-
Notifications
You must be signed in to change notification settings - Fork 0
/
sim_data_noise.py
237 lines (190 loc) · 7.49 KB
/
sim_data_noise.py
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
from deamidation.DSevidence import deamidationMatrix
import deamidation.deamidCorr as dc
import deamidation.reactionRates as rr
import deamidation.accFunctions as af
import matplotlib.pyplot as plt
import json
import numpy as np
from scipy.stats import truncnorm
from scipy.stats import truncexpon
from scipy.stats import spearmanr
import os
import warnings
warnings.filterwarnings("ignore")
plt.style.use('seaborn-dark-palette')
out_dir = '/home/ismael/palaeoproteomics/out/'
def create_deamid_mat(tripeps, D, t):
trps_data = np.array(
[tuple(['NA', 'NA', trp, 0, 0, trp]) for trp in tripeps],
dtype=[
('prot_id', '<U60'),
('prot_name', '<U10'),
('tripep', 'U3'),
('position', 'i4'),
('corr_pos', 'i4'),
('string', '<U45')
]
)
counts = None
rel_ints = None
Ydata = {'times': t}
sim_mat = deamidationMatrix(
D=D, Ydata=Ydata, counts=counts,
rel_ints=rel_ints, trps_data=trps_data
)
return sim_mat
def sigma_spearman_plot(corrs, sigmas, aa, out_dir, base_name):
fig = plt.figure(figsize=(9.6, 7.2), dpi=300)
for i in range(len(corrs)):
ax = fig.add_subplot(2, 1, i+1)
ax.boxplot(corrs[i], positions=sigmas, widths=0.03)
ax.set_xticks([0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1])
ax.set_xticklabels([0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1],
size='medium')
ax.set_xlim(-0.05, 1.05)
ax.set_xlabel(r'$\sigma$', size='xx-large')
ax.set_ylabel('Spearman\'s '+r'$\rho$', size='xx-large')
ax.set_title('X'+aa[i]+'Z')
if i == 0:
ax.xaxis.set_visible(False)
plt.savefig(out_dir + base_name + '_sigmaVSrho.png')
plt.close()
N_properties, Q_properties = af.readHalftimes('data/N_properties.json',
'data/Q_properties.json')
aa_properties = N_properties.copy()
aa_properties.update(Q_properties)
tripepsN = np.array([
'GNP','GNI','GNV','GNR','GNK','GNT','GND','GNA','GNS',
'ENG','VNG','PNG','HNG','LNG','ANG','KNG'
])
tripepsQ = np.array([
'GQP','PQP','GQD','GQR','PQE','GQA','GQT','GQM',
'DQG','EQG','SQG','LQG','FQG','KQG','VQG','PQG','IQG','AQG'
])
kN = np.array([aa_properties[trp]['lambda']
if trp in aa_properties else -1
for trp in tripepsN])
kQ = np.array([aa_properties[trp]['lambda']
if trp in aa_properties else -1
for trp in tripepsQ])
idxN = np.argsort(kN)
idxQ = np.argsort(kQ)
tripepsN = tripepsN[idxN]
kN = kN[idxN]
tripepsQ = tripepsQ[idxQ]
kQ = kQ[idxQ]
N = 20000
TmaxN = np.array([-np.log(1-0.99)/k for k in kN])
TmaxQ = np.array([-np.log(1-0.99)/k for k in kQ])
tN = [truncexpon.rvs(2.5, 0, tm/2.5, int(N/len(tripepsN))) for tm in TmaxN]
tN = np.concatenate(tN)
tQ = [truncexpon.rvs(2.5, 0, tm/2.5, int(N/len(tripepsQ))) for tm in TmaxQ]
tQ = np.concatenate(tQ)
sim_lim = 1e-7
deamid_cutoff = 0
# different levels of noise
sigmas = [1,0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2, 0.15, 0.1, 0.05, 0.005]
bootstraps = 500 # For each level of noise, repeat expetiment
bootstrap_size = 100
corrsN_1 = []; corrsQ_1 = [] # sigmas x bootstraps
corrsN_2 = []; corrsQ_2 = []
corrsN_3 = []; corrsQ_3 = []
for s in sigmas:
print('Simulating with sigma = {}'.format(s))
sigmasN = np.repeat(s, len(tripepsN))
sigmasQ = np.repeat(s, len(tripepsQ))
corrsN_1_b = []; corrsQ_1_b = []
corrsN_2_b = []; corrsQ_2_b = []
corrsN_3_b = []; corrsQ_3_b = []
b = 0 # Bootstrap counter
while b < bootstraps:
# Samples times
tN_b = tN[np.random.randint(tN.shape[0], size=bootstrap_size)]
tQ_b = tQ[np.random.randint(tQ.shape[0], size=bootstrap_size)]
DN_b = rr.simulate_deamidation(kN, sigmasN, tN_b, sim_lim)
DQ_b = rr.simulate_deamidation(kQ, sigmasQ, tQ_b, sim_lim)
RN = rr.calc_Ri(DN_b, num_tol=deamid_cutoff)
RQ = rr.calc_Ri(DQ_b, num_tol=deamid_cutoff)
argsortN = np.argsort([np.median(v) for v in RN])
argsortQ = np.argsort([np.median(v) for v in RQ])
simple_rankN = rr.calc_rankRj(DN_b, num_tol=deamid_cutoff)
simple_rankQ = rr.calc_rankRj(DQ_b, num_tol=deamid_cutoff)
values_rankN = rr.calc_rankj_values(DN_b, num_tol=deamid_cutoff)
values_rankQ = rr.calc_rankj_values(DQ_b, num_tol=deamid_cutoff)
sort_rankN = np.argsort([np.median(v) for v in RN])
sort_rankQ = np.argsort([np.median(v) for v in RQ])
# Use spearman correlation as "sortness" measure
corrN = spearmanr(argsortN, np.array(range(len(argsortN)))).correlation
corrQ = spearmanr(argsortQ, np.array(range(len(argsortQ)))).correlation
corrsN_1_b.append(corrN)
corrsQ_1_b.append(corrQ)
corrN = spearmanr(simple_rankN, np.array(range(len(simple_rankN)))).correlation
corrQ = spearmanr(simple_rankQ, np.array(range(len(simple_rankQ)))).correlation
corrsN_2_b.append(corrN)
corrsQ_2_b.append(corrQ)
corrN = spearmanr(sort_rankN, np.array(range(len(sort_rankN)))).correlation
corrQ = spearmanr(sort_rankQ, np.array(range(len(sort_rankQ)))).correlation
corrsN_3_b.append(corrN)
corrsQ_3_b.append(corrQ)
b += 1
corrsN_1.append(corrsN_1_b); corrsQ_1.append(corrsQ_1_b)
corrsN_2.append(corrsN_2_b); corrsQ_2.append(corrsQ_2_b)
corrsN_3.append(corrsN_3_b); corrsQ_3.append(corrsQ_3_b)
corrs_1 = [corrsN_1, corrsQ_1]
corrs_2 = [corrsN_2, corrsQ_2]
corrs_3 = [corrsN_3, corrsQ_3]
aa = ['N', 'Q']
# sigma_spearman_plot(corrs_1, sigmas, aa, out_dir, 'Ri')
# sigma_spearman_plot(corrs_2, sigmas, aa, out_dir, 'simple_rank')
# sigma_spearman_plot(corrs_3, sigmas, aa, out_dir, 'values_rank')
print('Done')
####### Simulate data at 0.15 noise
s = 0.15
print('Simulating with sigma = {}'.format(s))
sigmasN = np.repeat(s, len(tripepsN))
sigmasQ = np.repeat(s, len(tripepsQ))
# K, sigmas, Tmean, Tstd, Tmax, N, tol
tN_s = tN[np.random.randint(tN.shape[0], size=2000)]
DN = rr.simulate_deamidation(kN, sigmasN, tN_s, sim_lim)
sim_matN = create_deamid_mat(tripepsN, DN, tN_s)
RN = rr.calc_Ri(DN, num_tol=deamid_cutoff)
rankN = rr.calc_rankj_values(DN, num_tol=0)
tQ_s = tQ[np.random.randint(tQ.shape[0], size=2000)]
DQ = rr.simulate_deamidation(kQ, sigmasQ, tQ_s, sim_lim)
sim_matQ = create_deamid_mat(tripepsQ, DQ, tQ_s)
RQ = rr.calc_Ri(DQ, num_tol=deamid_cutoff)
rankQ = rr.calc_rankj_values(DQ, num_tol=0)
rr.Rlambda_distr(rankN, trps_data=sim_matN.trps_data,
sort_by='mean',
path=out_dir, base_name='simN_1_rank',
log=False)
rr.Rlambda_distr(rankQ, trps_data=sim_matQ.trps_data,
sort_by='mean',
path=out_dir, base_name='simQ_1_rank',
log=False)
print('Creating multiscatter...')
dc.multiscatter(
sim_matN, hts=kN,
t='norm', path=out_dir, base_name='simN_1',
low_counts=None, fontsize=10, fontpos=[0.10,0.35],
reg=False
)
dc.multiscatter(
sim_matQ, hts=kQ,
t='norm', path=out_dir, base_name='simQ_1',
low_counts=None, fontsize=10, fontpos=[0.10,0.35],
reg=False
)
print('Done')
print('Creating violin plots...')
print('For N...')
rr.Rlambda_distr(RN, trps_data=sim_matN.trps_data,
sort_by='median',
path=out_dir, base_name='simN_1',
log=False)
print('For Q...')
rr.Rlambda_distr(RQ, trps_data=sim_matQ.trps_data,
sort_by='median',
path=out_dir, base_name='simQ_1',
log=False)
print('Done')