如何使用cowplot为X轴和Y轴同时添加grob？

我学会了将PNG作为轴刻度标签插入ggplot，但该方法仅适用于单轴，不确定如何为图表同时绘制Y轴和X轴的Grob。

单轴添加Grob的示例代码 #

library(ggplot2)
library(cowplot)  # 原代码中cowplot2为笔误，正确包名为cowplot
library(png)
library(RCurl)  # 原代码中Rcurl大小写错误，正确包名为RCurl

# 统一数据框命名，避免后续报错
chiMate <- data.frame(
  pairType = c("assortative", "disassortative", "disassortative", "assortative", "disassortative", "disassortative", "assortative"),
  Male_Phenotype = c("metallic", "rufipennis", "metallic", "militaris-a", "metallic", "militaris-a", "rufipennis"),
  Female_Phenotype = c("metallic", "metallic", "militaris-a", "militaris-a", "rufipennis", "rufipennis", "rufipennis"),
  exp = c(0.10204082, 0.11224490, 0.28571429, 0.10204082, 0.02040816, 0.08163265, 0.29591837),
  obs = c(0.04081633, 0.02040816, 0.02040816, 0.03061224, 0.00000000, 0.00000000, 0.03061224)
)

rufiPNG<- "https://i.sstatic.net/Q8BqO.png"
rufipennis <- readPNG(getURLContent(rufiPNG))

militarisPNG<- "https://i.sstatic.net/EtdfR.png"
militaris_a <- readPNG(getURLContent(militarisPNG))

metallicPNG<- "https://i.sstatic.net/YIDoA.png"
metallic <- readPNG(getURLContent(metallicPNG))

# 绘制基础图表
p<- ggplot(chiMate, aes(x = `Male Phenotype`, y = `Female Phenotype`)) +
  geom_point(aes(color = pairType, size = 2.1*exp))+
  geom_point(color = "black", aes(size = 1.6*exp)) +
  geom_point(color = "white", aes(size = 1.3*exp)) +
  geom_point(aes(size = 1.25*obs), color= "salmon") +
  scale_color_manual("Pair type", values = c("assortative" = "cornflowerblue", "disassortative" = "aquamarine3")) +
  scale_size_continuous(range = c(10, 45))+
  scale_x_discrete(expand=c(.3,.3))+
  scale_y_discrete(expand=c(0.3,0.3))+
  geom_text(aes(label= round(obs, digits= 3)))+
  theme_minimal() +
  labs(x = "Male Phenotype", y = "Female Phenotype", size = "Mate Count")+
  theme(legend.position = 'bottom', legend.box.background = element_rect(color='black'),  axis.title = element_text(size= 14), axis.text = element_text(size= 10))+
  guides(color = guide_legend(override.aes = list(size = 12)),  size = "none")

# 创建Y轴图片画布
ystrip <- axis_canvas(p, axis = 'y') + 
  draw_image(rufipennis, y = 2.35, scale = 4) +
  draw_image(militaris_a, y = 1.5, scale = 4) +
  draw_image(metallic, y = .65, scale = 4)

# 仅添加Y轴Grob的效果
ggdraw(insert_yaxis_grob(p, ystrip, position = "left",width = grid::unit(0.05, "null")))

尝试过的无效方法 #

第一种：

ggdraw(p)+ 
insert_yaxis_grob(p, ystrip, position = "left",width = grid::unit(0.05, "null"))+
insert_xaxis_grob(p, xstrip, position = "bottom",height = grid::unit(0.1, "null"))

第二种：

ggdraw(insert_yaxis_grob(p, ystrip, position = "left",width = grid::unit(0.05, "null")),
insert_xaxis_grob(p, xstrip, position = "bottom",height = grid::unit(0.1, "null")))

解决方案 #

核心是逐步叠加Grob对象：先基于原始图插入Y轴Grob得到中间图，再基于这个中间图插入X轴Grob，最后用ggdraw()渲染。

完整代码如下：

# 创建X轴图片画布（对应X轴刻度顺序调整坐标）
xstrip <- axis_canvas(p, axis = 'x') + 
  draw_image(metallic, x = 0.3, scale = 4) +  # 对应X轴第一个值metallic
  draw_image(rufipennis, x = 1.3, scale = 4) +  # 对应X轴第二个值rufipennis
  draw_image(militaris_a, x = 2.3, scale = 4)  # 对应X轴第三个值militaris-a

# 第一步：给原始图插入Y轴Grob
p_with_y <- insert_yaxis_grob(p, ystrip, position = "left", width = grid::unit(0.05, "null"))

# 第二步：基于带Y轴Grob的图，插入X轴Grob
p_with_both <- insert_xaxis_grob(p_with_y, xstrip, position = "bottom", height = grid::unit(0.1, "null"))

# 最后渲染完整图表
ggdraw(p_with_both)

失效原因说明 #

• 第一种方法中，ggdraw(p)返回的是ggdraw对象，而insert_*_grob返回的是Grob对象，两者无法用+叠加，最终只会保留原始图。
• 第二种方法中，ggdraw()仅接受第一个参数作为渲染对象，第二个参数会被忽略，因此只有Y轴Grob生效。

内容的提问来源于stack exchange，提问作者paras