如何为ggplot绘制的肿瘤生长曲线添加P值标注括号

我用ggplot绘制了肿瘤生长曲线，想要给图添加GraphPad Prism那样的P值标注括号。但目前找到的R函数（比如ggpubr::stat_compare_means、ggpubr::stat_pvalue_manual或ggprism::add_pvalue）都要求对比分组在x轴上，而我的肿瘤生长曲线x轴是时间，对比分组在图例中。希望实现自动添加P值括号的效果，不需要为每张图手动指定坐标。

附示例图：

• GraphPad Prism目标效果：
• 我用ggplot生成的图：

示例代码：

library(tidyverse)
library(scales)

tumorDataExample <- data.frame(
  Mouse = rep(1:15, times = 5),
  Condition = rep(c("A", "B", "C"), each = 5, times = 5),
  Day = rep(c(0, 8, 10, 11, 14), each = 15),
  Volume = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
             55.419392, 63.819999, 9.622816, 43.9999665, 34.792254, 5.470524, 23.516325, 
             49.9453705, 9.7540625, 16.9136, 13.1383, 36.872528, 18.108342, 48.700748, 
             13.197388, 86.979584, 76.9613405, 36.452668, 20.9408, 26.978522, 24.628864, 
             6.397264, 132.420195, 18.3314305, 11.1797745, 5.0688, 29.738592, 17.87125, 
             32.8653, 19.9408, 135.7811505, 99.900178, 60.76815, 25.6632, 33.519087, 
             25.1935245, 12.581856, 294.5889, 38.69775, 15.056676, 7.7533775, 21.625065, 
             19.62, 55.118336, 16.1063615, 200.485611, 162.4776305, 104.56355, 40.106664, 
             66.2195385, 16.143108, 4.2439, 886.51492, 17.37468, 15.867072, 7.791552, 
             11.3016465, 37.626644, 74.365425, 4.080501)
)

tumorDataExampleSummary <- tumorDataExample %>%
  group_by(Day, Condition) %>%
  summarize(meanVolume = mean(Volume, na.rm = TRUE),
            se = sd(Volume) / sqrt(n()))

tumorSizeExample_Summary <- ggplot() +
  geom_line(data = tumorDataExampleSummary, aes(x = Day, y = meanVolume, color = Condition)) +
  geom_errorbar(data = tumorDataExampleSummary, aes(x = Day, ymin = meanVolume - se, ymax = meanVolume + se, color = Condition), width = 0.1) +
  geom_point(data = tumorDataExampleSummary, aes(x = Day, y = meanVolume, color = Condition)) +
  scale_y_continuous(breaks = pretty_breaks(n=8)) +
  scale_x_continuous(breaks = pretty_breaks(n=3)) +
  labs(title = "Tumor Volumes Over Time",
       x = "Days After Tumor Injection",
       y = expression("Tumor Volume (mm"^3*")")) +
  theme_minimal()

tumorSizeExample_Summary

要实现时间轴上每个时间点的组间P值标注，核心思路是先按时间分组计算组间差异的P值，再把P值和对应位置映射到图上。以下是两种可行的实现方式：

方法1：用ggsignif自动生成标注 #

ggsignif的geom_signif支持按分组变量指定对比，结合自动计算的位置适配时间点。

library(tidyverse)
library(scales)
library(ggsignif)
library(rstatix)

# 1. 计算每个Day的组间两两P值
p_values <- tumorDataExample %>%
  group_by(Day) %>%
  pairwise_t_test(Volume ~ Condition, p.adjust.method = "bonferroni") %>%
  # 整理标注所需的位置信息
  rowwise() %>%
  mutate(
    x = Day,
    xend = Day,
    # 自动计算每个时间点的标注起始y位置：最大均值+标准误的1.1倍
    y = tumorDataExampleSummary %>% filter(Day == !!Day) %>% pull(meanVolume + se) %>% max() * 1.1
  )

# 2. 绘制带P值标注的图
tumorSizeExample_Summary +
  geom_signif(
    data = p_values,
    aes(x = x, xend = xend, y = y, yend = y + 50,
        annotations = paste0("p = ", round(p.adj, 4))),
    manual = TRUE,
    tip_length = 0.01
  )

方法2：用ggprism::add_pvalue结合自定义位置 #

如果习惯使用ggprism工具链，可以先整理P值表格，再自动适配每个时间点的坐标：

library(tidyverse)
library(scales)
library(ggprism)
library(rstatix)

# 1. 整理P值表格，包含位置信息
p_table <- tumorDataExample %>%
  group_by(Day) %>%
  pairwise_t_test(Volume ~ Condition, p.adjust.method = "bonferroni") %>%
  rowwise() %>%
  mutate(
    group1 = Condition1,
    group2 = Condition2,
    x = Day, # 对应时间点的x坐标
    # 自动计算标注的y位置：最大均值+标准误的1.2倍
    y = tumorDataExampleSummary %>% filter(Day == !!Day) %>% pull(meanVolume + se) %>% max() * 1.2
  ) %>%
  select(x, y, group1, group2, p.adj)

# 2. 添加P值标注
tumorSizeExample_Summary +
  add_pvalue(
    p_table,
    x = "x",
    y = "y",
    label = "p = {round(p.adj, 4)}",
    bracket.size = 0.8,
    step.increase = 0.1 # 同一时间点多个对比时自动错开标注位置
  )

关键说明 #

• 代码中y的位置基于每个时间点的最大均值+标准误乘以系数（如1.1、1.2），实现自动适配曲线高度，无需手动指定坐标；可根据图表需求调整系数或偏移量。
• 可根据数据类型替换统计方法（如t.test换为wilcox.test），或调整P值校正方法。

内容的提问来源于stack exchange，提问作者Kevin Lenzi