GscTools.R

library("Haplin")
library("arm")
library("randomForest")

RawZeroRemove<-function(data,missratio=0.5){
  threshold<-(missratio)*ncol(data)
  NaRaw<-which(apply(data,1,function(x) sum(is.na(x))>=threshold))
  zero<-which(apply(data,1,function(x) sum(x==0)>=threshold))
  NaRAW<-c(NaRaw,zero)
  if(length(NaRAW)>0){
    output<-data[-NaRAW,]
  }else{
    output<-data;
  }
  output
}

manifest2barcode<-function(manifest){
x=read.table(manifest,header = T)
manifest_length= nrow(x)
id= toString(sprintf('"%s"', x$id))
Part1= '{"filters":{"op":"in","content":{"field":"files.file_id","value":[ '
Part2= '] }},"format":"TSV","fields":"file_id,file_name,cases.submitter_id,cases.case_id,data_category,data_type,cases.samples.tumor_descriptor,cases.samples.tissue_type,cases.samples.sample_type,cases.samples.submitter_id,cases.samples.sample_id,cases.samples.portions.analytes.aliquots.aliquot_id,cases.samples.portions.analytes.aliquots.submitter_id","size":'
Part3= paste0("\"",manifest_length, "\"", "}")
Sentence= paste(Part1,id,Part2,Part3, collapse=" ")
write.table(Sentence,"Payload.txt",quote=F,col.names=F,row.names=F)
system("curl --request POST --header \"Content-Type: application/json\" --data @Payload.txt \"https://api.gdc.cancer.gov/files\" > barcode.txt")
}

ENST2Symbol<-function(ENST){
  db<-read.table("https://raw.githubusercontent.com/Shicheng-Guo/AnnotationDatabase/master/ENSG.ENST.ENSP.Symbol.hg19.bed",sep="\t")
  Symbol<-db[match(ENST,db$V7),4]
  return(Symbol)
}

ENSG2Symbol<-function(ENSG){
  db<-read.table("https://raw.githubusercontent.com/Shicheng-Guo/AnnotationDatabase/master/ENSG.ENST.ENSP.Symbol.hg19.bed",sep="\t")
  Symbol<-db[match(ENSG,db$V8),4]
  ENSG<-unlist(lapply(strsplit(ENSG,split="[.]"),function(x) x[1]))
  Symbol<-db[match(as.character(ENSG),db$V8),4]
  return(Symbol)
}

ENSP2Symbol<-function(ENSP){
  db<-read.table("https://raw.githubusercontent.com/Shicheng-Guo/AnnotationDatabase/master/ENSG.ENST.ENSP.Symbol.hg19.bed",sep="\t")
  Symbol<-db[match(ENSP,db$V9),4]
  return(Symbol)
}

id2phen4<-function(filename){
  library("stringr")
  as.array(str_extract(filename,"TCGA-[0-9|a-z|A-Z]*-[0-9|a-z|A-Z]*-[0-9]*"))
}

id2phen3<-function(filename){
  library("stringr")
  as.array(str_extract(filename,"TCGA-[0-9|a-z|A-Z]*-[0-9|a-z|A-Z]*"))
}

id2bin<-function(filename){
  library("stringr")
  filename<-as.array(str_extract(filename,"TCGA-[0-9|a-z|A-Z]*-[0-9|a-z|A-Z]*-[0-9]*"))
  as.numeric(lapply(strsplit(filename,"-"),function(x) x[4]))
}

id2pid<-function(filename){
  library("stringr")
  filename<-as.array(str_extract(filename,"edu_...."))
  unlist(lapply(filename,function(x) unlist(strsplit(x,"[_]"))[2]))
}

cpg2symbol<-function(cpg){
map<-read.table("https://raw.githubusercontent.com/Shicheng-Guo/AnnotationDatabase/master/hg19/GPL13534_450K_hg19_V3.bed")
symbol<-map[match(cpg,map[,4]),5]
return(symbol)
}
                
                
id2phen4<-function(filename){
  library("stringr")
  as.array(str_extract(filename,"TCGA-[0-9|a-z|A-Z]*-[0-9|a-z|A-Z]*-[0-9]*"))
}

id2phen3<-function(filename){
  library("stringr")
  as.array(str_extract(filename,"TCGA-[0-9|a-z|A-Z]*-[0-9|a-z|A-Z]*"))
}

id2bin<-function(filename){
  library("stringr")
  filename<-as.array(str_extract(filename,"TCGA-[0-9|a-z|A-Z]*-[0-9|a-z|A-Z]*-[0-9]*"))
  as.numeric(lapply(strsplit(filename,"-"),function(x) x[4]))
}

id2pid<-function(filename){
  library("stringr")
  filename<-as.array(str_extract(filename,"edu_...."))
  unlist(lapply(filename,function(x) unlist(strsplit(x,"[_]"))[2]))
}

RawNARemove<-function(data,missratio=0.3){
  threshold<-(missratio)*ncol(data)
  NaRaw<-which(apply(data,1,function(x) sum(is.na(x))>=threshold))
  zero<-which(apply(data,1,function(x) all(x==0))==T)
  NaRAW<-c(NaRaw,zero)
  if(length(NaRAW)>0){
    output<-data[-NaRAW,]
  }else{
    output<-data;
  }
  output
}

cpg2symbol<-function(cpg){
  map<-read.table("https://raw.githubusercontent.com/Shicheng-Guo/AnnotationDatabase/master/hg19/GPL13534_450K_hg19_V3.bed")
  symbol<-map[match(cpg,map[,4]),5]
  return(symbol)
}

tsneplot<-function(mydata,phen,plot="tsne.plot.pdf"){
  library("tsne")
  data=data.frame(phen,mydata)
  pdf(plot)
  colors = rainbow(length(unique(data$phen)))
  names(colors) = unique(data$phen)
  ecb = function(x,y){plot(x,t='n'); text(x,labels=data$phen, col=colors[data$phen]) }
  tsne_iris = tsne(data, epoch_callback = ecb, perplexity=10)
  dev.off()
}