R语言ggline设置numeric.x.axis后stat_compare_means报错求助

问题场景 #

使用ggpubr的ggline绘制带统计检验的图形时，不设置numeric.x.axis = TRUE（X轴为离散型）代码能正常运行：

data("ToothGrowth")
library(ggplot2)
library(ggpubr)
ggline(ToothGrowth, x = "dose", y = "len", add = "mean_se",
       color = "supp", palette = "jco") + 
  stat_compare_means(aes(group = supp), label = "p.signif", 
                     label.y = c(16, 25, 29))

但开启numeric.x.axis = TRUE将X轴转为数值型后，运行以下代码报错：

ggline(ToothGrowth, x = "dose", y = "len", add = "mean_se",
       color = "supp", palette = "jco", numeric.x.axis = TRUE) + 
  stat_compare_means(aes(group = supp), label = "p.signif", 
                     label.y = c(16, 25, 29))

报错信息：

Computation failed in stat_compare_means(): '*.npc coord for x axis should be either a numeric value in [0-1] or a character strings including one of right, left, center, centre, middle

解决方法 #

当X轴为连续数值型时，stat_compare_means无法自动识别离散分组的X位置，需要手动指定比较的X坐标或调整位置对齐方式，以下两种方案都可行：

方案1：指定每个比较组的X轴数值位置

直接给每个dose对应的比较标签指定X轴数值（可轻微偏移避免重叠）：

ggline(ToothGrowth, x = "dose", y = "len", add = "mean_se",
       color = "supp", palette = "jco", numeric.x.axis = TRUE) +
  stat_compare_means(aes(group = supp), label = "p.signif",
                     x = c(0.5, 1.0, 2.0),  # 对应dose的三个数值
                     label.y = c(16, 25, 29))

方案2：使用position_dodge对齐分组位置

通过position_dodge让统计标签与分组线条对齐，无需手动指定X值：

ggline(ToothGrowth, x = "dose", y = "len", add = "mean_se",
       color = "supp", palette = "jco", numeric.x.axis = TRUE) +
  stat_compare_means(aes(group = supp), label = "p.signif",
                     position = position_dodge(width = 0.2),  # 调整宽度适配分组间距
                     label.y = c(16, 25, 29))

原理说明 #

X轴转为连续型后，stat_compare_means默认的离散分组位置计算逻辑失效，必须明确指定X坐标或通过位置调整函数对齐分组，避免npc坐标解析错误。

内容的提问来源于stack exchange，提问作者Janine Lux