双分组双变量场景下创建4条geom_hlines的实现问题

1. 提前计算分组均值数据 #

要实现按Responder分组，为Ncholest（Cholesterol）和Nldl_cho（LDL Cholesterol）各加两条均值线，首先得单独算出这4个均值：

library(dplyr)
library(tidyr)

# 计算每个Responder分组下，两个指标的均值
mean_lines <- your_data %>%
  group_by(Responder) %>%
  summarize(
    Cholesterol = mean(Ncholest, na.rm = TRUE),
    LDL_Cholesterol = mean(Nldl_cho, na.rm = TRUE)
  ) %>%
  pivot_longer(
    cols = -Responder,
    names_to = "Metric",
    values_to = "Mean_Value"
  )

2. 修改ggplot代码（兼容errorbar与水平线） #

如果你的原代码只处理了单个指标的errorbar，需要补充另一个指标的相关图层，同时调用提前计算好的均值数据添加水平线：

推荐写法：转长格式简化代码

把宽格式数据转成长格式，避免重复编写图层：

library(ggplot2)

# 将原数据转成长格式，统一处理两个指标
long_data <- your_data %>%
  select(Responder, Ncholest, Nldl_cho) %>%
  pivot_longer(
    cols = c(Ncholest, Nldl_cho),
    names_to = "Metric",
    values_to = "Value",
    names_transform = list(Metric = ~ recode(., Ncholest = "Cholesterol", Nldl_cho = "LDL Cholesterol"))
  )

# 绘图
ggplot(long_data, aes(x = Responder, y = Value, color = Metric)) +
  # 基础图层（箱线/散点等，根据你的需求替换）
  geom_boxplot() +
  # 同时添加两个指标的errorbar（以均值±标准差为例）
  stat_summary(
    fun.data = mean_sdl,
    fun.args = list(mult = 1),
    geom = "errorbar",
    width = 0.2
  ) +
  # 添加4条分组均值水平线
  geom_hline(
    data = mean_lines,
    aes(yintercept = Mean_Value, color = Metric, linetype = Responder),
    linewidth = 1
  ) +
  # 调整标签与图例
  labs(y = "Measurement Value", color = "Indicator", linetype = "Responder Group") +
  theme_minimal()

宽格式直接修改写法

如果不想转长格式，可分别为两个指标添加图层：

ggplot(your_data, aes(x = Responder)) +
  # Cholesterol相关图层
  geom_boxplot(aes(y = Ncholest, color = "Cholesterol")) +
  stat_summary(
    aes(y = Ncholest, color = "Cholesterol"),
    fun.data = mean_sdl,
    fun.args = list(mult = 1),
    geom = "errorbar",
    width = 0.2
  ) +
  # LDL Cholesterol相关图层
  geom_boxplot(aes(y = Nldl_cho, color = "LDL Cholesterol")) +
  stat_summary(
    aes(y = Nldl_cho, color = "LDL Cholesterol"),
    fun.data = mean_sdl,
    fun.args = list(mult = 1),
    geom = "errorbar",
    width = 0.2
  ) +
  # 添加4条均值水平线
  geom_hline(
    data = mean_lines,
    aes(yintercept = Mean_Value, color = Metric, linetype = Responder),
    linewidth = 1
  ) +
  labs(y = "Measurement Value", color = "Indicator", linetype = "Responder Group") +
  theme_minimal()

关键问题说明 #

你之前的errorbar部分只覆盖了单个指标，需要为两个指标分别配置统计图层；而geom_hline无法直接在图层内按双维度（指标+分组）计算均值，必须提前把4个分组均值整理成独立数据集，再通过data参数传入geom_hline才能正确生成4条线。

内容的提问来源于stack exchange，提问作者Molly Delzio