ggplot分组geom_line叠加geom_hline不显示图例问题咨询
你当前代码无法展示geom_hline图例的核心原因有两个:
- 三个图层的颜色映射变量名不同(分别是
intersect.nn、intersect.exp、seurat_clusters),ggplot会将其识别为三个独立的颜色标度,默认不会合并展示 geom_hline的linetype参数写在aes()外,属于固定赋值,不会被纳入图例生成逻辑
方案1:双图例展示(颜色区分细胞类型、线型区分数据类型)
该方案实现你提到的第二种需求,一个图例展示三种细胞类型的颜色,另一个图例单独展示三种线型对应的数据含义:
library(ggplot2) ggplot(acc.df, aes(x = K, y = prop)) + geom_line(aes(color = intersect.nn, linetype = "k梯度值")) + geom_point(aes(color = intersect.nn)) + geom_hline(data = acc.exp, aes(yintercept = prop, color = intersect.exp, linetype = "exp基准值"), show.legend = TRUE) + geom_hline(data = acc.clust, aes(yintercept = prop, color = seurat_clusters, linetype = "clust基准值"), show.legend = TRUE) + # 手动指定线型对应关系 scale_linetype_manual( values = c("k梯度值" = "solid", "exp基准值" = "dashed", "clust基准值" = "dotted"), name = "数据类型" ) + # 统一颜色标度名称,合并三个图层的颜色映射 scale_color_discrete(name = "细胞类型")
方案2:单图例合并展示(每个细胞类型对应三种线型)
该方案实现你提到的第一种需求,每个细胞类型的图例项中同时展示实线+点、虚线、点线三种样式:
library(ggplot2) library(dplyr) # 先整合所有数据集为统一长表 acc.df$type <- "k梯度值" acc.exp <- acc.exp %>% rename(intersect.nn = intersect.exp) %>% mutate(type = "exp基准值", K = NA) acc.clust <- acc.clust %>% rename(intersect.nn = seurat_clusters) %>% mutate(type = "clust基准值", K = NA) all_df <- bind_rows(acc.df, acc.exp, acc.clust) # 绘图 ggplot(all_df, aes(x = K, y = prop, color = intersect.nn, group = interaction(intersect.nn, type))) + geom_line(data = filter(all_df, type == "k梯度值"), aes(linetype = type)) + geom_point(data = filter(all_df, type == "k梯度值")) + geom_hline(data = filter(all_df, type != "k梯度值"), aes(yintercept = prop, linetype = type)) + scale_linetype_manual( values = c("k梯度值" = "solid", "exp基准值" = "dashed", "clust基准值" = "dotted"), name = "数据类型" ) + scale_color_discrete(name = "细胞类型") + # 调整图例样式,让每个细胞类型项同时展示三种线型 guides( color = guide_legend( override.aes = list( linetype = c("solid", "dashed", "dotted"), shape = c(16, NA, NA) ) ), linetype = "none" # 隐藏单独的线型图例 )
内容的提问来源于stack exchange,提问作者Gordon Beattie
相关产品推荐
相关产品推荐

