跟着Nature学作图：R语言ggplot2山脊图添加辅助线/图例添加到左下角

用户7010445

发布于 2023-01-06 19:09:30

7520

发布于 2023-01-06 19:09:30

文章被收录于专栏：小明的数据分析笔记本

论文

Graph pangenome captures missing heritability and empowers tomato breeding

https://www.nature.com/articles/s41586-022-04808-9#MOESM8

没有找到论文里的作图的代码，但是找到了部分组图数据，我们可以用论文中提供的原始数据模仿出论文中的图

今天的推文重复一下论文中的Figure2a

image.png

主要知识点

如何在山脊图上添加辅助线
另外一个知识点是如何把图例放到整个图的左下角

部分示例数据截图

image.png

读取数据

library(readxl)

dat.fig2a<-read_excel("data/20220711/41586_2022_4808_MOESM6_ESM.xlsx",
                      sheet = "Fig2a",
                      skip = 1)

数据转换为长格式

library(tidyverse)
library(stringr)
reshape2::melt(dat.fig2a) %>% 
  select(variable,value) %>% 
  mutate(new_col01 = str_split_fixed(variable,'_',2)[,1],
         new_col02 = str_split_fixed(variable,'_',2)[,2]) -> new.df

这里还有一个知识点是指定分隔符拆分字符串函数 str_split_fixed()可以指定拆分成多少个

赋予因子水平

new.df$new_col02<-factor(new.df$new_col02,
                         levels = c("snps","indels","svs",
                                    "snps_indels","snps_indels_svs"))

基本作图代码

ggplot(data=new.df,aes(x=value,y=new_col02))+
  geom_density_ridges(aes(fill=new_col01,color=new_col01),
                      alpha=0.4,
                      bandwidth=0.04,
                      quantile_lines=TRUE, 
                      quantile_fun=function(x,...)mean(x),
                      #linetype="dashed",
                      scale=1,
                      vline_linetype="dashed")+
  scale_fill_manual(values = c("graph"="#ca612d",
                               "linear"="#2772a7"))+
  scale_color_manual(values = c("graph"="#ca612d",
                               "linear"="#2772a7"))+
  theme_classic() +
  guides(fill="none",color="none") -> p1
p1

image.png

这里有一个问题是辅助线的位置是在平均值，这里通过一个求平均值的函数实现，如果是任意数值应该怎么做暂时想不到方法

添加文本

new.df %>% 
  group_by(new_col01,new_col02) %>% 
  summarise(mean_value=mean(value)) %>% 
  ungroup() %>% 
  mutate(new_col02 = fct_relevel(new_col02,
                                 c("snps","indels","svs",
                                   "snps_indels","snps_indels_svs"))) %>% 
  mutate(new_col03=as.numeric(new_col02)) -> new.df01

p1+
  scale_y_discrete(labels=c("SNP","Indel","SV","SNP + Indel","SNP + Indel + SV"))
  geom_text(data=new.df01 %>% filter(new_col01=="graph"),
            aes(y=new_col03+0.1,x=mean_value,
                label=round(mean_value,2)),
            hjust=-0.5,color="#ca612d")+
  geom_text(data=new.df01 %>% filter(new_col01=="linear"),
            aes(y=new_col03+0.1,x=mean_value,
                label=round(mean_value,2)),
            hjust=1.5,color="#2772a7")

绘制图例

ggplot(data=new.df,aes(x=value,y=new_col02))+
  geom_density_ridges(aes(fill=new_col01,color=new_col01),
                      alpha=0.4)+
  scale_fill_manual(values = c("graph"="#ca612d",
                               "linear"="#2772a7"),
                    name="",
                    label=c("TGG1.1-332","SL5.0-332"))+
  guides(color="none")-> p2

library(ggpubr)

as_ggplot(get_legend(p2)) -> p3

将图例和图组合到一起

library(latex2exp)

pdf(file = "Rplot13.pdf",
    width=9.4,height = 4)
p1+
    scale_y_discrete(labels=c("SNP","Indel","SV","SNP + Indel",
                              "SNP + Indel + SV"))+ 
  geom_text(data=new.df01 %>% filter(new_col01=="graph"),
            aes(y=new_col03+0.1,x=mean_value,
                label=round(mean_value,2)),
            hjust=-0.5,color="#ca612d")+
    geom_text(data=new.df01 %>% filter(new_col01=="linear"),
              aes(y=new_col03+0.1,x=mean_value,
                  label=round(mean_value,2)),
              hjust=1.5,color="#2772a7")+
  labs(x=TeX(r"(\textit{h}$^2$)"),y="")+
  annotation_custom(grob=ggplotGrob(p3),
                    xmin=-0.35,xmax = -0.35,ymin = 1,ymax = 1)+
    coord_cartesian(clip="on")+
  theme(plot.margin = unit(c(0.1,0.1,0.1,1),'cm'))+
  annotate(geom = "text",
           x=0.8,y=1.5,
           label=TeX(r"(\textit{P} = 1.70 \times 10$^{-217}$)"),
           vjust=-0.5)

dev.off()