TCGA数据统计

2021.6.27

HTSeq - Counts

点击下载

图片alt

HTSeq - FPKM

点击下载

图片alt

BCGSC miRNA Profiling

点击下载

图片alt

相关代码

library(TCGAbiolinks)
project_df <- (function(){
  projects <- TCGAbiolinks:::getGDCprojects()$project_id
  projects_full <- projects[grepl('^TCGA',projects,perl=T)]
  projects_ID <- sapply(projects_full,function(x){stringi::stri_sub(x,6,10)})
  res <- data.frame(cancer=projects_ID,TCGA_ID=projects_full)
  return(res)
})()
match.file.cases.all <- NULL
for(project in  project_df[,2]){
  #project <- project_df[5,2]
  query<- GDCquery(project = project,
                   data.category = "Transcriptome Profiling",
                   data.type = "Gene Expression Quantification",
                   workflow.type = "HTSeq - Counts")
  match.file.cases <- getResults(query,cols=c("cases","file_name"))
  match.file.cases$project <- project
  match.file.cases.all <- rbind(match.file.cases.all,match.file.cases)
}
readr::write_tsv(match.file.cases.all, path =  "filename_case.txt")
table(match.file.cases.all$project)

 query<- GDCquery(project = project,
                   data.category = "Transcriptome Profiling",
                   data.type = "Gene Expression Quantification",
                   workflow.type = "HTSeq - FPKM")

 query<- GDCquery(project = project,
                     data.category = "Transcriptome Profiling",
                     data.type = "miRNA Expression Quantification",
                     workflow.type = "BCGSC miRNA Profiling")