-
Notifications
You must be signed in to change notification settings - Fork 2
/
450k2gender.R
172 lines (156 loc) · 6.65 KB
/
450k2gender.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
#########################################################################################
#########################################################################################
########################## Shicheng Guo #############################################
#########################################################################################
## Prediction gender with 450K dataset
# Download json and transfer to csv (https://json-csv.com/)
# Only can be run in Linux since it requre bedtools
#########################################################################################
##########################Training with LIHC Dataset ####################################
#########################################################################################
#install.packages("Deducer")
#install.packages("stringr")
#install.packages("pROC")
library("stringr")
library("pROC")
#library("Deducer")
bed2cg<-function(bed1){
ref<-read.table("~/work/db/hg19/GPL13534.sort.bed",head=F,sep="\t")
cor2bed<-function(cor){
cor<-as.character(cor)
a<-unlist(lapply(strsplit(cor,split=c(":")),function(x) strsplit(x,"-")))
bed<-matrix(a,ncol=3,byrow=T)
bed<-data.frame(bed,cor)
return(data.frame(bed))
}
rbedintersect<-function(bed1,ref){
Rbedtools<-function(functionstring="intersectBed",bed1,bed2,opt.string=""){
#create temp files
a.file=tempfile()
b.file=tempfile()
out =tempfile()
options(scipen =99) # not to use scientific notation when writing out
#write bed formatted dataframes to tempfile
write.table(bed1,file=a.file,quote=F,sep="\t",col.names=F,row.names=F)
write.table(bed2,file=b.file,quote=F,sep="\t",col.names=F,row.names=F)
# create the command string and call the command using system()
command=paste(functionstring,"-a",a.file,"-b",b.file,opt.string,">",out,sep=" ")
cat(command,"\n")
try(system(command))
res=read.table(out,header=F)
unlink(a.file);unlink(b.file);unlink(out)
res=subset(res,V5!=".")
return(res)
}
merge<-Rbedtools(functionstring="intersectBed",bed1,ref,opt.string="-wao")
return(merge)
}
merge<-rbedintersect(bed1,ref)
return(merge)
}
cg2bed<-function(cg,extend=100){
bed2cor<-function(bed){
cor<-apply(bed,1,function(x) paste(x[1],":",as.numeric(x[2])-extend,"-",as.numeric(x[3])+extend,sep=""))
cor<-gsub(" ","",cor)
return(cor)
}
ref<-read.table("~/work/db/hg19/GPL13534.sort.bed",head=F,sep="\t")
bed<-ref[match(cg,ref[,4]),1:3]
bed[,2]=bed[,2]-extend
bed[,3]=bed[,3]+extend
cor<-bed2cor(bed)
rlt<-data.frame(bed,cor,cg)
return(rlt)
}
write.bed<-function(bed,file,extend=0){
bed[,2]<-as.numeric(as.character(bed[,2]))-extend
bed[,3]<-as.numeric(as.character(bed[,3]))+extend
if(ncol(bed)==3){
bed[,4]<-paste(bed[,1],":",bed[,2],"-",bed[,3],sep="")
}
if(ncol(bed)>=4){
write.table(bed,file=file,sep="\t",col.names=F,row.names=F,quote=F)
}
}
see<-function(x){
x[1:3,1:3]
}
readmeth450<-function(){
rlt<-list()
library("stringr")
file<-list.files(pattern="jhu*")
data<-c()
for(i in file){
tmp<-read.table(i,head=T,skip=1,row.names=1,sep="\t",check.names = FALSE,as.is=T)
data<-cbind(data,tmp[,1])
print(i)
}
#load("PancancerMethMatrix_March2016.RData")
#load("PancancerMethMatrix_March2016.Test.RData")
colnames(data)<-unlist(lapply(unlist(lapply(file,function(x) unlist(strsplit(x,"[.]"))[6])),function(x) substr(x,1,15)))
rownames(data)<-rownames(tmp)
cancertype<-unique(unlist(lapply(file,function(x) unlist(strsplit(x,"_|.Human"))[2])))
sampletype<-unlist(lapply(unlist(lapply(file,function(x) unlist(strsplit(x,"[.]"))[6])),function(x) substr(x,14,15)))
save(data,file=paste(cancertype,"meth.RData",sep="."))
rlt$data<-data
rlt$cancertype<-cancertype
rlt$sampletype<-sampletype
rlt$cpg<-rownames(data)
return(rlt)
}
setwd("/media/NAS3_volume2/shg047/HM450/TCGA/lihc")
saminfo<-read.csv("../clinical.project-TCGA-LIHC.2017-05-23T04-50-49.306129.csv")
saminfo<-data.frame(id=substr(as.character(saminfo$exposures__submitter_id),1,12),gender=as.character(saminfo$demographic__gender))
load("LIHC.meth.RData")
# idv<-as.array(str_extract(file,"TCGA-[0-9|a-z|A-Z]*-[0-9|a-z|A-Z]*-[0-9]*"))
# length(idv)
# colnames(data)<-idv
bed<-cg2bed(rownames(data))
newdata1<-data[which(bed[,1]=="chrX"),]
dim(newdata1)
gender<-saminfo[match(substr(colnames(data),1,12),saminfo[,1]),2]
delta<-data.frame(t(apply(newdata1,1,function(x) tapply(x,gender,function(x) mean(x,na.rm=T)))))
marker<-rownames(subset(delta,female<0.6 & female>0.4 & male<0.1))
test<-data[match(marker,rownames(data)),]
input<-data.frame(femaleScore=apply(test,2,function(x) sum(x>0.3,na.rm=T)/(length(na.omit(x)))),gender)
bed<-cg2bed(marker)
F1<-newdata1[,grep("female",gender)]
M1<-newdata1[,grep("male",gender)]
png("density.png")
plot(density(na.omit(as.numeric(M1))),col="blue",lwd=2)
lines(density(na.omit(as.numeric(F1))),col="red",lwd=2)
legend("topright",legend=c("female","male"),lwd=2,col=c("red","blue"))
dev.off()
## evaluation
fit<-glm(gender~score,input,family=binomial(link = "logit"))
prob=predict(fit,type=c("response"))
input$prob=prob
g <- roc(gender ~ prob, data = input)
jpeg("ROC1.jpg")
plot(g)
# give up since
#modelfit <- glm(formula=gender ~ score, family=binomial(), data=input, na.action=na.omit)
#jpeg("ROC2.jpg")
#rocplot(modelfit)
#dev.off()
#########################################################################################
##########################Test with ESCA Dataset ####################################
#########################################################################################
setwd("/media/NAS3_volume2/shg047/HM450/TCGA/chol")
saminfo<-read.csv("../clinical.project-TCGA-CHOL.2017-05-23T06-36-48.852956.csv")
saminfo<-data.frame(id=substr(as.character(saminfo$exposures__submitter_id),1,12),gender=as.character(saminfo$demographic__gender))
chol<-readmeth450()
data<-chol$data
gender<-data.frame(id=substr(as.character(colnames(data)),1,12),gender=unlist(saminfo[match(substr(colnames(data),1,12),saminfo[,1]),2]))
testdata2gender<-function(data){
probability<-apply(data[match(bed$cg,rownames(data)),],2,function(x) sum(na.omit(x)>0.3)/length(na.omit(x)))
gender<-c(0,length(probability))
gender[probability<0.3]<-"male"
gender[probability>0.3]<-"female"
prediction<-data.frame(id=substr(as.character(names(probability)),1,12),gender,probability)
return(prediction)
}
predict<-testdata2gender(data)
rlt<-merge(predict,gender,by="id")
rlt[which(! as.character(rlt$gender.x)==as.character(rlt$gender.y)),]
# only one sample (TCGA-W5-AA2T) prediction error, but I do think the sample was labelled to female, however, the truth of that sample is male samples.