如何让ggplot柱状图按定量变量降序排列？

你之前的尝试错误地修改了y轴参数，排序的核心是调整x轴gen的因子顺序，而非y轴数值。以下是两种可行的解决方案：

方案1：全局按RQ_gapdh均值降序排列 #

如果希望所有分面的基因保持统一的降序顺序（基于每个基因的RQ_gapdh均值），只需在reorder的数值参数前加负号，即可将默认的升序改为降序：

p1v <- ggplot(data = df_plot2, aes(x = reorder(gen, -RQ_gapdh), y = RQ_gapdh, ymin = after_scale(1), fill = Color)) +
  geom_bar(stat = "identity", width=.5)+ 
  scale_fill_manual(values = c("indianred1", "olivedrab3" )) +
  labs(x = "Genes", y = expression(2^-ddCt~"value"),
       title = "Gene expression",
       subtitle = "1 year post-intervention")+  # 修正原代码的subtitles拼写错误
  guides(fill = F) +
  coord_flip()+
  facet_wrap(~grup_int, labeller = labeller(grup_int = c("1" = "Olive oil", "2" = "Nuts", "3" = "Low-fat diet")))

reorder(gen, -RQ_gapdh)会自动计算每个基因对应的RQ_gapdh均值，并按均值从高到低排列基因。

方案2：分面内独立降序排列 #

如果需要每个干预组（分面）内的基因各自按RQ_gapdh降序排列，可以结合dplyr和forcats包预处理数据：

library(dplyr)
library(forcats)

# 按分组对基因排序
df_plot2_sorted <- df_plot2 %>%
  group_by(grup_int) %>%
  mutate(gen = fct_reorder(gen, -RQ_gapdh)) %>%
  ungroup()

# 绘制图形
p1v <- ggplot(data = df_plot2_sorted, aes(x = gen, y = RQ_gapdh, ymin = after_scale(1), fill = Color)) +
  geom_bar(stat = "identity", width=.5)+ 
  scale_fill_manual(values = c("indianred1", "olivedrab3" )) +
  labs(x = "Genes", y = expression(2^-ddCt~"value"),
       title = "Gene expression",
       subtitle = "1 year post-intervention")+
  guides(fill = F) +
  coord_flip()+
  facet_wrap(~grup_int, scales = "free_x",  # 开启x轴自由缩放，避免空因子显示
             labeller = labeller(grup_int = c("1" = "Olive oil", "2" = "Nuts", "3" = "Low-fat diet")))

额外修正说明 #

原代码中labs()的subtitles是拼写错误，正确参数名是subtitle；y轴的指数表达式建议用expression()包裹，以正确渲染上标格式。

内容的提问来源于stack exchange，提问作者Javier Hernando