热图进阶实战

生信喵

背景

热图进阶

场景一：sample间的相关性，用来评价生物学重复之间的重复性，或者体现不同组的差异

场景二：基因的表达变化聚类、sample聚类

输入数据

df<-read.table("mRNA.txt",row.names = 1,header = T,as.is = T)
head(df)
#查看一共有多少个基因，多少个sample
dim(df)

df前几行如下：

开始画图

场景一：sample间的相关性，用来评价生物学重复之间的重复性，或者不同组的差异

#计算每两个sample之间的相关系数，method可选"pearson" (default), "kendall", or "spearman"
cormat<-round(cor(df,method = "spearman"),2)

#install.packages("pheatmap")
library(pheatmap)
pheatmap(cormat,cellwidth = 8, cellheight = 8,fontsize = 8,
         color = colorRampPalette(c("#3C7DAF", "#EAF4F1","#FFFCBA", "#E83140"))(20),
         show_colnames=T,show_rownames =T,#显示sample的名字
         #border_color = "NA",#默认有边框，不要边框的话就加这行
         treeheight_row = "0",treeheight_col = "0")#不画树

输出pdf文件，只需加一行：filename="文件名"

pheatmap(cormat,cellwidth = 8, cellheight = 8,fontsize = 8,
         color = colorRampPalette(c("#3C7DAF", "#EAF4F1","#FFFCBA", "#E83140"))(20),
         show_colnames=T,show_rownames =T,
         #border_color = "NA",#默认有边框，不要边框的话就加这行
         treeheight_row = "0",treeheight_col = "0",
         filename="Correlation.pdf")

如果sample数量少，还可以在热图里显示数字

只需加一行：display_numbers = TRUE

此处用20个sample展示效果

pheatmap(cormat[5:25,5:25],cellwidth = 15, cellheight = 15,fontsize = 8,
         color = colorRampPalette(c("#3C7DAF", "#EAF4F1","#FFFCBA", "#E83140"))(20),
         show_colnames=T,show_rownames =T,
         display_numbers = TRUE,#显示数字
         treeheight_row = "0",treeheight_col = "0")#不画树

有些热图，同组sample用同一个颜色表示，那些色块也是同时画出来的

要为每组设置颜色

#先查看有哪些sample，顺序是怎样的
colnames(df)
#按照sample的顺序，告诉R，它是属于哪个组的
annotation_col = data.frame(
  type = factor(rep(c("DG","PFC","PCC","CA1","CB","OC","TC","PC"),c(8,8,8,8,8,8,8,8))))
rownames(annotation_col) = colnames(df)

annotation_row = data.frame(
  type = factor(rep(c("DG","PFC","PCC","CA1","CB","OC","TC","PC"),c(8,8,8,8,8,8,8,8))))
rownames(annotation_row) = colnames(df)

#然后给每个组设置颜色
ann_colors = list(
  type = c(DG = "blue", PFC = "green", PCC = "red", CA1 = "black",
  CB = "pink", OC = "grey", TC = "yellow", PC = "purple")
)

pheatmap(cormat,cellwidth = 8, cellheight = 8,
         fontsize = 8,
         #display_numbers = TRUE,
         color = colorRampPalette(c("navy", "white", "firebrick3"))(20),
         show_colnames=F,show_rownames =F,#不显示sample的名字
         annotation_col = annotation_col, annotation_row = annotation_row,
         annotation_colors = ann_colors,
         treeheight_row = "0",treeheight_col = "0",#不画树
         border_color = "NA")#不显示边框

场景二：基因的表达变化聚类、sample聚类

library(pheatmap)
#有时会有多种分组方式，那就分别告诉R，此处增加一个性别组
annotation_col = data.frame(
  Gender = factor(rep(c("F","M"), 32)), # 按性别分组
  type = factor(rep(c("DG","PFC","PCC","CA1","CB","OC","TC","PC"), c(8,8,8,8,8,8,8,8))))
rownames(annotation_col) = colnames(df)

ann_colors = list(
  Gender = c(F = "#FFA42D", M = "#A9D9DF"), # 给性别分组设置颜色
  type = c(DG = "blue", PFC = "green", PCC = "red", CA1 = "black",
           CB = "pink", OC = "grey", TC = "yellow", PC = "purple")
)

# 此处用前500行基因画图
# 实际作图时，先筛差异基因，再用差异基因画图；或者用变化大的Top几千个基因画图
pheatmap(df[1:500,], cellwidth = 8, cellheight = 1, fontsize = 8,
         method="spearman", # 计算gene或sample之间的相关性的方法，可选"pearson" (default), "kendall", or "spearman"
         scale="row", # 为基因做scale
         cluster_rows=T, # 为基因做聚类
         cluster_cols=T, # 为sample做聚类
         color = colorRampPalette(c("navy", "white", "firebrick3")) (20),
         show_colnames=F,show_rownames =F,
         annotation_col = annotation_col,
         annotation_colors = ann_colors,
         # treeheight_row = "0",treeheight_col = "0", # 不画树
         border_color = "NA")

生信喵

文中所用数据可以关注公众号“生信喵实验柴”

发送关键词“20250321”获取