-
Notifications
You must be signed in to change notification settings - Fork 0
/
4. Figure_4b.R
362 lines (326 loc) · 19.7 KB
/
4. Figure_4b.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
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
options("repos"= c(CRAN="https://mirrors.tuna.tsinghua.edu.cn/CRAN/"))
options(BioC_mirror="http://mirrors.tuna.tsinghua.edu.cn/bioconductor/")
library(ggplot2)
library(data.table)
library(survival)
library(ComplexHeatmap)
library(forestplot)
library(survminer)
library(circlize)
Sys.setenv(LANGUAGE = "en")
options(stringsAsFactors = FALSE)
#读取肿瘤注释文件#Read tumor annotation file
rawAnno <- read.delim("merged_sample_quality_annotations.tsv",sep = "\t",row.names = NULL,check.names = F,stringsAsFactors = F,header = T)
rawAnno$simple_barcode <- substr(rawAnno$aliquot_barcode,1,15)
samAnno <- rawAnno[!duplicated(rawAnno$simple_barcode),c("cancer type", "simple_barcode")]
samAnno <- samAnno[which(samAnno$`cancer type` != ""),]
write.table(samAnno,"output_simple_sample_annotation.txt",sep = "\t",row.names = F,col.names = T,quote = F)
# 读取生存数据# Read survival data
surv <- read.delim("Survival_SupplementalTable_S1_20171025_xena_sp", sep = "\t",row.names = 1,check.names = F,stringsAsFactors = F,header = T)
# 快速读取表达谱数据并做数据预处理# Quickly read expression profile data and do data preprocessing
#expr <- fread("222.tsv",sep = "\t",stringsAsFactors = F,check.names = F,header = T)
expr <- fread("EBPlusPlusAdjustPANCAN_IlluminaHiSeq_RNASeqV2.geneExp.tsv",sep = "\t",stringsAsFactors = F,check.names = F,header = T)
expr <- as.data.frame(expr); rownames(expr) <- expr[,1]; expr <- expr[,-1]
gene <- sapply(strsplit(rownames(expr),"|",fixed = T), "[",1)
expr$gene <- gene
expr <- expr[!duplicated(expr$gene),]
rownames(expr) <- expr$gene; expr <- expr[,-ncol(expr)]
expr[expr < 0] <- 0 # 对于这份泛癌数据，将略小于0的数值拉到0，否则不能取log（其他途径下载的泛癌数据可能不需要此操作）# For this pan-cancer data, pull the value slightly less than 0 to 0, otherwise the log cannot be taken (pan-cancer data downloaded through other channels may not require this operation)
colnames(expr) <- substr(colnames(expr),1,15)
gc()
# 设置感兴趣的基因# Set the genes of interest
geneOfInterest <- "TPSAB1"#or other four MC signature genes
if(!is.element(geneOfInterest, rownames(expr))) {
warning("The gene ", geneOfInterest," cannot be found!")
} else {
message("The gene ", geneOfInterest," can be matched!")
}
# 确定肿瘤样本以及对应肿瘤类型# Determine the tumor sample and corresponding tumor type
sam <- samAnno[which(samAnno$`cancer type` != "LAML"),"simple_barcode"] # 去掉白血病样本# Remove leukemia samples
comsam <- intersect(intersect(colnames(expr), sam), rownames(surv)) # 得到与表达谱以及生存的共有样本# Get shared samples with expression profiles and survival
tumsam <- comsam[substr(comsam,14,14) == "0"] # 仅提取肿瘤样本#Extract tumor samples only
tumAnno <- samAnno[which(samAnno$simple_barcode %in% tumsam),] # 获取这些肿瘤样本的注释信息# Get the annotation information of these tumor samples
tumAnno <- tumAnno[order(tumAnno$`cancer type`),] # 根据肿瘤类型排序# Sort by tumor type
tumors <- unique(tumAnno$`cancer type`) # 得到32个肿瘤# Get 32 tumors
# 合并表达和生存数据# Merge expression and survival data
exprSurv <- cbind.data.frame(expr = log2(as.numeric(expr[geneOfInterest,comsam]) + 1),
surv[comsam,c("OS","OS.time","DSS","DSS.time","DFI","DFI.time","PFI","PFI.time")])
write.table(exprSurv, "output_combined dataframe with both gene expression and survival information.txt",sep = "\t",row.names = T,col.names = NA,quote = F)
outTab.cox <- NULL
for(i in tumors) {
sam <- tumAnno[which(tumAnno$`cancer type` == i),"simple_barcode"]
exprSurvSub <- exprSurv[sam,]
## OS
coxres <- summary(coxph(Surv(OS.time, OS) ~ expr, data = exprSurvSub))
outTab.cox <- rbind.data.frame(outTab.cox,
data.frame(tumor = i,
event = "OS",
beta = coxres$coefficients[1,1],
hr = coxres$coefficients[1,2],
lower = coxres$conf.int[1,3],
upper = coxres$conf.int[1,4],
p = coxres$coefficients[1,5],
stringsAsFactors = F),
stringsAsFactors = F)
## DSS
coxres <- summary(coxph(Surv(DSS.time, DSS) ~ expr, data = exprSurvSub))
outTab.cox <- rbind.data.frame(outTab.cox,
data.frame(tumor = i,
event = "DSS",
beta = coxres$coefficients[1,1],
hr = coxres$coefficients[1,2],
lower = coxres$conf.int[1,3],
upper = coxres$conf.int[1,4],
p = coxres$coefficients[1,5],
stringsAsFactors = F),
stringsAsFactors = F)
## DFI
if(i %in% c("SKCM","THYM","UVM","GBM")) { # 前三个肿瘤没有对应的DFI数据，GBM只有3个样本有对应的DFI，也去除（和原文不符）# The first three tumors do not have corresponding DFI data. Only 3 samples of GBM have corresponding DFI data and are also removed (inconsistent with the original text)
outTab.cox <- rbind.data.frame(outTab.cox,
data.frame(tumor = i,
event = "DFI",
beta = NA,
hr = NA,
lower = NA,
upper = NA,
p = NA,
stringsAsFactors = F),
stringsAsFactors = F)
} else {
coxres <- summary(coxph(Surv(DFI.time, DFI) ~ expr, data = exprSurvSub))
outTab.cox <- rbind.data.frame(outTab.cox,
data.frame(tumor = i,
event = "DFI",
beta = coxres$coefficients[1,1],
hr = coxres$coefficients[1,2],
lower = coxres$conf.int[1,3],
upper = coxres$conf.int[1,4],
p = coxres$coefficients[1,5],
stringsAsFactors = F),
stringsAsFactors = F)
}
## PFI
coxres <- summary(coxph(Surv(PFI.time, PFI) ~ expr, data = exprSurvSub))
outTab.cox <- rbind.data.frame(outTab.cox,
data.frame(tumor = i,
event = "PFI",
beta = coxres$coefficients[1,1],
hr = coxres$coefficients[1,2],
lower = coxres$conf.int[1,4],
upper = coxres$conf.int[1,4],
p = coxres$coefficients[1,5],
stringsAsFactors = F),
stringsAsFactors = F)
}
write.table(outTab.cox, file = "output_summary of cox result in pancancer.txt",sep = "\r",row.names = F,col.names = T,quote = F)
minprop <- 0.2
outTab.km <- NULL
for(i in tumors) {
sam <- tumAnno[which(tumAnno$`cancer type` == i),"simple_barcode"]
exprSurvSub <- exprSurv[sam,]
## OS
bestcut <- surv_cutpoint(exprSurvSub,
time = "OS.time",
event = "OS",
variables = "expr",
minprop = minprop) #默认组内sample不能低于20%#The sample in the default group cannot be less than 20%
cutoff <- bestcut$cutpoint[1,1]
exprSurvSub$group <- factor(ifelse(exprSurvSub$expr > cutoff,"High","Low"), levels = c("Low","High"))
fitd <- survdiff(Surv(OS.time, OS) ~ group, data=exprSurvSub, na.action=na.exclude)
p.val <- 1 - pchisq(fitd$chisq, length(fitd$n) - 1)
HR <- (fitd$obs[2]/fitd$exp[2])/(fitd$obs[1]/fitd$exp[1])
outTab.km <- rbind.data.frame(outTab.km,
data.frame(tumor = i,
event = "OS",
hr = HR, # Hazard ratio
lower = exp(log(HR) - qnorm(0.975)*sqrt(1/fitd$exp[2]+1/fitd$exp[1])),
upper = exp(log(HR) + qnorm(0.975)*sqrt(1/fitd$exp[2]+1/fitd$exp[1])),
p = p.val,
stringsAsFactors = F),
stringsAsFactors = F)
## DSS
bestcut <- surv_cutpoint(exprSurvSub,
time = "DSS.time",
event = "DSS",
variables = "expr",
minprop = minprop)
cutoff <- bestcut$cutpoint[1,1]
exprSurvSub$group <- factor(ifelse(exprSurvSub$expr > cutoff,"High","Low"), levels = c("High","Low"))
fitd <- survdiff(Surv(DSS.time, DSS) ~ group, data=exprSurvSub, na.action=na.exclude)
p.val <- 1 - pchisq(fitd$chisq, length(fitd$n) - 1)
outTab.km <- rbind.data.frame(outTab.km,
data.frame(tumor = i,
event = "DSS",
hr = (fitd$obs[1]/fitd$exp[1])/(fitd$obs[2]/fitd$exp[2]), # Hazard ratio 我把1和2对调后结果才看起来对# Hazard ratio I swapped 1 and 2 and the result looked right.
lower = exp(log(HR) - qnorm(0.975)*sqrt(1/fitd$exp[2]+1/fitd$exp[1])),
upper = exp(log(HR) + qnorm(0.975)*sqrt(1/fitd$exp[2]+1/fitd$exp[1])),
p = p.val, # p值
stringsAsFactors = F),
stringsAsFactors = F)
## DFI
if(i %in% c("SKCM","THYM","UVM","GBM")) { # 前三个肿瘤没有对应的DFI数据，GBM只有3个样本有对应的DFI，也去除（和原文不符）# The first three tumors do not have corresponding DFI data. Only 3 samples of GBM have corresponding DFI data and are also removed (inconsistent with the original text)
outTab.km <- rbind.data.frame(outTab.km,
data.frame(tumor = i,
event = "DFI",
hr = NA,
lower = NA,
upper = NA,
p = NA,
stringsAsFactors = F),
stringsAsFactors = F)
} else {
bestcut <- surv_cutpoint(exprSurvSub,
time = "DFI.time",
event = "DFI",
variables = "expr",
minprop = minprop)
cutoff <- bestcut$cutpoint[1,1]
exprSurvSub$group <- factor(ifelse(exprSurvSub$expr > cutoff,"High","Low"), levels = c("High","Low"))
fitd <- survdiff(Surv(DFI.time, DFI) ~ group, data=exprSurvSub, na.action=na.exclude)
p.val <- 1 - pchisq(fitd$chisq, length(fitd$n) - 1)
outTab.km <- rbind.data.frame(outTab.km,
data.frame(tumor = i,
event = "DFI",
hr = (fitd$obs[1]/fitd$exp[1])/(fitd$obs[2]/fitd$exp[2]),# Hazard ratio
lower = exp(log(HR) - qnorm(0.975)*sqrt(1/fitd$exp[2]+1/fitd$exp[1])),
upper = exp(log(HR) + qnorm(0.975)*sqrt(1/fitd$exp[2]+1/fitd$exp[1])),
p = p.val,
stringsAsFactors = F),
stringsAsFactors = F)
}
## PFI
bestcut <- surv_cutpoint(exprSurvSub,
time = "PFI.time",
event = "PFI",
variables = "expr",
minprop = minprop)
cutoff <- bestcut$cutpoint[1,1]
exprSurvSub$group <- factor(ifelse(exprSurvSub$expr > cutoff,"High","Low"), levels = c("High","Low"))
fitd <- survdiff(Surv(PFI.time, PFI) ~ group, data=exprSurvSub, na.action=na.exclude)
p.val <- 1 - pchisq(fitd$chisq, length(fitd$n) - 1)
outTab.km <- rbind.data.frame(outTab.km,
data.frame(tumor = i,
event = "PFI",
hr = (fitd$obs[1]/fitd$exp[1])/(fitd$obs[2]/fitd$exp[2]),# Hazard ratio
lower = exp(log(HR) - qnorm(0.975)*sqrt(1/fitd$exp[2]+1/fitd$exp[1])), #
upper = exp(log(HR) + qnorm(0.975)*sqrt(1/fitd$exp[2]+1/fitd$exp[1])), #
p = p.val, #
stringsAsFactors = F),
stringsAsFactors = F)
}
write.table(outTab.km, file = "output_summary of km result in pancancer.txt",sep = "\r",row.names = F,col.names = T,quote = F)
# 设置颜色# Set color
blue <- "#A0CEE3"
yellow <- "#EFEFBE"
sea <- "#37BCDF"
green <- "#71BC5D"
cherry <- "#E5588C"
red <- "#E46A6B"
purple <- "#8959A3"
# 制作热图绘制数据# Make heat map drawing data
hmInput <- NULL
for (i in tumors) {
cox.res <- outTab.cox[which(outTab.cox$tumor == i),]
km.res <- outTab.km[which(outTab.km$tumor == i),]
cox.res$dirct <- ifelse(cox.res$hr > 1 & cox.res$p < 0.05, "Risky",
ifelse(cox.res$hr < 1 & cox.res$p < 0.05, "Protective","Nonsense"))
km.res$dirct <- ifelse(km.res$hr > 1 & km.res$p < 0.05, "Risky",
ifelse(km.res$hr < 1 & km.res$p < 0.05, "Protective","Nonsense"))
hmInput <- rbind.data.frame(hmInput,
data.frame(OS.cox = cox.res[1,"dirct"],
OS.km = km.res[1,"dirct"],
DSS.cox = cox.res[2,"dirct"],
DSS.km = km.res[2,"dirct"],
DFI.cox = cox.res[3,"dirct"],
DFI.km = km.res[3,"dirct"],
PFI.cox = cox.res[4,"dirct"],
PFI.km = km.res[4,"dirct"],
row.names = i,
stringsAsFactors = F),
stringsAsFactors = F)
}
hmInput[is.na(hmInput)] <- "N/A"
write.table(hmInput, file = "output_summary of gene prognositicationin pancancer.txt",sep = "\t",row.names = T,col.names = NA,quote = F)
# 注意：结果在某些肿瘤上有较大出入，可能的原因有：
# 1. 泛癌表达谱使用不一致
# 2. 最佳生存曲线cutoff的minprop选择不一致
# 3. 对KM分析中HR的计算不一致，原作者可能依然采用Cox回归估计二分类下的HR
# Note: The results vary greatly for some tumors. Possible reasons are:
# 1. Inconsistent use of pan-cancer expression profiles
# 2. The minprop selection of the optimal survival curve cutoff is inconsistent
# 3. The calculation of HR in KM analysis is inconsistent. The original author may still use Cox regression to estimate HR under two categories.
# 修改热图数据#Modify heat map data
indata <- hmInput
indata[indata == "Risky"] <- 0 # 为了使得risky出现在图例的顶部，所以设置一个最小值0To make risky appear at the top of the legend, set a minimum value of 0
indata[indata == "Protective"] <- 1 # 根据原文，依次设置为1# According to the original text, set to 1 in turn
indata[indata == "N/A"] <- 2 # 根据原文，依次设置为2
indata[indata == "Nonsense"] <- 3 # 根据原文，依次设置为3
# 构建列注释#Build column annotations
annCol <- data.frame("Method" = rep(c("Cox","Log-rank"),4),
"Survival Type" = rep(c("OS","DSS","DFI","PFI"), each = 2),
check.names = F,
row.names = colnames(indata))
annColors <- list("Method" = c("Cox" = blue, "Log-rank" = yellow),
"Survival Type" = c("OS" = sea, "DSS" = green, "DFI" = cherry, "PFI" = purple))
col_fun = circlize::colorRamp2(c(0, 1, 2, 3), c(red,green,"grey50","white"))
# 绘制热图#Draw a heat map
hm <- Heatmap(matrix = as.matrix(indata),
border = "black",
rect_gp = gpar(col = "black"),
name = "Prognostic role",
cluster_rows = F,
cluster_columns = F,
col = c(red,green,"grey50","white"),
show_row_names = T,
show_column_names = F,
row_names_side = "left",
top_annotation = HeatmapAnnotation(df = annCol,
col = annColors,
gp = gpar(col = "black"),
border = TRUE),
width = grid::unit(8, "cm"),
height = grid::unit(15, "cm"),
heatmap_legend_param = list(at = c(0, 1, 2, 3),
legend_gp = grid::gpar(fill = col_fun(c(0,1,2,3))),
labels = c("Risky", "Protective","N/A","Nonsense")))
pdf("prognostic heatmap.pdf", width = 10,height = 10)
draw(hm)
invisible(dev.off())
##################################################################
fpInput <- outTab.cox[which(outTab.cox$event == "OS"),]
hrtable <- fpInput[,c("tumor","event","beta","hr","lower","upper","p")]
tabletext <- cbind(c("Cancers",hrtable$tumor),
c("p value",ifelse(round(as.numeric(hrtable$p),3) < 0.001,"<0.001",round(as.numeric(hrtable$p),3))),
c("HR (95L-95H)",paste0(round(as.numeric(hrtable$hr),3), " (",
round(as.numeric(hrtable$lower),3),"-",
round(as.numeric(hrtable$upper),3),")")))
pdf("forestplot of os risk table in pancancer.pdf", width = 8, height = 8)
forestplot(labeltext=tabletext,
mean=c(NA,as.numeric(hrtable$hr)),# HR
lower=c(NA,as.numeric(hrtable$lower)), #
upper=c(NA,as.numeric(hrtable$upper)),#
graph.pos = 4,
graphwidth = unit(.3,"npc"),
fn.ci_norm="fpDrawDiamondCI",
col=fpColors(box="grey50", lines="grey50", zero = "black"),
boxsize=c(NA,ifelse(as.numeric(hrtable$p) < 0.05,0.8,0.4)),
lwd.ci=1,
ci.vertices.height = 0.1,ci.vertices=F,
zero=1,
lwd.zero=2,
xticks = c(0,1,2,4,8,13),
lwd.xaxis=2,
xlab="Hazard ratio",
hrzl_lines=list("1" = gpar(lwd=2, col="black"),
"34" = gpar(lwd=2, col="black")),#最后一行底部加黑线???""中数字为nrow(tabletext) + 1
txt_gp=fpTxtGp(label=gpar(cex=1),
ticks=gpar(cex=0.85),
xlab=gpar(cex=1),
title=gpar(cex=1.2)),
lineheight = unit(.55,"cm"),
colgap = unit(0.4,"cm"),
mar=unit(rep(1.5, times = 4), "cm"),
new_page = F
)
invisible(dev.off())
sessionInfo()