【原】跟着iMeta学做图｜ggplot2绘制多个饼图展示菌群物种组成

原始教程链接：https://github.com/iMetaScience/iMetaPlot/tree/main/221017multi-pieplot

写在前面

饼图 (Pie Plot) 在微生物组研究中可以用来展示菌群物种组成，可以起到与堆叠柱状图相同的展示效果。本期我们挑选2022年4月5日刊登在iMeta上的The impact of aquaculture system on the microbiome and gut metabolome of juvenile Chinese softshell turtle (Pelodiscus sinensis)- iMeta | 南昌大学丁霞等-水产养殖模式对水产动物皮肤、口腔和肠道微生物群落组装及宿主适应性的影响，选择文章的Figure 1D进行复现，讲解和探讨简单饼图以及同一图片中呈现多个饼图的方法，先上原图：

代码编写及注释：农心生信工作室

R包检测和安装

01

安装核心R包ggplot2以及一些功能辅助性R包，并载入所有R包

if (!require("ggplot2"))  install.packages('ggplot2') if (!require("dplyr"))  install.packages('dplyr') if (!require("ComplexHeatmap"))  BiocManager::install('ComplexHeatmap')#> #>   有二进制版本的，但源代码版本是后来的:#>      binary source needs_compilation#> clue 0.3-61 0.3-62              TRUE#> #> #> 下载的二进制程序包在#>  /var/folders/15/ywvz065n3jl4qm8jygpq5bz80000gn/T//RtmpLnQspz/downloaded_packages里# 加载包library(ggplot2)library(dplyr)library(ComplexHeatmap)library(grid)

读取或生成数据

02

设该图数据来自文章的补充文件supplementary table3，大家可以根据链接自行下载。在这里，我们下载它的补充文件后，导出为test.CSV进行读取。

#读取数据df<-read.csv("test.CSV",header = T)#创建一个向量，包含了图中将要展示的丰度最高的7个科top_phylum=c("Actinobacteria","Bacteroidetes","Firmicutes","Fusobacteria","Proteobacteria","Verrucomicrobia","[Thermi]")#将其他低丰度的科统一命名为Otherdf[!(df$Taxonomyt %in% top_phylum),]$Taxonomyt = "Other"

饼图预览

03

选择样本abdomen，使用ggplot2包绘制一个最简单的饼图，方法与绘制柱状图一样，只是将直角坐标系转化为了极坐标系：

#选择样本abdomenper_df<-df[df$Sample=="abdomen",]#使用aggregate函数，根据Taxonomyt的分组，将科名称相同的丰度求和，即最终得到Other的丰度的和per_df<-aggregate(per_df$Abundance,by=list(Taxonomyt=per_df$Taxonomyt),sum) %>% rename(Abundance=x)p<-ggplot(per_df,aes("",Abundance,fill=Taxonomyt))+geom_bar(stat = "identity")+  coord_polar(theta = "y")

04

根据原图，我们固定图例中科名显示的顺序，并美化图片，设置颜色，去除背景，坐标轴，图例：

mycol<-c("#F7F114","#168936","#F2C8AD","#0BE9F4","#0D5ED3","#F2AD3D","#757272","#EA1313")#设置颜色per_df<-df[df$Sample=="abdomen",]per_df<-aggregate(per_df$Abundance,by=list(Taxonomyt=per_df$Taxonomyt),sum) %>% rename(Abundance=x)per_df$Taxonomyt<-factor(per_df$Taxonomyt,levels =c("Actinobacteria","Bacteroidetes","Firmicutes","Fusobacteria","Proteobacteria","Verrucomicrobia","[Thermi]","Other"))#固定图例顺序p<-ggplot(per_df,aes("",Abundance,fill=Taxonomyt))+geom_bar(stat = "identity")+  coord_polar(theta = "y")+  scale_fill_manual(values = mycol)+  guides(fill="none")+  theme(axis.text.y = element_blank(),panel.background = element_blank(),        line = element_blank(),axis.ticks.y = element_blank(),axis.text.x = element_blank())

05

接下来是全文的重点，如何在一张画布上绘制多个饼图，并将它们有序排列。这里，我们要用assign函数，使每个样本对应一副饼图：

#首先生成一个向量，包含全部八个样本名sample_name<-unique(df$Sample)#unique()函数去除重复数据#设置颜色mycol<-c("#F7F114","#168936","#F2C8AD","#0BE9F4","#0D5ED3","#F2AD3D","#757272","#EA1313")#编写函数getPieplot()，该函数包含两个参数，第一个参数为数据框df，第二个参数为样本名称，最终返回饼图getPieplot<-function(dfname,Samplename){  per_df<-dfname[dfname$Sample==Samplename,]  per_df<-aggregate(per_df$Abundance,by=list(Taxonomyt=per_df$Taxonomyt),sum) %>% rename(Abundance=x)  per_df$Taxonomyt<-factor(per_df$Taxonomyt,levels =c("Actinobacteria","Bacteroidetes","Firmicutes","Fusobacteria","Proteobacteria","Verrucomicrobia","[Thermi]","Other"))  p<-ggplot(per_df,aes("",Abundance,fill=Taxonomyt))+    geom_bar(stat = "identity")+    coord_polar(theta = "y")+    scale_fill_manual(values = mycol)+    labs(x=Samplename,y="")+    guides(fill="none")+    theme(axis.text.y = element_blank(),panel.background = element_blank(),          line = element_blank(),axis.ticks.y = element_blank(),axis.text.x = element_blank())    return(p)}#for循环遍历八个样本名，利用assign函数，以样本名作变量名，将饼图分别赋给对应的变量for (i in sample_name){  assign(i,getPieplot(df,i))}

06

使用底层绘图包grid，按顺序将饼图一一排列：

#创建新一个新的画布grid.newpage()#创建一个4行2列的布局pushViewport(viewport(layout = grid.layout(nrow = 4, ncol = 2),width = 0.3))#编写一个函数，方便定义每一个饼图在画布布局中的具体位置vp_value <- function(row, col){  viewport(layout.pos.row = row, layout.pos.col = col)} #print将每个图形输出到布局的不同区域中print(abdomen,vp = vp_value(row = 1, col = 1))print(duodenum,vp = vp_value(row = 1, col = 2))print(back,vp = vp_value(row = 2, col = 1))print(ileum,vp = vp_value(row = 2, col = 2))print(limb,vp = vp_value(row = 3, col = 1))print(colon,vp = vp_value(row = 3, col = 2))print(oral,vp = vp_value(row = 4, col = 1))print(rectum,vp = vp_value(row = 4, col = 2))

07

最后，我们利用顾祖光博士开发的ComplexHeatmap包（关于ComplexHeatmap包的使用，可以参考往期推文跟着iMeta学做图｜ComplexHeatmap绘制多样的热图），绘制一个单独的图例，并置于画布的最下方：

pdf("Figure1D.pdf",width = 8, height = 6)grid.newpage()#重新创建一个5行2列的布局，最后一行用于放置图例pushViewport(viewport(layout = grid.layout(nrow = 5, ncol = 2),width = 0.3))vp_value <- function(row, col){  viewport(layout.pos.row = row, layout.pos.col = col)} print(abdomen,vp = vp_value(row = 1, col = 1))print(duodenum,vp = vp_value(row = 1, col = 2))print(back,vp = vp_value(row = 2, col = 1))print(ileum,vp = vp_value(row = 2, col = 2))print(limb,vp = vp_value(row = 3, col = 1))print(colon,vp = vp_value(row = 3, col = 2))print(oral,vp = vp_value(row = 4, col = 1))print(rectum,vp = vp_value(row = 4, col = 2))#创建图例lgd_points = Legend(at = c("Actinobacteria","Bacteroidetes","Firmicutes","Fusobacteria","Proteobacteria","Verrucomicrobia","[Thermi]","Other"), type = "points", pch = 15,                    legend_gp = gpar(col = mycol),                     title = "",background = mycol)#将图例与饼图合并draw(lgd_points, x = unit(45, "mm"), y = unit(5, "mm"), just = c( "bottom"))dev.off()#> quartz_off_screen #>                 2

完整代码

if (!require("ggplot2"))  install.packages('ggplot2') if (!require("dplyr"))  install.packages('dplyr') if (!require("ComplexHeatmap"))  BiocManager::install('ComplexHeatmap')#> #>   有二进制版本的，但源代码版本是后来的:#>      binary source needs_compilation#> clue 0.3-61 0.3-62              TRUE#> #> #> 下载的二进制程序包在#>  /var/folders/15/ywvz065n3jl4qm8jygpq5bz80000gn/T//RtmpLnQspz/downloaded_packages里# 加载包library(ggplot2)library(dplyr)library(ComplexHeatmap)library(grid)
#读取数据df<-read.csv("test.CSV",header = T)#创建一个向量，包含了图中将要展示的丰度最高的7个科top_phylum=c("Actinobacteria","Bacteroidetes","Firmicutes","Fusobacteria","Proteobacteria","Verrucomicrobia","[Thermi]")#将其他低丰度的科统一命名为Otherdf[!(df$Taxonomyt %in% top_phylum),]$Taxonomyt = "Other"#选择样本abdomenper_df<-df[df$Sample=="abdomen",]#使用aggregate函数，根据Taxonomyt的分组，将科名称相同的丰度求和，即最终得到Other的丰度的和per_df<-aggregate(per_df$Abundance,by=list(Taxonomyt=per_df$Taxonomyt),sum) %>% rename(Abundance=x)p<-ggplot(per_df,aes("",Abundance,fill=Taxonomyt))+geom_bar(stat = "identity")+  coord_polar(theta = "y")mycol<-c("#F7F114","#168936","#F2C8AD","#0BE9F4","#0D5ED3","#F2AD3D","#757272","#EA1313")#设置颜色per_df<-df[df$Sample=="abdomen",]per_df<-aggregate(per_df$Abundance,by=list(Taxonomyt=per_df$Taxonomyt),sum) %>% rename(Abundance=x)per_df$Taxonomyt<-factor(per_df$Taxonomyt,levels =c("Actinobacteria","Bacteroidetes","Firmicutes","Fusobacteria","Proteobacteria","Verrucomicrobia","[Thermi]","Other"))#固定图例顺序p<-ggplot(per_df,aes("",Abundance,fill=Taxonomyt))+geom_bar(stat = "identity")+  coord_polar(theta = "y")+  scale_fill_manual(values = mycol)+  guides(fill="none")+  theme(axis.text.y = element_blank(),panel.background = element_blank(),        line = element_blank(),axis.ticks.y = element_blank(),axis.text.x = element_blank())p#首先生成一个向量，包含全部八个样本名sample_name<-unique(df$Sample)#unique()函数去除重复数据#设置颜色mycol<-c("#F7F114","#168936","#F2C8AD","#0BE9F4","#0D5ED3","#F2AD3D","#757272","#EA1313")#编写函数getPieplot()，该函数包含两个参数，第一个参数为数据框df，第二个参数为样本名称，最终返回饼图getPieplot<-function(dfname,Samplename){  per_df<-dfname[dfname$Sample==Samplename,]  per_df<-aggregate(per_df$Abundance,by=list(Taxonomyt=per_df$Taxonomyt),sum) %>% rename(Abundance=x)  per_df$Taxonomyt<-factor(per_df$Taxonomyt,levels =c("Actinobacteria","Bacteroidetes","Firmicutes","Fusobacteria","Proteobacteria","Verrucomicrobia","[Thermi]","Other"))  p<-ggplot(per_df,aes("",Abundance,fill=Taxonomyt))+    geom_bar(stat = "identity")+    coord_polar(theta = "y")+    scale_fill_manual(values = mycol)+    labs(x=Samplename,y="")+    guides(fill="none")+    theme(axis.text.y = element_blank(),panel.background = element_blank(),          line = element_blank(),axis.ticks.y = element_blank(),axis.text.x = element_blank())    return(p)}#for循环遍历八个样本名，利用assign函数，以样本名作变量名，将饼图分别赋给对应的变量for (i in sample_name){  assign(i,getPieplot(df,i))}
#创建新一个新的画布grid.newpage()#创建一个4行2列的布局pushViewport(viewport(layout = grid.layout(nrow = 4, ncol = 2),width = 0.3))#编写一个函数，方便定义每一个饼图在画布布局中的具体位置vp_value <- function(row, col){  viewport(layout.pos.row = row, layout.pos.col = col)} #print将每个图形输出到布局的不同区域中print(abdomen,vp = vp_value(row = 1, col = 1))print(duodenum,vp = vp_value(row = 1, col = 2))print(back,vp = vp_value(row = 2, col = 1))print(ileum,vp = vp_value(row = 2, col = 2))print(limb,vp = vp_value(row = 3, col = 1))print(colon,vp = vp_value(row = 3, col = 2))print(oral,vp = vp_value(row = 4, col = 1))print(rectum,vp = vp_value(row = 4, col = 2))pdf("Figure1D.pdf",width = 8, height = 6)grid.newpage()#重新创建一个5行2列的布局，最后一行用于放置图例pushViewport(viewport(layout = grid.layout(nrow = 5, ncol = 2),width = 0.3))vp_value <- function(row, col){  viewport(layout.pos.row = row, layout.pos.col = col)} print(abdomen,vp = vp_value(row = 1, col = 1))print(duodenum,vp = vp_value(row = 1, col = 2))print(back,vp = vp_value(row = 2, col = 1))print(ileum,vp = vp_value(row = 2, col = 2))print(limb,vp = vp_value(row = 3, col = 1))print(colon,vp = vp_value(row = 3, col = 2))print(oral,vp = vp_value(row = 4, col = 1))print(rectum,vp = vp_value(row = 4, col = 2))#创建图例lgd_points = Legend(at = c("Actinobacteria","Bacteroidetes","Firmicutes","Fusobacteria","Proteobacteria","Verrucomicrobia","[Thermi]","Other"), type = "points", pch = 15,                    legend_gp = gpar(col = mycol),                     title = "",background = mycol)#将图例与饼图合并draw(lgd_points, x = unit(45, "mm"), y = unit(5, "mm"), just = c( "bottom"))dev.off()#> quartz_off_screen #>                 2

以上数据和代码仅供大家参考，如有不完善之处，欢迎大家指正！