导读

堆叠图或pheatmap图的另一种展现形式----和弦图。这里用的是矩阵数据melt成含对应关系的三列表，有对应关系即可，是非矩阵都可以。另外circos函数博大精深，很值得学习哦。

一、模拟输入

library(circlize)
library(reshape2)

data = data.frame(matrix(sample(100, 16, replace=F), 4, 4))
rownames(data) = paste("sample", 1:4, sep=".")
colnames(data) = paste("bacteria", 1:4, sep=".")

data = data.frame(data, taxonomy = rownames(data))

input = melt(data, id='taxonomy')

二、和弦图

pdf("circ.pdf")
chordDiagram(input)
dev.off()

三、自定义顺序

倒序

queue = c(paste("sample", c(4,3,2,1), sep="."), paste("bacteria", c(4,3,2,1), sep="."))
pdf("circ_order.pdf")
chordDiagram(input, order = queue)
dev.off()

四、自定义颜色

1 连线颜色

pdf("circ_line_color.pdf")
chordDiagram(input, col = c("red", "skyblue", "pink", "yellow"))
dev.off()

2 标签颜色

pdf("circ_label_color.pdf")
chordDiagram(input, grid.col = c(sample.1="red", sample.2="skyblue", sample.3="pink", sample.4="yellow", bacteria.1="purple", bacteria.2="green", bacteria.3="grey", bacteria.4="orange"))
dev.off()

五、选择展示标签、刻度、方格

1 去标签

pdf("circ_no_name.pdf")
chordDiagram(input, annotationTrack = c("axis", "grid"), grid.col = c(sample.1="red", sample.2="skyblue", sample.3="pink", sample.4="yellow", bacteria.1="purple", bacteria.2="green", bacteria.3="grey", bacteria.4="orange"))
dev.off()

2 去刻度

pdf("circ_no_axis.pdf")
chordDiagram(input, annotationTrack = c("name", "grid"), grid.col = c(sample.1="red", sample.2="skyblue", sample.3="pink", sample.4="yellow", bacteria.1="purple", bacteria.2="green", bacteria.3="grey", bacteria.4="orange"))
dev.off()

3 去方格

刻度一起没了

pdf("circ_no_grid.pdf")
chordDiagram(input, annotationTrack = c("name", "axis"), grid.col = c(sample.1="red", sample.2="skyblue", sample.3="pink", sample.4="yellow", bacteria.1="purple", bacteria.2="green", bacteria.3="grey", bacteria.4="orange"))
dev.off()

更多参数：

small.gap = 0.001  # 组内空隙，默认是两个组
big.gap = 1  # 组建空隙
order = c()  # 自定义label的顺序
label = c(paste("u", 1:73, sep="."),
          paste("c", 1:50, sep="."), "unco_c", "unco_u")
col_set = c(rep("red", 73), rep("blue", 50), "grey", "grey")
names(col_set) = label
grid.col = col_set  # 自定义配色

比较基因组数据结构：