R绘图_热图包pheatmap

pheatmap，全称Pretty Heatmaps，直译“完美的热图”，是一款非常优秀的热图绘制R包。网上也有很多可供参考的教程^[1]^[2]，在初学R后，我也曾写过了一篇教程^[3]^[4]，但是对于形形色色的绘图参数，存在堆砌罗列，无章可循，使之不利于后来者学习。

官方示例

根据对pheatmap pdf说明文档^[5]示例的解读，参考其他网络教程，加以归纳整理，确定了绘制热图的基本元素和绘图框架。

构建数据集：创建数据、设置行名、列名

绘图参数：颜色、边框线、图例、热图块大小、标签字体大小、热图块文本、字符格式及颜色、自定义行/列名、标签方向、标题

构建注释参数：注释数据框、图例、注释颜色

聚类相关参数：归一化、聚类数目、行列聚类、聚类树高度、添加gap、分块聚类、聚类方法、自定义聚类

后面参数教程和绘图代码也是按照此流程、框架来进行演示：
由于图片较多，故只列出部分图片，其余可按代码运行依次生成

1. 构建数据集

rm(list = ls())
library(pheatmap)
#############构建数据集###############
test <-  matrix(rnorm(200), 20, 10)
test[1:10, seq(1, 10, 2)]  <-  test[1:10, seq(1, 10, 2)] + 3
test[11:20, seq(2, 10, 2)] <-  test[11:20, seq(2, 10, 2)] + 2
test[15:20, seq(2, 10, 2)] <-  test[15:20, seq(2, 10, 2)] + 4
# 设置行名、列名
rownames(test) <-  paste("Gene", 1:20, sep = "")
colnames(test) <-  paste("Test", 1:10, sep = "")

2. 绘图参数

# 默认绘图
pheatmap(test)
# 自定义热图颜色
pheatmap(test, color = colorRampPalette(c("navy", "white", "firebrick3"))(50))
# 去掉热块边框线
pheatmap(test,  color = colorRampPalette(c("navy", "white", "firebrick3"))(50), 
         border = FALSE)
# 设定边框线色
pheatmap(test,  border_color = "red")
# 不绘制图例
pheatmap(test, legend = FALSE)
# legend_breaks参数设定图例断点范围，legend_labels参数添加图例断点标签
pheatmap(test, cluster_row = FALSE, legend_breaks = -1:4, 
         legend_labels = c("0","1e-4", "1e-3", "1e-2", "1e-1", "1"))

图片：Rplot01---Rplot06

Rplot02.png

# 设置热图块大小
pheatmap(test, cellwidth = 15, cellheight = 12)
#fontsize参数设定标签字体大小
pheatmap(test, cellwidth = 15, cellheight = 12, fontsize = 8)

# 热图块内显示数值
pheatmap(test, display_numbers = TRUE)
# 设定数值字符串的格式及颜色
pheatmap(test, display_numbers = TRUE, number_format = "%.1e",number_color = "red")
# 构建矩阵文本，填充热图块
pheatmap(test, display_numbers = matrix(ifelse(test > 5, "*", ""), nrow(test)))
# 等同
pheatmap(test, display_numbers = ifelse(test > 5, "*", ""))

ifelse(test > 5, "", "") test矩阵内数字显著性判定
nrow(test) test矩阵内行数
matrix(ifelse(test > 5, "", ""), nrow(test))) 将标记显著性的矩阵构建矩阵，行数为test行数

图片：Rplot07---Rplot012

Rplot07.png
Rplot09.png
Rplot11.png

# 参数设定是否显示行名和列名
pheatmap(test,show_rownames = F,show_colnames = F)
# 自定义行标签名
labels_row = c("", "", "", "", "", "", "", "", "", "", "", "", "", "", "", 
               "", "", "Il10", "Il15", "Il1b")
pheatmap(test, labels_row = labels_row)
# 设定列标签的方向
pheatmap(test, angle_col = "45")
pheatmap(test, angle_col = "0")
# 设置标题
pheatmap(test,  main = "The Pretty Heatmaps by myself")

图片：Rplot13---Rplot18

Rplot17.png

3. 构建注释参数

############构建注释参数###############
# 构建列注释数据框
annotation_col <-  data.frame(
  CellType = factor(rep(c("CT1", "CT2"), 5)), 
  Time = 1:5
)
# 注释数据框'列名'必须和绘图矩阵'列名'一致
rownames(annotation_col) = paste("Test", 1:10, sep = "")
head(annotation_col)
# 构建行注释数据框
annotation_row = data.frame(
  GeneClass = factor(rep(c("Path1", "Path2", "Path3"), c(10, 4, 6)))
)
#注释数据框'行名'必须和绘图矩阵'行名'一致
rownames(annotation_row) = paste("Gene", 1:20, sep = "")
head(annotation_row)

# 显示列注释数据框
pheatmap(test, annotation_col = annotation_col)
# 显示列注释数据框和行注释数据框
pheatmap(test, annotation_col = annotation_col, annotation_row = annotation_row)
# 取消注释的图例
pheatmap(test, annotation_col = annotation_col, annotation_row = annotation_row,
         annotation_legend = FALSE)

图片：Rplot19---Rplot21

Rplot19.png
Rplot20.png

# 列表指定注释行和注释列的颜色
ann_colors = list(
  Time = c("white", "firebrick"),
  CellType = c(CT1 = "#1B9E77", CT2 = "#D95F02"),
  GeneClass = c(Path1 = "#7570B3", Path2 = "#E7298A", Path3 = "#66A61E")
)
#自定义注释数据框的颜色
pheatmap(test, annotation_col = annotation_col, annotation_colors = ann_colors,
         main = "Title")
pheatmap(test, annotation_col = annotation_col, annotation_row = annotation_row, 
         annotation_colors = ann_colors)
#只指定cellType颜色
pheatmap(test, annotation_col = annotation_col, annotation_colors = ann_colors[2])

图片：Rplot22---Rplot24

Rplot23.png

4. 聚类参数

############聚类参数###############
# 参数对行进行归一化（中心化）
pheatmap(test, scale = "row")
#聚类数目设定为4
pheatmap(test, kmeans_k = 4)
# 不对行进行聚类
pheatmap(test, cluster_row = FALSE)
#参数设定行和列聚类树的高度，默认为50
pheatmap(test, treeheight_row = 30, treeheight_col = 50)
# 取消按行聚类，在第10和14行处添加gap
pheatmap(test, annotation_col = annotation_col, cluster_rows = FALSE, 
         gaps_row = c(10, 14))
# 取消按行聚类,将列分成两部分再进行聚类
pheatmap(test, annotation_col = annotation_col, cluster_rows = FALSE, 
         gaps_row = c(10, 14), 
         cutree_col = 2)

图片：Rplot25---Rplot30

Rplot29.png
Rplot30.png

# 设定不同聚类方法
pheatmap(test,scale = "row", clustering_method = "average")
# 设定聚类距离的方法
pheatmap(test, scale = "row", clustering_distance_rows = "correlation")

# 自定义聚类的距离方法
drows = dist(test, method = "minkowski")
dcols = dist(t(test), method = "minkowski")
pheatmap(test, clustering_distance_rows = drows, clustering_distance_cols = dcols)
# 修饰聚类的回调函数
callback = function(hc, mat){
  sv = svd(t(mat))$v[,1]
  dend = reorder(as.dendrogram(hc), wts = sv)
  as.hclust(dend)
}
pheatmap(test, clustering_callback = callback)

图片：Rplot31---Rplot34

效果图：
按照数据-绘图参数-注释参数-聚类参数的顺序，思路清晰，可读性强，亦便于后续修改。

#按照数据-绘图参数-注释参数-聚类参数的顺序，思路清晰，可读性强，亦便于后续修改
pheatmap(test,
         color = colorRampPalette(c("navy","white","firebrick3"))(256),
         cellwidth = 20, cellheight = 20, border_color = "grey60", fontsize = 8, 
         display_numbers = matrix(ifelse(test > 5, "*", ""), nrow(test)),
         main = "The Pretty Heatmaps By Myself",
         annotation_col = annotation_col,annotation_row = annotation_row, 
         annotation_colors = ann_colors,
         scale = "row", treeheight_row = 20, treeheight_col =30,
         cutree_cols = 2, cutree_rows = 2,
         clustering_method = "average")

Rplot35.png

参考资料：