如何用for循环向分组柱状图手动添加多个geom_signif()标记

需求：创建分组柱状图，基于预计算的统计结果在存在显著性差异的位置添加星号，通过for循环结合ggplot2的geom_signif()实现多列间的标注。

原始数据与绘图代码 #

df <- data.frame(
  mouse = rep(1:5, each = 4),
  treatment = rep(rep(c("A", "B"), each = 2), times = 5),
  poke_type = rep(rep(c("Left", "Right"), times = 2), times = 5),
  pokes = rpois(20, lambda = 5)
)

# 修正：生成均值统计框（原代码缺少赋值步骤）
mean_df <- df %>%
  group_by(treatment, poke_type) %>%
  summarize(mean = mean(pokes), .groups = "drop")

# 预定义显著性标注信息
annotations_df <- data.frame(
  main=c("A","B","Left","Right"),
  group1=c("Left","Right","A","B"),
  group2=c("Right","Left","B","A"),
  pval_symbol=c("*","n.s","**","n.s"),
  xmin_index=c(0.8,2.8,0.8,2.8),
  xmax_index=c(1.2,3.2,2.8,3.2),
  y=c(7,7,7,7)  # 修正：调整y值至数据均值上方，避免标注超出图范围
)

# 绘制分组柱状图
plot <- mean_df %>%
  ggplot(aes(x=treatment,y=mean, fill=poke_type))+
  geom_bar(stat='identity',position='dodge',color= 'black',size=0.7)

尝试的标注代码 #

if(any(annotations_df$pval_symbol!="n.s")){
  row_index<<-which(annotations_df$pval_symbol != "n.s")
  for(i in row_index){
    index <- as.numeric(i)
    plot<- plot+geom_signif(
      data = annotations_df,
      aes(xmin=annotations_df$xmin_index[index],
          xmax=annotations_df$xmax_index[index],
          annotations=annotations_df$pval_symbol[index],
          y_position=annotations_df$y[index]),
      textsize=10,size=0.8,manual=TRUE,inherit.aes = FALSE, tip_length = 0.1)
  }
} else {}

修正后的标注实现 #

问题点说明 #

1. 原代码使用全局变量<<-易污染环境，完全无需使用
2. geom_signif()在manual=TRUE模式下，直接传入标量值即可，无需嵌套在aes()中
3. 原y值（210）远高于数据均值，导致标注超出图范围
4. 原textsize=10过大，影响视觉效果

优化代码 #

# 过滤掉无需标注的行（n.s）
sig_annot <- annotations_df %>% filter(pval_symbol != "n.s")

# 循环添加显著性标注
for (i in seq(nrow(sig_annot))) {
  plot <- plot +
    geom_signif(
      xmin = sig_annot$xmin_index[i],
      xmax = sig_annot$xmax_index[i],
      annotations = sig_annot$pval_symbol[i],
      y_position = sig_annot$y[i],
      textsize = 5,  # 调整字号至合理大小
      size = 0.8,
      manual = TRUE,
      inherit.aes = FALSE,
      tip_length = 0.1
    )
}

# 调整y轴范围，确保标注完整显示
plot <- plot + ylim(0, max(mean_df$mean) + 2)

运行后即可在分组柱状图的对应位置显示显著性星号，标注位置与样式均符合需求。

内容的提问来源于stack exchange，提问作者JLit98