/
TEA_pipeline.R
182 lines (149 loc) · 7.91 KB
/
TEA_pipeline.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
### module load gcc/6.2.0
### module load R/3.3.0
# .libPaths("/mnt/work1/users/bhklab/Rlib")
# source("TEA_pipeline.R")
########################
### clean workspace
rm(list = ls())
########################
### load libraries
require(gdata)
require(e1071)
require(genefu)
library(Biobase)
require(xtable)
library(biomaRt)
library(gplots)
library(devtools)
library(preprocessCore)
library(rgl)
library(qpcR)
library(data.table)
library(piano)
library(snowfall)
library(utils)
library(Hmisc)
library(WriteXLS)
# devtools::install_github(repo="bhklab/PharmacoGx")
library(PharmacoGx)
library(survcomp)
########################
### global parameters
### set global R options
options(stringsAsFactors=FALSE)
### create directory for all the results and intermediate files
OutDir <- file.path("Output")
dir.create(OutDir, showWarnings=FALSE, recursive=TRUE)
### create directory for enrichment results
GSEADir <- file.path(OutDir, "Drug_Tissue_Associations")
dir.create(GSEADir, showWarnings=FALSE, recursive=TRUE)
### quantile for AUC in each tissue type of interest
quantileAUC <- 0.75
### Min and max number of cell lines in a tissue type
TissueSize <- c(15, 200)
FDRcutoff <- 0.05
### number of permutations for enrichment analysis
nperm <- 1000000
### number of CPU cores used for parallelization, use NULL for all the cores minus one
nbcore <- NULL
availcore <- parallel::detectCores()
if (is.null(nbcore) || nbcore > availcore) { nbcore <- availcore - 1 }
options("mc.cores"=nbcore)
########################
### load list of in vitro drug screening datasets
PsetList <- list("CCLE", "gCSI", "CTRPv2", "GDSC1000")
myfn <- file.path(OutDir, "psets.rds")
if (!file.exists(myfn)) {
PsetVec <- lapply(PsetList, PharmacoGx::downloadPSet, saveDir=OutDir)
names(PsetVec) <- PsetList
saveRDS(object=PsetVec, file=myfn)
} else {
PsetVec <- readRDS(file=myfn)
}
########################
### update the PSets for tissue enrichment analysis
### split lung into NSCLC and SCLC
NSCLC_cellines <- PsetVec$GDSC1000@cell$Sample.Name[which(PsetVec$GDSC1000@cell$GDSC.Tissue.descriptor.1 == "lung_NSCLC" & PsetVec$GDSC1000@cell$tissueid == "lung")]
SCLC_cellines <- PsetVec$GDSC1000@cell$Sample.Name[which(PsetVec$GDSC1000@cell$GDSC.Tissue.descriptor.1 == "lung_SCLC" & PsetVec$GDSC1000@cell$tissueid == "lung")]
### CCLE
PsetVec$CCLE@cell <- cbind(PsetVec$CCLE@cell, "tissueid_TEA"=PsetVec$CCLE@cell[ , "tissueid"])
PsetVec$CCLE@cell[ , "tissueid_TEA"] <- as.character(PsetVec$CCLE@cell[ , "tissueid_TEA"])
PsetVec$CCLE@cell[!is.na(PsetVec$CCLE@cell[ , "tissueid_TEA"]) & PsetVec$CCLE@cell[ , "tissueid_TEA"] == "", "tissueid_TEA"] <- NA
PsetVec$CCLE@cell[!is.na(PsetVec$CCLE@cell[ , "tissueid_TEA"]) & PsetVec$CCLE@cell[ , "tissueid_TEA"] == "lung", "tissueid_TEA"] <- NA
PsetVec$CCLE@cell[which((PsetVec$CCLE@cell[ ,"Hist.Subtype1"] == "adenocarcinoma" |
PsetVec$CCLE@cell[ ,"Hist.Subtype1"] == "non_small_cell_carcinoma" |
PsetVec$CCLE@cell[ ,"Hist.Subtype1"] == "squamous_cell_carcinoma") &
PsetVec$CCLE@cell[ ,"tissueid"] == "lung"), "tissueid_TEA"] <- "NSCLC"
PsetVec$CCLE@cell[which(PsetVec$CCLE@cell[ ,"Hist.Subtype1"] == "small_cell_carcinoma" &
PsetVec$CCLE@cell[ ,"tissueid"] == "lung"),"tissueid_TEA"] <- "SCLC"
### gCSI
PsetVec$gCSI@cell <- cbind(PsetVec$gCSI@cell, "tissueid_TEA"=PsetVec$gCSI@cell[ , "tissueid"])
PsetVec$gCSI@cell[ , "tissueid_TEA"] <- as.character(PsetVec$gCSI@cell[ , "tissueid_TEA"])
PsetVec$gCSI@cell[!is.na(PsetVec$gCSI@cell[ , "tissueid_TEA"]) & PsetVec$gCSI@cell[ , "tissueid_TEA"] == "", "tissueid_TEA"] <- NA
PsetVec$gCSI@cell[!is.na(PsetVec$gCSI@cell[ , "tissueid_TEA"]) & PsetVec$gCSI@cell[ , "tissueid_TEA"] == "lung", "tissueid_TEA"] <- NA
PsetVec$gCSI@cell[which(PsetVec$gCSI@cell[ ,"CellLineName"] %in% NSCLC_cellines & PsetVec$gCSI@cell[ ,"tissueid"] == "lung"), "tissueid_TEA"] <- "NSCLC"
PsetVec$gCSI@cell[which(PsetVec$gCSI@cell[ ,"CellLineName"] %in% SCLC_cellines & PsetVec$gCSI@cell[,"tissueid"] == "lung"), "tissueid_TEA"] <- "SCLC"
### CTRPv2
PsetVec$CTRPv2@cell <- cbind(PsetVec$CTRPv2@cell, "tissueid_TEA"=PsetVec$CTRPv2@cell[ , "tissueid"])
PsetVec$CTRPv2@cell[ , "tissueid_TEA"] <- as.character(PsetVec$CTRPv2@cell[ , "tissueid_TEA"])
PsetVec$CTRPv2@cell[!is.na(PsetVec$CTRPv2@cell[ , "tissueid_TEA"]) & PsetVec$CTRPv2@cell[ , "tissueid_TEA"] == "", "tissueid_TEA"] <- NA
PsetVec$CTRPv2@cell[!is.na(PsetVec$CTRPv2@cell[ , "tissueid_TEA"]) & PsetVec$CTRPv2@cell[ , "tissueid_TEA"] == "lung", "tissueid_TEA"] <- NA
PsetVec$CTRPv2@cell[which((PsetVec$CTRPv2@cell[,"ccle_hist_subtype_1"] == "adenocarcinoma" |
PsetVec$CTRPv2@cell[ , "ccle_hist_subtype_1"] == "non_small_cell_carcinoma" |
PsetVec$CTRPv2@cell[ , "ccle_hist_subtype_1"] == "squamous_cell_carcinoma") &
PsetVec$CTRPv2@cell[ , "tissueid"] == "lung"), "tissueid_TEA"] <- "NSCLC"
PsetVec$CTRPv2@cell[which(PsetVec$CTRPv2@cell[ , "ccle_hist_subtype_1"] == "small_cell_carcinoma" &
PsetVec$CTRPv2@cell[ ,"tissueid"] == "lung"), "tissueid_TEA"] <- "SCLC"
### GDSC1000
PsetVec$GDSC1000@cell <- cbind(PsetVec$GDSC1000@cell, "tissueid_TEA"=PsetVec$GDSC1000@cell[ , "tissueid"])
PsetVec$GDSC1000@cell[ , "tissueid_TEA"] <- as.character(PsetVec$GDSC1000@cell[ , "tissueid_TEA"])
PsetVec$GDSC1000@cell[!is.na(PsetVec$GDSC1000@cell[ , "tissueid_TEA"]) & PsetVec$GDSC1000@cell[ , "tissueid_TEA"] == "", "tissueid_TEA"] <- NA
PsetVec$GDSC1000@cell[!is.na(PsetVec$GDSC1000@cell[ , "tissueid_TEA"]) & PsetVec$GDSC1000@cell[ , "tissueid_TEA"] == "lung", "tissueid_TEA"] <- NA
PsetVec$GDSC1000@cell[which(PsetVec$GDSC1000@cell[,"GDSC.Tissue.descriptor.1"] == "lung_NSCLC" & PsetVec$GDSC1000@cell[ ,"tissueid"] == "lung"), "tissueid_TEA"] <- "NSCLC"
PsetVec$GDSC1000@cell[which(PsetVec$GDSC1000@cell[,"GDSC.Tissue.descriptor.1"] == "lung_SCLC" & PsetVec$GDSC1000@cell[ ,"tissueid"] == "lung"), "tissueid_TEA"] <- "SCLC"
########################
### list drugs and cell lines and drugs in all datasets
### drugs
drugs <- sapply(PsetVec, PharmacoGx::drugNames)
drugs <- sort(unique(do.call(c, drugs)))
drugsMat <- matrix(NA, nrow=length(drugs), ncol=length(PsetVec), dimnames=list(drugs, names(PsetVec)))
for (PsetIter in 1:length(PsetVec)) {
drugsMat[PharmacoGx::drugNames(PsetVec[[PsetIter]]), names(PsetVec)[PsetIter]] <- "YES"
}
### cell lines
cells <- sapply(PsetVec, PharmacoGx::cellNames)
cells <- sort(unique(do.call(c, cells)))
cellsMat <- matrix(NA, nrow=length(cells), ncol=length(PsetVec), dimnames=list(cells, names(PsetVec)))
for (PsetIter in 1:length(PsetVec)) {
cellsMat[PharmacoGx::cellNames(PsetVec[[PsetIter]]), names(PsetVec)[PsetIter]] <- "YES"
}
### tissues
tissues <- sapply(PsetVec, function(x) { return (sort(unique(PharmacoGx::cellInfo(x)[ , "tissueid_TEA"]))) })
tissues <- sort(unique(do.call(c, tissues)))
tissuesMat <- matrix(NA, nrow=length(tissues), ncol=length(PsetVec), dimnames=list(tissues, names(PsetVec)))
for (PsetIter in 1:length(PsetVec)) {
tt <- sort(unique(PharmacoGx::cellInfo(PsetVec[[PsetIter]])[ , "tissueid_TEA"]))
tissuesMat[tt, names(PsetVec)[PsetIter]] <- "YES"
}
### save
ll <- list("Drugs"=data.frame(drugsMat), "Cell lines"=data.frame(cellsMat), "Tissue Types"=data.frame(tissuesMat))
for(ll_c in 1:length(ll)){
xlsx::write.xlsx(ll[[ll_c]], file = file.path(OutDir, sprintf("Dataset_Info.xlsx")), row.names = TRUE, append = TRUE, sheetName = names(ll)[ll_c], showNA = FALSE)
}
#WriteXLS::WriteXLS("ll", ExcelFileName=file.path(OutDir, sprintf("Dataset_Info.xlsx")), row.names=TRUE)
########################
### run tissue enrichment analysis with original AUC
Adjustment <- FALSE
source("TEA_analysis.R")
### run tissue enrichment analysis with AUC values adjusted for genel level of drug sensitivity
Adjustment <- TRUE
source("TEA_analysis.R")
### combine the results
source("TEA_postprocess.R")
### compute predictability of significant interactions
source("TEA_predictability.R")
source("TEA_figure_generator.R")
########################
### save session info
write(toLatex(sessionInfo(), locale=FALSE), file=file.path(OutDir, "sessionInfoR.tex"), append=FALSE)
### end