如何在分组箱线图中连接同受试者的偏移散点？

问题背景 #

需要绘制带position_dodge的箱线图和散点，并用线条连接同一Marker组内同一Subject的散点，但原代码未达到预期效果。

原代码问题分析 #

• geom_line仅按Subject分组，会错误地跨不同Marker连接点，违背“同一Marker组内连接”的需求
• 未给geom_line设置匹配的position_dodge，导致线条位置无法对齐dodge后的散点

修正后的代码 #

library(tidyverse)
mock_data <- tibble::tribble(
  ~Subject, ~Marker, ~Cell_type, ~Value,
        "1",     "A",        "B",    70L,
        "2",     "A",        "B",    80L,
        "3",     "A",        "B",    90L,
        "1",     "A",        "T",     5L,
        "2",     "A",        "T",    10L,
        "3",     "A",        "T",    15L,
        "1",     "B",        "B",     1L,
        "2",     "B",        "B",     2L,
        "3",     "B",        "B",     3L,
        "1",     "B",        "T",    99L,
        "2",     "B",        "T",    90L,
        "3",     "B",        "T",    99L
  )

mock_data |> 
  ggplot(aes(x = Marker, y = Value)) +
  # 箱线图明确设置dodge宽度，与后续元素保持一致
  geom_boxplot(aes(fill = Cell_type), alpha = 0.2, position = position_dodge(width = 0.75)) +
  # 散点按Cell_type分组并dodge，确保与箱线图对齐
  geom_point(aes(col = Subject, group = Cell_type), position = position_dodge(width = 0.75), size = 3) +
  # 线条按Subject+Marker分组，仅连接同一Marker下同一受试者的点，同时设置dodge对齐
  geom_line(aes(group = interaction(Subject, Marker)), 
            lty = 2, col = "black", 
            position = position_dodge(width = 0.75))

关键改动说明 #

• interaction(Subject, Marker)作为分组依据：确保线条只连接同一受试者在同一标记类型下的两个细胞类型散点，避免跨组连接
• 统一position_dodge(width = 0.75)：让箱线图、散点、线条的偏移位置完全匹配，保证视觉对齐
• 保留geom_point的group = Cell_type：确保散点随箱线图按细胞类型偏移排列

内容的提问来源于stack exchange，提问作者biomiha