如何在配对堆叠柱状图的对应细胞组间添加连线？

我在绘制配对堆叠柱状图时遇到麻烦：没法在Pre和Post对应的细胞组之间添加连线。我要比较治疗前后（Pre→Post）的细胞组成变化，现在用下面的R脚本生成了普通堆叠柱状图，但想改成带连线的样式，作为新手，我搜过谷歌和Stack Overflow的示例，但没法把代码整合到自己的脚本里，求帮忙。

原代码：

data <- data.frame(
  Group = rep(c("Responder", "Non-responder"), each = 16), # 16 x 2 = 32
  Time = rep(c("Pre", "Post"), times = 16),               # 16 x 2 = 32
  CellType = rep(c("B", "Mono", "DC", "CD4_T", "CD8_T", "NK", "Other", "Platelet"), each = 2, times = 2), # 길이 32
  Proportion = runif(32, 0.01, 0.2)                       
)

data <- data %>%
  group_by(Group, Time) %>%
  mutate(Proportion = Proportion / sum(Proportion)) %>% 
  ungroup()

# Stacked Bar Plot 
ggplot(data, aes(x = Time, y = Proportion, fill = CellType)) +
  geom_bar(stat = "identity", position = "stack") + # Stacked bar plot
  facet_wrap(~ Group, ncol = 2) +                 
  scale_fill_brewer(palette = "Paired") +         
  theme_minimal() +                                
  labs(
    title = "Cell Type Proportions Before and After Treatment",
    x = "Time",
    y = "Cell Proportions",
    fill = "Cell Type"
  )

要实现Pre与Post对应细胞组间的连线，核心是先计算每个细胞类型在堆叠图中的上下边界位置，再用geom_line连接对应位置的中点。以下是修改后的完整代码及关键解释：

修改后的完整代码 #

library(tidyverse)

data <- data.frame(
  Group = rep(c("Responder", "Non-responder"), each = 16), 
  Time = rep(c("Pre", "Post"), times = 16),               
  CellType = rep(c("B", "Mono", "DC", "CD4_T", "CD8_T", "NK", "Other", "Platelet"), each = 2, times = 2), 
  Proportion = runif(32, 0.01, 0.2)                       
)

# 计算比例及堆叠位置参数
data <- data %>%
  group_by(Group, Time) %>%
  mutate(
    Proportion = Proportion / sum(Proportion),
    # 堆叠顶部位置：累积求和
    y_end = cumsum(Proportion),
    # 堆叠底部位置：顶部位置减去当前比例
    y_start = y_end - Proportion,
    # 取中间值作为连线的锚点，确保连线穿过细胞组中心
    y_mid = (y_start + y_end) / 2
  ) %>% 
  ungroup()

# 绘制带连线的堆叠柱状图
ggplot(data, aes(x = Time, y = Proportion, fill = CellType)) +
  # 堆叠柱状图主体
  geom_bar(stat = "identity", position = "stack") +
  # 添加组间连线：按Group+CellType分组，连接Pre/Post的中点
  geom_line(aes(y = y_mid, group = interaction(Group, CellType)), color = "black", linewidth = 0.5) +
  facet_wrap(~ Group, ncol = 2) +
  scale_fill_brewer(palette = "Paired") +
  theme_minimal() +
  labs(
    title = "治疗前后细胞类型比例变化",
    x = "时间点",
    y = "细胞比例",
    fill = "细胞类型"
  )

关键逻辑说明 #

• 计算堆叠位置：通过cumsum(Proportion)得到每个细胞类型在堆叠柱中的顶部边界，再减去自身比例得到底部边界，取两者中点作为连线的锚点，保证连线精准对应到每个细胞组的中心。
• 连线分组：用interaction(Group, CellType)作为group参数，确保每个分组内的同一细胞类型，能在Pre和Post两个时间点之间建立唯一的连线。

内容的提问来源于stack exchange，提问作者juyong seong