ggplot2绘制一个高端的venn图

venn图在论文中的作用与地位也许不如其它图形那么重要，因此本文主要不是教大家如何画一个高端的venn图，而是通过整个画图过程中的代码来介绍一些重要的画图函数而已，希望各位观众老爷能够喜欢

library(ggvenn)
library(tidyverse)
library(ggtext)

创建数据

P <- tibble(
  A = sample(LETTERS, 15),
  B = sample(LETTERS, 15), 
  C = sample(LETTERS, 15),
  D = sample(LETTERS, 15))

绘制venn图

list(A=P$A,B=P$B,C=P$C,D=P$D) %>% 
  ggvenn(show_percentage = T,show_elements = F,label_sep = ",",
             digits = 1,stroke_color = "white",
         fill_color = c("#E41A1C", "#1E90FF", "#FF8C00",
                        "#4DAF4A", "#984EA3"),
         set_name_color = c("#E41A1C", "#1E90FF","#FF8C00","#984EA3"))

image

先将字符格式转化为数据框，因为我们想先通过inner_join()函数来进行一番操作

A <- P$A %>% as.data.frame()
B <- P$B %>% as.data.frame()
C <- P$C %>% as.data.frame()
D <- P$D %>% as.data.frame()

统计各组之间的交集元素

inner_join(A,B) %>% inner_join(.,C) %>% inner_join(.,D)

可以看到如果是2组数据之间通过inner_join( )函数进行元素统计很方便，但是多组之间该如何进行统计？

下面让我们通过自定义函数来解决这一问题

Intersect <- function (x) {  
  if (length(x) == 1) {
    unlist(x)
  } else if (length(x) == 2) {
    intersect(x[[1]], x[[2]])
  } else if (length(x) > 2){
    intersect(x[[1]], Intersect(x[-1]))
  }
}

Union <- function (x) {  
  if (length(x) == 1) {
    unlist(x)
  } else if (length(x) == 2) {
    union(x[[1]], x[[2]])
  } else if (length(x) > 2) {
    union(x[[1]], Union(x[-1]))
  }
}

diff <- function (x, y) {
  xx <- Intersect(x)
  yy <- Union(y)
  setdiff(xx, yy)
}

xx <- list(A=P$A,B=P$B,C=P$C,D=P$D)

四组中共有的

Intersect(xx) 

C,D中存在;A,B中不存在

diff(xx[c("C", "D")], xx[c("A", "B")])

B,C,D中存在;A中不存在的

diff(xx[c("B","C","D")], xx[("A")]) 

通过这种方式我们可以获得任意组合之间的元素名称,但是值得一提的是ggvenn提供交集元素展示这一功能

list(A=P$A,B=P$B,C=P$C,D=P$D) %>% 
  ggvenn(show_percentage = T,show_elements = T,label_sep = ",",
         digits = 1,stroke_color = "white",
         fill_color = c("#E41A1C", "#1E90FF", "#FF8C00",
                        "#4DAF4A", "#984EA3"),
         set_name_color = c("#E41A1C", "#1E90FF","#FF8C00","#984EA3"))

image

show_elements = T 即可展示共有的元素，但是如果元素文本过长就会影响美感,由于此图支持ggplot2语法，那我们就可以对图形进行一些特殊的注释

下面来进行一番骚操作

定义文本的位置信息

txt <- data.frame(
  x = c(1.3,2,-2),
  y = c(1.5,-0.2,-0.2),
  label = c("Y","D,F,U,X","Q,M"))  

geom_curve( )函数来绘制曲线，geom_richtext()来添加文本

list(A=P$A,B=P$B,C=P$C,D=P$D) %>% 
  ggvenn(show_percentage = T,show_elements = F,label_sep = ",",
         digits = 1,stroke_color = "white",
         fill_color = c("#E41A1C", "#1E90FF", "#FF8C00",
                        "#4DAF4A", "#984EA3"),
         set_name_color = c("#E41A1C", "#1E90FF","#FF8C00","#984EA3"))+
  geom_curve(aes(x = 0, y = 0.2,xend = 1.3,yend = 1.5),
             arrow = arrow(length = unit(0.07, "inch"),
                           ends="first"),size = 0.3,
             color = "grey30", curvature = 0.2) +
  geom_curve(aes(x = 0.1, y = -0.6,xend = 2,yend = -0.3),
             arrow = arrow(length = unit(0.07, "inch"),
                           ends="first"),size = 0.3,
             color = "grey30", curvature = -0.1) +
  geom_curve(aes(x = -0.5, y = -0.4,xend = -2,yend = -0.3),
             arrow = arrow(length = unit(0.07, "inch"),
                           ends="first"),size = 0.3,
             color = "grey30", curvature = -0.1) +
  geom_richtext(
    data = txt,
    aes(x, y, label = label),
    hjust = 1, vjust = 1,angle = 30 )

image