SMOTE for Imbalanced Data in R (With Example) Adjusting the loss function- Loss functions in machine learning quantify the penalty for a bad prediction. They can be adjusted to where the minority class is penalized more forcing the model to learn to make fewer mistakes when predicting the minority class. Alternative Performance Metrics- When training the model, alternative metrics to overall accuracy may yield a more robust model capable of better predicting the minority class. Example alternatives may include balanced accuracy or an F1-score. The caret package further allows for custom, user-defined metrics to be evaluated during training by specifying the summaryFunction parameter in the trainControl() function, as seen below, in addition to the defaultSummary() and twoClassSummary() functions. In the example code below, we’re creating a function (f1) that will calculate the F1 score and find the optimal model with the highest F1 score as opposed to the highest accuracy as we did above. install.packages(&quot;MLmetrics&quot;) library(MLmetrics) f1 &lt;- function(data, lev = NULL, model = NULL) { # Creating a function to calculate the F1 score f1_val &lt;- F1_Score(y_pred = data$pred, y_true = data$obs, positive = lev[1]) c(F1 = f1_val) } # 5 fold CV ctrl &lt;- trainControl( method = &quot;cv&quot;, number = 5, classProbs = TRUE, summaryFunction = f1 ) # Training the RF model mod &lt;- train(x = X, y = Y, trControl = ctrl, method = &quot;rf&quot;, tuneGrid = tunegrid_rf, importance = TRUE, # Basing the best model performance off of the F1 score within 5 CV metric = &quot;F1&quot;) For more in-depth information and additional ways to address class imbalance check out How to Deal with Imbalanced Data in Classification. Answer to Environmental Health Question 4 With this, we can answer Environmental Health Question #4: How could this RF model be improved upon, acknowledging that there is class imbalance? Answer: We can implement SMOTE to increase the number of training data points for the minority class thereby reducing the class imbalance. In conjunction with using SMOTE, another approach includes selecting an alternative performance metric during training that does a better job taking the existing class imbalance into consideration, such as balanced accuracy or an F1-score, improves our predictive ability for the minority class. Concluding Remarks In conclusion, this training module has provided an introduction to supervised machine learning using classification techniques in R. Machine learning is a powerful tool that can help researchers gain new insights and improve models to analyze complex datasets faster and in a more comprehensive way. The example we’ve explored demonstrates the utility of supervised machine learning models on an environmentally relevant dataset. Additional Resources To learn more check out the following resources: IBM - What is Machine Learning Curate List of AI and Machine Learning Resources Introduction to Machine Learning in R Machine Learning by Mueller, J. P. (2021). Machine learning for dummies. John Wiley &amp; Sons. Test Your Knowledge Using the “Module5_2TYKInput.xlsx”, use RF to determine if well water data can be accurate predictors of Manganese detection. The data is structured similarly to the “Module5_2_InputData.xlsx” used in this module, however it now includes 4 additional features: Longitude: Longitude of address (decimal degrees) Latitude: Latitude of address (decimal degrees) Stream_Distance: Euclidean distance to the nearest stream (feet) Elevation: Surface elevation of the sample location (feet) "],["supervised-machine-learning-model-interpretation.html", "Supervised Machine Learning Model Interpretation Introduction to Training Module Variable Importance Decision Boundary Introduction to Example Dataset and Activity Predicting iAs Detection with a Random Forest (RF) Model Variable Importance Plot Decision Boundary Plot Decision Boundary Plot Incorporating SMOTE Concluding Remarks", " Supervised Machine Learning Model Interpretation This training module was developed by Alexis Payton, Lauren E. Koval, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module Supervised machine learning (ML) represents a subset of ML methods wherein the outcome variable is known or assigned prior to training a model to be able to predict said outcome. As we discussed in previous modules, ML methods are advantageous in that they easily incorporate a multitude of potential predictor variables, which allows these models to more closely consider real-world, complex environmental health scenarios and offer new insights through a more holistic consideration of available data inputs. However, one disadvantage of ML is that it is often not as easily interpretable as traditional statistics (e.g., regression based methods with defined beta coefficients for each input predictor variable). With this limitation in mind, there are methods and concepts that can be applied to supervised ML algorithms to aid in the understanding of their predictions including variable (feature) importance and decision boundaries, which we will cover in this module. We will also include example visualization techniques of these methods, representing important aspects contributing to model interpretability, since visualizing helps convey concepts faster and across a broader target audience. In addition, this module addresses methods to communicate these findings in a paper so that a wider span of readers can understand overall take-home points. As with other data analyses, we advise to focus just as much on the why components of a study’s research question(s) as opposed to only focusing on the what or how. To elaborate, we explain through this module that it is not as important to explain all the intricacies of how a model works and how its parameters were tuned; rather, it is more important to focus on why a particular model was selected and how it will be leveraged to answer your research questions. This can all be a bit subjective and requires expertise within your research field. As a first step, let’s first learn about some model interpretation methodologies highlighting Variable Importance and Decision Boundaries as important examples relevant to environmental health research. Then, this training module will further describe approaches to summarize these methods and communicate supervised ML findings to a broader audience. Variable Importance When a supervised ML algorithm makes predictions, it relies more heavily on some variables than others. How much a variable contributes to classifying data is known as variable (feature) importance. Often times, this is thought of as the impact on overall model performance if a variable were to be removed from the model. There are many methods that are used to measure feature importance, including… SHapley Additive exPlanations (SHAP): based on game theory where each variable is considered a “player” where we’re seeking to determine each player’s contribution to the outcome of a “game” or overall model performance. It divides the model performance metric amongst all the variables, so that the sum of the shapley values for all the predictors is equal to the overall model performance. For more information on SHAP, see A Novel Approach to Feature Importance. Mean decrease gini (gini impurity): quantifies the improvement of predictivity with the addition of each predictor in a decision tree, which is then averaged over all the decision trees tested. The higher the value the greater the importance on the algorithm. This metric can easily be extracted from classification-based models, including random forest (RF) classifications, which is what we will focus on in this module. Note for RF regression-based models, node purity can be extracted as a measure of feature importance. For more information, please see the following resources regarding Feature Importance and Mean Decrease Gini. Decision Boundary Another concept that is pertinent to a model’s interpretability is understanding a decision boundary and how visualizing it can further aid in understanding how the model classifies new data points. A decision boundary is a line (or a hyperplane) that seeks to separate the training data by class. This line can be linear or non-linear and is formed in n-dimensional space. To clarify, although support vector machine (SVM) specifically uses decision boundaries to classify training data and make predictions on test data, decision boundaries can still be drawn for other algorithms. A decision boundary can be visualized to convey how well an algorithm is able to classify an outcome based on the data given. It is important to note that most ML models make use of datasets that contain three or more predictors, and it is difficult to visualize a plot in more than three dimensions. Therefore, the number of features and which features to plot need to be narrowed down to two variables. For this reason, the resulting visualization is not a true representation of the decision boundary from the initial model using all predictors, since the visualization only relies on prediction results from two variables. Nevertheless, decision boundary plots can be powerful visualizations to determine thresholds between the outcome classes. When choosing variables for decision boundary plots, features that have the most influence on the model are often selected, but that is not always the case. Sometimes predictors are selected based upon the environmental health implications relevant to the research question. For example in Perryman et. al, lung response following ozone exposure was investigated by sampling derivatives of cholesterol biosynthesis in human subjects. In this paper, these sterol metabolites were used to predict whether a subject would be classified as having a lung response that was considered non-responsive or responsive. A decision boundary plot was made using two predictors: Cholesterol, given that it had the highest variable importance and Vitamin D, given its synthesis can be affected by ozone despite it having a lower variable importance in the paper’s models. Figure 5. Decision boundary plot for SVM model predicting lung response class. Cholesterol and 25-hydroxyvitamin D were used as predictors visualizing responder status [non-responders(green) and responders (yellow)] and disease status [non-asthmatics (triangles) and asthmatics (circles)]. The shaded regions are the model’s prediction of a subject’s lung response class at a given cholesterol and 25-hydroxyvitamin D concentration. Takeaways from this decision boundary plot: Subjects with more lung inflammation (“responders”) after ozone exposure tended to have higher Vitamin D levels (&gt; 35pmol/mL) and lower Cholesterol levels (&lt; 675nmol/mL). These “responder” subjects were more likely to be non-asthmatics. Introduction to Example Dataset and Activity In the previous module, we investigated whether a classification-based RF model using well water variables would be accurate predictors of inorganic arsenic (iAs) contamination. While it is helpful to know if certain variables are able to be used to construct a model that accurately predict detectability, from a public health standpoint, it is also helpful to know which of those features contribute the most to a model’s accuracy. Therefore, if we can identify the features that are associated with having lower arsenic detection, we can use that information to inform policies when new wells are constructed. In addition to identifying variables with the greatest importance to the algorithm, it is also pertinent to understand the ranges of when a well is more or less likely to have arsenic detected. For example, are wells with a lower flow rate more likely to have arsenic detected? In this module, this will be addressed by extracting variable importance from the same algorithm and plotting it. The two features with the highest variable importance will be identified and used to construct a decision boundary plot to determine how features are associated with iAs detection. The data to be used in this module was described and referenced previously in TAME 2.0 Module 5.2 Supervised Machine Learning. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: After plotting variable importance from highest to lowest, which two predictors have the highest variable importance on the predictive accuracy of iAs detection from a RF algorithm? Using the two features with the highest variable importance, under what conditions are we more likely to predict detectable iAs in wells based on a decision boundary plot? How do the decision boundaries shift after incorporating SMOTE to address class imbalance? Script Preparations Cleaning the global environment rm(list=ls()) Installing required R packages If you already have these packages installed, you can skip this step, or you can run the below code which checks installation status for you if (!requireNamespace(&quot;readxl&quot;)) install.packages(&quot;readxl&quot;); if (!requireNamespace(&quot;lubridate&quot;)) install.packages(&quot;lubridate&quot;); if (!requireNamespace(&quot;tidyverse&quot;)) install.packages(&quot;tidyverse&quot;); if (!requireNamespace(&quot;caret&quot;)) install.packages(&quot;caret&quot;); if (!requireNamespace(&quot;randomForest&quot;)) install.packages(&quot;randomForest&quot;); if (!requireNamespace(&quot;themis&quot;)) install.packages(&quot;themis&quot;); Loading R packages required for this session library(readxl) library(lubridate) library(tidyverse) library(caret) library(randomForest) library(e1071) library(ggsci) library(themis) Set your working directory setwd(&quot;/filepath to where your input files are&quot;) Importing example dataset # Load the data arsenic_data &lt;- data.frame(read_excel(&quot;Module5_3_Input/Module5_3_InputData.xlsx&quot;)) # View the top of the dataset head(arsenic_data) ## Well_ID Water_Sample_Date Casing_Depth Well_Depth Static_Water_Depth ## 1 W_1 9/24/12 52 165 41 ## 2 W_2 12/17/15 40 445 42 ## 3 W_3 2/2/15 45 160 40 ## 4 W_4 10/22/12 42 440 57 ## 5 W_5 1/3/11 48 120 42 ## 6 W_6 12/15/15 60 280 32 ## Flow_Rate pH Detect_Concentration ## 1 60.0 7.7 ND ## 2 2.0 7.3 ND ## 3 40.0 7.4 ND ## 4 1.5 8.0 D ## 5 25.0 7.1 ND ## 6 10.0 8.2 D Changing Data Types First, Detect_Concentration needs to be converted from a character to a factor so that Random Forest knows that the non-detect class is the baseline or “negative” class, while the detect class will be the “positive” class. Water_Sample_Date will be converted from a character to a date type using the mdy() function from the lubridate package. This is done so that the model understands this column contains dates. arsenic_data &lt;- arsenic_data %&gt;% # Converting `Detect_Concentration` from a character to a factor mutate(Detect_Concentration = relevel(factor(Detect_Concentration), ref = &quot;ND&quot;), # Converting water sample date from a character to a date type Water_Sample_Date = mdy(Water_Sample_Date)) %&gt;% # Removing well id and only keeping the predictor and outcome variables in the dataset # This allows us to put the entire dataframe as is into RF select(-Well_ID) # View the top of the current dataset head(arsenic_data) ## Water_Sample_Date Casing_Depth Well_Depth Static_Water_Depth Flow_Rate pH ## 1 2012-09-24 52 165 41 60.0 7.7 ## 2 2015-12-17 40 445 42 2.0 7.3 ## 3 2015-02-02 45 160 40 40.0 7.4 ## 4 2012-10-22 42 440 57 1.5 8.0 ## 5 2011-01-03 48 120 42 25.0 7.1 ## 6 2015-12-15 60 280 32 10.0 8.2 ## Detect_Concentration ## 1 ND ## 2 ND ## 3 ND ## 4 D ## 5 ND ## 6 D Setting up Cross Validation Note that the code below is different than the code presented in the previous module, TAME 2.0 Module 5.2 Supervised Machine Learning. Both coding methods are valid and produce comparable results, however we wanted to present another way to run k-fold cross validation and random forest. In 5-fold cross validation (CV), there are 5 equally-sized folds (ideally!). This means that 80% of the original dataset is split into the 4 folds that comprise the training set and the remaining 20% in the last fold is reserved for the test set. Previously, the trainControl() function was used for CV. This time we’ll use the createFolds() function also from the caret package. # Setting seed for reproducibility set.seed(12) # 5-fold cross validation arsenic_index = createFolds(arsenic_data$Detect_Concentration, k = 5) # Seeing if about 20% of the records are in the testing set kfold1 = arsenic_index[[1]] length(kfold1)/nrow(arsenic_data) ## [1] 0.1991585 # Creating vectors for parameters to be tuned ntree_values = c(50, 250, 500) # number of decision trees p = dim(arsenic_data)[2] - 1 # number of predictor variables in the dataset mtry_values = c(sqrt(p), p/2, p) # number of predictors to be used in the model Predicting iAs Detection with a Random Forest (RF) Model Notice that in the code below we are choosing the final RF model to be the one with the lowest out of bag (OOB) error. In the previous module, the final model was chosen based on the highest accuracy, however this is a similar approach here given that OOB error = 1 - Accuracy. # Setting the seed again so the predictions are consistent set.seed(12) # Creating an empty dataframe to save the confusion matrix metrics and variable importance metrics = data.frame() variable_importance_df = data.frame() # Iterating through the cross validation folds for (i in 1:length(arsenic_index)){ # Training data data_train = arsenic_data[-arsenic_index[[i]],] # Test data data_test = arsenic_data[arsenic_index[[i]],] # Creating empty lists and dataframes to store errors reg_rf_pred_tune = list() rf_OOB_errors = list() rf_error_df = data.frame() # Tuning parameters: using ntree and mtry values to determine which combination yields the smallest OOB error # from the validation datasets for (j in 1:length(ntree_values)){ for (k in 1:length(mtry_values)){ # Running RF to tune parameters reg_rf_pred_tune[[k]] = randomForest(Detect_Concentration ~ ., data = data_train, ntree = ntree_values[j], mtry = mtry_values[k]) # Obtaining the OOB error rf_OOB_errors[[k]] = data.frame(&quot;Tree Number&quot; = ntree_values[j], &quot;Variable Number&quot; = mtry_values[k], &quot;OOB_errors&quot; = reg_rf_pred_tune[[k]]$err.rate[ntree_values[j],1]) # Storing the values in a dataframe rf_error_df = rbind(rf_error_df, rf_OOB_errors[[k]]) } } # Finding the lowest OOB error from the 5 folds using best number of predictors at split best_oob_errors &lt;- which(rf_error_df$OOB_errors == min(rf_error_df$OOB_errors)) # Now running RF on the entire training set with the tuned parameters # This will be done 5 times for each fold reg_rf &lt;- randomForest(Detect_Concentration ~ ., data = data_train, ntree = rf_error_df$Tree.Number[min(best_oob_errors)], mtry = rf_error_df$Variable.Number[min(best_oob_errors)]) # Predicting on test set and adding the predicted values as an additional column to the test data data_test$Pred_Detect_Concentration = predict(reg_rf, newdata = data_test, type = &quot;response&quot;) matrix = confusionMatrix(data = data_test$Pred_Detect_Concentration, reference = data_test$Detect_Concentration, positive = &quot;D&quot;) # Extracting accuracy, sens, spec, PPV, NPV and adding to the dataframe to take mean later matrix_values = data.frame(t(c(matrix$byClass[11])), t(c(matrix$byClass[1:4]))) metrics = rbind(metrics, matrix_values) # Extracting variable importance variable_importance_values = data.frame(importance(reg_rf)) %&gt;% rownames_to_column(var = &quot;Predictor&quot;) variable_importance_df = rbind(variable_importance_df, variable_importance_values) } # Taking average across the 5 folds metrics = metrics %&gt;% summarise(`Balanced Accuracy` = mean(Balanced.Accuracy), Sensitivity = mean(Sensitivity), Specificity = mean(Specificity), PPV = mean(Pos.Pred.Value), NPV = mean(Neg.Pred.Value)) variable_importance_df = variable_importance_df %&gt;% group_by(Predictor) %&gt;% summarise(MeanDecreaseGini = mean(MeanDecreaseGini)) %&gt;% # Sorting from highest to lowest arrange(-MeanDecreaseGini) The confusion matrix results from the previous module are shown below. Now let’s double check that when using this new method, our results are still comparable. # First comparing results to the previous module round(metrics, 2) ## Balanced Accuracy Sensitivity Specificity PPV NPV ## 1 0.64 0.41 0.87 0.55 0.79 They are! Now we’ll take a look at the model’s variable importance. variable_importance_df ## # A tibble: 6 × 2 ## Predictor MeanDecreaseGini ## &lt;chr&gt; &lt;dbl&gt; ## 1 Casing_Depth 50.6 ## 2 pH 42.3 ## 3 Water_Sample_Date 37.6 ## 4 Flow_Rate 33.5 ## 5 Well_Depth 32.2 ## 6 Static_Water_Depth 31.7 Although we have the results we need, let’s take it a step further and plot the data. Reformatting the dataframe for plotting First, the dataframe will be transformed so that the figure is more legible. Specifically, spaces will be added between the variables, and the Predictor column will be put into a factor to rearrange the order of the variables from lowest to highest mean decrease gini. For additional information on tricks like this to make visualizations easier to read, see TAME 2.0 Module 3.2 Improving Data Visualizations. # Adding spaces between the variables that need the space modified_variable_importance_df = variable_importance_df %&gt;% mutate(Predictor = gsub(&quot;_&quot;, &quot; &quot;, Predictor)) # Saving the order of the variables from lowest to highest mean decrease gini by putting into a factor predictor_order = rev(modified_variable_importance_df$Predictor) modified_variable_importance_df$Predictor = factor(modified_variable_importance_df$Predictor, levels = predictor_order) head(modified_variable_importance_df) ## # A tibble: 6 × 2 ## Predictor MeanDecreaseGini ## &lt;fct&gt; &lt;dbl&gt; ## 1 Casing Depth 50.6 ## 2 pH 42.3 ## 3 Water Sample Date 37.6 ## 4 Flow Rate 33.5 ## 5 Well Depth 32.2 ## 6 Static Water Depth 31.7 Variable Importance Plot ggplot(data = modified_variable_importance_df , aes(x = MeanDecreaseGini, y = Predictor, size = 2)) + geom_point() + theme_light() + theme(axis.line = element_line(color = &quot;black&quot;), #making x and y axes black axis.text = element_text(size = 12), #changing size of x axis labels axis.title = element_text(face = &quot;bold&quot;, size = rel(1.7)), #changes axis titles legend.title = element_text(face = &#39;bold&#39;, size = 14), #changes legend title legend.text = element_text(size = 12), #changes legend text strip.text.x = element_text(size = 15, face = &quot;bold&quot;), #changes size of facet x axis strip.text.y = element_text(size = 15, face = &quot;bold&quot;)) + #changes size of facet y axis labs(x = &#39;Variable Importance&#39;, y = &#39;Predictor&#39;) + #changing axis labels guides(size = &quot;none&quot;)#removing size legend An appropriate title for this figure could be: “Figure X. Variable importance from random forest models predicting iAs detection. Variable importance is derived from mean decrease gini values extracted from random forest models. Features are listed on the y axis from greatest (top) to least (bottom) mean decrease gini.” Answer to Environmental Health Question 1 With this, we can answer Environmental Health Question #1: After plotting variable importance from highest to lowest, which two predictors have the highest variable importance on the predictive accuracy of iAs detection from a RF algorithm? Answer: From the variable importance dataframe and plot, we can see that casing depth and pH had the greatest impact on RF followed by water sample date, flow rate, static water depth, and well depth in descending order. Since casing depth and pH have been identified as the predictors with the highest variable importance, they will be prioritized as the two predictors included in the decision boundary plot example below. Decision Boundary Calculation First, models will be trained using only casing depth and pH as variables. Since, the decision boundary plot will be used for visualization purposes, and a 2-D figure can only plot two variables, we will not worry about tuning the parameters as was previously done. In this module, we’re creating a decision boundary based on a random forest model, however we’ll also explore what decision boundaries look like for other algorithms including support vector machine (SVM), and k nearest neighbor (KNN), logistic regression. Each supervised ML method has its advantages and performance is dependent upon the situation and the dataset. Therefore, it is common to see multiple models used to predict an outcome of interest in a publication. Let’s create additional boundary plots still using casing depth and pH, but this time we will use logistic regression, SVM, and KNN as comparisons to RF. # Creating a dataframe with variables based on the highest predictors highest_pred_data = data.frame(arsenic_data[,c(&quot;Casing_Depth&quot;, &quot;pH&quot;, &quot;Detect_Concentration&quot;)]) # Training RF rf_detect_arsenic = randomForest(Detect_Concentration~., data = highest_pred_data) # Logistic regression lr_detect_arsenic = glm(Detect_Concentration~., data = highest_pred_data, family = binomial(link = &#39;logit&#39;)) # SVM with a radial kernel (hyperplane) svm_detect_arsenic = svm(Detect_Concentration~., data = highest_pred_data, kernel = &quot;radial&quot;) # KNN knn_detect_arsenic = knn3(Detect_Concentration~., data = highest_pred_data) # specifying 2 classes From these predictions, decision boundaries will be calculated. This will be done by predicting Detect_Concentration between a grid of values - specifically the minimum and maximum of the two predictors (casing depth and pH). A non-linear line will be drawn on the plot to separate the two classes. get_grid_df &lt;- function(classification_model, data, resolution = 100, predict_type) { # This function predicts the outcome (Detect_Concentration) at evenly spaced data points using the two variables (pH and casing depth) # to create a decision boundary between the outcome classes (detect and non-detect samples). # :parameters: a classification-based supervised machine learning model, dataset containing the predictors and outcome variable, # specifies the number of data points to make between the minimum and maximum predictor values, prediction type # :output: a grid of values for both predictors and their corresponding predicted outcome class # Grabbing only the predictor data predictor_data &lt;- data[,1:2] # Creating a dataframe that contains the min and max for both features min_max_df &lt;- sapply(predictor_data, range, na.rm = TRUE) # Creating a vector of evenly spaced points between the min and max for the first variable (casing depth) variable1_vector &lt;- seq(min_max_df[1,1], min_max_df[2,1], length.out = resolution) # Creating a vector of evenly spaced points between the min and max for the second variable (pH) variable2_vector &lt;- seq(min_max_df[1,2], min_max_df[2,2], length.out = resolution) # Creating a dataframe of grid values by combining the two vectors grid_df &lt;- data.frame(cbind(rep(variable1_vector, each = resolution), rep(variable2_vector, time = resolution))) colnames(grid_df) &lt;- colnames(min_max_df) # Predicting class label based on all the predictor pairs of data grid_df$Pred_Class = predict(classification_model, grid_df, type = predict_type) return(grid_df) } # calling function # RF grid_df_rf = get_grid_df(rf_detect_arsenic, highest_pred_data, predict_type = &quot;class&quot;) %&gt;% # Adding in a column that indicates the model so all the dataframes can be combined mutate(Model = &quot;A. Random Forest&quot;) # SVM with a radial kernel (hyperplane) grid_df_svm = get_grid_df(svm_detect_arsenic, highest_pred_data, predict_type = &quot;class&quot;) %&gt;% mutate(Model = &quot;B. Support Vector Machine&quot;) # KNN grid_df_knn = get_grid_df(knn_detect_arsenic, highest_pred_data, predict_type = &quot;class&quot;) %&gt;% mutate(Model = &quot;C. K Nearest Neighbor&quot;) # Logistic regression grid_df_lr = get_grid_df(lr_detect_arsenic, highest_pred_data, predict_type = &quot;response&quot;) %&gt;% # First specifying the cutoff point for logistic regression predictions # If the response is &gt;= 0.5 it will be classified as a detect prediction mutate(Pred_Class = relevel(factor(ifelse(Pred_Class &gt;= 0.5, &quot;D&quot;, &quot;ND&quot;)), ref = &quot;ND&quot;), Model = &quot;D. Logistic Regression&quot;) # Creating 1 dataframe grid_df = rbind(grid_df_rf, grid_df_lr, grid_df_svm, grid_df_knn) # Viewing the dataframe to be plotted head(grid_df) ## Casing_Depth pH Pred_Class Model ## 1 27 5.400000 ND A. Random Forest ## 2 27 5.433333 ND A. Random Forest ## 3 27 5.466667 ND A. Random Forest ## 4 27 5.500000 ND A. Random Forest ## 5 27 5.533333 ND A. Random Forest ## 6 27 5.566667 ND A. Random Forest Decision Boundary Plot Now let’s plot the grid of predictions with the sampled data. # choosing palette from package ggsci_colors = pal_npg()(5) ggplot() + geom_point(data = arsenic_data, aes(x = pH, y = Casing_Depth, color = Detect_Concentration), position = position_jitter(w = 0.1, h = 0.1), size = 4, alpha = 0.8) + geom_contour(data = grid_df, aes(x = pH, y = Casing_Depth, z = as.numeric(Pred_Class == &quot;D&quot;)), color = &quot;black&quot;, breaks = 0.5) + # adds contour line geom_point(data = grid_df, aes(x = pH, y = Casing_Depth, color = Pred_Class), size = 0.1) + # shades plot xlim(5.9, NA) + # changes the limits of the x axis facet_wrap(~Model, scales = &#39;free&#39;) + theme_light() + theme(axis.line = element_line(color = &quot;black&quot;), #making x and y axes black axis.text = element_text(size = 10), #changing size of x axis labels axis.title = element_text(face = &quot;bold&quot;, size = rel(1.7)), #changes axis titles legend.title = element_text(face = &#39;bold&#39;, size = 12), #changes legend title legend.text = element_text(size = 12), #changes legend text legend.position = &quot;bottom&quot;, # move legend to top left corner legend.background = element_rect(color = &#39;black&#39;, fill = &#39;white&#39;, linetype = &#39;solid&#39;), # changes legend background strip.text = element_text(size = 15, face = &quot;bold&quot;)) + #changes size of facet x axis labs(y = &#39;Casing Depth (ft)&#39;) + #changing axis labels scale_color_manual(name = &quot;Arsenic Detection&quot;, # renaming the legend values = ggsci_colors[c(4,5)], labels = c(&#39;Non-Detect&#39;,&#39;Detect&#39;)) # renaming the classes Answer to Environmental Health Question 2 With this, we can answer Environmental Health Question #2: Using the two features with the highest variable importance, under what conditions are we more likely to predict detectable iAs in wells based on a decision boundary plot? Answer: There is some overlap between detect and non-detect iAs samples; however, it is evident that wells with detectable levels of iAs were more likely to have lower (&lt;80 ft) casing depths and a more basic pH (&gt; 7) based on RF and KNN models. It seems like SVM and logistic regression could have potentially captured a greater “detect” region indicating that the models likely struggled to predict “detect” values. In the next section, SMOTE will be used to see if these decision boundaries can be improved. Decision Boundary Plot Incorporating SMOTE Here, we will create a decision boundary plot still using casing depth and pH, but this time we will make our dataset more balance to see how improve model performance visually. The Synthetic Minority Oversampling Technique (SMOTE) was introduced in TAME 2.0 Module 5.2 Supervised Machine Learning and will be used to make the dataset more balanced by oversampling the minority class (detect values) and undersampling the majority class (non-detect values). Starting by training each model: # Using SMOTE first to balance classes balanced_highest_pred_data = smotenc(highest_pred_data, &quot;Detect_Concentration&quot;) # Training RF rf_detect_arsenic = randomForest(Detect_Concentration~., data = balanced_highest_pred_data) # Logistic regression lr_detect_arsenic = glm(Detect_Concentration~., data = balanced_highest_pred_data, family = binomial(link = &#39;logit&#39;)) # SVM with a radial kernel (hyperplane) svm_detect_arsenic = svm(Detect_Concentration~., data = balanced_highest_pred_data, kernel = &quot;radial&quot;) # KNN knn_detect_arsenic = knn3(Detect_Concentration~., data = balanced_highest_pred_data) # specifying 2 classes Now calling the get_grid_df() function we created above to create a grid of predictions. # Calling function # RF balanced_grid_df_rf = get_grid_df(rf_detect_arsenic, balanced_highest_pred_data, predict_type = &quot;class&quot;) %&gt;% # Adding in a column that indicates the model so all the dataframes can be combined mutate(Model = &quot;A. Random Forest&quot;) # SVM with a radial kernel (hyperplane) balanced_grid_df_svm = get_grid_df(svm_detect_arsenic, balanced_highest_pred_data, predict_type = &quot;class&quot;) %&gt;% mutate(Model = &quot;B. Support Vector Machine&quot;) # KNN balanced_grid_df_knn = get_grid_df(knn_detect_arsenic, balanced_highest_pred_data, predict_type = &quot;class&quot;) %&gt;% mutate(Model = &quot;C. K Nearest Neighbor&quot;) # Logistic regression balanced_grid_df_lr = get_grid_df(lr_detect_arsenic, balanced_highest_pred_data, predict_type = &quot;response&quot;) %&gt;% # First specifying the cutoff point for logistic regression predictions # If the response is &gt;= 0.5 it will be classified as a detect prediction mutate(Pred_Class = relevel(factor(ifelse(Pred_Class &gt;= 0.5, &quot;D&quot;, &quot;ND&quot;)), ref = &quot;ND&quot;), Model = &quot;D. Logistic Regression&quot;) # Creating 1 dataframe balanced_grid_df = rbind(balanced_grid_df_rf, balanced_grid_df_lr, balanced_grid_df_svm, balanced_grid_df_knn) # Viewing the dataframe to be plotted head(balanced_grid_df) ## Casing_Depth pH Pred_Class Model ## 1 27 5.400000 ND A. Random Forest ## 2 27 5.433333 ND A. Random Forest ## 3 27 5.466667 ND A. Random Forest ## 4 27 5.500000 ND A. Random Forest ## 5 27 5.533333 ND A. Random Forest ## 6 27 5.566667 ND A. Random Forest # choosing palette from package ggsci_colors = pal_npg()(5) ggplot() + geom_point(data = arsenic_data, aes(x = pH, y = Casing_Depth, color = Detect_Concentration), position = position_jitter(w = 0.1, h = 0.1), size = 4, alpha = 0.8) + geom_contour(data = balanced_grid_df, aes(x = pH, y = Casing_Depth, z = as.numeric(Pred_Class == &quot;D&quot;)), color = &quot;black&quot;, breaks = 0.5) + # adds contour line geom_point(data = balanced_grid_df, aes(x = pH, y = Casing_Depth, color = Pred_Class), size = 0.1) + # shades plot xlim(5.9, NA) + # changes the limits of the x axis facet_wrap(~Model, scales = &#39;free&#39;) + theme_light() + theme(axis.line = element_line(color = &quot;black&quot;), #making x and y axes black axis.text = element_text(size = 10), #changing size of x axis labels axis.title = element_text(face = &quot;bold&quot;, size = rel(1.7)), #changes axis titles legend.title = element_text(face = &#39;bold&#39;, size = 12), #changes legend title legend.text = element_text(size = 12), #changes legend text legend.position = &quot;bottom&quot;, # move legend to top left corner legend.background = element_rect(color = &#39;black&#39;, fill = &#39;white&#39;, linetype = &#39;solid&#39;), # changes legend background strip.text = element_text(size = 15, face = &quot;bold&quot;)) + #changes size of facet x axis labs(y = &#39;Casing Depth (ft)&#39;) + #changing axis labels scale_color_manual(name = &quot;Arsenic Detection&quot;, # renaming the legend values = ggsci_colors[c(4,5)], labels = c(&#39;Non-Detect&#39;,&#39;Detect&#39;)) # renaming the classes An appropriate title for this figure could be: “Figure X. Decision boundary plots from supervised machine learning models predicting iAs detection. The top two predictors on model performance, casing depth and pH, were used to visualize arsenic detection [non-detect (red) and detect (blue)]. The shaded regions represent prediction of a well’s detection class based on varying casing depth and pH values using (A) Random Forest, (B) Support Vector Machine, (C) K Nearest Neighbor, and (D) Logistic Regression. Answer to Environmental Health Question 3 With this, we can answer Environmental Health Question #3: How do the decision boundaries shift after incorporating SMOTE to address class imbalance? Answer: It is still evident that wells with detectable levels of iAs were more likely to have lower (&lt;80 ft) casing depths and a more basic pH (&gt; 7). However, we see the greatest shifts in the decision boundaries of SVM and logistic regression with both models now predicting greater regions to detectable iAs levels. Concluding Remarks In conclusion, this training module provided methodologies to aid in the interpretation of supervised ML with variable importance and decision boundary plots. Variable importance helps quantify the impact of each feature’s importance on an algorithm’s predictivity. The most important or environmentally-relevant predictors can be selected in a decision boundary plot to further understand and visualize the features impact on the model’s classification. Additional Resources Christoph Molnar. (2019, August 27). Interpretable Machine Learning. Github.io. https://christophm.github.io/interpretable-ml-book/ Variable Importance Decision Boundary Test Your Knowledge Using the “Module5_2_TYKInput.xlsx”, use RF to determine if well water data can be accurate predictors of manganese detection as was done in the previous module. However, this time, incorporate SMOTE in the model. Feel free to use either the trainControl() or createFolds() function for CV. Extract the variable importance for each predictor on a RF model. What two features have the highest variable importance? Hint: Regardless of the cross validation function you choose, run SMOTE on the training dataset only to create a more balanced training set while the test set will remain unchanged. Using casing depth and the feature with the highest variable importance, construct a decision boundary plot. Under what conditions are a well more likely to predict detectable manganese levels based on a decision boundary plot? "],["unsupervised-machine-learning-part-1-k-means-clustering-pca.html", "Unsupervised Machine Learning Part 1: K-Means Clustering &amp; PCA Introduction to Training Module Introduction to Unsupervised Machine Learning Introduction to Example Data Identifying Clusters of Chemicals through K-Means Principal Component Analysis (PCA) Incorporating K-Means into PCA for Predictive Modeling Concluding Remarks", " Unsupervised Machine Learning Part 1: K-Means Clustering &amp; PCA This training module was developed by David M. Reif with contributions from Alexis Payton, Lauren E. Koval, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module To reiterate what has been discussed in the previous module, machine learning is a field that has great utility in environmental health sciences, often to investigate high-dimensional datasets. The two main classifications of machine learning discussed throughout the TAME Toolkit are supervised and unsupervised machine learning, though additional classifications exist. Previously, we discussed artificial intelligence and supervised machine learning in TAME 2.0 Module 5.1 Introduction to Machine Learning &amp; Artificial Intelligence, TAME 2.0 Module 5.2 Supervised Machine Learning, and TAME 2.0 Module 5.3 Supervised Machine Learning Model Interpretation. In this module, we’ll cover background information on unsupervised machine learning and then work through a scripted example of an unsupervised machine learning analysis. Introduction to Unsupervised Machine Learning Unsupervised machine learning, as opposed to supervised machine learning, involves training a model on a dataset lacking ground truths or response variables. In this regard, unsupervised approaches are often used to identify underlying patterns amongst data in a more unbiased manner. This can provide the analyst with insights into the data that may not otherwise be apparent. Unsupervised machine learning has been used for understanding differences in gene expression patterns of breast cancer patients (Jezequel et. al, 2015) and evaluating metabolomic signatures of patients with and without cystic fibrosis (Laguna et. al, 2015). Note: Unsupervised machine learning is used for exploratory purposes, and just because it can find relationships between data points, that doesn’t necessarily mean that those relationships have merit, are indicative of causal relationships, or have direct biological implications. Rather, these methods can be used to find new patterns that can also inform future studies testing direct relationships. Langs, G., Röhrich, S., Hofmanninger, J., Prayer, F., Pan, J., Herold, C., &amp; Prosch, H. (2018). Machine learning: from radiomics to discovery and routine. Der Radiologe, 58(S1), 1–6. PMID: 34013136. Figure regenerated here in alignment with its published Creative Commons Attribution 4.0 International License Unsupervised machine learning includes: Clustering: Involves grouping elements in a dataset such that the elements in the same group are more similar to each other than to the elements in the other groups. Exclusive (K-means) Overlapping Hierarchical Probabilistic Dimensionality reduction: Focuses on taking high-dimensional data and transforming it into a lower-dimensional space that has fewer features while preserving important information inherent to the original dataset. This is useful because reducing the number of features makes the data easier to visualize while trying to maintain the initial integrity of the dataset. Principal Component Analysis (PCA) Singular Value Decomposition (SVD) t-Distributed Stochastic Neighbor Embedding (t-SNE) Uniform Manifold Approximation and Projection (UMAP) Partial Least Squares-Discriminant Analysis (PLS-DA) In this module, we’ll focus on methods for K-means clustering and Principal Component Analysis described in more detail in the following sections. In the next module, TAME 2.0 Module 5.5 Unsupervised Machine Learning II: Hierarchical Clustering, we’ll focus on hierarchical clustering. For further information on types of unsupervised machine learning, check out Unsupervised Learning. K-Means Clustering K-means is a common clustering algorithm used to partition quantitative data. This algorithm works by first randomly selecting a pre-specified number of clusters, k, across the data space with each cluster having a data centroid. When using a standard Euclidean distance metric, the distance is calculated from an observation to each centroid, then the observation is assigned to the cluster of the closest centroid. After all observations have been assigned to one of the k clusters, the average of all observations in a cluster is calculated, and the centroid for the cluster is moved to the location of the mean. The process then repeats, with the distance computed between the observations and the updated centroids. Observations may be reassigned to the same cluster or moved to a different cluster if it is closer to another centroid. These iterations continue until there are no longer changes between cluster assignments for observations, resulting in the final cluster assignments that are then carried forward for analysis/interpretation. Helpful resources on k-means clustering include the following: The Elements of Statistical Learning &amp; Towards Data Science. Principal Component Analysis (PCA) Principal Component Analysis, or PCA, is a dimensionality-reduction technique used to transform high-dimensional data into a lower dimensional space while trying to preserve as much of the variability in the original data as possible. PCA has strong foundations in linear algebra, so background knowledge of eigenvalues and eigenvectors is extremely useful. Though the mathematics of PCA is beyond the scope of this module, a variety of more in-depth resources on PCA exist including this Towards Data Science Blog, and this Sartorius Blog. At a higher level, important concepts in PCA include: PCA partitions variance in a dataset into linearly uncorrelated principal components (PCs), which are weighted combinations of the original features. Each PC (starting from the first one) summarizes a decreasing percentage of variance. Every instance (e.g. chemical) in the original dataset has a “weight” or score” on each PC. Any combination of PCs can be compared to summarize relationships amongst the instances (e.g. chemicals), but typically it’s the first two eigenvectors that capture a majority of the variance. Introduction to Example Data In this activity, we are going to analyze an example dataset of physicochemical property information for chemicals spanning per- and polyfluoroalkyl substances (PFAS) and statins. PFAS represent a ubiquitous and pervasive class of man-made industrial chemicals that are commonly used in food packaging, commercial household products such as Teflon, cleaning products, and flame retardants. PFAS are recognized as highly stable compounds that, upon entering the environment, can persist for many years and act as harmful sources of exposure. Statins represent a class of lipid-lowering compounds that are commonly used as pharmaceutical treatments for patients at risk of cardiovascular disease. Because of their common use amongst patients, statins can also end up in water and wastewater effluent, making them environmentally relevant as well. This example analysis was designed to evaluate the chemical space of these diverse compounds and to illustrate the utility of unsupervised machine learning methods to differentiate chemical class and make associations between chemical groupings that can inform a variety of environmental and toxicological applications. The two types of machine learning methods that will be employed are k-means and PCA (as described in the introduction). Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: Can we differentiate between PFAS and statin chemical classes when considering just the raw physicochemical property variables without applying unsupervised machine learning techniques? If substances are able to be clustered, what are some of the physicochemical properties that seem to be driving chemical clustering patterns derived through k-means? How do the data compare when physicochemical properties are reduced using PCA? Upon reducing the data through PCA, which physicochemical property contributes the most towards informing data variance captured in the primary principal component? If we did not have information telling us which chemical belonged to which class, could we use PCA and k-means to inform whether a chemical is more similar to a PFAS or a statin? What kinds of applications/endpoints can be better understood and/or predicted because of these derived chemical groupings? Script Preparations Cleaning the global environment rm(list=ls()) Installing required R packages If you already have these packages installed, you can skip this step, or you can run the below code which checks installation status for you if (!requireNamespace(&quot;factoextra&quot;)) install.packages(&quot;factoextra&quot;); if (!requireNamespace(&quot;pheatmap&quot;)) install.packages(&quot;pheatmap&quot;); if (!requireNamespace(&quot;cowplot&quot;)) install.packages(&quot;cowplot&quot;); Loading required R packages library(tidyverse) library(factoextra) library(pheatmap) #used to make heatmaps library(cowplot) Getting help with packages and functions ?tidyverse # Package documentation for tidyverse ?kmeans # Package documentation for kmeans (a part of the standard stats R package, automatically uploaded) ?prcomp # Package documentation for deriving principal components within a PCA (a part of the standard stats R package, automatically uploaded) ?pheatmap # Package documentation for pheatmap Set your working directory setwd(&quot;/filepath to where your input files are&quot;) Loading the Example Dataset Let’s start by loading the datasets needed for this training module. We are going to use a dataset of substances that have a diverse chemical space of PFAS and statin compounds. This list of chemicals will be uploaded alongside physicochemical property data. The chemical lists for ‘PFAS’ and ‘Statins’ were obtained from the EPA’s Computational Toxicology Dashboard Chemical Lists. The physicochemical properties were obtained by uploading these lists into the National Toxicology Program’s Integrated Chemical Environment (ICE). dat &lt;- read.csv(&quot;Module5_4_Input/Module5_4_InputData.csv&quot;, fileEncoding = &quot;UTF-8-BOM&quot;) Data Viewing Starting with the overall dimensions: dim(dat) ## [1] 144 14 Then looking at the first four rows and five columns of data: dat[1:4,1:5] ## List Substance.Name CASRN ## 1 PFAS Perfluoro-2-(trifluoromethyl)propanesulphonic acid 93762-09-5 ## 2 PFAS Potassium perfluoroheptanesulfonate 60270-55-5 ## 3 PFAS Bis(2-hydroxyethyl)ammonium perfluoroheptanesulfonate 70225-15-9 ## 4 PFAS Potassium perfluoro-p-ethylcyclohexanesulfonate 335-24-0 ## DTXSID Molecular.Weight ## 1 DTXSID90239569 300.100 ## 2 DTXSID9069392 488.212 ## 3 DTXSID60880946 555.258 ## 4 DTXSID50880117 500.223 Note that the first column, List, designates the following two larger chemical classes: unique(dat$List) ## [1] &quot;PFAS&quot; &quot;Statins&quot; Let’s lastly view all of the column headers: colnames(dat) ## [1] &quot;List&quot; ## [2] &quot;Substance.Name&quot; ## [3] &quot;CASRN&quot; ## [4] &quot;DTXSID&quot; ## [5] &quot;Molecular.Weight&quot; ## [6] &quot;OPERA..Boiling.Point&quot; ## [7] &quot;OPERA..Henry.s.Law.Constant&quot; ## [8] &quot;OPERA..Melting.Point&quot; ## [9] &quot;OPERA..Negative.Log.of.Acid.Dissociation.Constant&quot; ## [10] &quot;OPERA..Octanol.Air.Partition.Coefficient&quot; ## [11] &quot;OPERA..Octanol.Water.Distribution.Coefficient&quot; ## [12] &quot;OPERA..Octanol.Water.Partition.Coefficient&quot; ## [13] &quot;OPERA..Vapor.Pressure&quot; ## [14] &quot;OPERA..Water.Solubility&quot; In the data file, the first four columns represent chemical identifier information. All remaining columns represent different physicochemical properties derived from OPERA via Integrated Chemical Environment (ICE). Because the original titles of these physicochemical properties contained commas and spaces, R automatically converted these into periods. Hence, titles like OPERA..Boiling.Point. For ease of downstream data analyses, let’s create a more focused dataframe option containing only one chemical identifier (CASRN) as row names and then just the physicochemical property columns. # Creating a new dataframe that contains the physiocochemical properties chemical_prop_df &lt;- dat[,5:ncol(dat)] rownames(chemical_prop_df) &lt;- dat$CASRN Now explore this data subset: dim(chemical_prop_df) # overall dimensions ## [1] 144 10 chemical_prop_df[1:4,1:5] # viewing the first four rows and five columns ## Molecular.Weight OPERA..Boiling.Point OPERA..Henry.s.Law.Constant ## 93762-09-5 300.100 213.095 -3.60 ## 60270-55-5 488.212 223.097 -9.75 ## 70225-15-9 555.258 223.097 -9.75 ## 335-24-0 500.223 220.578 -7.56 ## OPERA..Melting.Point ## 93762-09-5 96.455 ## 60270-55-5 273.228 ## 70225-15-9 182.152 ## 335-24-0 231.827 ## OPERA..Negative.Log.of.Acid.Dissociation.Constant ## 93762-09-5 0.175 ## 60270-55-5 -1.810 ## 70225-15-9 1.000 ## 335-24-0 1.000 colnames(chemical_prop_df) ## [1] &quot;Molecular.Weight&quot; ## [2] &quot;OPERA..Boiling.Point&quot; ## [3] &quot;OPERA..Henry.s.Law.Constant&quot; ## [4] &quot;OPERA..Melting.Point&quot; ## [5] &quot;OPERA..Negative.Log.of.Acid.Dissociation.Constant&quot; ## [6] &quot;OPERA..Octanol.Air.Partition.Coefficient&quot; ## [7] &quot;OPERA..Octanol.Water.Distribution.Coefficient&quot; ## [8] &quot;OPERA..Octanol.Water.Partition.Coefficient&quot; ## [9] &quot;OPERA..Vapor.Pressure&quot; ## [10] &quot;OPERA..Water.Solubility&quot; Evaluating the Original Physicochemical Properties across Substances Let’s first plot two physicochemical properties to determine if and how substances group together without any fancy data reduction or other machine learning techniques. This will answer Environmental Health Question #1: Can we differentiate between PFAS and statin chemical classes when considering just the raw physicochemical property variables without applying unsupervised machine learning techniques? Let’s put molecular weight (Molecular.Weight) as one axis and boiling point (OPERA..Boiling.Point) on the other. We’ll also color by the chemical classes using the List column from the original dataframe. ggplot(chemical_prop_df[,1:2], aes(x = Molecular.Weight, y = OPERA..Boiling.Point, color = dat$List)) + geom_point(size = 2) + theme_bw() + ggtitle(&#39;Version A: Bivariate Plot of Two Original Physchem Variables&#39;) + xlab(&quot;Molecular Weight&quot;) + ylab(&quot;Boiling Point&quot;) Let’s plot two other physicochemical property variables, Henry’s Law constant (OPERA..Henry.s.Law.Constant) and melting point (OPERA..Melting.Point), to see if the same separation of chemical classes is apparent. ggplot(chemical_prop_df[,3:4], aes(x = OPERA..Henry.s.Law.Constant, y = OPERA..Melting.Point, color = dat$List)) + geom_point(size = 2) + theme_bw() + ggtitle(&#39;Version B: Bivariate Plot of Two Other Original Physchem Variables&#39;) + xlab(&quot;OPERA..Henry.s.Law.Constant&quot;) + ylab(&quot;OPERA..Melting.Point&quot;) Answer to Environmental Health Question 1 With these, we can answer Environmental Health Question #1: Can we differentiate between PFAS and statin chemical classes when considering just the raw physicochemical property variables without applying machine learning techniques? Answer: Only in part. From the first plot, we can see that PFAS tend to have lower molecular weight ranges in comparison to the statins, though other property variables clearly overlap in ranges of values making the groupings not entirely clear. Identifying Clusters of Chemicals through K-Means Let’s turn our attention to Environmental Health Question #2: If substances are able to be clustered, what are some of the physicochemical properties that seem to be driving chemical clustering patterns derived through k-means? This will be done deriving clusters of chemicals based on ALL underlying physicochemical property data using k-means clustering. For this example, let’s coerce the k-means algorithms to calculate 2 distinct clusters (based on their corresponding mean centered values). Here, we choose to derive two distinct clusters, because we are ultimately going to see if we can use this information to predict each chemical’s classification into two distinct chemical classes (i.e., PFAS vs statins). Note that we can derive more clusters using similar code depending on the question being addressed. We can give a name to this variable to easily provide the number of clusters in the next lines of code, num.centers: num.centers &lt;- 2 Here we derive chemical clusters using k-means: clusters &lt;- kmeans(chemical_prop_df, # input dataframe centers = num.centers, # number of cluster centers to calculate iter.max = 1000, # the maximum number of iterations allowed nstart = 50) # the number of rows used as the random set for the initial centers (during the first iteration) The resulting property values that were derived as the final cluster centers can be pulled using: clusters$centers ## Molecular.Weight OPERA..Boiling.Point OPERA..Henry.s.Law.Constant ## 1 690.1443 233.0402 -9.589444 ## 2 395.0716 281.4445 -8.655185 ## OPERA..Melting.Point OPERA..Negative.Log.of.Acid.Dissociation.Constant ## 1 183.7980 0.01658333 ## 2 157.5036 1.33226852 ## OPERA..Octanol.Air.Partition.Coefficient ## 1 5.940861 ## 2 6.629556 ## OPERA..Octanol.Water.Distribution.Coefficient ## 1 -2.541750 ## 2 -1.271315 ## OPERA..Octanol.Water.Partition.Coefficient OPERA..Vapor.Pressure ## 1 4.000639 -5.538889 ## 2 3.010302 -6.762009 ## OPERA..Water.Solubility ## 1 -3.760222 ## 2 -3.450750 Let’s add the cluster assignments to the physicochemical data and create a new dataframe, which can then be used in a heatmap visualization to see how these physicochemical data distributions clustered according to k-means. These cluster assignments can be pulled from the cluster list output, where chemicals are designated to each cluster with either a 1 or 2. You can view these using: clusters$cluster ## 93762-09-5 60270-55-5 70225-15-9 335-24-0 647-29-0 68259-12-1 ## 2 2 1 2 2 1 ## 68259-09-6 68259-07-4 60453-92-1 357-31-3 441296-91-9 749786-16-1 ## 2 2 2 2 1 1 ## 93762-10-8 135524-36-6 93894-55-4 34642-43-8 2706-91-4 791563-89-8 ## 2 2 1 2 2 1 ## 742-73-4 29420-49-3 3871-99-6 29359-39-5 3872-25-1 126105-34-8 ## 2 2 2 1 2 1 ## 630402-22-1 2274731-07-4 98789-57-2 85963-79-7 375-73-5 108427-53-8 ## 2 1 1 2 2 2 ## 4021-47-0 117806-54-9 67906-42-7 68555-66-8 92982-03-1 375-92-8 ## 2 2 1 2 1 2 ## 175905-36-9 102061-82-5 134615-58-0 174675-49-1 79780-39-5 91036-71-4 ## 2 2 1 1 1 1 ## 70225-17-1 6401-03-2 374-58-3 646-83-3 86525-30-6 3916-24-3 ## 2 2 2 2 1 2 ## 42409-05-2 474511-07-4 2795-39-3 45187-15-3 82382-12-5 79963-95-4 ## 2 1 1 2 2 2 ## 45298-90-6 134615-57-9 927670-12-0 2806-15-7 70225-14-8 131651-65-5 ## 2 2 2 1 1 2 ## 343629-46-9 144797-51-3 29081-56-9 80988-54-1 1379460-39-5 343629-43-6 ## 1 2 2 2 1 1 ## 146689-46-5 29457-72-5 355-46-4 3107-18-4 70259-86-8 1036375-28-6 ## 2 2 2 2 2 2 ## 70225-18-2 70225-16-0 84224-48-6 507453-86-3 40365-28-4 110676-15-8 ## 2 2 2 1 1 2 ## 70259-85-7 2106-55-0 1997344-07-6 423-41-6 115416-68-7 17202-41-4 ## 2 2 2 2 2 1 ## 93894-73-6 134615-56-8 134615-59-1 68259-08-5 68259-10-9 374-62-9 ## 1 2 1 2 2 2 ## 68555-67-9 2806-16-8 36913-91-4 85187-17-3 803688-15-5 55120-77-9 ## 2 1 1 1 2 2 ## 335-77-3 141263-54-9 95465-60-4 130200-44-1 144535-22-8 130468-11-0 ## 1 2 2 2 2 2 ## 93957-54-1 126059-69-6 153463-20-8 154417-69-3 147511-69-1 141263-69-6 ## 2 2 2 2 2 2 ## 77517-29-4 80799-31-1 73390-02-0 503-49-1 117678-63-4 145599-86-6 ## 2 2 2 2 2 2 ## 147098-20-2 85798-96-5 120551-59-9 13552-81-3 90761-31-2 79691-18-2 ## 1 2 2 1 2 2 ## 73573-88-3 114801-27-3 151106-12-6 129443-92-1 134523-03-8 122254-45-9 ## 2 2 2 2 1 2 ## 75330-75-5 137023-81-5 136320-61-1 87770-13-6 85551-06-0 144501-27-9 ## 2 2 2 2 2 2 ## 159014-70-7 153321-50-7 133983-25-2 78366-44-6 148750-02-1 79902-63-9 ## 2 2 2 2 2 2 ## 120185-34-4 120171-12-2 141267-47-2 94061-80-0 141240-46-2 81093-37-0 ## 2 2 2 2 2 2 Because these results are listed in the exact same order as the inputted dataframe, we can simply add these assignments to the chemical_prop_df dataframe. dat_wclusters &lt;- cbind(chemical_prop_df,clusters$cluster) colnames(dat_wclusters)[11] &lt;- &quot;Cluster&quot; # renaming this new column &quot;Custer&quot; dat_wclusters &lt;- dat_wclusters[order(dat_wclusters$Cluster),] # sorting data by cluster assignments To generate a heatmap, we need to first create a separate dataframe for the cluster assignments, ordered in the same way as the physicochemical data: hm_cluster &lt;- data.frame(dat_wclusters$Cluster, row.names = row.names(dat_wclusters)) # creating the dataframe colnames(hm_cluster) &lt;- &quot;Cluster&quot; # reassigning the column name hm_cluster$Cluster &lt;- as.factor(hm_cluster$Cluster) # coercing the cluster numbers into factor variables, to make the heatmap prettier head(hm_cluster) # viewing this new cluster assignment dataframe ## Cluster ## 70225-15-9 1 ## 68259-12-1 1 ## 441296-91-9 1 ## 749786-16-1 1 ## 93894-55-4 1 ## 791563-89-8 1 We’re going to go ahead and clean up the physiocochemical property names to make the heatmap a bit tidier. clean_names1 = gsub(&quot;OPERA..&quot;, &quot;&quot;, colnames(dat_wclusters)) # &quot;\\\\.&quot; denotes a period clean_names2 = gsub(&quot;\\\\.&quot;, &quot; &quot;, clean_names1) # Reassigning the cleaner names back to the df colnames(dat_wclusters) = clean_names2 # Going back to add in the apostrophe in &quot;Henry&#39;s Law Constant&quot; colnames(dat_wclusters)[3] = &quot;Henry&#39;s Law Constant&quot; Then we can call this dataframe (data_wclusters) into the following heatmap visualization code leveraging the pheatmap() function. This function was designed specifically to enable clustered heatmap visualizations. Check out pheatmap Documenation for additional information. Heatmap Visualization of the Resulting K-Means Clusters pheatmap(dat_wclusters[,1:10], cluster_rows = FALSE, cluster_cols = FALSE, # no further clustering, for simplicity scale = &quot;column&quot;, # scaling the data to make differences across chemicals more apparent annotation_row = hm_cluster, # calling the cluster assignment dataframe as a separate color bar annotation_names_row = FALSE, # adding removing the annotation name (&quot;Cluster&quot;) from the x axis angle_col = 45, fontsize_col = 7, fontsize_row = 3, # adjusting size/ orientation of axes labels cellheight = 3, cellwidth = 25, # setting height and width for cells border_color = FALSE # specify no border surrounding the cells ) An appropriate title for this figure could be: “Figure X. Heatmap of physicochemical properties with k-means cluster assignments. Shown are the relative values for each physicochemical property labeled on the x axis. Individual chemical names are listed on the y axis. The chemicals are grouped based on their k-means cluster assignment as denoted by the color bar on the left.” Notice that the pheatmap() function does not add axes or legend titles. Adding those can provide clarity, however those can be added to the figure after exporting from R in MS Powerpoint or Adobe. Answer to Environmental Health Question 2 With this, we can answer Environmental Health Question #2: What are some of the physicochemical properties that seem to be driving chemical clustering patterns derived through k-means? Answer: Properties with values that show obvious differences between resulting clusters including molecular weight, boiling point, negative log of acid dissociation constant, octanol air partition coefficient, and octanol water distribution coefficient. Principal Component Analysis (PCA) Next, we will run through some example analyses applying the common data reduction technique of PCA. We’ll start by determining how much of the variance is able to be captured within the first two principal components to answer Environmental Health Question #3: How do the data compare when physicochemical properties are reduced using PCA? We can calculate the principal components across ALL physicochemical data across all chemicals using the prcomp() function. Always make sure your data is centered and scaled prior to running to PCA, since it’s sensitive to variables having different scales. my.pca &lt;- prcomp(chemical_prop_df, # input dataframe of physchem data scale = TRUE, center = TRUE) We can see how much of the variance was able to be captured in each of the eigenvectors or dimensions using a scree plot. fviz_eig(my.pca, addlabels = TRUE) We can also calculate these values and pull them into a dataframe for future use. For example, to pull the percentage of variance explained by each principal component, we can run the following calculations, where first eigenvalues (eigs) are calculated and then used to calculate percent of variance per principal component: eigs &lt;- my.pca$sdev^2 Comp.stats &lt;- data.frame(eigs, eigs/sum(eigs), row.names = names(eigs)) colnames(Comp.stats) &lt;- c(&quot;Eigen_Values&quot;, &quot;Percent_of_Variance&quot;) head(Comp.stats) ## Eigen_Values Percent_of_Variance ## 1 4.0925925 0.40925925 ## 2 2.3840022 0.23840022 ## 3 1.5415308 0.15415308 ## 4 0.5873057 0.05873057 ## 5 0.4769190 0.04769190 ## 6 0.3659181 0.03659181 Answer to Environmental Health Question 3 With this, we can answer Environmental Health Question #3: How do the data compare when physicochemical properties are reduced using PCA? Answer: Principal Component 1 captures ~41% of the variance and Principal Component 2 captures ~24% across all physicochemical property values across all chemicals. These two components together describe ~65% of data. Next, we’ll use PCA to answer Environmental Health Question #4: Upon reducing the data through PCA, which physicochemical property contributes the most towards informing data variance captured in the primary principal component (Comp.1)? Here are the resulting scores for each chemical’s contribution towards each principal component (shown here as components PC1-PC10). head(my.pca$x) ## PC1 PC2 PC3 PC4 PC5 ## 93762-09-5 -2.0354310 -1.4824239 1.29328367 0.04865893 0.25305471 ## 60270-55-5 -1.2249015 2.2857092 -0.24261251 -0.40140814 0.62781108 ## 70225-15-9 -1.0944361 1.3915069 -0.03340358 -0.90391749 0.05735985 ## 335-24-0 -1.1334897 1.0675553 0.14299978 -1.08713268 0.21172964 ## 647-29-0 -0.4830620 0.1259827 -1.16147937 1.11383217 -0.29571642 ## 68259-12-1 -0.3264762 0.2369031 -1.31984895 0.47512052 -1.17555774 ## PC6 PC7 PC8 PC9 PC10 ## 93762-09-5 -0.67755281 -0.14546495 1.25520979 0.230936852 0.14075496 ## 60270-55-5 -1.01515519 -0.11316504 -0.34335060 -0.383682123 0.09849160 ## 70225-15-9 0.14334979 0.38355763 0.01717491 -0.006702116 -0.02715723 ## 335-24-0 -0.93947357 -0.22877517 0.11457280 -0.107720120 0.17701035 ## 647-29-0 0.19174355 -0.18228398 -0.48013543 0.074967840 0.22750496 ## 68259-12-1 0.05910151 0.01399123 0.03790455 0.043309249 -0.18032083 And the resulting loading factors of each property’s contribution towards each principal component. my.pca$rotation ## PC1 PC2 ## Molecular.Weight 0.09825313 0.108454961 ## OPERA..Boiling.Point 0.46350428 0.029650863 ## OPERA..Henry.s.Law.Constant -0.17856542 -0.502116638 ## OPERA..Melting.Point 0.20645719 0.474473487 ## OPERA..Negative.Log.of.Acid.Dissociation.Constant 0.32172963 -0.119465105 ## OPERA..Octanol.Air.Partition.Coefficient 0.45329804 -0.008918089 ## OPERA..Octanol.Water.Distribution.Coefficient 0.33026147 -0.437053508 ## OPERA..Octanol.Water.Partition.Coefficient 0.16152270 -0.342905461 ## OPERA..Vapor.Pressure -0.35205065 -0.349618523 ## OPERA..Water.Solubility -0.36456684 0.254809066 ## PC3 PC4 ## Molecular.Weight -0.679740448 -0.48432419 ## OPERA..Boiling.Point 0.199365885 0.03108544 ## OPERA..Henry.s.Law.Constant 0.179876660 -0.27695374 ## OPERA..Melting.Point -0.214857945 -0.09449999 ## OPERA..Negative.Log.of.Acid.Dissociation.Constant 0.286239544 -0.58268278 ## OPERA..Octanol.Air.Partition.Coefficient 0.132157694 0.04820475 ## OPERA..Octanol.Water.Distribution.Coefficient -0.005904541 0.15115855 ## OPERA..Octanol.Water.Partition.Coefficient -0.466833374 0.48460961 ## OPERA..Vapor.Pressure -0.195008167 -0.24988717 ## OPERA..Water.Solubility 0.253557563 0.12980079 ## PC5 PC6 ## Molecular.Weight 0.17351578 0.357367954 ## OPERA..Boiling.Point 0.22224554 -0.018507534 ## OPERA..Henry.s.Law.Constant 0.30566003 -0.470666686 ## OPERA..Melting.Point -0.08063905 -0.686723561 ## OPERA..Negative.Log.of.Acid.Dissociation.Constant -0.66998767 0.029248038 ## OPERA..Octanol.Air.Partition.Coefficient 0.20778895 0.205757894 ## OPERA..Octanol.Water.Distribution.Coefficient 0.06378821 0.169005123 ## OPERA..Octanol.Water.Partition.Coefficient -0.48515616 -0.162279640 ## OPERA..Vapor.Pressure -0.06782464 -0.007031883 ## OPERA..Water.Solubility -0.29062543 0.284060941 ## PC7 PC8 ## Molecular.Weight -0.11763362 0.32938640 ## OPERA..Boiling.Point -0.12503355 0.09718690 ## OPERA..Henry.s.Law.Constant -0.21138163 0.44526650 ## OPERA..Melting.Point -0.34342931 -0.10233816 ## OPERA..Negative.Log.of.Acid.Dissociation.Constant 0.09083446 0.03113686 ## OPERA..Octanol.Air.Partition.Coefficient -0.44434707 -0.29734602 ## OPERA..Octanol.Water.Distribution.Coefficient -0.22039378 0.05052653 ## OPERA..Octanol.Water.Partition.Coefficient -0.08082351 0.17794120 ## OPERA..Vapor.Pressure -0.34958097 -0.65367959 ## OPERA..Water.Solubility -0.65183980 0.34989539 ## PC9 PC10 ## Molecular.Weight 0.03295675 -0.02698233 ## OPERA..Boiling.Point 0.03336277 -0.81709497 ## OPERA..Henry.s.Law.Constant 0.19706729 0.10099077 ## OPERA..Melting.Point -0.24532148 0.10229774 ## OPERA..Negative.Log.of.Acid.Dissociation.Constant 0.02576652 0.03380215 ## OPERA..Octanol.Air.Partition.Coefficient 0.49672303 0.39565984 ## OPERA..Octanol.Water.Distribution.Coefficient -0.74417871 0.19866516 ## OPERA..Octanol.Water.Partition.Coefficient 0.30648450 -0.09852386 ## OPERA..Vapor.Pressure -0.03202641 -0.31333091 ## OPERA..Water.Solubility -0.05123040 -0.07696401 Answer to Environmental Health Question 4 With these results, we can answer Environmental Health Question #4: Upon reducing the data through PCA, which physicochemical property contributes the most towards informing data variance captured in the primary principal component (Comp.1)? Answer: Boiling point contributes the most towards principal component 1, as it has the largest magnitude (0.464). Visualizing PCA Results Let’s turn our attention to Environmental Health Question #5: If we did not have information telling us which chemical belonged to which class, could we use PCA and k-means to inform whether a chemical is more similar to a PFAS or a statin? We can start by answering this question by visualizing the first two principal components and coloring each chemical according to class (i.e. PFAS vs statins). ggplot(data.frame(my.pca$x), aes(x = PC1, y = PC2, color = dat$List)) + geom_point(size = 2) + theme_bw() + ggtitle(&#39;Version C: PCA Plot of the First 2 PCs, colored by Chemical Class&#39;) + # it&#39;s good practice to put the percentage of the variance captured in the axes titles xlab(&quot;Principal Component 1 (40.9%)&quot;) + ylab(&quot;Principal Component 2 (23.8%)&quot;) Answer to Environmental Health Question 5 With this, we can answer Environmental Health Question #5: If we did not have information telling us which chemical belonged to which class, could we use PCA and k-means to inform whether a chemical is more similar to a PFAS or a statin? Answer: Data become more compressed and variables reduce across principal components capturing the majority of the variance from the original dataset (~65%). This results in improved data visualizations, where all dimensions of the physiochemical dataset are compressed and captured across the displayed components. In addition, the figure above shows a clear separation between PFAS and statin chemical when visualizing the reduced dataset. Incorporating K-Means into PCA for Predictive Modeling We can also identify cluster-based trends within data that are reduced after running PCA. This example analysis does so, expanding upon the previously generated PCA results. Estimate K-Means Clusters from PCA Results Let’s first run code similar to the previous k-means analysis and associated parameters, though instead here we will use data reduced values from the PCA analysis. Specifically, clusters across PCA “scores” values will be derived, where scores represent the relative amount each chemical contributed to each principal component. clusters_PCA &lt;- kmeans(my.pca$x, centers = num.centers, iter.max = 1000, nstart = 50) The resulting PCA score values that were derived as the final cluster centers can be pulled using: clusters_PCA$centers ## PC1 PC2 PC3 PC4 PC5 PC6 ## 1 2.612553 -0.3054641 0.6219575 -0.008818739 0.18490367 0.05041185 ## 2 -1.187524 0.1388473 -0.2827080 0.004008518 -0.08404712 -0.02291448 ## PC7 PC8 PC9 PC10 ## 1 -0.03417184 -0.05270363 -0.018330999 -0.011904302 ## 2 0.01553265 0.02395620 0.008332272 0.005411046 Viewing the final cluster assignment per chemical: head(cbind(rownames(chemical_prop_df),clusters_PCA$cluster)) ## [,1] [,2] ## 93762-09-5 &quot;93762-09-5&quot; &quot;2&quot; ## 60270-55-5 &quot;60270-55-5&quot; &quot;2&quot; ## 70225-15-9 &quot;70225-15-9&quot; &quot;2&quot; ## 335-24-0 &quot;335-24-0&quot; &quot;2&quot; ## 647-29-0 &quot;647-29-0&quot; &quot;2&quot; ## 68259-12-1 &quot;68259-12-1&quot; &quot;2&quot; Visualizing K-Means Clusters from PCA Results Let’s now view, again, the results of the main PCA focusing on the first two principal components; though this time let’s color each chemical according to k-means cluster. ggplot(data.frame(my.pca$x), aes(x = PC1, y = PC2, color = as.factor(clusters_PCA$cluster))) + geom_point(size = 2) + theme_bw() + ggtitle(&#39;Version D: PCA Plot of the First 2 PCs, colored by k-means Clustering&#39;) + # it&#39;s good practice to put the percentage of the variance capture in the axes titles xlab(&quot;Principal Component 1 (40.9%)&quot;) + ylab(&quot;Principal Component 2 (23.8%)&quot;) Let’s put these two PCA plots side by side to compare them more easily. We’ll also tidy up the figures a bit so they’re closer to publication-ready. # PCA plot colored by chemical class pcaplot1 = ggplot(data.frame(my.pca$x), aes(x = PC1, y = PC2, color = dat$List)) + geom_point(size = 2) + theme_light() + theme(axis.text = element_text(size = 9), # changing size of axis labels axis.title = element_text(face = &quot;bold&quot;, size = rel(1.3)), # changes axis titles legend.title = element_text(face = &#39;bold&#39;, size = 10), # changes legend title legend.text = element_text(size = 9)) + # changes legend text labs(x = &#39;Principal Component 1 (40.9%)&#39;, y = &#39;Principal Component 2 (23.8%)&#39;, color = &quot;Chemical Class&quot;) # changing axis labels # PCA Plot by k means clusters pcaplot2 = ggplot(data.frame(my.pca$x), aes(x = PC1, y = PC2, color = as.factor(clusters_PCA$cluster))) + geom_point(size = 2) + theme_light() + theme(axis.text = element_text(size = 9), # changing size of axis labels axis.title = element_text(face = &quot;bold&quot;, size = rel(1.3)), # changes axis titles legend.text = element_text(size = 9)) + # changes legend text labs(x = &#39;Principal Component 1 (40.9%)&#39;, y = &#39;Principal Component 2 (23.8%)&#39;, color = expression(bold(bolditalic(K)-Means~Cluster))) # changing axis labels # Creating 1 figure plot_grid(pcaplot1, pcaplot2, # Adding labels, changing size their size and position labels = &quot;AUTO&quot;, label_size = 15, label_x = 0.03) An appropriate title for this figure could be: “Figure X. Principal Component Analysis (PCA) plots highlight similarities between chemical class and k-means clusters. These PCA plots are based on physiochemical properties and compare (A) chemical class categories and the (B) K-means derived cluster assignments.” Answer to Environmental Health Question 6 With this we can answer Environmental Health Question #6: What kinds of applications/endpoints can be better understood and/or predicted because of these derived chemical groupings? Answer: With these well-informed chemical groupings, we can now better understand the variables that attribute to the chemical classifications. We can also use this information to better understand data trends and predict environmental fate and transport for these chemicals. The reduced variables derived through PCA, and/or k-means clustering patterns can also be used as input variables to predict toxicological outcomes. Concluding Remarks In conclusion, this training module provides an example exercise on organizing physicochemical data and analyzing trends within these data to determine chemical groupings. Results are compared from those produced using just the original data vs. clustered data from k-means vs. reduced data from PCA. These methods represent common tools that are used in high dimensional data analyses within the field of environmental health sciences. Additional Resources Detailed study of Principal Component Analysis Practical Guide to Cluster Analysis in R Test Your Knowledge In this training module, we presented an unsupervised machine learning example that was based on defining k-means clusters based on chemical class where k = 2. Often times, analyses are conducted to explore potential clustering relationships without a preexisting idea of what k or the number of clusters should be. In this test your knowledge section, we’ll go through an example like that. Using the accompanying flame retardant and pesticide physicochemical property variables found in the file (“Module5_4_TYKInput.csv”), answer the following questions: What are some of the physicochemical properties that seem to be driving chemical clustering patterns derived through k-means? Upon reducing the data through PCA, which physicochemical property contributes the most towards informing data variance captured in the primary principal component? If we did not have information telling us which chemical belonged to which class, could we use PCA and k-means to accurately predict whether a chemical is a PFAS or a statin? "],["unsupervised-machine-learning-part-2-additional-clustering-applications.html", "Unsupervised Machine Learning Part 2: Additional Clustering Applications Introduction to Training Module K-Means Clustering Hierarchical Clustering Introduction to Example Data Hierarchical Clustering Clustering Plot Hierarchical Clustering Visualization Variable Contributions Concluding Remarks", " Unsupervised Machine Learning Part 2: Additional Clustering Applications This training module was developed by Alexis Payton, Lauren E. Koval, David M. Reif, and Julia E. Rager. All input files (script, data, and figures) can be downloaded from the UNC-SRP TAME2 GitHub website. Introduction to Training Module The previous module TAME 2.0 Module 5.4 Unsupervised Machine Learning, served as introduction to unsupervised machine learning (ML). Unsupervised ML involves training a model on a dataset lacking ground truths or response variables. However, in the previous module, the number of clusters was selected based on prior information (i.e., chemical class), but what if you’re in a situation where you don’t know how many clusters to investigate a priori? This commonly occurs, particularly in the field of environmental health research in instances when investigators want to take a more unbiased view of their data and/or do not have information that can be used to inform the optimal number of clusters to select. In these instances,unsupervised ML techniques can be very helpful, and in this module, we’ll explore the following concepts to further understand unsupervised ML: K-Means and hierarchical clustering Deriving the optimal number of clusters Visualizing clusters through a PCA-based plot, dendrograms, and heatmaps Determining each variable’s contribution to the clusters K-Means Clustering As mentioned in the previous module, K-means is a common clustering algorithm used to partition quantitative data. This algorithm works by first, randomly selecting a pre-specified number of clusters, k, across the data space, with each cluster having a data centroid. When using a standard Euclidean distance metric, the distance is calculated from an observation to each centroid, then the observation is assigned to the cluster of the closest centroid. After all observations have been assigned to one of the k clusters, the average of all observations in a cluster is calculated, and the centroid for the cluster is moved to the location of the mean. The process then repeats, with the distance computed between the observations and the updated centroids. Observations may be reassigned to the same cluster or moved to a different cluster if it is closer to another centroid. These iterations continue until there are no longer changes between cluster assignments for observations, resulting in the final cluster assignments that are then carried forward for analysis/interpretation. Helpful resources on k-means clustering include the following: The Elements of Statistical Learning &amp; Towards Data Science. Hierarchical Clustering Hierarchical clustering groups objects into clusters by repetitively joining similar observations until there is one large cluster (aka agglomerative or bottom-up) or repetitively splitting one large cluster until each observation stands alone (aka divisive or top-down). Regardless of whether agglomerative or divisive hierarchical clustering is used, the results can be visually represented in a tree-like figure called a dendrogram. The dendrogram below is based on the USArrests dataset available in R. The datset contains statistics on violent crimes rates (murder, assault, and rape) per capita (per 100,000 residents) for each state in the United States in 1973. For more information on the USArrests dataset, check out its associated RDocumentation. An appropriate title for this figure could be: “Figure X. Hierarchical clustering of states based on violent crime. The dendogram uses violent crime data including murder, assault, and rape rates per 100,000 residents for each state in 1973.” Takeaways from this dendogram: The 50 states can be grouped into 4 clusters based on violent crime statistics from 1973 The dendogram can only show us clusters of states but not the data trends that led to the clustering patterns that we see. Yes, it is useful to know what states have similar violent crime patterns overall, but it is important to pinpoint the variables (ie. murder, assault, and rape) that are responsible for the clustering patterns we’re seeing. This idea will be explored later in the module with an environmentally-relevant dataset. Going back to hiearchical clustering, during the repetitive splitting or joining of observations, the similarity between existing clusters is calculated after each iteration. This value informs the formation of subsequent clusters. Different methods, or linkage functions, can be considered when calculating this similarity, particularly for agglomerative clustering which is often the preferred approach. Some example methods include: Complete Linkage: the maximum distance between two data points located in separate clusters. Single Linkage: the minimum distance between two data points located in separate clusters. Average Linkage: the average pairwise distance between all pairs of data points in separate clusters. Centroid Linkage: the distance between the centroids or centers of each cluster. Ward Linkage: seeks to minimize the variance between clusters. Each method has its advantages and disadvantages and more information on all distance calculations between clusters can be found at the following resource: Hierarchical Clustering. Deriving the Optimal Number of Clusters Before clustering can be performed, the function needs to be informed of the number of clusters to group the objects into. In the previous module, an example was explored to see if k-means clustering would group the chemicals similarly to their chemical class (either a PFAS or statin). Therefore, we told the k-means function to cluster into 2 groups. In situations where there is little to no prior knowledge regarding the “correct” number of clusters to specify, methods exist for deriving the optimal number of clusters. Three common methods to find the optimal k, or number of clusters, for both k-means and hierarchical clustering include: the elbow method, silhouette method, and the gap statistic method. These techniques help us in determining the optimal k using visual inspection. Elbow Method: uses a plot of the within cluster sum of squares (WCSS) on the y axis and different values of k on the x axis. The location where we no longer observe a significant reduction in WCSS, or where an “elbow” can be seen, is the optimal k value. As we can see, after a certain point, having more clusters does not lead to a significant reduction in WCSS. Looking at the figures above, the elbow point is much clearer in the first plot versus the second, however, elbow curves from real-world datasets typically resemble the second figure. This is why it’s recommended to consider more than one method to determine the optimal number of clusters. Silhouette Method: uses a plot of the average silhouette width (score) on the y axis and different values of k on the x axis. The silhouette score is measure of each object’s similarity to its own cluster and how dissimilar it is to other clusters. The location where the average silhouette width is maximized is the optimal k value. Based on the figure above, the optimal number of clusters is 2 using the silhouette method. Gap Statistic Method: uses a plot of the gap statistic on the y axis and different values of k on the x axis. The gap statistic evaluates the intracluster variation in comparison to expected values derived from a Monte Carlo generated, null reference data distribution for varying values of k. The optimal number of clusters is the smallest value where the gap statistic of k is greater than or equal to the gap statistic of k+1 minus the standard deviation of k+1. More details can be found here. Based on the figure above, the optimal number of clusters is 2 using the gap statistic method. For additional information and code on all three methods, check out Determining the Optimal Number of Clusters: 3 Must Know Methods. It is also worth mentioning that while these methods are useful, further interpreting the result in the context of your problem can be beneficial, for example, checking whether clusters make biological sense when working with a genomic dataset. Introduction to Example Data We will apply these techniques using an example dataset from a previously published study where 22 cytokine concentrations were derived from 44 subjects with varying smoking statuses (14 non-smokers, 17 e-cigarette users, and 13 cigarette smokers) from 4 different sampling regions in the body. These samples were derived from nasal lavage fluid (NLF), nasal epithelieum fluid (NELF), sputum, and serum as pictured below. Samples were taken from different regions in the body to compare cytokine expression in the upper respiratory tract, lower respiratory tract, and systemic circulation. A research question that we had was “Does cytokine expression change based on a subject’s smoking habits? If so, does cigarette smoke or e-cigarette vapor induce cytokine suppression or proliferation?” Traditionally these questions would have been answered by analyzing each biomarker individually using a two-group comparison test like a t test (which we completed in this study). However, biomarkers do not work in isolation in the body, suggesting that individual biomarker statistical approaches may not capture the full biological responses occurring. Therefore we used a clustering approach to group cytokines as an attempt to more closely simulate interactions that occur in vivo. From there, statistical tests were run to assess the effects of smoking status on each cluster. For the purposes of this training exercise, we will focus solely on the nasal epithelieum lining fluid, or NELF, samples. In addition, we’ll use k-means and hierarchical clustering to compare how cytokines cluster at baseline. Full methods are further described in the publication below: Payton AD, Perryman AN, Hoffman JR, Avula V, Wells H, Robinette C, Alexis NE, Jaspers I, Rager JE, Rebuli ME. Cytokine signature clusters as a tool to compare changes associated with tobacco product use in upper and lower airway samples. American Journal of Physiology-Lung Cellular and Molecular Physiology 2022 322:5, L722-L736. PMID: 35318855 Let’s read in and view the dataset we’ll be working with. Script Preparations Cleaning the global environment rm(list=ls()) Installing required R packages If you already have these packages installed, you can skip this step, or you can run the below code which checks installation status for you if (!requireNamespace(&quot;vegan&quot;)) install.packages(&quot;vegan&quot;); if (!requireNamespace(&quot;ggrepel&quot;)) install.packages(&quot;ggrepel&quot;); if (!requireNamespace(&quot;dendextend&quot;)) install.packages(&quot;dendextend&quot;); if (!requireNamespace(&quot;ggsci&quot;)) install.packages(&quot;ggsci&quot;); if (!requireNamespace(&quot;FactoMineR&quot;)) install.packages(&quot;FactoMineR&quot;); Loading required R packages library(readxl) library(factoextra) library(FactoMineR) library(tidyverse) library(vegan) library(ggrepel) library(reshape2) library(pheatmap) library(ggsci) suppressPackageStartupMessages(library(dendextend)) Set your working directory setwd(&quot;/filepath to where your input files are&quot;) Importing example dataset Then let’s read in our example dataset. As mentioned in the introduction, this example dataset contains cytokine concentrations derived from 44 subjects. Let’s import and view these data: # Reading in file cytokines_df &lt;- data.frame(read_excel(&quot;Module5_5_Input/Module5_5_InputData.xlsx&quot;, sheet = 2)) # Viewing data head(cytokines_df) ## Original_Identifier Group SubjectNo SubjectID Compartment Protein Conc ## 1 E_C_F_002 NS 1 NS_1 NELF IFNg 17.642316 ## 2 E_C_F_002 NS 1 NS_1 NELF IL10 2.873724 ## 3 E_C_F_002 NS 1 NS_1 NELF IL12p70 1.625272 ## 4 E_C_F_002 NS 1 NS_1 NELF IL13 36.117692 ## 5 E_C_F_002 NS 1 NS_1 NELF IL1b 104.409217 ## 6 E_C_F_002 NS 1 NS_1 NELF IL6 21.159536 ## Conc_pslog2 ## 1 4.220509 ## 2 1.953721 ## 3 1.392467 ## 4 5.214035 ## 5 6.719857 ## 6 4.469856 These data contain the following information: Original_Identifier: initial identifier given to each subject by our wet bench colleagues Group: denotes the smoking status of the subject (“NS” = “non-smoker”, “Ecig” = “E-cigarette user”, “CS” = “cigarette smoker”) SubjectNo: ordinal subject number assigned to each subject after the dataset was wrangled (1-44) SubjectID: unique subject identifier that combines the group and subject number Compartment: region of the body from which the sample was taken (“NLF” = “nasal lavage fluid sample”, “NELF” = “nasal epithelieum lining fluid sample”, “Sputum” = “induced sputum sample”, “Serum” = “blood serum sample”) Protein: cytokine name Conc: concentration (pg/mL) Conc_pslog2: psuedo-log2 concentration Now that the data has been read in, we can start by asking some initial questions about the data. Training Module’s Environmental Health Questions This training module was specifically developed to answer the following environmental health questions: What is the optimal number of clusters the cytokines can be grouped into that were derived from nasal epithelium fluid (NELF) in non-smokers using k-means clustering? After selecting a cluster number, which cytokines were assigned to each k-means cluster? What is the optimal number of clusters the cytokines can be grouped into that were derived from nasal epithelium fluid (NELF) in non-smokers using hierarchical clustering? How do the hierarchical cluster assignments compare to the k-means cluster assignments? Which cytokines have the greatest contributions to the first two eigenvectors? To answer the first environmental health question, let’s start by filtering to include only NELF derived samples and non-smokers. baseline_df &lt;- cytokines_df %&gt;% filter(Group == &quot;NS&quot;, Compartment == &quot;NELF&quot;) head(baseline_df) ## Original_Identifier Group SubjectNo SubjectID Compartment Protein Conc ## 1 E_C_F_002 NS 1 NS_1 NELF IFNg 17.642316 ## 2 E_C_F_002 NS 1 NS_1 NELF IL10 2.873724 ## 3 E_C_F_002 NS 1 NS_1 NELF IL12p70 1.625272 ## 4 E_C_F_002 NS 1 NS_1 NELF IL13 36.117692 ## 5 E_C_F_002 NS 1 NS_1 NELF IL1b 104.409217 ## 6 E_C_F_002 NS 1 NS_1 NELF IL6 21.159536 ## Conc_pslog2 ## 1 4.220509 ## 2 1.953721 ## 3 1.392467 ## 4 5.214035 ## 5 6.719857 ## 6 4.469856 The functions we use will require us to cast the data wider. We will accomplish this using the dcast() function from the reshape2 package. wider_baseline_df &lt;- reshape2::dcast(baseline_df, Protein ~ SubjectID, value.var = &quot;Conc_pslog2&quot;) %&gt;% column_to_rownames(&quot;Protein&quot;) head(wider_baseline_df) ## NS_1 NS_10 NS_11 NS_12 NS_13 NS_14 ## Eotaxin 7.215729 8.016193 6.264252 6.938344 5.152683 7.937892 ## Eotaxin3 4.396783 4.006488 5.289509 3.597945 3.051951 3.063025 ## Fractalkine 13.584554 13.679882 12.952220 13.464652 12.110742 12.785656 ## I309 0.000000 0.000000 0.000000 0.000000 0.000000 1.617372 ## IFNg 4.220509 1.990018 2.656691 3.439682 2.318340 3.439850 ## IL10 1.953721 1.145162 1.348623 1.801954 1.176498 1.800412 ## NS_2 NS_3 NS_4 NS_5 NS_6 NS_7 ## Eotaxin 9.773407 6.12290913 5.758602 4.19868540 6.832578 7.998467 ## Eotaxin3 9.022390 2.50161364 2.721195 2.21084320 3.820379 5.549164 ## Fractalkine 14.961872 12.71538464 12.057105 10.38711624 12.917180 14.636190 ## I309 1.103140 0.00000000 0.000000 0.00000000 0.000000 0.000000 ## IFNg 4.394705 0.00000000 2.638213 0.04830196 1.914092 3.251501 ## IL10 2.652120 0.08945805 2.020155 0.84649523 1.119272 2.007941 ## NS_8 NS_9 ## Eotaxin 9.242464 7.127746 ## Eotaxin3 13.724108 3.553097 ## Fractalkine 14.047382 13.146599 ## I309 0.000000 0.000000 ## IFNg 4.104820 2.940248 ## IL10 2.576814 1.970627 Now we can derive clusters using the fviz_nbclust() function to determine the optimal k based on suggestions from the elbow, silhouette, and gap statistic methods. We can use this code for both k-means and hierarchical clustering by changing the FUNcluster parameter. Lets start with k-means:. # Elbow method fviz_nbclust(wider_baseline_df, FUNcluster = kmeans, method = &quot;wss&quot;) + labs(subtitle = &quot;Elbow method&quot;) # Silhouette method fviz_nbclust(wider_baseline_df, FUNcluster = kmeans, method = &quot;silhouette&quot;) + labs(subtitle = &quot;Silhouette method&quot;) # Gap statistic method fviz_nbclust(wider_baseline_df, FUNcluster = kmeans, method = &quot;gap_stat&quot;) + labs(subtitle = &quot;Gap Statisitc method&quot;) The elbow method is suggesting 2 or 3 clusters, the silhouette method is suggesting 2, and the gap statistic method is suggesting 1. Since each of these methods is recommending different k values, we can go ahead and run k-means to visualize the clusters and test those different k’s. K-means clusters will be visualized using the fviz_cluster() function. # Choosing to iterate through 2 or 3 clusters using i as our iterator for (i in 2:3){ # nstart = number of random starting partitions, it&#39;s recommended for nstart &gt; 1 cluster_k &lt;- kmeans(wider_baseline_df, centers = i, nstart = 25) cluster_plot &lt;- fviz_cluster(cluster_k, data = wider_baseline_df) + ggtitle(paste0(&quot;k = &quot;, i)) print(cluster_plot) } Answer to Environmental Health Question 1 With this, we can answer Environmental Health Question #1: What is the optimal number of clusters the cytokines can be grouped into that were derived from nasal epithelium fluid (NELF) in non-smokers using k-means clustering? Answer: 2 or 3 clusters can be justified here, based on using the elbow or silhouette method or if k-means happens to group cytokines together that were implicated in similar biological pathways. In the final paper, we moved forward with 3 clusters, because it was justifiable from the methods and provided more granularity in the clusters. The final cluster assignments can easily be obtained using the kmeans() function from the stats package. cluster_kmeans_3 &lt;- kmeans(wider_baseline_df, centers = 3, nstart = 25) cluster_kmeans_df &lt;- data.frame(cluster_kmeans_3$cluster) %&gt;% rownames_to_column(&quot;Cytokine&quot;) %&gt;% rename(`K-Means Cluster` = cluster_kmeans_3.cluster) %&gt;% # Ordering the dataframe for easier comparison arrange(`K-Means Cluster`) cluster_kmeans_df ## Cytokine K-Means Cluster ## 1 Fractalkine 1 ## 2 IL8 1 ## 3 IP10 1 ## 4 MIG 1 ## 5 Eotaxin 2 ## 6 Eotaxin3 2 ## 7 IL13 2 ## 8 IL17B 2 ## 9 IL1b 2 ## 10 IL6 2 ## 11 ITAC 2 ## 12 MCP1 2 ## 13 MIP1a 2 ## 14 MIP1b 2 ## 15 TARC 2 ## 16 I309 3 ## 17 IFNg 3 ## 18 IL10 3 ## 19 IL12p70 3 ## 20 IL17D 3 ## 21 IL4 3 ## 22 TNFa 3 Answer to Environmental Health Question 2 With this, we can answer Environmental Health Question #2: After selecting a cluster number, which cytokines were assigned to each k-means cluster? Answer: After choosing the number of clusters to be 3, the cluster assignments are as follows: Hierarchical Clustering Next, we’ll turn our attention to answering environmental health questions 3 and 4: What is the optimal number of clusters the cytokines can be grouped into that were derived from nasal epithelium fluid (NELF) in non-smokers using hierarchical clustering? How do the hierarchical cluster assignments compare to the k-means cluster assignments? Just as we used the elbow method, silhouette profile, and gap statistic to determine the optimal number of clusters for k-means, we can leverage the same approaches for hierarchical by changing the FUNcluster parameter. # Elbow method fviz_nbclust(wider_baseline_df, FUNcluster = hcut, method = &quot;wss&quot;) + labs(subtitle = &quot;Elbow method&quot;) # Silhouette method fviz_nbclust(wider_baseline_df, FUNcluster = hcut, method = &quot;silhouette&quot;) + labs(subtitle = &quot;Silhouette method&quot;) # Gap statistic method fviz_nbclust(wider_baseline_df, FUNcluster = hcut, method = &quot;gap_stat&quot;) + labs(subtitle = &quot;Gap Statisitc method&quot;) We can see the results are quite similar with 2-3 clusters appearing optimal. Answer to Environmental Health Question 3 With this, we can answer Environmental Health Question #3: What is the optimal number of clusters the cytokines can be grouped into that were derived from nasal epithelium fluid (NELF) in non-smokers using hierarchical clustering? Answer: Again, 2 or 3 clusters can be justified here, but for the same reasons mentioned for the first environmental health question, we landed on 3 clusters. Now we can perform the clustering and visualize and extract the results. We’ll start by using the dist() function to calculate the euclidean distance between the clusters followed by the hclust() function to obtain the hierarchical clustering assignments. # Viewing the wider dataframe we&#39;ll be working with head(wider_baseline_df) ## NS_1 NS_10 NS_11 NS_12 NS_13 NS_14 ## Eotaxin 7.215729 8.016193 6.264252 6.938344 5.152683 7.937892 ## Eotaxin3 4.396783 4.006488 5.289509 3.597945 3.051951 3.063025 ## Fractalkine 13.584554 13.679882 12.952220 13.464652 12.110742 12.785656 ## I309 0.000000 0.000000 0.000000 0.000000 0.000000 1.617372 ## IFNg 4.220509 1.990018 2.656691 3.439682 2.318340 3.439850 ## IL10 1.953721 1.145162 1.348623 1.801954 1.176498 1.800412 ## NS_2 NS_3 NS_4 NS_5 NS_6 NS_7 ## Eotaxin 9.773407 6.12290913 5.758602 4.19868540 6.832578 7.998467 ## Eotaxin3 9.022390 2.50161364 2.721195 2.21084320 3.820379 5.549164 ## Fractalkine 14.961872 12.71538464 12.057105 10.38711624 12.917180 14.636190 ## I309 1.103140 0.00000000 0.000000 0.00000000 0.000000 0.000000 ## IFNg 4.394705 0.00000000 2.638213 0.04830196 1.914092 3.251501 ## IL10 2.652120 0.08945805 2.020155 0.84649523 1.119272 2.007941 ## NS_8 NS_9 ## Eotaxin 9.242464 7.127746 ## Eotaxin3 13.724108 3.553097 ## Fractalkine 14.047382 13.146599 ## I309 0.000000 0.000000 ## IFNg 4.104820 2.940248 ## IL10 2.576814 1.970627 # First scaling data with each subject (down columns) scaled_df &lt;- data.frame(apply(wider_baseline_df, 2, scale)) rownames(scaled_df) = rownames(wider_baseline_df) head(scaled_df) ## NS_1 NS_10 NS_11 NS_12 NS_13 NS_14 ## Eotaxin 0.1563938 0.5913460 0.20280481 0.3812918 0.04218372 0.3889166 ## Eotaxin3 -0.4836295 -0.3751558 -0.03031899 -0.4135856 -0.51644074 -0.7745030 ## Fractalkine 1.6023933 1.9565250 1.80232784 1.9342844 1.89246390 1.5458678 ## I309 -1.4818900 -1.3408820 -1.29538062 -1.2697485 -1.32801259 -1.1195175 ## IFNg -0.5236512 -0.8612069 -0.65999509 -0.4512458 -0.71152186 -0.6845711 ## IL10 -1.0383105 -1.0648515 -0.97283813 -0.8409577 -1.01515934 -1.0758338 ## NS_2 NS_3 NS_4 NS_5 NS_6 NS_7 ## Eotaxin 0.5142129 0.5748344 0.3352251 0.1297382 0.6605088 0.4509596 ## Eotaxin3 0.3434700 -0.3390521 -0.4516462 -0.3710044 -0.1481961 -0.1213017 ## Fractalkine 1.6938060 2.2385415 1.9669167 1.6886200 2.2940816 2.0018134 ## I309 -1.4569648 -0.9703707 -1.1565995 -0.9279215 -1.1738781 -1.4178220 ## IFNg -0.7086302 -0.9703707 -0.4731435 -0.9157541 -0.6599894 -0.6581332 ## IL10 -1.1048056 -0.9477947 -0.6332576 -0.7146872 -0.8733797 -0.9486817 ## NS_8 NS_9 ## Eotaxin 0.5504372 0.6258899 ## Eotaxin3 1.4999690 -0.3063521 ## Fractalkine 1.5684615 2.1955622 ## I309 -1.4077759 -1.2329735 ## IFNg -0.5380823 -0.4661787 ## IL10 -0.8618228 -0.7190487 The dist() function is initially used to calculate the Euclidean distance between each cytokine. Next, the hclust() function is used to run the hierarchical clustering analysis using the complete method by default. The method can be changed in the function using the method parameter. # Calculating euclidean dist dist_matrix &lt;- dist(scaled_df, method = &#39;euclidean&#39;) # Hierarchical clustering cytokines_hc &lt;- hclust(dist_matrix) Now we can generate a dendrogram to help us evaluate the results using the fviz_dend() function from the factoextra package. We use k=3 to be consistent with the k-means analysis. fviz_dend(cytokines_hc, k = 3, # Specifying k cex = 0.85, # Label size palette = &quot;futurama&quot;, # Color palette see ?ggpubr::ggpar rect = TRUE, rect_fill = TRUE, # Add rectangle around groups horiz = TRUE, # Changes the orientation of the dendogram rect_border = &quot;futurama&quot;, # Rectangle color labels_track_height = 0.8 # Changes the room for labels ) We can also extract those cluster assignments using the cutree() function from the stats package. hc_assignments_df &lt;- data.frame(cutree(cytokines_hc, k = 3)) %&gt;% rownames_to_column(&quot;Cytokine&quot;) %&gt;% rename(`Hierarchical Cluster` = cutree.cytokines_hc..k...3.) %&gt;% # Ordering the dataframe for easier comparison arrange(`Hierarchical Cluster`) # Combining the dataframes to compare the cluster assignments from each approach comp &lt;- full_join(cluster_kmeans_df, hc_assignments_df, by = &quot;Cytokine&quot;) comp ## Cytokine K-Means Cluster Hierarchical Cluster ## 1 Fractalkine 1 2 ## 2 IL8 1 2 ## 3 IP10 1 2 ## 4 MIG 1 2 ## 5 Eotaxin 2 1 ## 6 Eotaxin3 2 1 ## 7 IL13 2 1 ## 8 IL17B 2 1 ## 9 IL1b 2 1 ## 10 IL6 2 1 ## 11 ITAC 2 1 ## 12 MCP1 2 1 ## 13 MIP1a 2 1 ## 14 MIP1b 2 1 ## 15 TARC 2 1 ## 16 I309 3 3 ## 17 IFNg 3 3 ## 18 IL10 3 3 ## 19 IL12p70 3 3 ## 20 IL17D 3 3 ## 21 IL4 3 3 ## 22 TNFa 3 3 For additional resources on running hierarchical clustering in R, see Visualizing Clustering Dendrogram in R and Hiearchical Clustering on Principal Components. Answer to Environmental Health Question 4 With this, we can answer Environmental Health Question #4: How do the hierarchical cluster assignments compare to the k-means cluster assignments? Answer: Though this may not always be the case, in this instance, we see that k-means and hierarchical clustering with k=3 clusters yield the same groupings despite the clusters being presented in a different order. Clustering Plot One additional way to visualize clustering is to plot the first two principal components on the axes and color the data points based on their corresponding cluster. This visualization can be used for both k-means and hierarchical clustering using the fviz_cluster() function. This figure is essentially a PCA plot with shapes drawn around each cluster to make them distinct from each other. fviz_cluster(cluster_kmeans_3, data = wider_baseline_df) Rather than using the fviz_cluster() function as shown in the figure above, we’ll extract the data to recreate the sample figure using ggplot(). For the manuscript this was necessary, since it was important to facet the plots for each compartment (i.e., NLF, NELF, sputum, and serum). For a single plot, this data extraction isn’t required, and the figure above can be further customized within the fviz_cluster() function. However, we’ll go through the steps of obtaining the indices to recreate the same polygons in ggplot() directly. K-means actually uses principal component analysis (PCA) to reduce a dataset’s dimensionality prior to obtaining the cluster assignments and plotting those clusters. Therefore, to obtain the coordinates of each cytokine within their respective clusters, PCA will need to be run first. # First running PCA pca_cytokine &lt;- prcomp(wider_baseline_df, scale = TRUE, center = TRUE) # Only need PC1 and PC2 for plotting, so selecting the first two columns baseline_scores_df &lt;- data.frame(scores(pca_cytokine)[,1:2]) baseline_scores_df$Cluster &lt;- cluster_kmeans_3$cluster baseline_scores_df$Protein &lt;- rownames(baseline_scores_df) # Changing cluster to a character for plotting baseline_scores_df$Cluster = as.character(baseline_scores_df$Cluster) head(baseline_scores_df) ## PC1 PC2 Cluster Protein ## Eotaxin -1.4837936 -0.37325290 2 Eotaxin ## Eotaxin3 0.7505718 -1.04251406 2 Eotaxin3 ## Fractalkine -7.0500961 -0.49034208 1 Fractalkine ## I309 4.6869896 0.42263944 3 I309 ## IFNg 2.4935464 -0.21910385 3 IFNg ## IL10 3.4361668 -0.04689913 3 IL10 Within each cluster, the chull() function is used to compute the indices of the points on the convex hull. These are needed for ggplot() to create the polygon shapes of each cluster. # hull values for cluster 1 cluster_1 &lt;- baseline_scores_df[baseline_scores_df$Cluster == 1, ][chull(baseline_scores_df %&gt;% filter(Cluster == 1)),] # hull values for cluster 2 cluster_2 &lt;- baseline_scores_df[baseline_scores_df$Cluster == 2, ][chull(baseline_scores_df %&gt;% filter(Cluster == 2)),] # hull values for cluster 3 cluster_3 &lt;- baseline_scores_df[baseline_scores_df$Cluster == 3, ][chull(baseline_scores_df %&gt;% filter(Cluster == 3)),] all_hulls_baseline &lt;- rbind(cluster_1, cluster_2, cluster_3) # Changing cluster to a character for plotting all_hulls_baseline$Cluster = as.character(all_hulls_baseline$Cluster) head(all_hulls_baseline) ## PC1 PC2 Cluster Protein ## MIG -4.7353654 -0.2498662 1 MIG ## IL8 -7.7200078 -0.8317172 1 IL8 ## IP10 -3.8660925 4.2481655 1 IP10 ## Eotaxin3 0.7505718 -1.0425141 2 Eotaxin3 ## MIP1b -0.8699488 -0.8834532 2 MIP1b ## MCP1 -2.4837364 -0.3652235 2 MCP1 Now plotting the clusters using ggplot(). ggplot() + geom_point(data = baseline_scores_df, aes(x = PC1, y = PC2, color = Cluster, shape = Cluster), size = 4) + # Adding cytokine names geom_text_repel(data = baseline_scores_df, aes(x = PC1, y = PC2, color = Cluster, label = Protein), show.legend = FALSE, size = 4.5) + # Creating polygon shapes of the clusters geom_polygon(data = all_hulls_baseline, aes(x = PC1, y = PC2, group = as.factor(Cluster), fill = Cluster, color = Cluster), alpha = 0.25, show.legend = FALSE) + theme_light() + theme(axis.text.x = element_text(vjust = 0.5), #rotating x labels/ moving x labels slightly to the left axis.line = element_line(colour=&quot;black&quot;), #making x and y axes black axis.text = element_text(size = 13), #changing size of x axis labels axis.title = element_text(face = &quot;bold&quot;, size = rel(1.7)), #changes axis titles legend.title = element_text(face = &#39;bold&#39;, size = 17), #changes legend title legend.text = element_text(size = 14), #changes legend text legend.position = &#39;bottom&#39;, # moving the legend to the bottom legend.background = element_rect(colour = &#39;black&#39;, fill = &#39;white&#39;, linetype = &#39;solid&#39;), #changes the legend background strip.text.x = element_text(size = 18, face = &quot;bold&quot;), #changes size of facet x axis strip.text.y = element_text(size = 18, face = &quot;bold&quot;)) + #changes size of facet y axis xlab(&#39;Dimension 1 (85.1%)&#39;) + ylab(&#39;Dimension 2 (7.7%)&#39;) + #changing axis labels # Using colors from the startrek palette from ggsci scale_color_startrek(name = &#39;Cluster&#39;) + scale_fill_startrek(name = &#39;Cluster&#39;) An appropriate title for this figure could be: “Figure X. K-means clusters of cytokines at baseline. Cytokines samples are derived from nasal epithelium (NELF) samples in 14 non-smoking subjects. Cytokine concentration values were transformed using a data reduction technique known as Principal Component Analysis (PCA). The first two eigenvectors plotted on the axes were able to capture a majority of the variance across all samples from the original dataset.” Takeaways from this clustering plot: PCA was able to capture almost all (~93%) of the variance from the original dataset The 22 cytokines were able to be clustered into 3 distinct clusters using k-means Hierarchical Clustering Visualization We can also build a heatmap using the pheatmap() function that has the capability to display hierarchical clustering dendrograms. To do so, we’ll need to go back and use the wider_baseline_df dataframe. pheatmap(wider_baseline_df, cluster_cols = FALSE, # hierarchical clustering of cytokines scale = &quot;column&quot;, # scaling the data to make differences across cytokines more apparent cutree_row = 3, # adds a space between the 3 largest clusters display_numbers = TRUE, number_color = &quot;black&quot;, fontsize = 12, # adding average concentration values angle_col = 45, fontsize_col = 12, fontsize_row = 12, # adjusting size/ orientation of axes labels cellheight = 17, cellwidth = 30 # setting height and width for cells ) An appropriate title for this figure could be: “Figure X. Hierarchical clustering of cytokines at baseline. Cytokines samples are derived from nasal epithelium (NELF) samples in 14 non-smoking subjects. The heatmap visualizes psuedo log2 cytokine concentrations that were scaled within each subject.” It may be helpful to add axes titles like “Subject ID” for the x axis, “Cytokine” for the y axis, and “Scaled pslog2 Concentration” for the legend after exporting from R. The pheatmap() function does not have the functionality to add those titles. Nevertheless, let’s identify some key takeaways from this heatmap: The 22 cytokines were able to be clustered into 3 distinct clusters using hierarchical clustering These clusters are based on cytokine concentration levels with the first cluster having the highest expression, the second cluster having the lowest expression, and the last cluster having average expression Variable Contributions To answer our final environmental health question: Which cytokines have the greatest contributions to the first two eigenvectors, we’ll use the fviz_contrib() function that plots the percentage of each variable’s contribution to the principal component(s). It also displays a red dashed line, and variables that fall above are considered to have significant contributions to those principal components. For a refresher on PCA and variable contributions, see the previous module, TAME 2.0 Module 5.4 Unsupervised Machine Learning. # kmeans contributions fviz_contrib(pca_cytokine, choice = &quot;ind&quot;, addlabels = TRUE, axes = 1:2) # specifies to show contribution percentages for first 2 PCs An appropriate title for this figure could be: “Figure X. Cytokine contributions to principal components. The bar chart displays each cytokine’s contribution to the first two eigenvectors in descending order from left to right. The red dashed line represents the expected contribution of each cytokine if all inputs were uniform, therefore the seven cytokines that fall above this reference line are considered to have significant contributions to the first two principal components.” Answer to Environmental Health Question 5 With this, we can answer Environmental Health Question #5: Which cytokines have the greatest contributions to the first two eigenvectors? Answer: The cytokines that have significant contributions to the first two principal components include IL-8, Fractalkine, IP-10, IL-4, MIG, I309, and IL-12p70. Concluding Remarks In this module, we explored scenarios where clustering would be appropriate but lack contextual details informing the number of clusters that should be considered, thus resulting in the need to derive such a number. In addition, methodology for k-means and hierarchical clustering was presented, along with corresponding visualizations. Lastly, variable contributions to the eigenvectors were introduced as a means to determine the most influential variables on the principal components’ composition. Additional Resources K-Means Cluster Analysis K-Means Clustering in R Hierarchical Clustering in R Test Your Knowledge Using the same dataset, answer the questions below. Determine the optimal number of k-means clusters of cytokines derived from the nasal epithelieum lining fluid of e-cigarette users. How do those clusters compare to the ones that were derived at baseline (in non-smokers)? Which cytokines have the greatest contributions to the first two eigenvectors? "]]
diff --git a/supervised-machine-learning-model-interpretation.html b/supervised-machine-learning-model-interpretation.html
index 6177f20..be1ddd1 100644
--- a/supervised-machine-learning-model-interpretation.html
+++ b/supervised-machine-learning-model-interpretation.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
diff --git a/supervised-machine-learning.html b/supervised-machine-learning.html
index 121f211..276278c 100644
--- a/supervised-machine-learning.html
+++ b/supervised-machine-learning.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
@@ -651,7 +660,7 @@ <h2>Testing for Differences in Predictor Variables across the Outcome Classes<a
 ##   statistic: The package &quot;cardx&quot; (&gt;= 0.2.1) is required.
 ## ✖ For variable `pH` (`Detect_Concentration`) and &quot;p.value&quot; statistic:
 ##   The package &quot;cardx&quot; (&gt;= 0.2.1) is required.</code></pre>
-<div class="tabwid"><style>.cl-0244e6de{}.cl-023f84c8{font-family:'Arial';font-size:11pt;font-weight:bold;font-style:normal;text-decoration:none;color:rgba(0, 0, 0, 1.00);background-color:transparent;}.cl-023f84d2{font-family:'Arial';font-size:6.6pt;font-weight:bold;font-style:normal;text-decoration:none;color:rgba(0, 0, 0, 1.00);background-color:transparent;position: relative;bottom:3.3pt;}.cl-023f84dc{font-family:'Arial';font-size:11pt;font-weight:normal;font-style:normal;text-decoration:none;color:rgba(0, 0, 0, 1.00);background-color:transparent;}.cl-023f84dd{font-family:'Arial';font-size:6.6pt;font-weight:normal;font-style:normal;text-decoration:none;color:rgba(0, 0, 0, 1.00);background-color:transparent;position: relative;bottom:3.3pt;}.cl-02418cfa{margin:0;text-align:left;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);padding-bottom:2pt;padding-top:2pt;padding-left:5pt;padding-right:5pt;line-height: 1;background-color:transparent;}.cl-02418cfb{margin:0;text-align:center;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);padding-bottom:2pt;padding-top:2pt;padding-left:5pt;padding-right:5pt;line-height: 1;background-color:transparent;}.cl-02418d04{margin:0;text-align:left;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);padding-bottom:5pt;padding-top:5pt;padding-left:5pt;padding-right:5pt;line-height: 1;background-color:transparent;}.cl-02418d05{margin:0;text-align:center;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);padding-bottom:5pt;padding-top:5pt;padding-left:5pt;padding-right:5pt;line-height: 1;background-color:transparent;}.cl-02418d06{margin:0;text-align:left;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);padding-bottom:5pt;padding-top:5pt;padding-left:5pt;padding-right:5pt;line-height: 1;background-color:transparent;}.cl-0241a1b8{width:1.703in;background-color:transparent;vertical-align: middle;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 1pt solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1b9{width:0.54in;background-color:transparent;vertical-align: middle;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 1pt solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1c2{width:2.528in;background-color:transparent;vertical-align: middle;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 1pt solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1c3{width:0.82in;background-color:transparent;vertical-align: middle;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 1pt solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1cc{width:1.703in;background-color:transparent;vertical-align: top;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1cd{width:0.54in;background-color:transparent;vertical-align: top;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1ce{width:2.528in;background-color:transparent;vertical-align: top;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1cf{width:0.82in;background-color:transparent;vertical-align: top;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1d6{width:1.703in;background-color:transparent;vertical-align: top;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1d7{width:0.54in;background-color:transparent;vertical-align: top;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1d8{width:2.528in;background-color:transparent;vertical-align: top;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1d9{width:0.82in;background-color:transparent;vertical-align: top;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1da{width:1.703in;background-color:transparent;vertical-align: middle;border-bottom: 0 solid rgba(255, 255, 255, 0.00);border-top: 0 solid rgba(255, 255, 255, 0.00);border-left: 0 solid rgba(255, 255, 255, 0.00);border-right: 0 solid rgba(255, 255, 255, 0.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1e0{width:0.54in;background-color:transparent;vertical-align: middle;border-bottom: 0 solid rgba(255, 255, 255, 0.00);border-top: 0 solid rgba(255, 255, 255, 0.00);border-left: 0 solid rgba(255, 255, 255, 0.00);border-right: 0 solid rgba(255, 255, 255, 0.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1e1{width:2.528in;background-color:transparent;vertical-align: middle;border-bottom: 0 solid rgba(255, 255, 255, 0.00);border-top: 0 solid rgba(255, 255, 255, 0.00);border-left: 0 solid rgba(255, 255, 255, 0.00);border-right: 0 solid rgba(255, 255, 255, 0.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-0241a1e2{width:0.82in;background-color:transparent;vertical-align: middle;border-bottom: 0 solid rgba(255, 255, 255, 0.00);border-top: 0 solid rgba(255, 255, 255, 0.00);border-left: 0 solid rgba(255, 255, 255, 0.00);border-right: 0 solid rgba(255, 255, 255, 0.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}</style><table data-quarto-disable-processing='true' class='cl-0244e6de'><thead><tr style="overflow-wrap:break-word;"><th class="cl-0241a1b8"><p class="cl-02418cfa"><span class="cl-023f84c8">Characteristic</span></p></th><th class="cl-0241a1b9"><p class="cl-02418cfb"><span class="cl-023f84c8">N</span></p></th><th class="cl-0241a1c2"><p class="cl-02418cfb"><span class="cl-023f84c8">ND</span><span class="cl-023f84c8">  </span><br><span class="cl-023f84c8">N = 515</span><span class="cl-023f84d2">1</span></p></th><th class="cl-0241a1c2"><p class="cl-02418cfb"><span class="cl-023f84c8">D</span><span class="cl-023f84c8">  </span><br><span class="cl-023f84c8">N = 198</span><span class="cl-023f84d2">1</span></p></th><th class="cl-0241a1c3"><p class="cl-02418cfb"><span class="cl-023f84c8">p-value</span></p></th></tr></thead><tbody><tr style="overflow-wrap:break-word;"><td class="cl-0241a1cc"><p class="cl-02418d04"><span class="cl-023f84dc">Water_Sample_Date</span></p></td><td class="cl-0241a1cd"><p class="cl-02418d05"><span class="cl-023f84dc">713</span></p></td><td class="cl-0241a1ce"><p class="cl-02418d05"><span class="cl-023f84dc">2013-06-05 (979.174260670888)</span></p></td><td class="cl-0241a1ce"><p class="cl-02418d05"><span class="cl-023f84dc">2013-03-05 (957.843005291701)</span></p></td><td class="cl-0241a1cf"><p class="cl-02418d05"><span class="cl-023f84dc"></span></p></td></tr><tr style="overflow-wrap:break-word;"><td class="cl-0241a1cc"><p class="cl-02418d04"><span class="cl-023f84dc">Casing_Depth</span></p></td><td class="cl-0241a1cd"><p class="cl-02418d05"><span class="cl-023f84dc">713</span></p></td><td class="cl-0241a1ce"><p class="cl-02418d05"><span class="cl-023f84dc">74 (33)</span></p></td><td class="cl-0241a1ce"><p class="cl-02418d05"><span class="cl-023f84dc">55 (23)</span></p></td><td class="cl-0241a1cf"><p class="cl-02418d05"><span class="cl-023f84dc"></span></p></td></tr><tr style="overflow-wrap:break-word;"><td class="cl-0241a1cc"><p class="cl-02418d04"><span class="cl-023f84dc">Well_Depth</span></p></td><td class="cl-0241a1cd"><p class="cl-02418d05"><span class="cl-023f84dc">713</span></p></td><td class="cl-0241a1ce"><p class="cl-02418d05"><span class="cl-023f84dc">301 (144)</span></p></td><td class="cl-0241a1ce"><p class="cl-02418d05"><span class="cl-023f84dc">334 (128)</span></p></td><td class="cl-0241a1cf"><p class="cl-02418d05"><span class="cl-023f84dc"></span></p></td></tr><tr style="overflow-wrap:break-word;"><td class="cl-0241a1cc"><p class="cl-02418d04"><span class="cl-023f84dc">Static_Water_Depth</span></p></td><td class="cl-0241a1cd"><p class="cl-02418d05"><span class="cl-023f84dc">713</span></p></td><td class="cl-0241a1ce"><p class="cl-02418d05"><span class="cl-023f84dc">35 (12)</span></p></td><td class="cl-0241a1ce"><p class="cl-02418d05"><span class="cl-023f84dc">36 (13)</span></p></td><td class="cl-0241a1cf"><p class="cl-02418d05"><span class="cl-023f84dc"></span></p></td></tr><tr style="overflow-wrap:break-word;"><td class="cl-0241a1cc"><p class="cl-02418d04"><span class="cl-023f84dc">Flow_Rate</span></p></td><td class="cl-0241a1cd"><p class="cl-02418d05"><span class="cl-023f84dc">713</span></p></td><td class="cl-0241a1ce"><p class="cl-02418d05"><span class="cl-023f84dc">25 (33)</span></p></td><td class="cl-0241a1ce"><p class="cl-02418d05"><span class="cl-023f84dc">14 (16)</span></p></td><td class="cl-0241a1cf"><p class="cl-02418d05"><span class="cl-023f84dc"></span></p></td></tr><tr style="overflow-wrap:break-word;"><td class="cl-0241a1d6"><p class="cl-02418d04"><span class="cl-023f84dc">pH</span></p></td><td class="cl-0241a1d7"><p class="cl-02418d05"><span class="cl-023f84dc">713</span></p></td><td class="cl-0241a1d8"><p class="cl-02418d05"><span class="cl-023f84dc">7.45 (0.55)</span></p></td><td class="cl-0241a1d8"><p class="cl-02418d05"><span class="cl-023f84dc">7.82 (0.40)</span></p></td><td class="cl-0241a1d9"><p class="cl-02418d05"><span class="cl-023f84dc"></span></p></td></tr></tbody><tfoot><tr style="overflow-wrap:break-word;"><td  colspan="5"class="cl-0241a1da"><p class="cl-02418d06"><span class="cl-023f84dd">1</span><span class="cl-023f84dc">Mean (SD)</span></p></td></tr></tfoot></table></div>
+<div class="tabwid"><style>.cl-74794e7e{}.cl-7473d14c{font-family:'Arial';font-size:11pt;font-weight:bold;font-style:normal;text-decoration:none;color:rgba(0, 0, 0, 1.00);background-color:transparent;}.cl-7473d156{font-family:'Arial';font-size:6.6pt;font-weight:bold;font-style:normal;text-decoration:none;color:rgba(0, 0, 0, 1.00);background-color:transparent;position: relative;bottom:3.3pt;}.cl-7473d160{font-family:'Arial';font-size:11pt;font-weight:normal;font-style:normal;text-decoration:none;color:rgba(0, 0, 0, 1.00);background-color:transparent;}.cl-7473d161{font-family:'Arial';font-size:6.6pt;font-weight:normal;font-style:normal;text-decoration:none;color:rgba(0, 0, 0, 1.00);background-color:transparent;position: relative;bottom:3.3pt;}.cl-7475e59a{margin:0;text-align:left;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);padding-bottom:2pt;padding-top:2pt;padding-left:5pt;padding-right:5pt;line-height: 1;background-color:transparent;}.cl-7475e59b{margin:0;text-align:center;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);padding-bottom:2pt;padding-top:2pt;padding-left:5pt;padding-right:5pt;line-height: 1;background-color:transparent;}.cl-7475e5a4{margin:0;text-align:left;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);padding-bottom:5pt;padding-top:5pt;padding-left:5pt;padding-right:5pt;line-height: 1;background-color:transparent;}.cl-7475e5a5{margin:0;text-align:center;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);padding-bottom:5pt;padding-top:5pt;padding-left:5pt;padding-right:5pt;line-height: 1;background-color:transparent;}.cl-7475e5a6{margin:0;text-align:left;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);padding-bottom:5pt;padding-top:5pt;padding-left:5pt;padding-right:5pt;line-height: 1;background-color:transparent;}.cl-7475fa58{width:1.703in;background-color:transparent;vertical-align: middle;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 1pt solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa62{width:0.54in;background-color:transparent;vertical-align: middle;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 1pt solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa63{width:2.528in;background-color:transparent;vertical-align: middle;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 1pt solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa6c{width:0.82in;background-color:transparent;vertical-align: middle;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 1pt solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa6d{width:1.703in;background-color:transparent;vertical-align: top;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa76{width:0.54in;background-color:transparent;vertical-align: top;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa77{width:2.528in;background-color:transparent;vertical-align: top;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa78{width:0.82in;background-color:transparent;vertical-align: top;border-bottom: 0 solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa80{width:1.703in;background-color:transparent;vertical-align: top;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa81{width:0.54in;background-color:transparent;vertical-align: top;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa82{width:2.528in;background-color:transparent;vertical-align: top;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa83{width:0.82in;background-color:transparent;vertical-align: top;border-bottom: 1pt solid rgba(0, 0, 0, 1.00);border-top: 0 solid rgba(0, 0, 0, 1.00);border-left: 0 solid rgba(0, 0, 0, 1.00);border-right: 0 solid rgba(0, 0, 0, 1.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa8a{width:1.703in;background-color:transparent;vertical-align: middle;border-bottom: 0 solid rgba(255, 255, 255, 0.00);border-top: 0 solid rgba(255, 255, 255, 0.00);border-left: 0 solid rgba(255, 255, 255, 0.00);border-right: 0 solid rgba(255, 255, 255, 0.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa8b{width:0.54in;background-color:transparent;vertical-align: middle;border-bottom: 0 solid rgba(255, 255, 255, 0.00);border-top: 0 solid rgba(255, 255, 255, 0.00);border-left: 0 solid rgba(255, 255, 255, 0.00);border-right: 0 solid rgba(255, 255, 255, 0.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa8c{width:2.528in;background-color:transparent;vertical-align: middle;border-bottom: 0 solid rgba(255, 255, 255, 0.00);border-top: 0 solid rgba(255, 255, 255, 0.00);border-left: 0 solid rgba(255, 255, 255, 0.00);border-right: 0 solid rgba(255, 255, 255, 0.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}.cl-7475fa8d{width:0.82in;background-color:transparent;vertical-align: middle;border-bottom: 0 solid rgba(255, 255, 255, 0.00);border-top: 0 solid rgba(255, 255, 255, 0.00);border-left: 0 solid rgba(255, 255, 255, 0.00);border-right: 0 solid rgba(255, 255, 255, 0.00);margin-bottom:0;margin-top:0;margin-left:0;margin-right:0;}</style><table data-quarto-disable-processing='true' class='cl-74794e7e'><thead><tr style="overflow-wrap:break-word;"><th class="cl-7475fa58"><p class="cl-7475e59a"><span class="cl-7473d14c">Characteristic</span></p></th><th class="cl-7475fa62"><p class="cl-7475e59b"><span class="cl-7473d14c">N</span></p></th><th class="cl-7475fa63"><p class="cl-7475e59b"><span class="cl-7473d14c">ND</span><span class="cl-7473d14c">  </span><br><span class="cl-7473d14c">N = 515</span><span class="cl-7473d156">1</span></p></th><th class="cl-7475fa63"><p class="cl-7475e59b"><span class="cl-7473d14c">D</span><span class="cl-7473d14c">  </span><br><span class="cl-7473d14c">N = 198</span><span class="cl-7473d156">1</span></p></th><th class="cl-7475fa6c"><p class="cl-7475e59b"><span class="cl-7473d14c">p-value</span></p></th></tr></thead><tbody><tr style="overflow-wrap:break-word;"><td class="cl-7475fa6d"><p class="cl-7475e5a4"><span class="cl-7473d160">Water_Sample_Date</span></p></td><td class="cl-7475fa76"><p class="cl-7475e5a5"><span class="cl-7473d160">713</span></p></td><td class="cl-7475fa77"><p class="cl-7475e5a5"><span class="cl-7473d160">2013-06-05 (979.174260670888)</span></p></td><td class="cl-7475fa77"><p class="cl-7475e5a5"><span class="cl-7473d160">2013-03-05 (957.843005291701)</span></p></td><td class="cl-7475fa78"><p class="cl-7475e5a5"><span class="cl-7473d160"></span></p></td></tr><tr style="overflow-wrap:break-word;"><td class="cl-7475fa6d"><p class="cl-7475e5a4"><span class="cl-7473d160">Casing_Depth</span></p></td><td class="cl-7475fa76"><p class="cl-7475e5a5"><span class="cl-7473d160">713</span></p></td><td class="cl-7475fa77"><p class="cl-7475e5a5"><span class="cl-7473d160">74 (33)</span></p></td><td class="cl-7475fa77"><p class="cl-7475e5a5"><span class="cl-7473d160">55 (23)</span></p></td><td class="cl-7475fa78"><p class="cl-7475e5a5"><span class="cl-7473d160"></span></p></td></tr><tr style="overflow-wrap:break-word;"><td class="cl-7475fa6d"><p class="cl-7475e5a4"><span class="cl-7473d160">Well_Depth</span></p></td><td class="cl-7475fa76"><p class="cl-7475e5a5"><span class="cl-7473d160">713</span></p></td><td class="cl-7475fa77"><p class="cl-7475e5a5"><span class="cl-7473d160">301 (144)</span></p></td><td class="cl-7475fa77"><p class="cl-7475e5a5"><span class="cl-7473d160">334 (128)</span></p></td><td class="cl-7475fa78"><p class="cl-7475e5a5"><span class="cl-7473d160"></span></p></td></tr><tr style="overflow-wrap:break-word;"><td class="cl-7475fa6d"><p class="cl-7475e5a4"><span class="cl-7473d160">Static_Water_Depth</span></p></td><td class="cl-7475fa76"><p class="cl-7475e5a5"><span class="cl-7473d160">713</span></p></td><td class="cl-7475fa77"><p class="cl-7475e5a5"><span class="cl-7473d160">35 (12)</span></p></td><td class="cl-7475fa77"><p class="cl-7475e5a5"><span class="cl-7473d160">36 (13)</span></p></td><td class="cl-7475fa78"><p class="cl-7475e5a5"><span class="cl-7473d160"></span></p></td></tr><tr style="overflow-wrap:break-word;"><td class="cl-7475fa6d"><p class="cl-7475e5a4"><span class="cl-7473d160">Flow_Rate</span></p></td><td class="cl-7475fa76"><p class="cl-7475e5a5"><span class="cl-7473d160">713</span></p></td><td class="cl-7475fa77"><p class="cl-7475e5a5"><span class="cl-7473d160">25 (33)</span></p></td><td class="cl-7475fa77"><p class="cl-7475e5a5"><span class="cl-7473d160">14 (16)</span></p></td><td class="cl-7475fa78"><p class="cl-7475e5a5"><span class="cl-7473d160"></span></p></td></tr><tr style="overflow-wrap:break-word;"><td class="cl-7475fa80"><p class="cl-7475e5a4"><span class="cl-7473d160">pH</span></p></td><td class="cl-7475fa81"><p class="cl-7475e5a5"><span class="cl-7473d160">713</span></p></td><td class="cl-7475fa82"><p class="cl-7475e5a5"><span class="cl-7473d160">7.45 (0.55)</span></p></td><td class="cl-7475fa82"><p class="cl-7475e5a5"><span class="cl-7473d160">7.82 (0.40)</span></p></td><td class="cl-7475fa83"><p class="cl-7475e5a5"><span class="cl-7473d160"></span></p></td></tr></tbody><tfoot><tr style="overflow-wrap:break-word;"><td  colspan="5"class="cl-7475fa8a"><p class="cl-7475e5a6"><span class="cl-7473d161">1</span><span class="cl-7473d160">Mean (SD)</span></p></td></tr></tfoot></table></div>
 <p>Note that N refers to the total sample number; ND refers to the samples that contained non-detectable levels of iAs; and D refers to the samples that contained detectable levels of iAs.</p>
 <div id="answer-to-environmental-health-question-1-6" class="section level3 hasAnchor">
 <h3>Answer to Environmental Health Question 1<a href="supervised-machine-learning.html#answer-to-environmental-health-question-1-6" class="anchor-section" aria-label="Anchor link to header"></a></h3>
diff --git a/two-group-comparisons-and-visualizations.html b/two-group-comparisons-and-visualizations.html
index 50ab0fd..f883549 100644
--- a/two-group-comparisons-and-visualizations.html
+++ b/two-group-comparisons-and-visualizations.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
@@ -434,11 +443,11 @@ <h3>Workspace Preparation and Data Import<a href="two-group-comparisons-and-visu
 <span id="cb408-4"><a href="two-group-comparisons-and-visualizations.html#cb408-4" tabindex="-1"></a></span>
 <span id="cb408-5"><a href="two-group-comparisons-and-visualizations.html#cb408-5" tabindex="-1"></a><span class="co"># View data</span></span>
 <span id="cb408-6"><a href="two-group-comparisons-and-visualizations.html#cb408-6" tabindex="-1"></a><span class="fu">datatable</span>(biomarker_data)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-99702a877ea426fdaef3" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-99702a877ea426fdaef3">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80"],["D10","D10","D10","D10","D10","D11","D11","D11","D11","D11","D12","D12","D12","D12","D12","D13","D13","D13","D13","D13","D14","D14","D14","D14","D14","D15","D15","D15","D15","D15","D16","D16","D16","D16","D16","D1","D1","D1","D1","D1","D2","D2","D2","D2","D2","D3","D3","D3","D3","D3","D4","D4","D4","D4","D4","D5","D5","D5","D5","D5","D6","D6","D6","D6","D6","D7","D7","D7","D7","D7","D8","D8","D8","D8","D8","D9","D9","D9","D9","D9"],["0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4"],[3.71808758396052,3.7235585614721,3.70597790168252,3.76553474636298,3.6724253419715,3.83491342970104,3.93073733756289,3.92884403671257,3.80632405739003,3.81762325751143,3.36036427654347,3.27649666564036,3.57167680997093,3.54101915313356,3.10097764772482,3.35755200461808,3.33628338786443,3.31469652565629,3.50589092972996,3.69041709553247,3.31614574229336,3.43562859405209,3.57773093149008,3.87578006306849,3.83086375675176,3.42894634476948,3.49825086752783,3.57288966842058,3.63807383718072,3.6915341649192,3.38681094647222,3.49825086752783,3.58856473740135,3.66220549965362,3.69376571221778,3.76871365703048,3.77820857639809,3.74953426766926,3.84096670448742,3.50589092972996,3.53480866123175,3.60880924267552,3.58856473740135,3.2585189247113,3.75167794568758,3.58014548442338,3.73876783680019,3.61823865559545,3.58736499093646,3.39643353125099,3.52732060787705,3.50842865253186,3.62175885670973,3.23266075679027,4.13996056954546,3.16671544496642,3.18269229751619,3.13422093976063,3.7279204545632,3.799087306074,2.86195536414487,3.08746284125034,3.64961545906341,3.66220549965362,3.64616265715789,3.04614178164472,3.1811025507538,3.27202318906105,3.54101915313356,3.42088657497553,3.37712374912949,3.57531233068744,3.58976348698498,3.83289001416474,3.8972404255748,3.21878116778407,3.30888505741176,3.51853513898218,3.5765221379205,3.7279204545632],[8.68422263328295,8.773963368433559,8.561326136453919,9.710324307072939,7.55028504908841,10.2222394489088,10.1590009172826,9.94747622373011,9.81610366284939,8.03815081026938,9.27052894238072,9.05077375519723,8.99740532549677,8.27635707925924,7.61889934790406,8.15790268278964,8.089053172156429,7.41515020542639,7.37920505996061,4.10265813136374,7.56094417806009,7.5894638929561,7.9047851297316,7.83857365266832,6.50969584193339,9.254131033541951,9.422148885880191,9.165032591502291,9.09402488708305,8.202809492125009,8.68422263328295,8.773963368433559,8.77360079528254,8.60099074240823,7.71163256182322,10.1099481540673,10.3357141430951,9.686097776967401,10.0151641270811,10.1498741117182,8.88962564051876,9.317005526387209,9.03875441393145,8.60099074240823,4.87479696589866,9.23160520814165,9.677085350559841,9.23208509872283,9.583628142323651,9.560179957764181,8.34837407519116,8.351469324267081,8.77360079528254,9.33695280016051,7.70355756192103,3.9429835981871,4.12018602753186,4.24868662687383,5.4110872267986,5.17871464117544,7.43212427264156,6.85112436288675,7.38076407498511,6.96220195703608,8.36382813184704,8.12861303870471,8.0503384381673,7.86530024272912,7.3639595068907,8.90926307937532,11.4238722598684,11.7288813640812,11.4570986742783,10.878112237463,8.52575569282949,7.01077983875324,7.32804655627074,7.03452387533439,7.27537959625362,4.70099449416827],[15.0940690235359,15.1807289234454,15.1191518828101,16.3458996676138,16.7241460980745,15.0028356680746,15.5465951732949,15.4894995973234,16.169128959283,16.0996069083821,13.9520295507825,14.0492691434224,13.1502883461063,14.8168226706095,16.3560077655342,14.30563571243,14.2750718138596,14.2603087281481,14.3170903457307,16.6038859325635,13.9567554773633,14.1371656727442,14.4431119040824,15.1950634099335,16.3613300730261,14.4305440239947,14.6030328264996,14.6583565484435,15.3569594653045,16.5368999542995,14.3291992954213,14.4239037658366,14.5587322032289,15.1950634099335,16.3717050181953,15.1620702745062,15.5056676741312,15.1429673906615,15.4079647745043,16.9008155214109,14.6079451777076,14.6753941971444,13.1497550568455,14.9206710618732,17.2934204566841,14.5987126147926,14.8967320984425,14.6970451993658,15.0929012215711,17.1264130896721,14.542155252148,14.4584290430224,14.8614654668171,15.9922404459517,15.3426552970282,14.1001424218123,13.9841362654046,14.0738746130993,15.2370937904406,16.3510586187167,13.4034767756747,13.4654490662827,13.7572472746771,15.5483372770406,16.2848127161883,13.5861396108168,13.8384840329337,13.8699889688218,14.4670316946933,17.0361832734254,14.1156143059506,14.4239037658366,14.4497857465895,15.0182445232662,15.1484821432167,14.3291992954213,14.1946037445951,13.1529914376826,14.708543993667,16.3449160294487],[0.189033824390017,-0.3040061868901,0.286881147788162,0.298658315564515,1.0703893278914,2.3305584000308,1.37851162325373,1.30451104180995,0.978195629681652,1.64154602908752,0.526068811667588,0.238786859587116,0.713695814843359,1.50589092972996,0.333423733725192,1.24488705912353,-0.3040061868901,1.56559717585423,1.09761079662642,0.855989697308481,1.06350294230616,0.20163386116965,0.895302621333307,1.1243281350022,1.37295209791183,1.27500704749987,0.632268215499513,1.16992500144231,0.910732661902913,1.16992500144231,0.8155754288625729,0.333423733725192,0.748461233004036,0.903038270112912,1.13750352374994,0.367371065648529,0.137503523749935,0.464668267003444,0.910732661902913,0.454175893185802,1.35614381022528,0.298658315564515,0.632268215499513,0.748461233004036,1.40599235967584,0.31034012061215,0.124328135002202,0.53605290024021,0.495695162624069,0.321928094887362,0.879705766282288,0.298658315564515,1.01435529297707,1.19534759832222,2.03210084316702,0.933572638261024,0.20163386116965,2.68706068833989,1.13750352374994,0.84799690655495,1.86393845042397,1.41142624572647,0.992768430768924,1.0071955014042,0.443606651475615,0.650764559116902,0.8718436485093179,1.19534759832222,0.16349873228288,1.17632277264046,1.01435529297707,0.941106310946431,1.04264433740849,0.8718436485093179,1.57531233068744,2.51349074558812,0.604071323668861,0.411426245726465,0.584962500721156,1.47508488294878],[1.73552217729654,1.55581615506164,1.93734439215023,2.08406426478847,2.83390207666916,3.23572705983806,3.24031432933371,3.15218341905505,3.16510798514454,3.48284828306847,1.63226821549951,1.91073266190291,2.35049724708413,3.24031432933371,1.57046293102604,1.32192809488736,1.20789285164133,1.20789285164133,2.75274859140713,2.59931779369823,1.91456452349394,1.72246602447109,1.96347412397489,2.81352468929781,3.27202318906105,2.08746284125034,2.02502879449152,2.37016428054021,2.63923216324928,2.94485844580754,1.83187724119167,1.91073266190291,2.0496307677246,2.64154602908752,2.93168305705981,1.86393845042397,1.790772037862,1.86393845042397,1.88362081628567,2.88362081628567,1.83187724119167,2.0071955014042,2.02147972741045,2.48542682717024,2.98913900736823,2.52606881166759,1.49569516262407,1.85598969730848,1.50080205305716,2.06695024392463,2.47508488294878,2.31034012061215,2.39231742277876,2.99457972421575,3.73226919950145,2.20476675065461,2.34482849699744,1.96347412397489,2.40053792958373,2.85798099512757,1.790772037862,2.11103131238874,2.39506279951758,3.0232553523003,2.16671544496642,1.56559717585423,1.32192809488736,3.59693514238723,1.73552217729654,2.75702324650746,2.53106949272595,2.48284828306847,2.7158933705476,2.99819550315325,3.47118746038699,1.0703893278914,1.33342373372519,1.97819562968165,2.58496250072116,3.08406426478847],[10.4475275409266,10.4320198468125,10.4523853997561,10.7211590148677,11.2458081856384,9.02901108665907,9.22583403553784,9.412739192116421,9.91233549843777,9.316688823436939,9.428025110276749,9.412146396540679,9.620146497339491,10.9260809830959,7.3039633562453,9.247642742898019,9.32424806586749,10.2153509830024,11.2601080099627,10.8557639471917,10.6537676442859,10.6473145095471,10.8149028512154,11.0078962674403,10.9341110109735,9.933439857942711,10.009954550333,10.2267734234482,10.7809497987497,10.646036617491,9.797969484059131,9.8490610025615,10.2090268302306,10.7211590148677,10.8557639471917,9.696480751820189,9.72723977748481,9.49325503512067,10.0461417816447,11.0193335777127,10.534828110721,10.8215188973143,10.8050102485124,11.7900226112658,11.031563370024,9.35887093968562,9.18935511254853,9.033670771393091,9.36430978186441,11.0217850685075,8.283551423174311,8.22211866974429,8.617467465294199,9.414283550503709,8.973180744513019,9.425320822759881,9.308885057411761,9.61402881967061,10.4624204732526,10.7035402591663,10.7662472988462,10.9813603523445,11.5188793273505,11.3227900730377,9.89590846480624,10.3333342781343,10.3711558797312,10.374854737178,10.7157046495212,11.7319596151152,9.92478281821723,9.8490610025615,10.0477374043091,10.6360535917989,9.40224465525408,9.797969484059131,10.0853263897027,10.4360035526972,11.191225812124,10.9299462991725]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Donor<\/th>\n      <th>Dose<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[3,4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Donor","targets":1},{"name":"Dose","targets":2},{"name":"IL1B","targets":3},{"name":"IL6","targets":4},{"name":"IL8","targets":5},{"name":"IL10","targets":6},{"name":"TNFa","targets":7},{"name":"VEGF","targets":8}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-1172e7f46d65e8d58fed" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-1172e7f46d65e8d58fed">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80"],["D10","D10","D10","D10","D10","D11","D11","D11","D11","D11","D12","D12","D12","D12","D12","D13","D13","D13","D13","D13","D14","D14","D14","D14","D14","D15","D15","D15","D15","D15","D16","D16","D16","D16","D16","D1","D1","D1","D1","D1","D2","D2","D2","D2","D2","D3","D3","D3","D3","D3","D4","D4","D4","D4","D4","D5","D5","D5","D5","D5","D6","D6","D6","D6","D6","D7","D7","D7","D7","D7","D8","D8","D8","D8","D8","D9","D9","D9","D9","D9"],["0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4","0","0.6","1","2","4"],[3.71808758396052,3.7235585614721,3.70597790168252,3.76553474636298,3.6724253419715,3.83491342970104,3.93073733756289,3.92884403671257,3.80632405739003,3.81762325751143,3.36036427654347,3.27649666564036,3.57167680997093,3.54101915313356,3.10097764772482,3.35755200461808,3.33628338786443,3.31469652565629,3.50589092972996,3.69041709553247,3.31614574229336,3.43562859405209,3.57773093149008,3.87578006306849,3.83086375675176,3.42894634476948,3.49825086752783,3.57288966842058,3.63807383718072,3.6915341649192,3.38681094647222,3.49825086752783,3.58856473740135,3.66220549965362,3.69376571221778,3.76871365703048,3.77820857639809,3.74953426766926,3.84096670448742,3.50589092972996,3.53480866123175,3.60880924267552,3.58856473740135,3.2585189247113,3.75167794568758,3.58014548442338,3.73876783680019,3.61823865559545,3.58736499093646,3.39643353125099,3.52732060787705,3.50842865253186,3.62175885670973,3.23266075679027,4.13996056954546,3.16671544496642,3.18269229751619,3.13422093976063,3.7279204545632,3.799087306074,2.86195536414487,3.08746284125034,3.64961545906341,3.66220549965362,3.64616265715789,3.04614178164472,3.1811025507538,3.27202318906105,3.54101915313356,3.42088657497553,3.37712374912949,3.57531233068744,3.58976348698498,3.83289001416474,3.8972404255748,3.21878116778407,3.30888505741176,3.51853513898218,3.5765221379205,3.7279204545632],[8.68422263328295,8.773963368433559,8.561326136453919,9.710324307072939,7.55028504908841,10.2222394489088,10.1590009172826,9.94747622373011,9.81610366284939,8.03815081026938,9.27052894238072,9.05077375519723,8.99740532549677,8.27635707925924,7.61889934790406,8.15790268278964,8.089053172156429,7.41515020542639,7.37920505996061,4.10265813136374,7.56094417806009,7.5894638929561,7.9047851297316,7.83857365266832,6.50969584193339,9.254131033541951,9.422148885880191,9.165032591502291,9.09402488708305,8.202809492125009,8.68422263328295,8.773963368433559,8.77360079528254,8.60099074240823,7.71163256182322,10.1099481540673,10.3357141430951,9.686097776967401,10.0151641270811,10.1498741117182,8.88962564051876,9.317005526387209,9.03875441393145,8.60099074240823,4.87479696589866,9.23160520814165,9.677085350559841,9.23208509872283,9.583628142323651,9.560179957764181,8.34837407519116,8.351469324267081,8.77360079528254,9.33695280016051,7.70355756192103,3.9429835981871,4.12018602753186,4.24868662687383,5.4110872267986,5.17871464117544,7.43212427264156,6.85112436288675,7.38076407498511,6.96220195703608,8.36382813184704,8.12861303870471,8.0503384381673,7.86530024272912,7.3639595068907,8.90926307937532,11.4238722598684,11.7288813640812,11.4570986742783,10.878112237463,8.52575569282949,7.01077983875324,7.32804655627074,7.03452387533439,7.27537959625362,4.70099449416827],[15.0940690235359,15.1807289234454,15.1191518828101,16.3458996676138,16.7241460980745,15.0028356680746,15.5465951732949,15.4894995973234,16.169128959283,16.0996069083821,13.9520295507825,14.0492691434224,13.1502883461063,14.8168226706095,16.3560077655342,14.30563571243,14.2750718138596,14.2603087281481,14.3170903457307,16.6038859325635,13.9567554773633,14.1371656727442,14.4431119040824,15.1950634099335,16.3613300730261,14.4305440239947,14.6030328264996,14.6583565484435,15.3569594653045,16.5368999542995,14.3291992954213,14.4239037658366,14.5587322032289,15.1950634099335,16.3717050181953,15.1620702745062,15.5056676741312,15.1429673906615,15.4079647745043,16.9008155214109,14.6079451777076,14.6753941971444,13.1497550568455,14.9206710618732,17.2934204566841,14.5987126147926,14.8967320984425,14.6970451993658,15.0929012215711,17.1264130896721,14.542155252148,14.4584290430224,14.8614654668171,15.9922404459517,15.3426552970282,14.1001424218123,13.9841362654046,14.0738746130993,15.2370937904406,16.3510586187167,13.4034767756747,13.4654490662827,13.7572472746771,15.5483372770406,16.2848127161883,13.5861396108168,13.8384840329337,13.8699889688218,14.4670316946933,17.0361832734254,14.1156143059506,14.4239037658366,14.4497857465895,15.0182445232662,15.1484821432167,14.3291992954213,14.1946037445951,13.1529914376826,14.708543993667,16.3449160294487],[0.189033824390017,-0.3040061868901,0.286881147788162,0.298658315564515,1.0703893278914,2.3305584000308,1.37851162325373,1.30451104180995,0.978195629681652,1.64154602908752,0.526068811667588,0.238786859587116,0.713695814843359,1.50589092972996,0.333423733725192,1.24488705912353,-0.3040061868901,1.56559717585423,1.09761079662642,0.855989697308481,1.06350294230616,0.20163386116965,0.895302621333307,1.1243281350022,1.37295209791183,1.27500704749987,0.632268215499513,1.16992500144231,0.910732661902913,1.16992500144231,0.8155754288625729,0.333423733725192,0.748461233004036,0.903038270112912,1.13750352374994,0.367371065648529,0.137503523749935,0.464668267003444,0.910732661902913,0.454175893185802,1.35614381022528,0.298658315564515,0.632268215499513,0.748461233004036,1.40599235967584,0.31034012061215,0.124328135002202,0.53605290024021,0.495695162624069,0.321928094887362,0.879705766282288,0.298658315564515,1.01435529297707,1.19534759832222,2.03210084316702,0.933572638261024,0.20163386116965,2.68706068833989,1.13750352374994,0.84799690655495,1.86393845042397,1.41142624572647,0.992768430768924,1.0071955014042,0.443606651475615,0.650764559116902,0.8718436485093179,1.19534759832222,0.16349873228288,1.17632277264046,1.01435529297707,0.941106310946431,1.04264433740849,0.8718436485093179,1.57531233068744,2.51349074558812,0.604071323668861,0.411426245726465,0.584962500721156,1.47508488294878],[1.73552217729654,1.55581615506164,1.93734439215023,2.08406426478847,2.83390207666916,3.23572705983806,3.24031432933371,3.15218341905505,3.16510798514454,3.48284828306847,1.63226821549951,1.91073266190291,2.35049724708413,3.24031432933371,1.57046293102604,1.32192809488736,1.20789285164133,1.20789285164133,2.75274859140713,2.59931779369823,1.91456452349394,1.72246602447109,1.96347412397489,2.81352468929781,3.27202318906105,2.08746284125034,2.02502879449152,2.37016428054021,2.63923216324928,2.94485844580754,1.83187724119167,1.91073266190291,2.0496307677246,2.64154602908752,2.93168305705981,1.86393845042397,1.790772037862,1.86393845042397,1.88362081628567,2.88362081628567,1.83187724119167,2.0071955014042,2.02147972741045,2.48542682717024,2.98913900736823,2.52606881166759,1.49569516262407,1.85598969730848,1.50080205305716,2.06695024392463,2.47508488294878,2.31034012061215,2.39231742277876,2.99457972421575,3.73226919950145,2.20476675065461,2.34482849699744,1.96347412397489,2.40053792958373,2.85798099512757,1.790772037862,2.11103131238874,2.39506279951758,3.0232553523003,2.16671544496642,1.56559717585423,1.32192809488736,3.59693514238723,1.73552217729654,2.75702324650746,2.53106949272595,2.48284828306847,2.7158933705476,2.99819550315325,3.47118746038699,1.0703893278914,1.33342373372519,1.97819562968165,2.58496250072116,3.08406426478847],[10.4475275409266,10.4320198468125,10.4523853997561,10.7211590148677,11.2458081856384,9.02901108665907,9.22583403553784,9.412739192116421,9.91233549843777,9.316688823436939,9.428025110276749,9.412146396540679,9.620146497339491,10.9260809830959,7.3039633562453,9.247642742898019,9.32424806586749,10.2153509830024,11.2601080099627,10.8557639471917,10.6537676442859,10.6473145095471,10.8149028512154,11.0078962674403,10.9341110109735,9.933439857942711,10.009954550333,10.2267734234482,10.7809497987497,10.646036617491,9.797969484059131,9.8490610025615,10.2090268302306,10.7211590148677,10.8557639471917,9.696480751820189,9.72723977748481,9.49325503512067,10.0461417816447,11.0193335777127,10.534828110721,10.8215188973143,10.8050102485124,11.7900226112658,11.031563370024,9.35887093968562,9.18935511254853,9.033670771393091,9.36430978186441,11.0217850685075,8.283551423174311,8.22211866974429,8.617467465294199,9.414283550503709,8.973180744513019,9.425320822759881,9.308885057411761,9.61402881967061,10.4624204732526,10.7035402591663,10.7662472988462,10.9813603523445,11.5188793273505,11.3227900730377,9.89590846480624,10.3333342781343,10.3711558797312,10.374854737178,10.7157046495212,11.7319596151152,9.92478281821723,9.8490610025615,10.0477374043091,10.6360535917989,9.40224465525408,9.797969484059131,10.0853263897027,10.4360035526972,11.191225812124,10.9299462991725]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Donor<\/th>\n      <th>Dose<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[3,4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Donor","targets":1},{"name":"Dose","targets":2},{"name":"IL1B","targets":3},{"name":"IL6","targets":4},{"name":"IL8","targets":5},{"name":"IL10","targets":6},{"name":"TNFa","targets":7},{"name":"VEGF","targets":8}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <div class="sourceCode" id="cb409"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb409-1"><a href="two-group-comparisons-and-visualizations.html#cb409-1" tabindex="-1"></a><span class="fu">datatable</span>(demographic_data)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-a6c6a624cf442750ada9" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-a6c6a624cf442750ada9">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16"],["D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16"],[22,35,27,60,43,38,26,36,40,44,46,52,58,65,37,25],["M","F","M","F","F","M","M","F","F","M","M","M","M","M","F","F"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Donor<\/th>\n      <th>Age_yr<\/th>\n      <th>Sex<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":2},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Donor","targets":1},{"name":"Age_yr","targets":2},{"name":"Sex","targets":3}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-0ceadb6a955cc93976a6" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-0ceadb6a955cc93976a6">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16"],["D1","D2","D3","D4","D5","D6","D7","D8","D9","D10","D11","D12","D13","D14","D15","D16"],[22,35,27,60,43,38,26,36,40,44,46,52,58,65,37,25],["M","F","M","F","F","M","M","F","F","M","M","M","M","M","F","F"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Donor<\/th>\n      <th>Age_yr<\/th>\n      <th>Sex<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":2},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Donor","targets":1},{"name":"Age_yr","targets":2},{"name":"Sex","targets":3}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p><br></p>
 </div>
 </div>
@@ -505,8 +514,8 @@ <h2>Unpaired Test Example<a href="two-group-comparisons-and-visualizations.html#
 <div class="sourceCode" id="cb411"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb411-1"><a href="two-group-comparisons-and-visualizations.html#cb411-1" tabindex="-1"></a>biomarker_data_malevsfemale <span class="ot">&lt;-</span> biomarker_data_malevsfemale <span class="sc">%&gt;%</span> <span class="fu">left_join</span>(demographic_data <span class="sc">%&gt;%</span> <span class="fu">select</span>(Donor, Sex), <span class="at">by =</span> <span class="st">&quot;Donor&quot;</span>) </span></code></pre></div>
 <p>Here is what our data look like now:</p>
 <div class="sourceCode" id="cb412"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb412-1"><a href="two-group-comparisons-and-visualizations.html#cb412-1" tabindex="-1"></a><span class="fu">datatable</span>(biomarker_data_malevsfemale)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-c7cfbb62ec68f15e1824" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-c7cfbb62ec68f15e1824">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16"],["D10","D11","D12","D13","D14","D15","D16","D1","D2","D3","D4","D5","D6","D7","D8","D9"],["0","0","0","0","0","0","0","0","0","0","0","0","0","0","0","0"],[3.71808758396052,3.83491342970104,3.36036427654347,3.35755200461808,3.31614574229336,3.42894634476948,3.38681094647222,3.76871365703048,3.53480866123175,3.58014548442338,3.52732060787705,3.16671544496642,2.86195536414487,3.04614178164472,3.37712374912949,3.21878116778407],[8.68422263328295,10.2222394489088,9.27052894238072,8.15790268278964,7.56094417806009,9.254131033541951,8.68422263328295,10.1099481540673,8.88962564051876,9.23160520814165,8.34837407519116,3.9429835981871,7.43212427264156,8.12861303870471,11.4238722598684,7.01077983875324],[15.0940690235359,15.0028356680746,13.9520295507825,14.30563571243,13.9567554773633,14.4305440239947,14.3291992954213,15.1620702745062,14.6079451777076,14.5987126147926,14.542155252148,14.1001424218123,13.4034767756747,13.5861396108168,14.1156143059506,14.3291992954213],[0.189033824390017,2.3305584000308,0.526068811667588,1.24488705912353,1.06350294230616,1.27500704749987,0.8155754288625729,0.367371065648529,1.35614381022528,0.31034012061215,0.879705766282288,0.933572638261024,1.86393845042397,0.650764559116902,1.01435529297707,2.51349074558812],[1.73552217729654,3.23572705983806,1.63226821549951,1.32192809488736,1.91456452349394,2.08746284125034,1.83187724119167,1.86393845042397,1.83187724119167,2.52606881166759,2.47508488294878,2.20476675065461,1.790772037862,1.56559717585423,2.53106949272595,1.0703893278914],[10.4475275409266,9.02901108665907,9.428025110276749,9.247642742898019,10.6537676442859,9.933439857942711,9.797969484059131,9.696480751820189,10.534828110721,9.35887093968562,8.283551423174311,9.425320822759881,10.7662472988462,10.3333342781343,9.92478281821723,9.797969484059131],["M","M","M","M","M","F","F","M","F","M","F","F","M","M","F","F"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Donor<\/th>\n      <th>Dose<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n      <th>Sex<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[3,4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Donor","targets":1},{"name":"Dose","targets":2},{"name":"IL1B","targets":3},{"name":"IL6","targets":4},{"name":"IL8","targets":5},{"name":"IL10","targets":6},{"name":"TNFa","targets":7},{"name":"VEGF","targets":8},{"name":"Sex","targets":9}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-ed7c9ebcbe2c4f369ed1" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-ed7c9ebcbe2c4f369ed1">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16"],["D10","D11","D12","D13","D14","D15","D16","D1","D2","D3","D4","D5","D6","D7","D8","D9"],["0","0","0","0","0","0","0","0","0","0","0","0","0","0","0","0"],[3.71808758396052,3.83491342970104,3.36036427654347,3.35755200461808,3.31614574229336,3.42894634476948,3.38681094647222,3.76871365703048,3.53480866123175,3.58014548442338,3.52732060787705,3.16671544496642,2.86195536414487,3.04614178164472,3.37712374912949,3.21878116778407],[8.68422263328295,10.2222394489088,9.27052894238072,8.15790268278964,7.56094417806009,9.254131033541951,8.68422263328295,10.1099481540673,8.88962564051876,9.23160520814165,8.34837407519116,3.9429835981871,7.43212427264156,8.12861303870471,11.4238722598684,7.01077983875324],[15.0940690235359,15.0028356680746,13.9520295507825,14.30563571243,13.9567554773633,14.4305440239947,14.3291992954213,15.1620702745062,14.6079451777076,14.5987126147926,14.542155252148,14.1001424218123,13.4034767756747,13.5861396108168,14.1156143059506,14.3291992954213],[0.189033824390017,2.3305584000308,0.526068811667588,1.24488705912353,1.06350294230616,1.27500704749987,0.8155754288625729,0.367371065648529,1.35614381022528,0.31034012061215,0.879705766282288,0.933572638261024,1.86393845042397,0.650764559116902,1.01435529297707,2.51349074558812],[1.73552217729654,3.23572705983806,1.63226821549951,1.32192809488736,1.91456452349394,2.08746284125034,1.83187724119167,1.86393845042397,1.83187724119167,2.52606881166759,2.47508488294878,2.20476675065461,1.790772037862,1.56559717585423,2.53106949272595,1.0703893278914],[10.4475275409266,9.02901108665907,9.428025110276749,9.247642742898019,10.6537676442859,9.933439857942711,9.797969484059131,9.696480751820189,10.534828110721,9.35887093968562,8.283551423174311,9.425320822759881,10.7662472988462,10.3333342781343,9.92478281821723,9.797969484059131],["M","M","M","M","M","F","F","M","F","M","F","F","M","M","F","F"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Donor<\/th>\n      <th>Dose<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n      <th>Sex<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[3,4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Donor","targets":1},{"name":"Dose","targets":2},{"name":"IL1B","targets":3},{"name":"IL6","targets":4},{"name":"IL8","targets":5},{"name":"IL10","targets":6},{"name":"TNFa","targets":7},{"name":"VEGF","targets":8},{"name":"Sex","targets":9}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>We can demonstrate the basic anatomy of the Wilcoxon test function <code>wilcox.test()</code> by running the function on just one variable.</p>
 <div class="sourceCode" id="cb413"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb413-1"><a href="two-group-comparisons-and-visualizations.html#cb413-1" tabindex="-1"></a><span class="fu">wilcox.test</span>(IL1B <span class="sc">~</span> Sex, <span class="at">data =</span> biomarker_data_malevsfemale)</span></code></pre></div>
 <pre><code>## 
@@ -587,8 +596,8 @@ <h2>Paired Test Example<a href="two-group-comparisons-and-visualizations.html#pa
 <span id="cb419-2"><a href="two-group-comparisons-and-visualizations.html#cb419-2" tabindex="-1"></a>  <span class="fu">filter</span>(Dose <span class="sc">==</span> <span class="st">&quot;0&quot;</span> <span class="sc">|</span> Dose <span class="sc">==</span> <span class="st">&quot;4&quot;</span>)</span></code></pre></div>
 <p>Let’s view the dataframe. Note how the measurements for each donor are next to each other - this an important element of the default handling of the paired analysis in R. The dataframe should have the donors in the same order for the 0 and 4 ppm data.</p>
 <div class="sourceCode" id="cb420"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb420-1"><a href="two-group-comparisons-and-visualizations.html#cb420-1" tabindex="-1"></a><span class="fu">datatable</span>(biomarker_data_0vs4)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-33d9fa8f6519f319168e" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-33d9fa8f6519f319168e">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32"],["D10","D10","D11","D11","D12","D12","D13","D13","D14","D14","D15","D15","D16","D16","D1","D1","D2","D2","D3","D3","D4","D4","D5","D5","D6","D6","D7","D7","D8","D8","D9","D9"],["0","4","0","4","0","4","0","4","0","4","0","4","0","4","0","4","0","4","0","4","0","4","0","4","0","4","0","4","0","4","0","4"],[3.71808758396052,3.6724253419715,3.83491342970104,3.81762325751143,3.36036427654347,3.10097764772482,3.35755200461808,3.69041709553247,3.31614574229336,3.83086375675176,3.42894634476948,3.6915341649192,3.38681094647222,3.69376571221778,3.76871365703048,3.50589092972996,3.53480866123175,3.75167794568758,3.58014548442338,3.39643353125099,3.52732060787705,4.13996056954546,3.16671544496642,3.799087306074,2.86195536414487,3.64616265715789,3.04614178164472,3.42088657497553,3.37712374912949,3.8972404255748,3.21878116778407,3.7279204545632],[8.68422263328295,7.55028504908841,10.2222394489088,8.03815081026938,9.27052894238072,7.61889934790406,8.15790268278964,4.10265813136374,7.56094417806009,6.50969584193339,9.254131033541951,8.202809492125009,8.68422263328295,7.71163256182322,10.1099481540673,10.1498741117182,8.88962564051876,4.87479696589866,9.23160520814165,9.560179957764181,8.34837407519116,7.70355756192103,3.9429835981871,5.17871464117544,7.43212427264156,8.36382813184704,8.12861303870471,8.90926307937532,11.4238722598684,8.52575569282949,7.01077983875324,4.70099449416827],[15.0940690235359,16.7241460980745,15.0028356680746,16.0996069083821,13.9520295507825,16.3560077655342,14.30563571243,16.6038859325635,13.9567554773633,16.3613300730261,14.4305440239947,16.5368999542995,14.3291992954213,16.3717050181953,15.1620702745062,16.9008155214109,14.6079451777076,17.2934204566841,14.5987126147926,17.1264130896721,14.542155252148,15.3426552970282,14.1001424218123,16.3510586187167,13.4034767756747,16.2848127161883,13.5861396108168,17.0361832734254,14.1156143059506,15.1484821432167,14.3291992954213,16.3449160294487],[0.189033824390017,1.0703893278914,2.3305584000308,1.64154602908752,0.526068811667588,0.333423733725192,1.24488705912353,0.855989697308481,1.06350294230616,1.37295209791183,1.27500704749987,1.16992500144231,0.8155754288625729,1.13750352374994,0.367371065648529,0.454175893185802,1.35614381022528,1.40599235967584,0.31034012061215,0.321928094887362,0.879705766282288,2.03210084316702,0.933572638261024,0.84799690655495,1.86393845042397,0.443606651475615,0.650764559116902,1.17632277264046,1.01435529297707,1.57531233068744,2.51349074558812,1.47508488294878],[1.73552217729654,2.83390207666916,3.23572705983806,3.48284828306847,1.63226821549951,1.57046293102604,1.32192809488736,2.59931779369823,1.91456452349394,3.27202318906105,2.08746284125034,2.94485844580754,1.83187724119167,2.93168305705981,1.86393845042397,2.88362081628567,1.83187724119167,2.98913900736823,2.52606881166759,2.06695024392463,2.47508488294878,3.73226919950145,2.20476675065461,2.85798099512757,1.790772037862,2.16671544496642,1.56559717585423,2.75702324650746,2.53106949272595,3.47118746038699,1.0703893278914,3.08406426478847],[10.4475275409266,11.2458081856384,9.02901108665907,9.316688823436939,9.428025110276749,7.3039633562453,9.247642742898019,10.8557639471917,10.6537676442859,10.9341110109735,9.933439857942711,10.646036617491,9.797969484059131,10.8557639471917,9.696480751820189,11.0193335777127,10.534828110721,11.031563370024,9.35887093968562,11.0217850685075,8.283551423174311,8.973180744513019,9.425320822759881,10.7035402591663,10.7662472988462,9.89590846480624,10.3333342781343,11.7319596151152,9.92478281821723,9.40224465525408,9.797969484059131,10.9299462991725]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Donor<\/th>\n      <th>Dose<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[3,4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Donor","targets":1},{"name":"Dose","targets":2},{"name":"IL1B","targets":3},{"name":"IL6","targets":4},{"name":"IL8","targets":5},{"name":"IL10","targets":6},{"name":"TNFa","targets":7},{"name":"VEGF","targets":8}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-536342f6cb312f798270" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-536342f6cb312f798270">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32"],["D10","D10","D11","D11","D12","D12","D13","D13","D14","D14","D15","D15","D16","D16","D1","D1","D2","D2","D3","D3","D4","D4","D5","D5","D6","D6","D7","D7","D8","D8","D9","D9"],["0","4","0","4","0","4","0","4","0","4","0","4","0","4","0","4","0","4","0","4","0","4","0","4","0","4","0","4","0","4","0","4"],[3.71808758396052,3.6724253419715,3.83491342970104,3.81762325751143,3.36036427654347,3.10097764772482,3.35755200461808,3.69041709553247,3.31614574229336,3.83086375675176,3.42894634476948,3.6915341649192,3.38681094647222,3.69376571221778,3.76871365703048,3.50589092972996,3.53480866123175,3.75167794568758,3.58014548442338,3.39643353125099,3.52732060787705,4.13996056954546,3.16671544496642,3.799087306074,2.86195536414487,3.64616265715789,3.04614178164472,3.42088657497553,3.37712374912949,3.8972404255748,3.21878116778407,3.7279204545632],[8.68422263328295,7.55028504908841,10.2222394489088,8.03815081026938,9.27052894238072,7.61889934790406,8.15790268278964,4.10265813136374,7.56094417806009,6.50969584193339,9.254131033541951,8.202809492125009,8.68422263328295,7.71163256182322,10.1099481540673,10.1498741117182,8.88962564051876,4.87479696589866,9.23160520814165,9.560179957764181,8.34837407519116,7.70355756192103,3.9429835981871,5.17871464117544,7.43212427264156,8.36382813184704,8.12861303870471,8.90926307937532,11.4238722598684,8.52575569282949,7.01077983875324,4.70099449416827],[15.0940690235359,16.7241460980745,15.0028356680746,16.0996069083821,13.9520295507825,16.3560077655342,14.30563571243,16.6038859325635,13.9567554773633,16.3613300730261,14.4305440239947,16.5368999542995,14.3291992954213,16.3717050181953,15.1620702745062,16.9008155214109,14.6079451777076,17.2934204566841,14.5987126147926,17.1264130896721,14.542155252148,15.3426552970282,14.1001424218123,16.3510586187167,13.4034767756747,16.2848127161883,13.5861396108168,17.0361832734254,14.1156143059506,15.1484821432167,14.3291992954213,16.3449160294487],[0.189033824390017,1.0703893278914,2.3305584000308,1.64154602908752,0.526068811667588,0.333423733725192,1.24488705912353,0.855989697308481,1.06350294230616,1.37295209791183,1.27500704749987,1.16992500144231,0.8155754288625729,1.13750352374994,0.367371065648529,0.454175893185802,1.35614381022528,1.40599235967584,0.31034012061215,0.321928094887362,0.879705766282288,2.03210084316702,0.933572638261024,0.84799690655495,1.86393845042397,0.443606651475615,0.650764559116902,1.17632277264046,1.01435529297707,1.57531233068744,2.51349074558812,1.47508488294878],[1.73552217729654,2.83390207666916,3.23572705983806,3.48284828306847,1.63226821549951,1.57046293102604,1.32192809488736,2.59931779369823,1.91456452349394,3.27202318906105,2.08746284125034,2.94485844580754,1.83187724119167,2.93168305705981,1.86393845042397,2.88362081628567,1.83187724119167,2.98913900736823,2.52606881166759,2.06695024392463,2.47508488294878,3.73226919950145,2.20476675065461,2.85798099512757,1.790772037862,2.16671544496642,1.56559717585423,2.75702324650746,2.53106949272595,3.47118746038699,1.0703893278914,3.08406426478847],[10.4475275409266,11.2458081856384,9.02901108665907,9.316688823436939,9.428025110276749,7.3039633562453,9.247642742898019,10.8557639471917,10.6537676442859,10.9341110109735,9.933439857942711,10.646036617491,9.797969484059131,10.8557639471917,9.696480751820189,11.0193335777127,10.534828110721,11.031563370024,9.35887093968562,11.0217850685075,8.283551423174311,8.973180744513019,9.425320822759881,10.7035402591663,10.7662472988462,9.89590846480624,10.3333342781343,11.7319596151152,9.92478281821723,9.40224465525408,9.797969484059131,10.9299462991725]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Donor<\/th>\n      <th>Dose<\/th>\n      <th>IL1B<\/th>\n      <th>IL6<\/th>\n      <th>IL8<\/th>\n      <th>IL10<\/th>\n      <th>TNFa<\/th>\n      <th>VEGF<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":[3,4,5,6,7,8]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Donor","targets":1},{"name":"Dose","targets":2},{"name":"IL1B","targets":3},{"name":"IL6","targets":4},{"name":"IL8","targets":5},{"name":"IL10","targets":6},{"name":"TNFa","targets":7},{"name":"VEGF","targets":8}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>We can now run the same type of loop that we ran before, changing the independent variable in the formula to <code>~ Dose</code> and adding <code>paired = TRUE</code> to the <code>wilcox_test()</code> function.</p>
 <div class="sourceCode" id="cb421"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb421-1"><a href="two-group-comparisons-and-visualizations.html#cb421-1" tabindex="-1"></a><span class="co"># Create a vector with the names of the variables you want to run the test on</span></span>
 <span id="cb421-2"><a href="two-group-comparisons-and-visualizations.html#cb421-2" tabindex="-1"></a>endpoints <span class="ot">&lt;-</span> <span class="fu">colnames</span>(biomarker_data_0vs4 <span class="sc">%&gt;%</span> <span class="fu">select</span>(IL1B<span class="sc">:</span>VEGF)) </span>
@@ -629,8 +638,8 @@ <h2>Paired Test Example<a href="two-group-comparisons-and-visualizations.html#pa
 <span id="cb423-4"><a href="two-group-comparisons-and-visualizations.html#cb423-4" tabindex="-1"></a>  <span class="fu">rename</span>(<span class="st">&quot;Variable&quot;</span> <span class="ot">=</span> <span class="st">&quot;.y.&quot;</span>, <span class="st">&quot;P-Value&quot;</span> <span class="ot">=</span> <span class="st">&quot;p&quot;</span>)</span>
 <span id="cb423-5"><a href="two-group-comparisons-and-visualizations.html#cb423-5" tabindex="-1"></a></span>
 <span id="cb423-6"><a href="two-group-comparisons-and-visualizations.html#cb423-6" tabindex="-1"></a><span class="fu">datatable</span>(dose_wilcoxres)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-5ab3bfc745b3cccdc15c" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-5ab3bfc745b3cccdc15c">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6"],["IL1B","IL6","IL8","IL10","TNFa","VEGF"],["7.63e-03","1.55e-02","3.05e-05","9.00e-01","3.05e-04","3.86e-02"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Variable<\/th>\n      <th>P-Value<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Variable","targets":1},{"name":"P-Value","targets":2}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-c785def214d390c3650f" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-c785def214d390c3650f">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6"],["IL1B","IL6","IL8","IL10","TNFa","VEGF"],["7.63e-03","1.55e-02","3.05e-05","9.00e-01","3.05e-04","3.86e-02"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Variable<\/th>\n      <th>P-Value<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Variable","targets":1},{"name":"P-Value","targets":2}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <div class="question">
 <p><i>With this, we can answer <strong>Environmental Health Question #2</strong>:</i></p>
 <p>Are there significant differences in inflammatory biomarker concentrations between cells exposed to 0 and 4 ppm acrolein?</p>
@@ -692,8 +701,8 @@ <h3>Multiple plots<a href="two-group-comparisons-and-visualizations.html#multipl
 <span id="cb427-2"><a href="two-group-comparisons-and-visualizations.html#cb427-2" tabindex="-1"></a>  <span class="fu">pivot_longer</span>(<span class="sc">-</span><span class="fu">c</span>(Donor, Dose), <span class="at">names_to =</span> <span class="st">&quot;variable&quot;</span>, <span class="at">values_to =</span> <span class="st">&quot;value&quot;</span>)</span>
 <span id="cb427-3"><a href="two-group-comparisons-and-visualizations.html#cb427-3" tabindex="-1"></a></span>
 <span id="cb427-4"><a href="two-group-comparisons-and-visualizations.html#cb427-4" tabindex="-1"></a><span class="fu">datatable</span>(biomarker_data_0vs4_long)</span></code></pre></div>
-<div class="datatables html-widget html-fill-item" id="htmlwidget-ef842e085dd072377902" style="width:100%;height:auto;"></div>
-<script type="application/json" data-for="htmlwidget-ef842e085dd072377902">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80","81","82","83","84","85","86","87","88","89","90","91","92","93","94","95","96","97","98","99","100","101","102","103","104","105","106","107","108","109","110","111","112","113","114","115","116","117","118","119","120","121","122","123","124","125","126","127","128","129","130","131","132","133","134","135","136","137","138","139","140","141","142","143","144","145","146","147","148","149","150","151","152","153","154","155","156","157","158","159","160","161","162","163","164","165","166","167","168","169","170","171","172","173","174","175","176","177","178","179","180","181","182","183","184","185","186","187","188","189","190","191","192"],["D10","D10","D10","D10","D10","D10","D10","D10","D10","D10","D10","D10","D11","D11","D11","D11","D11","D11","D11","D11","D11","D11","D11","D11","D12","D12","D12","D12","D12","D12","D12","D12","D12","D12","D12","D12","D13","D13","D13","D13","D13","D13","D13","D13","D13","D13","D13","D13","D14","D14","D14","D14","D14","D14","D14","D14","D14","D14","D14","D14","D15","D15","D15","D15","D15","D15","D15","D15","D15","D15","D15","D15","D16","D16","D16","D16","D16","D16","D16","D16","D16","D16","D16","D16","D1","D1","D1","D1","D1","D1","D1","D1","D1","D1","D1","D1","D2","D2","D2","D2","D2","D2","D2","D2","D2","D2","D2","D2","D3","D3","D3","D3","D3","D3","D3","D3","D3","D3","D3","D3","D4","D4","D4","D4","D4","D4","D4","D4","D4","D4","D4","D4","D5","D5","D5","D5","D5","D5","D5","D5","D5","D5","D5","D5","D6","D6","D6","D6","D6","D6","D6","D6","D6","D6","D6","D6","D7","D7","D7","D7","D7","D7","D7","D7","D7","D7","D7","D7","D8","D8","D8","D8","D8","D8","D8","D8","D8","D8","D8","D8","D9","D9","D9","D9","D9","D9","D9","D9","D9","D9","D9","D9"],["0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4"],["IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF"],[3.71808758396052,8.68422263328295,15.0940690235359,0.189033824390017,1.73552217729654,10.4475275409266,3.6724253419715,7.55028504908841,16.7241460980745,1.0703893278914,2.83390207666916,11.2458081856384,3.83491342970104,10.2222394489088,15.0028356680746,2.3305584000308,3.23572705983806,9.02901108665907,3.81762325751143,8.03815081026938,16.0996069083821,1.64154602908752,3.48284828306847,9.316688823436939,3.36036427654347,9.27052894238072,13.9520295507825,0.526068811667588,1.63226821549951,9.428025110276749,3.10097764772482,7.61889934790406,16.3560077655342,0.333423733725192,1.57046293102604,7.3039633562453,3.35755200461808,8.15790268278964,14.30563571243,1.24488705912353,1.32192809488736,9.247642742898019,3.69041709553247,4.10265813136374,16.6038859325635,0.855989697308481,2.59931779369823,10.8557639471917,3.31614574229336,7.56094417806009,13.9567554773633,1.06350294230616,1.91456452349394,10.6537676442859,3.83086375675176,6.50969584193339,16.3613300730261,1.37295209791183,3.27202318906105,10.9341110109735,3.42894634476948,9.254131033541951,14.4305440239947,1.27500704749987,2.08746284125034,9.933439857942711,3.6915341649192,8.202809492125009,16.5368999542995,1.16992500144231,2.94485844580754,10.646036617491,3.38681094647222,8.68422263328295,14.3291992954213,0.8155754288625729,1.83187724119167,9.797969484059131,3.69376571221778,7.71163256182322,16.3717050181953,1.13750352374994,2.93168305705981,10.8557639471917,3.76871365703048,10.1099481540673,15.1620702745062,0.367371065648529,1.86393845042397,9.696480751820189,3.50589092972996,10.1498741117182,16.9008155214109,0.454175893185802,2.88362081628567,11.0193335777127,3.53480866123175,8.88962564051876,14.6079451777076,1.35614381022528,1.83187724119167,10.534828110721,3.75167794568758,4.87479696589866,17.2934204566841,1.40599235967584,2.98913900736823,11.031563370024,3.58014548442338,9.23160520814165,14.5987126147926,0.31034012061215,2.52606881166759,9.35887093968562,3.39643353125099,9.560179957764181,17.1264130896721,0.321928094887362,2.06695024392463,11.0217850685075,3.52732060787705,8.34837407519116,14.542155252148,0.879705766282288,2.47508488294878,8.283551423174311,4.13996056954546,7.70355756192103,15.3426552970282,2.03210084316702,3.73226919950145,8.973180744513019,3.16671544496642,3.9429835981871,14.1001424218123,0.933572638261024,2.20476675065461,9.425320822759881,3.799087306074,5.17871464117544,16.3510586187167,0.84799690655495,2.85798099512757,10.7035402591663,2.86195536414487,7.43212427264156,13.4034767756747,1.86393845042397,1.790772037862,10.7662472988462,3.64616265715789,8.36382813184704,16.2848127161883,0.443606651475615,2.16671544496642,9.89590846480624,3.04614178164472,8.12861303870471,13.5861396108168,0.650764559116902,1.56559717585423,10.3333342781343,3.42088657497553,8.90926307937532,17.0361832734254,1.17632277264046,2.75702324650746,11.7319596151152,3.37712374912949,11.4238722598684,14.1156143059506,1.01435529297707,2.53106949272595,9.92478281821723,3.8972404255748,8.52575569282949,15.1484821432167,1.57531233068744,3.47118746038699,9.40224465525408,3.21878116778407,7.01077983875324,14.3291992954213,2.51349074558812,1.0703893278914,9.797969484059131,3.7279204545632,4.70099449416827,16.3449160294487,1.47508488294878,3.08406426478847,10.9299462991725]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Donor<\/th>\n      <th>Dose<\/th>\n      <th>variable<\/th>\n      <th>value<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":4},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Donor","targets":1},{"name":"Dose","targets":2},{"name":"variable","targets":3},{"name":"value","targets":4}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
+<div class="datatables html-widget html-fill-item" id="htmlwidget-084f78e73bdb6c3301a0" style="width:100%;height:auto;"></div>
+<script type="application/json" data-for="htmlwidget-084f78e73bdb6c3301a0">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31","32","33","34","35","36","37","38","39","40","41","42","43","44","45","46","47","48","49","50","51","52","53","54","55","56","57","58","59","60","61","62","63","64","65","66","67","68","69","70","71","72","73","74","75","76","77","78","79","80","81","82","83","84","85","86","87","88","89","90","91","92","93","94","95","96","97","98","99","100","101","102","103","104","105","106","107","108","109","110","111","112","113","114","115","116","117","118","119","120","121","122","123","124","125","126","127","128","129","130","131","132","133","134","135","136","137","138","139","140","141","142","143","144","145","146","147","148","149","150","151","152","153","154","155","156","157","158","159","160","161","162","163","164","165","166","167","168","169","170","171","172","173","174","175","176","177","178","179","180","181","182","183","184","185","186","187","188","189","190","191","192"],["D10","D10","D10","D10","D10","D10","D10","D10","D10","D10","D10","D10","D11","D11","D11","D11","D11","D11","D11","D11","D11","D11","D11","D11","D12","D12","D12","D12","D12","D12","D12","D12","D12","D12","D12","D12","D13","D13","D13","D13","D13","D13","D13","D13","D13","D13","D13","D13","D14","D14","D14","D14","D14","D14","D14","D14","D14","D14","D14","D14","D15","D15","D15","D15","D15","D15","D15","D15","D15","D15","D15","D15","D16","D16","D16","D16","D16","D16","D16","D16","D16","D16","D16","D16","D1","D1","D1","D1","D1","D1","D1","D1","D1","D1","D1","D1","D2","D2","D2","D2","D2","D2","D2","D2","D2","D2","D2","D2","D3","D3","D3","D3","D3","D3","D3","D3","D3","D3","D3","D3","D4","D4","D4","D4","D4","D4","D4","D4","D4","D4","D4","D4","D5","D5","D5","D5","D5","D5","D5","D5","D5","D5","D5","D5","D6","D6","D6","D6","D6","D6","D6","D6","D6","D6","D6","D6","D7","D7","D7","D7","D7","D7","D7","D7","D7","D7","D7","D7","D8","D8","D8","D8","D8","D8","D8","D8","D8","D8","D8","D8","D9","D9","D9","D9","D9","D9","D9","D9","D9","D9","D9","D9"],["0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4","0","0","0","0","0","0","4","4","4","4","4","4"],["IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF","IL1B","IL6","IL8","IL10","TNFa","VEGF"],[3.71808758396052,8.68422263328295,15.0940690235359,0.189033824390017,1.73552217729654,10.4475275409266,3.6724253419715,7.55028504908841,16.7241460980745,1.0703893278914,2.83390207666916,11.2458081856384,3.83491342970104,10.2222394489088,15.0028356680746,2.3305584000308,3.23572705983806,9.02901108665907,3.81762325751143,8.03815081026938,16.0996069083821,1.64154602908752,3.48284828306847,9.316688823436939,3.36036427654347,9.27052894238072,13.9520295507825,0.526068811667588,1.63226821549951,9.428025110276749,3.10097764772482,7.61889934790406,16.3560077655342,0.333423733725192,1.57046293102604,7.3039633562453,3.35755200461808,8.15790268278964,14.30563571243,1.24488705912353,1.32192809488736,9.247642742898019,3.69041709553247,4.10265813136374,16.6038859325635,0.855989697308481,2.59931779369823,10.8557639471917,3.31614574229336,7.56094417806009,13.9567554773633,1.06350294230616,1.91456452349394,10.6537676442859,3.83086375675176,6.50969584193339,16.3613300730261,1.37295209791183,3.27202318906105,10.9341110109735,3.42894634476948,9.254131033541951,14.4305440239947,1.27500704749987,2.08746284125034,9.933439857942711,3.6915341649192,8.202809492125009,16.5368999542995,1.16992500144231,2.94485844580754,10.646036617491,3.38681094647222,8.68422263328295,14.3291992954213,0.8155754288625729,1.83187724119167,9.797969484059131,3.69376571221778,7.71163256182322,16.3717050181953,1.13750352374994,2.93168305705981,10.8557639471917,3.76871365703048,10.1099481540673,15.1620702745062,0.367371065648529,1.86393845042397,9.696480751820189,3.50589092972996,10.1498741117182,16.9008155214109,0.454175893185802,2.88362081628567,11.0193335777127,3.53480866123175,8.88962564051876,14.6079451777076,1.35614381022528,1.83187724119167,10.534828110721,3.75167794568758,4.87479696589866,17.2934204566841,1.40599235967584,2.98913900736823,11.031563370024,3.58014548442338,9.23160520814165,14.5987126147926,0.31034012061215,2.52606881166759,9.35887093968562,3.39643353125099,9.560179957764181,17.1264130896721,0.321928094887362,2.06695024392463,11.0217850685075,3.52732060787705,8.34837407519116,14.542155252148,0.879705766282288,2.47508488294878,8.283551423174311,4.13996056954546,7.70355756192103,15.3426552970282,2.03210084316702,3.73226919950145,8.973180744513019,3.16671544496642,3.9429835981871,14.1001424218123,0.933572638261024,2.20476675065461,9.425320822759881,3.799087306074,5.17871464117544,16.3510586187167,0.84799690655495,2.85798099512757,10.7035402591663,2.86195536414487,7.43212427264156,13.4034767756747,1.86393845042397,1.790772037862,10.7662472988462,3.64616265715789,8.36382813184704,16.2848127161883,0.443606651475615,2.16671544496642,9.89590846480624,3.04614178164472,8.12861303870471,13.5861396108168,0.650764559116902,1.56559717585423,10.3333342781343,3.42088657497553,8.90926307937532,17.0361832734254,1.17632277264046,2.75702324650746,11.7319596151152,3.37712374912949,11.4238722598684,14.1156143059506,1.01435529297707,2.53106949272595,9.92478281821723,3.8972404255748,8.52575569282949,15.1484821432167,1.57531233068744,3.47118746038699,9.40224465525408,3.21878116778407,7.01077983875324,14.3291992954213,2.51349074558812,1.0703893278914,9.797969484059131,3.7279204545632,4.70099449416827,16.3449160294487,1.47508488294878,3.08406426478847,10.9299462991725]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Donor<\/th>\n      <th>Dose<\/th>\n      <th>variable<\/th>\n      <th>value<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"columnDefs":[{"className":"dt-right","targets":4},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Donor","targets":1},{"name":"Dose","targets":2},{"name":"variable","targets":3},{"name":"value","targets":4}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>
 <p>Then, we can use similar code to what we used to make our single graph, with a few modifications to plot multiple panels simultaneously and adjust the style of the plot. Although it is beyond the scope of this module to explain the mechanics of each line of code, here are a few specific things to note about the code below that may be helpful when constructing similar plots:</p>
 <ul>
 <li>To create the plot with all six endpoints instead of just one, we:
diff --git a/unsupervised-machine-learning-part-1-k-means-clustering-pca.html b/unsupervised-machine-learning-part-1-k-means-clustering-pca.html
index 3f0ca2c..3c9805a 100644
--- a/unsupervised-machine-learning-part-1-k-means-clustering-pca.html
+++ b/unsupervised-machine-learning-part-1-k-means-clustering-pca.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">
diff --git a/unsupervised-machine-learning-part-2-additional-clustering-applications.html b/unsupervised-machine-learning-part-2-additional-clustering-applications.html
index 35ede8e..53dfe47 100644
--- a/unsupervised-machine-learning-part-2-additional-clustering-applications.html
+++ b/unsupervised-machine-learning-part-2-additional-clustering-applications.html
@@ -65,6 +65,15 @@
 <link href="libs/lightable-0.0.1/lightable.css" rel="stylesheet" />
 <link href="libs/tabwid-1.1.3/tabwid.css" rel="stylesheet" />
 <script src="libs/tabwid-1.1.3/tabwid.js"></script>
+<!-- Google tag (gtag.js) -->
+<script async src="https://www.googletagmanager.com/gtag/js?id=G-FGFKVX2X3G"></script>
+<script>
+  window.dataLayer = window.dataLayer || [];
+  function gtag(){dataLayer.push(arguments);}
+  gtag('js', new Date());
+
+  gtag('config', 'G-FGFKVX2X3G');
+</script>
 
 
 <style type="text/css">