我来帮你搞定这个颜色映射错位的问题，核心原因是你当前通过多个geom_point分别指定颜色标签，容易让ggplot的颜色映射顺序出现偏差。下面给你两种解决方案，推荐第一种更规范的做法：

方案一：重构数据（推荐，符合ggplot最佳实践） #

ggplot更适合处理「长格式」数据，把四个子集合并成一个带分组列的数据框，既能简化代码，又能从根源避免颜色映射混乱：

1. 先给每个子集添加分组标识列：

# 为每个子集添加分组列
y_mers$kmer_group <- "y_mers"
x_mers$kmer_group <- "x_mers"
z_mers$kmer_group <- "z_mers"
w_mers$kmer_group <- "w_mers"

# 合并所有子集到一个数据框
merged_subsets <- rbind(y_mers, x_mers, z_mers, w_mers)

2. 修改ggplot代码，用合并后的数据集映射颜色：

ggplot(regression_with_pred, aes(x= countF, y= countM)) + 
  geom_point(alpha = 0.5, size=0.1) + 
  geom_smooth(method = lm, se=F, color ="#e08214") + 
  geom_ribbon(aes(ymin=lwr, ymax=upr, fill="#e08214"), alpha = 0.2) + 
  ggtitle(samplename) + 
  theme_bw() + 
  # 用合并后的数据集，将color映射到分组列
  geom_point(data=merged_subsets, aes(x= countF, y= countM, color = kmer_group), alpha = 0.5, size=0.1) + 
  scale_color_manual(
    name = "Kmer association", 
    values = c(
      "y_mers" = "#053061", 
      "x_mers" = "#b2182b", 
      "z_mers" = "#2166ac", 
      "w_mers"="#d6604d"
    ),
    # 可选：指定图例顺序（和values的键对应即可）
    breaks = c("y_mers","x_mers","z_mers","w_mers")
  ) + 
  scale_fill_identity()

方案二：在现有代码基础上修复 #

如果不想合并数据，只需给scale_color_manual添加breaks参数，强制指定类别顺序，确保颜色和标签一一对应：

ggplot(regression_with_pred, aes(x= countF, y= countM)) + 
  geom_point(alpha = 0.5, size=0.1) + 
  geom_smooth(method = lm, se=F, color ="#e08214") + 
  geom_ribbon(aes(ymin=lwr, ymax=upr, fill="#e08214"), alpha = 0.2) + 
  ggtitle(samplename) + 
  theme_bw() + 
  geom_point(data=y_mers, aes(x= countF, y= countM, color = "y_mers"), alpha = 0.5, size=0.1) + 
  geom_point(data=x_mers, aes(x= countF, y= countM, color = "x_mers"), alpha = 0.5, size=0.1) + 
  geom_point(data=z_mers, aes(x= countF, y= countM, color = "z_mers"), alpha = 0.5, size=0.1) + 
  geom_point(data=w_mers, aes(x= countF, y= countM, color = "w_mers"), alpha = 0.5, size=0.1) + 
  scale_color_manual(
    name = "Kmer association", 
    values = c("y_mers" = "#053061", "x_mers" = "#b2182b", "z_mers" = "#2166ac", "w_mers"="#d6604d"),
    # 关键：用breaks指定顺序，确保颜色和标签严格对应
    breaks = c("y_mers","x_mers","z_mers","w_mers"),
    labels = c("y_mers","x_mers","z_mers","w_mers")
  ) + 
  scale_fill_identity()

问题根源解释 #

你之前的代码中，多个geom_point分别设置color = "类别名"，ggplot会自动把这些字符串转成因子，但如果没有明确指定breaks，它的映射顺序可能和你预期的不一致。添加breaks后，就能强制让每个类别按你指定的顺序匹配对应的颜色值，彻底解决错位问题。

内容的提问来源于stack exchange，提问作者N. Lichilín