如何使用ggplot2绘制y轴两侧分开展示的非堆叠分组条形图

数据老陈

2026-9-1

问题原因

代码输出堆叠条形图，是两个关键步骤缺失导致的：

geom_bar()默认采用position = "stack"堆叠模式，同位置的多组条形会自动堆叠
要实现y轴两侧分列的双向条形图，必须将其中一个分组的数值转为负值，否则两组都是正值无法分布在0基准线两侧

修正方案

绘图时将其中一组的表达量取负值，让两组数值分别落在0轴两侧
显式指定geom_bar()的位置参数为position = "identity"，关闭默认堆叠逻辑
调整y轴刻度标签，将负刻度转换为正数显示，符合常规阅读习惯
补充0基准线，明确分隔两侧分组

完整可运行修正代码

library(tidyverse)

# 原始数据
Data <- structure(list(SampleID = c("Sample_1", "Sample_2", "Sample_3", 
                            "Sample_4", "Sample_5", "Sample_6"), Sustained_CR = c("NO", "NO", 
                                                                                  "NO", "YES", "YES", "YES"), Gene_A = c(2.245937679, 0, 0, 1.128343065, 
                                                                                                                         0, 0), Gene_CCC = c(83.09969414, 16.10799215, 122.2752332, 57.54549633, 
                                                                                                                                             13.12780895, 62.70159708), Gene_XED = c(6.737813038, 1.006749509, 
                                                                                                                                                                                     12.80978634, 1.128343065, 0, 2.508063883), RNA = c(580.5748901, 
                                                                                                                                                                                                                                        312.0923478, 347.0287571, 292.2408539, 355.7636226, 125.4031942
                                                                                                                                                                                     )), row.names = c("Sample_1", "Sample_2", "Sample_3", "Sample_4", 
                                                                                                                                                                                                       "Sample_5", "Sample_6"), class = "data.frame")

# 长格式转换、均值计算（和原有逻辑一致，补充.groups参数消除版本提示）
Data_long <- Data[, -c(1)] %>% 
  pivot_longer(cols = -"Sustained_CR",names_to="Gene_Symbols",values_to="Normalized expression values")
Data_mean <- Data_long %>% 
  group_by(Gene_Symbols, Sustained_CR) %>% 
  summarize(Norm_exp=mean(`Normalized expression values`, na.rm = T), .groups = "drop")

# 修正后的绘图代码
ggplot(Data_mean, aes(
  x = reorder(Gene_Symbols, Norm_exp, FUN = max),
  y = ifelse(Sustained_CR == "YES", -Norm_exp, Norm_exp),
  fill = Sustained_CR
)) +
  geom_bar(stat = "identity", position = "identity", width = 0.7) +
  # 将y轴负刻度转为正数显示
  scale_y_continuous(labels = function(x) abs(x)) +
  coord_flip() +
  # 添加中间0值基准线
  geom_hline(yintercept = 0, linewidth = 0.5) +
  labs(x = "Gene_Symbols", y = "Normalized expression values", fill = "Sustained_CR") +
  theme_bw()

自定义调整提示

若要调换两个分组的左右展示位置，修改ifelse(Sustained_CR == "YES", -Norm_exp, Norm_exp)里的分组判断逻辑，将NO组设为负值即可
若要调整基因排序规则，可以修改reorder()内的排序函数，比如按两组表达量总和排序
可通过scale_fill_manual()自定义两个分组的填充颜色，匹配目标可视化风格

效果对照

当前错误堆叠效果：

预期双向条形效果：

内容的提问来源于stack exchange，提问作者MOHAMMED TOUFIQ