-
Notifications
You must be signed in to change notification settings - Fork 0
/
TCGA_workflow.R
332 lines (235 loc) · 11.1 KB
/
TCGA_workflow.R
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
# TCGA Cancer -- Workflow
# Required packages
library(TopKSignal)
#library(ggplot2)
library(limma) # installed from Bioconductor
print("---")
#### SUBFUNCTIONS ###############################################################
#################################################################################
readTCGA <- function(cancertype=c("kidney"), loc="~/TCGA_data/NAR Data/"){
TCGA_death <- list()
TCGA_surv <- list()
for(xx in 1:length(cancertype)){
ctype <- cancertype[xx]
# NAR TCGA -------------------------
path1 <- paste(loc, ctype, "/exp", sep="")
path2 <- paste(loc, ctype, "/survival", sep="")
tcga_exp <- read.table(path1)
tcga_survival <- read.table(path2)
# Fix names
colnames(tcga_exp) <- gsub(".", "-", colnames(tcga_exp), fixed=TRUE)
# Get the Death and Survived Outcomes
death_ids <- which(tcga_survival[,3]==1)
surv_ids <- which(tcga_survival[,3]==0)
death_pat <- as.character(tcga_survival[,1][death_ids])
surv_pat <- as.character(tcga_survival[,1][surv_ids])
# death
iii <- match(death_pat, colnames(tcga_exp))
iii <- iii[!is.na(iii)]
# survived
jjj <- match(surv_pat, colnames(tcga_exp))
jjj <- jjj[!is.na(jjj)]
tcga_exp_death <- tcga_exp[,iii]
tcga_exp_surv <- tcga_exp[,jjj]
# fix the rownames
rnames <- rownames(tcga_exp_death)
rnames <- gsub(".", "|", rnames, fixed=TRUE)
rownames(tcga_exp_death) <- rnames
rnames <- rownames(tcga_exp_surv)
rnames <- gsub(".", "|", rnames, fixed=TRUE)
rownames(tcga_exp_surv) <- rnames
TCGA_death[[xx]] <- tcga_exp_death
TCGA_surv[[xx]] <- tcga_exp_surv
}
# Concatenate
## Intersection of genes across cancer types
if(length(cancertype)==1){
return(list(FULL_TCGA_DEATH=TCGA_death[[1]], FULL_TCGA_SURV=TCGA_surv[[1]]))
}
genes_inter <- NULL
for(xx in 1:(length(cancertype)-1)){
genes_inter <- c(genes_inter,intersect(rownames(TCGA_death[[xx]]),
rownames(TCGA_death[[xx+1]])))
}
# Sample the same number of patients accross the cancer types
n.samples <- sapply(TCGA_death, function(x){dim(x)[2]})
n.samples <- min(n.samples)
# Concatenate cancer types (exp) DEATH
FULL_TCGA_DEATH <- TCGA_death[[1]][genes_inter,1:n.samples]
for(xx in 2:length(cancertype)){
FULL_TCGA_DEATH <- cbind(FULL_TCGA_DEATH, TCGA_death[[xx]][genes_inter,1:n.samples])
}
# Concatenate cancer types (exp) SURVIVED
FULL_TCGA_SURV <- TCGA_surv[[1]][genes_inter,1:n.samples]
for(xx in 2:length(cancertype)){
FULL_TCGA_SURV <- cbind(FULL_TCGA_SURV, TCGA_surv[[xx]][genes_inter,1:n.samples])
}
return(list(FULL_TCGA_DEATH=FULL_TCGA_DEATH,FULL_TCGA_SURV=FULL_TCGA_SURV))
}# end of function
#### MAIN WORKFLOW #####
### READ IN DATA ############################################
#############################################################
# KIDNEY Gene expression values and clinical data can be downloaded from here:
# http://acgt.cs.tau.ac.il/multi_omic_benchmark/download.html.
# Location of the file
loc="~/TCGA_data/NAR Data/"
# Set parameter for dimension reduction
n.patients <- 50
cat("Reading the data ...\n")
# READ IN!
tcga <- readTCGA("kidney", loc)
## Read in known genes
Known_Genes <- read.table("multiomics_features.tsv", sep="\t")
### END OF READ IN DATA #####################################
#############################################################
cat("Remove genes with ZERO expression ...\n")
DATA_full_DEATH <- tcga$FULL_TCGA_DEATH # for death outcomes
# Remove genes with ZERO expression
ids <- which(apply(DATA_full_DEATH,1,function(x){any(x==0)}))
DATA_full_DEATH <- DATA_full_DEATH[-ids,]
# -------------------------------
DATA_full_SURV <- tcga$FULL_TCGA_SURV # for survived outcomes
# Remove genes with ZERO expression
ids <- which(apply(DATA_full_SURV,1,function(x){any(x==0)}))
DATA_full_SURV <- DATA_full_SURV[-ids,]
# -------------------------------
# Take the genes which have no zero expression in one of the pops
common_GENES <- intersect(rownames(DATA_full_DEATH),rownames(DATA_full_SURV))
DATA_full_DEATH <- DATA_full_DEATH[common_GENES,]
DATA_full_SURV <- DATA_full_SURV[common_GENES,]
cat("Dimensions are:\n")
print(dim(DATA_full_DEATH))
print(dim(DATA_full_SURV))
cat("Filter by known cancer genes ...\n")
known_genes <- Known_Genes[,1]
current_genes <- rownames(DATA_full_DEATH)
current_genes <- strsplit(current_genes,split="|", fixed=TRUE)
current_genes <- sapply(current_genes,function(x){x[1]})
ids <- match(current_genes, known_genes)
del.genes <- which(is.na(ids))
DATA_full_DEATH <- DATA_full_DEATH[-del.genes,]
DATA_full_SURV <- DATA_full_SURV[-del.genes,]
cat("Dimensions are:\n")
print(dim(DATA_full_DEATH))
print(dim(DATA_full_SURV))
### End of filter by known_genes
cat("Subsample patients ...\n")
# Subsample patients ---------------------------- #
ids <- sample(1:dim(DATA_full_DEATH)[2], n.patients, replace=FALSE)
DATA_full_DEATH <- DATA_full_DEATH[,ids]
ids <- sample(1:dim(DATA_full_SURV)[2], n.patients, replace=FALSE)
DATA_full_SURV <- DATA_full_SURV[,ids]
# ------------------------------------------ #
# save
#save(DATA_full_DEATH, file="DATA_full_DEATH.RData")
#save(DATA_full_SURV, file="DATA_full_SURV.RData")
# load Patient subsample
load("DATA_full_DEATH.RData")
load("DATA_full_SURV.RData")
cat("Dimensions are:\n")
print(dim(DATA_full_DEATH))
print(dim(DATA_full_SURV))
cat("Voom transformation by limma ... \n")
COUNTS <- cbind(DATA_full_DEATH, DATA_full_SURV)
V <- voom(COUNTS)
COUNTS_norm <- V$E
VAR <- apply(COUNTS_norm,1,var)
cat("Select genes by variances ... \n")
# 5% high, 5% low, 5% random
n.genes <- floor((length(VAR)/100)*5) # 10% # 5%
# top and bottom VARIANCE genes
topgenes1 <- names(head(sort(VAR, decreasing=TRUE), n.genes))
topgenes2 <- names(tail(sort(VAR, decreasing=TRUE), n.genes))
# random genes
randgenes <- sample(names(VAR),n.genes)
topgenes <- unique(c(topgenes1, topgenes2, randgenes))
# without random genes
#topgenes <- unique(c(topgenes1, topgenes2))
# Reduce data - VOOM
DATA_top_DEATH <- COUNTS_norm[topgenes, colnames(DATA_full_DEATH)]#DEATH_patients]
DATA_top_SURV <- COUNTS_norm[topgenes, colnames(DATA_full_SURV)]#SURV_patients]
cat("Final dimensions are:\n")
print(dim(DATA_top_DEATH))
print(dim(DATA_top_SURV))
save(DATA_top_DEATH, file="DATA_top_DEATH.RData")
save(DATA_top_SURV, file="DATA_top_SURV.RData")
# ---------------------- CONSENSUS SIGNALS ------------------------------ #
###########################################################################
# ----------------------------------------------------------------------- #
## ---------------------- DEATH OUTCOMES -------------------------------- #
###########################################################################
cat("Signal estimation (non-survived outcomes)\n")
rankMatrix_top <- apply(DATA_top_DEATH, 2, function(x) rank(-x, ties.method = "random"))
rownames(rankMatrix_top) <- paste("r",1:dim(rankMatrix_top)[1], sep="")
colnames(rankMatrix_top) <- paste("c",1:dim(rankMatrix_top)[2], sep="")
# Peform estimation via TopKSignal
res_DEATH <- estimateTheta(R.input = rankMatrix_top, b = 0.1, num.boot = 100, solver = "gurobi",type ="restrictedQuadratic", bootstrap.type = "classic.bootstrap", nCore=2)
save(res_DEATH, file="res_DEATH.RData")
TKSrank <- rank(-res_DEATH$estimation$signal.estimate)
names(TKSrank) <- rownames(DATA_top_DEATH)
TKSrank_DEATH <- TKSrank
signal_DEATH <- res_DEATH$estimation$signal.estimate
signal_SE_DEATH <- res_DEATH$estimation$SE
names(signal_DEATH) <- rownames(DATA_top_DEATH)
names(signal_SE_DEATH) <- rownames(DATA_top_DEATH)
signal_sorted_DEATH <- sort(signal_DEATH, decreasing=TRUE)
ids <- match(names(signal_sorted_DEATH),names(signal_SE_DEATH))
signal_sorted_SE_DEATH <- signal_SE_DEATH[ids]
# Plots - Elbow Plot #####################
##########################################
## ggplot
NAM = names(signal_sorted_DEATH[1:20])
NAM = sapply(strsplit(NAM,"|", fixed=TRUE), function(x){x[1]})
library(ggplot2)
#pdf("DEATH_ELBOW_quadratic.pdf")
plot.elbow <- ggplot(data = data.frame(),aes(x = 1:length(signal_sorted_DEATH[1:20]), y = signal_sorted_DEATH[1:20])) + geom_point() + geom_line() +
geom_errorbar(aes(ymin=signal_sorted_DEATH[1:20]-signal_sorted_SE_DEATH[1:20], ymax=signal_sorted_DEATH[1:20] + signal_sorted_SE_DEATH[1:20]), width=.2,
position=position_dodge(0.05),colour="black") +
ylab("Signal estimate (theta)") +
ylim(c(2,2.5)) +
ggtitle("Top-20 genes of non-surviving patients") +
theme(axis.text.x = element_text(size=14, angle=45)) +
scale_x_discrete(name ="Genes (ordered)",
limits=NAM)
#theme_light()
plot.elbow #
#dev.off()
## ---------------------- SURVIVED OUTCOMES -------------------------------- #
###########################################################################
cat("Signal estimation (survived outcomes)\n")
rankMatrix_top <- apply(DATA_top_SURV, 2, function(x) rank(-x, ties.method = "random"))
rownames(rankMatrix_top) <- paste("r",1:dim(rankMatrix_top)[1], sep="")
colnames(rankMatrix_top) <- paste("c",1:dim(rankMatrix_top)[2], sep="")
# Peform estimation via TopKSignal
res_SURV <- estimateTheta(R.input = rankMatrix_top, b = 0.1, num.boot = 100, solver = "gurobi",type ="restrictedQuadratic", bootstrap.type = "classic.bootstrap", nCore=2)
save(res_SURV, file="res_SURV.RData")
TKSrank <- rank(-res_SURV$estimation$signal.estimate)
names(TKSrank) <- rownames(DATA_top_SURV)
TKSrank_SURV <- TKSrank
signal_SURV <- res_SURV$estimation$signal.estimate
signal_SE_SURV <- res_SURV$estimation$SE
names(signal_SURV) <- rownames(DATA_top_SURV)
names(signal_SE_SURV) <- rownames(DATA_top_SURV)
signal_sorted_SURV <- sort(signal_SURV, decreasing=TRUE)
ids <- match(names(signal_sorted_SURV),names(signal_SE_SURV))
signal_sorted_SE_SURV <- signal_SE_SURV[ids]
# Plots - Elbow Plot #####################
##########################################
## ggplot
NAM = names(signal_sorted_SURV[1:20])
NAM = sapply(strsplit(NAM,"|", fixed=TRUE), function(x){x[1]})
library(ggplot2)
#pdf("SURVIVED_ELBOW_quadratic.pdf")
plot.elbow <- ggplot(data = data.frame(),aes(x = 1:length(signal_sorted_SURV[1:20]),y = signal_sorted_SURV[1:20])) + geom_point() + geom_line() +
geom_errorbar(aes(ymin=signal_sorted_SURV[1:20]-signal_sorted_SE_SURV[1:20], ymax=signal_sorted_SURV[1:20] + signal_sorted_SE_SURV[1:20]), width=.2,
position=position_dodge(0.05),colour="black") +
#xlab("Genes (ordered)") +
ylab("Signal estimate (theta)") +
ylim(c(2,2.5)) +
ggtitle("Top-20 genes of surviving patients")+
theme(axis.text.x = element_text(size=14, angle=45))+
scale_x_discrete(name ="Genes (ordered)",
limits=NAM)
#theme_light()
plot.elbow #
#dev.off()