树状结构的数据在生活中非常常见，比如层次聚类的结果，这种数据通常有一种包含关系，上面一层可以分为多个分支，每个分支又可以继续分。
这种数据的可视化方法非常多，今天介绍使用ggraph包可视化，功能很多，布局也很多，很多常见的网络图都是这个包画出来的。
加载R包 # libraries library (ggraph)## Loading required package: ggplot2 library (igraph)## ## Attaching package: 'igraph' ## The following objects are masked from 'package:stats': ## ## decompose, spectrum ## The following object is masked from 'package:base': ## ## union library (tidyverse))数据结构 数据结构一般是2列，表示从哪里到哪里。
比如下面这个例子，最上面是根节点origin，然后下面有7个分支，分别是group1到group5，每个group下面都有7个sub_group。
d1 <- data.frame(from="origin" , to=paste("group" , seq(1 ,7 ), sep="" ))7 ), to=paste("subgroup" , seq(1 ,49 ), sep="_" ))## from to ## 1 origin group1 ## 2 origin group2 ## 3 origin group3 ## 4 origin group4 ## 5 origin group5 ## 6 origin group6 ## 7 origin group7 ## 8 group1 subgroup_1 ## 9 group1 subgroup_2 ## 10 group1 subgroup_3 ## 11 group1 subgroup_4 ## 12 group1 subgroup_5 ## 13 group1 subgroup_6 ## 14 group1 subgroup_7 ## 15 group2 subgroup_8 ## 16 group2 subgroup_9 ## 17 group2 subgroup_10 ## 18 group2 subgroup_11 ## 19 group2 subgroup_12 ## 20 group2 subgroup_13 ## 21 group2 subgroup_14 ## 22 group3 subgroup_15 ## 23 group3 subgroup_16 ## 24 group3 subgroup_17 ## 25 group3 subgroup_18 ## 26 group3 subgroup_19 ## 27 group3 subgroup_20 ## 28 group3 subgroup_21 ## 29 group4 subgroup_22 ## 30 group4 subgroup_23 ## 31 group4 subgroup_24 ## 32 group4 subgroup_25 ## 33 group4 subgroup_26 ## 34 group4 subgroup_27 ## 35 group4 subgroup_28 ## 36 group5 subgroup_29 ## 37 group5 subgroup_30 ## 38 group5 subgroup_31 ## 39 group5 subgroup_32 ## 40 group5 subgroup_33 ## 41 group5 subgroup_34 ## 42 group5 subgroup_35 ## 43 group6 subgroup_36 ## 44 group6 subgroup_37 ## 45 group6 subgroup_38 ## 46 group6 subgroup_39 ## 47 group6 subgroup_40 ## 48 group6 subgroup_41 ## 49 group6 subgroup_42 ## 50 group7 subgroup_43 ## 51 group7 subgroup_44 ## 52 group7 subgroup_45 ## 53 group7 subgroup_46 ## 54 group7 subgroup_47 ## 55 group7 subgroup_48 ## 56 group7 subgroup_49 画图 首先转换数据为ggraph需要的格式。
mygraph <- graph_from_data_frame(edges)'dendrogram' , circular = FALSE ) + plot of chunk unnamed-chunk-3 一个简单的树状图就画好了。接下来进行一些美化操作，看看这个包的强大之处。
美化 通过组合不同的布局和线型，可以得到非常多神奇的结果。比如大家比较喜欢的圆形布局。
theme_set(theme_void())'dendrogram' , circular = T ) + 'linear' ,circular=T ) + 'partition' ,circular=T ) + 'circlepack' ) + 'circlepack' ) + 'treemap' ) + 'tree' ) + 'unrooted' ) + 'dendrogram' , circular = TRUE ) + library (patchwork)3 )plot of chunk unnamed-chunk-4 如果给数据增加一些属性列，就会变成大家在论文中常见的各种图形了！
就是这么简单！
给节点安排一些属性信息，把这些信息放在另一个数据框里。
name <- unique(c(as.character(edges$from), as.character(edges$to)))NA ,8 ) , rep( paste("group" , seq(1 ,7 ), sep="" ), each=7 )),1 :4 ], length(name), replace=T ),10 ,30 ), length(name), replace=T )## name group cluster value ## 1 origin <NA> b 19 ## 2 group1 <NA> b 29 ## 3 group2 <NA> d 17 ## 4 group3 <NA> a 15 ## 5 group4 <NA> b 23 ## 6 group5 <NA> b 12 ## 7 group6 <NA> b 20 ## 8 group7 <NA> b 17 ## 9 subgroup_1 group1 b 30 ## 10 subgroup_2 group1 d 10 ## 11 subgroup_3 group1 d 12 ## 12 subgroup_4 group1 c 27 ## 13 subgroup_5 group1 c 21 ## 14 subgroup_6 group1 b 19 ## 15 subgroup_7 group1 c 14 ## 16 subgroup_8 group2 c 22 ## 17 subgroup_9 group2 b 17 ## 18 subgroup_10 group2 d 30 ## 19 subgroup_11 group2 a 20 ## 20 subgroup_12 group2 c 15 ## 21 subgroup_13 group2 c 16 ## 22 subgroup_14 group2 b 19 ## 23 subgroup_15 group3 b 26 ## 24 subgroup_16 group3 c 25 ## 25 subgroup_17 group3 b 14 ## 26 subgroup_18 group3 d 22 ## 27 subgroup_19 group3 a 23 ## 28 subgroup_20 group3 c 10 ## 29 subgroup_21 group3 d 28 ## 30 subgroup_22 group4 b 25 ## 31 subgroup_23 group4 b 15 ## 32 subgroup_24 group4 d 11 ## 33 subgroup_25 group4 a 18 ## 34 subgroup_26 group4 a 19 ## 35 subgroup_27 group4 c 19 ## 36 subgroup_28 group4 b 10 ## 37 subgroup_29 group5 b 22 ## 38 subgroup_30 group5 b 28 ## 39 subgroup_31 group5 d 19 ## 40 subgroup_32 group5 c 19 ## 41 subgroup_33 group5 d 11 ## 42 subgroup_34 group5 c 20 ## 43 subgroup_35 group5 d 28 ## 44 subgroup_36 group6 b 17 ## 45 subgroup_37 group6 d 19 ## 46 subgroup_38 group6 a 22 ## 47 subgroup_39 group6 b 27 ## 48 subgroup_40 group6 a 23 ## 49 subgroup_41 group6 a 24 ## 50 subgroup_42 group6 d 10 ## 51 subgroup_43 group7 c 14 ## 52 subgroup_44 group7 d 19 ## 53 subgroup_45 group7 c 30 ## 54 subgroup_46 group7 c 12 ## 55 subgroup_47 group7 d 24 ## 56 subgroup_48 group7 d 21 ## 57 subgroup_49 group7 c 20 重新再画一下：
mygraph <- graph_from_data_frame(edges, vertices=vertices)theme_set(theme_void())'dendrogram' , circular = T ) + 0.6 ) +'linear' ,circular=T ) + 'partition' ,circular=T ) + 'circlepack' ) + 'circlepack' ) + 'treemap' ) + 'dendrogram' ) + 90 , hjust=1 , nudge_y=-0.1 ) +0.6 )'unrooted' ) + 'dendrogram' , circular = TRUE ) + 4 )library (patchwork)3 )plot of chunk unnamed-chunk-7 简单，喜欢哪种就用哪种，但是要注意，我们提供的信息是点的，并不是线条的，所以如果你想让线条也有属性映射的话，需要在最开始构建from-to数据时添加线条属性哦~
如果你还不会，可以参考另一篇推文：R语言生信图表学习之网络图
