用R写了个函数实现GDC文件合并，记录一下

image.png

library(dplyr)
library(readr)
merge_column <- function(input_dir, column_name, output_dir) {
  # 加载必要的包
  library(dplyr)
  library(readr)
  
  # 创建输出目录（如果不存在）
  if (!dir.exists(output_dir)) {
    dir.create(output_dir, recursive = TRUE)
  }
  
  cat("output_dir:", output_dir, "\n")
  # 获取所有子文件夹路径
  sub_dirs <- list.dirs(input_dir, full.names = TRUE, recursive = FALSE)
  
  # 初始化合并结果数据框
  merged_data <- NULL
  print("merging data...")
  for (i in 1:length(sub_dirs)) {
    cat(i, ":",sub_dirs[i])
    # 获取子文件夹中的所有文件
    files <- list.files(sub_dirs[i], pattern = "\\.tsv$", full.names = TRUE)
    
    for (file in files) {
      # 读取文件
      exp_data <- read_tsv(file, skip = 6, col_names = FALSE, show_col_types = FALSE) # 从第7行开始读取
      colnames(exp_data) <- c("gene_id",    "gene_name",    "gene_type",    "unstranded",   "stranded_first",   "stranded_second",  "tpm_unstranded",   "fpkm_unstranded", "fpkm_uq_unstranded")
      exp_data <- as.data.frame(exp_data)
      # 检查文件中是否有指定的列
      if (!(column_name %in% colnames(exp_data))) {
        stop(paste("文件中缺少列：", column_name, "，文件：", file))
      }
      
      # 提取Ensembl基因ID、基因名称和目标列
      sample_data <- exp_data %>%
        select(gene_id = 1, gene_name = 2, gene_type = 3, target_column = !!sym(column_name))
      
      # 为目标列命名为当前文件名（不含扩展名）
      colnames(sample_data)[4] <- tools::file_path_sans_ext(basename(file))
      
      # 合并数据
      if (is.null(merged_data)) {
        merged_data <- sample_data
      } else {
        merged_data <- full_join(merged_data, sample_data, by = c("gene_id", "gene_name", "gene_type"))
      }
    }
  }
  
  # 保存合并结果
  output_file <- file.path(output_dir, paste0(column_name, "_merged.tsv"))
  write_tsv(merged_data, output_file)
  cat("保存合并表到：", output_file, "\n")
  
  return(merged_data) # 返回合并后的数据框
}
input_dir <- getwd()
output_dir <- "/Users/zhengyiyi/Desktop/res"
column_name <- "unstranded"
# 调用函数合并 unstranded 列
res <- list()
for (column_i in c("unstranded",    "stranded_first",   "stranded_second",  "tpm_unstranded",   "fpkm_unstranded", "fpkm_uq_unstranded")){
  print(column_i)
  res[[column_i]] <- merge_column(input_dir, column_i, output_dir)
}