单细胞小记1：将多个样本的表达矩阵，合并成seurat格式

1. 数据准备

# 下载所需的数据集
$ mkdir GSE20652
$  wget https://ftp.ncbi.nlm.nih.gov/geo/series/GSE206nnn/GSE206528/suppl/GSE206528_RAW.tar

# 解压
$ tar -xf GSE206528_RAW.tar -C GSE20652/

$ cd  GSE20652
$ gunzip *.gz

2. 数据整合

# 加载所需的R包
library(Seurat)
library(dplyr)
library(readr)

# 设置文件夹路径
file_path <- "/data/shumin/ED/GSE206528/"

# 1. 列出所有 CSV 文件
file_names <- list.files(path = file_path, pattern = "*.csv", full.names = TRUE)

# 2. 创建函数来读取文件和创建 meta 数据框
process_file <- function(file) {
  # 读取 CSV 文件，设置 row.names = 1 将第一列作为行名
  data <- read.csv(file, row.names = 1)
  
  # 计算列数（即重复次数）
  num_columns <- ncol(data)
  
  # 获取文件名并提取样本名和分组信息
  file_name <- basename(file)
  split_name <- strsplit(file_name, "_")[[1]]
  sample_name <- split_name[1]
  group <- split_name[2]
  
  # 创建一个 meta 数据框，每行记录一个样本的重复次数
  meta <- data.frame(
    Sample = rep(sample_name, num_columns),
    Group = rep(group, num_columns),
    stringsAsFactors = FALSE
  )
  
  return(list(data = data, meta = meta))
}

# 3. 使用 lapply 处理所有文件
results <- lapply(file_names, process_file)

# 提取所有数据和 meta 数据框
data_list <- lapply(results, `[[`, "data")
meta_list <- do.call(rbind, lapply(results, `[[`, "meta"))

# 4. 合并所有数据框
combined_data <- do.call(cbind, data_list)

# 5. 创建 Seurat 对象
seurat_obj <- CreateSeuratObject(counts = combined_data)

# 6. 将 meta 数据添加到 Seurat 对象
seurat_obj <- AddMetaData(seurat_obj, metadata = meta_list)

head(seurat_obj)
##                               orig.ident nCount_RNA nFeature_RNA     Sample  Group
## LZ037.AAACCCAAGAGGGTGG.1_1 SeuratProject       4407         1673 GSM6255907 Normal
## LZ037.AAACCCAAGATACGAT.1_1 SeuratProject       6786         2275 GSM6255907 Normal
## LZ037.AAACCCAAGATGGCAC.1_1 SeuratProject       5277         2130 GSM6255907 Normal
## LZ037.AAACCCAAGGGAGTGG.1_1 SeuratProject       5773         2050 GSM6255907 Normal
## LZ037.AAACCCAAGTTGGACG.1_1 SeuratProject       6518         2099 GSM6255907 Normal
## LZ037.AAACCCACAACTTCTT.1_1 SeuratProject       1595          838 GSM6255907 Normal
## LZ037.AAACCCACAATCTCGA.1_1 SeuratProject       6017         2047 GSM6255907 Normal
## LZ037.AAACCCACAATGCTCA.1_1 SeuratProject       3954         1401 GSM6255907 Normal
## LZ037.AAACCCACAGCCGGTT.1_1 SeuratProject      11695         3281 GSM6255907 Normal
## LZ037.AAACCCAGTATGATCC.1_1 SeuratProject       6469         2092 GSM6255907 Normal

# 数据标准化
seurat_obj <- NormalizeData(seurat_obj) %>% FindVariableFeatures() %>% ScaleData() %>% RunPCA(verbose=FALSE)

# harmony及降维
scRNA <- RunHarmony(seurat_obj, group.by.vars = "Sample")
scRNA <- FindNeighbors(scRNA, reduction = "harmony", dims = 1:20) %>% FindClusters(resolution = 0.6)
scRNA <- RunUMAP(scRNA, reduction = "harmony")
scRNA <- RunTSNE(scRNA, reduction = "harmony")