关注公众号，发送R语言 或python,可获取资料

加入我们，一起探索数据科学如何赋能生物医学 ，开启精准医疗的新篇章。🚀🌐🧬

前言

之前的推文中，卡卡已经带大家学习了Pan-cancer T cell atlas links a cellular stress response state to immunotherapy resistance文献中的CD8T细胞功能基因集评分，今天卡卡继续带大家学习这篇推文中的CD4T细胞功能基因集评分。往期推文如下：

跟着顶刊学习单细胞CD8T细功能基因集评分！

回复20250411获取本期数据的CD4T细胞功能基因集评分！

先上图：

❝

c, Heat map displaying expression of 16 curated gene signatures (as listed in Supplementary Table 6) across CD4+ T cell clusters.

实战

加载单细胞数据集

# 加载必要的库
library(Seurat)
library(tidyverse)
library(pheatmap)
library(scales) # 用于rescale函数
##这里加载自己的CD8T细胞处理好的Seurat对象即可
load("CD4_Obj.Rdata")
CD4_Obj$seurat_clusters=as.factor(CD4_Obj$seurat_clusters)

加载CD4T细胞功能基因集

library(scales)
# 设置输出路径
figurePath <- "./CD4_figures"# 输出图表的目录
dir.create(figurePath, showWarnings = FALSE, recursive = TRUE)
Idents(CD4_Obj) <- CD4_Obj$seurat_clusters
# 读取标记基因 - 正确处理宽格式CSV
CD4_markers <- read_csv("CD4-markers.csv", skip = 1)  # 如果第一行是标题行，跳过它

# 将宽格式转换为标记基因列表
marker.list <- list()
for(col_name in colnames(CD4_markers)) {
# 提取非NA的基因
  genes <- CD4_markers[[col_name]]
  genes <- genes[!is.na(genes) & genes != ""]

# 替换列名中的特殊字符
  clean_name <- gsub("/", ":", col_name)

# 添加到列表
if(length(genes) > 0) {
    marker.list[[clean_name]] <- genes
  }
}

基因集评分

CD4_Obj <-  AddModuleScore(CD4_Obj,
                           features = marker.list,
                           ctrl = 5,
                           name = "FunctionScore")
for(i in1:length(marker.list)) {
  colnames(CD4_Obj@meta.data)[colnames(CD4_Obj@meta.data) == paste0("FunctionScore", i)] <- names(marker.list)[i]
}
Idents(CD4_Obj) <- CD4_Obj$seurat_clusters
Differentiation <- c("Naive", "Activation:Effector function")
Function <- c("TCR signaling", "Cytotoxicity", "Cytokine:Cytokine receptor",
              "Chemokine:Chemokine receptor", "Stress response", "Adhesion",
              "IFN response", "Treg signature", "Costimulatory molecules")
Metabolism <- c("OXPHOS", "Glycolysis", "Lipid metabolism")
Apoptosis <- c("Pro-apoptosis", "Anti-apoptosis")
MarkerNameVector <- c(Differentiation, Function, Metabolism, Apoptosis)
FunctionScoreMatrix <- matrix(0,
                              ncol = length(unique(CD4_Obj$seurat_clusters)),
                              nrow = length(MarkerNameVector))
colnames(FunctionScoreMatrix) <- paste0("CD4_c", unique(CD4_Obj$seurat_clusters))
rownames(FunctionScoreMatrix) <- MarkerNameVector
for(ci in1:ncol(FunctionScoreMatrix)) {
for(ri in1:nrow(FunctionScoreMatrix)) {
    FunctionVec <- as_tibble(CD4_Obj@meta.data) %>% pull(MarkerNameVector[ri])
    fv <- mean(FunctionVec[CD4_Obj$seurat_clusters == unique(CD4_Obj$seurat_clusters)[ci]])
    FunctionScoreMatrix[ri, ci] <- fv
  }
}

热图绘制

FunctionScoreMatrix <- t(apply(FunctionScoreMatrix, 1, rescale, to=c(-1, 1)))
my.breaks <- c(seq(-1, 0, by=0.1), seq(0.1, 1, by=0.1))
my.colors <- c(
  colorRampPalette(colors = c("#6DCCFD", "white"))(length(my.breaks)/2),
  colorRampPalette(colors = c("white", "#FD9AA0"))(length(my.breaks)/2))
signatureType_row <- data.frame(Signature.type = c(
  rep("Differentiation", length(Differentiation)),
  rep("Function", length(Function)),
  rep("Metabolism", length(Metabolism)),
  rep("Apoptosis", length(Apoptosis))))
rownames(signatureType_row) <- MarkerNameVector
col_name_v <- paste0("CD4_c", c(1,3,8,7,9,10,2,6,4,11,0,5))
FunctionScoreMatrix <- FunctionScoreMatrix[, col_name_v]
pheatmap(FunctionScoreMatrix,
         show_colnames = T,
         show_rownames = T,
         annotation_row = signatureType_row,
         gaps_row = c(2, 11, 14),
         cluster_rows = F,
         cluster_cols = F,
         breaks = my.breaks,
         color = my.colors,
         border_color = "NA",
         fontsize = 8,
         width = 5,
         height = 3.8,
         filename = file.path(figurePath, paste0("CD4_FunctionScore_heatmap.pdf")))

后续AI修整图片

结束语

卡卡学习这篇文献收获了很多内容，不仅是通路基因集的评分，里面的marker也可以收集当作自己去定义细胞亚型的基因。后台回复20250411获取本期数据的CD4T细胞功能基因集评分！