调整 R 中 ggarrange 上每个单独图形的图例
Adjust legends for each individual graph on ggarrange in R
我有三个 geom_point 图表，绘制了 3 个不同的变量，这些变量与 ggarrange 排列在一起。最终输出显示了相互叠加的图例。尝试 common_legend = TRUE 时，它只显示第一个的图例。是否可以安排图例，以便我为每个图表（右侧）提供三种颜色比例，然后在每个图表的某处也有变量名称。
这是一个可重现的例子：
数据集：
Samples <- c(1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17)
X = c(1.16, 1.16, 0.96, 0.96, 0.96, 0.67, 0.67, 0.67, 0.78, 0.78, 0.55, 0.3, 0.3, 0.3, 0.26, 0.26, 0.26)
Y = c(75.45, 75.45, 86.66, 86.66, 86.66, 103.36, 103.36, 103.36, NA, NA, 107.53, NA, NA, NA, 128.49, 128.49, 128.49)
AA = c(0.003437318, 0.005842468, 0.005573348, 0.006074338, 0.002537367, 0.006583666, 0.006015314, 0.010983784, 0.009116288, 0.010872489, 0.010924257, 0.009359167, 0.009068434, 0.00601658, 0.017616501, 0.014813675, 0.018048576)
BB = c(0.007614672, 0.007632451, 0.007066506, 0.007524053, 0.008337992, 0.012520277, 0.012249, 0.011351902, 0.01263021, 0.009969673, 0.008850031, 0.007290232, 0.00724349, 0.007161781, 0.004299581, 0.004896156, 0.005970637)
CC = c(0.002133046, 0.00168291, 0.001580502, 0.001491037, 0.001295399, 0.001644785, 0.001738881, 0.001496376, 0.00140218, 0.001247361, 0.001364975, 0.001209774, 0.000933038, 0.002034014, 0.000665552, 0.000855588, 0.000878233)
这是用于创建单个数据框并使用三个合并的 gig-ggplots 绘图的代码：
library(ggplot2)
library(ggpubr)
coex1 = data.frame(Samples, X, Y, AA)
coex1 <- data.frame(X,Y,value = c(AA), letters = rep(c("AA"), each = length(AA)))
coex2 = data.frame(Samples, X, Y, BB)
coex2 <- data.frame(X,Y,value = c(BB), letters = rep(c("BB"), each = length(BB)))
coex3 = data.frame(Samples, X, Y, CC)
coex3 <- data.frame(X,Y,value = c(CC), letters = rep(c("CC"), each = length(CC)))
p1 <- ggplot(coex1,aes(x=X,y=Y,shape=letters,col=value, col=value))+geom_jitter(width=0.05) + scale_color_gradient(low="blue", high="red")
p2 <- ggplot(coex2,aes(x=X,y=Y,shape=letters,col=value, size=value))+geom_jitter(width=0.05) + scale_color_gradient(low="blue", high="red")
p3 <- ggplot(coex3,aes(x=X,y=Y,shape=letters,col=value, size=value))+geom_jitter(width=0.05) + scale_color_gradient(low="blue", high="red")
ggarrange(p1, p2, p3, rremove("x.text"),
labels = c("a", "b", "c"),
ncol = 1, nrow = 3, legend = "right")
输出：
我们可以添加 legend.box = "horizontal 并为每个 ggplot 上的图例设置正确的顺序。
然后，在 ggarrange
上添加 width 和 align = "v"
p1 <- ggplot(coex1,aes(x=X, y=Y, shape=letters, col=value, col=value)) +
geom_jitter(width=0.05) +
scale_color_gradient(low="blue", high="red") +
theme(legend.box = "horizontal")+
guides(color = guide_legend(order=1),
size = guide_legend(order=2),
shape = guide_legend(order=3))
p2 <- ggplot(coex2,aes(x=X, y=Y, shape=letters, col=value, size=value)) +
geom_jitter(width=0.05) +
scale_color_gradient(low="orange", high="yellow") +
theme(legend.box = "horizontal")+
guides(color = guide_legend(order=1),
size = guide_legend(order=2),
shape = guide_legend(order=3))
p3 <- ggplot(coex3,aes(x=X, y=Y, shape=letters, col=value, size=value)) +
geom_jitter(width=0.05) +
scale_color_gradient(low="green", high="cyan") +
theme(legend.box = "horizontal")+
guides(color = guide_legend(order=1),
size = guide_legend(order=2),
shape = guide_legend(order=3))
ggarrange(p1, p2, p3 + rremove("x.text"),
labels = c("a", "b", "c"),
ncol = 1, nrow = 3,
legend = "right",
widths = c(2, 2, 3),
align = "v"
)
你用这种方法让自己的生活变得很困难。使用 ggplot 出色的分组功能，然后就不需要绘图组合包了。
您可以按美学 and/or 方面进行分组。见下文。关键是让你的数据变长，另见下文。了解它如何显着减少您的代码。
我还介绍了如何将颜色和尺寸组合成一个图例。
library(tidyverse)
mydat <- data.frame(Samples, X, Y, AA, BB, CC) %>%
pivot_longer(names_to = "letters", values_to = "value", cols = AA:CC) %>%
group_by(letters) %>%
mutate(scaled_val = scale(value)) %>%
ungroup()
ggplot(mydat, aes(X, Y, col = scaled_val, size = scaled_val)) +
geom_jitter(width = 0.05) +
scale_color_gradient(low = "blue", high = "red") +
facet_grid(~letters) +
guides(color=guide_legend(), size = guide_legend())
#> Warning: Removed 15 rows containing missing values (geom_point).
由 reprex package (v0.3.0)
于 2020-03-30 创建
数据
Samples <- c(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17)
X <- c(1.16, 1.16, 0.96, 0.96, 0.96, 0.67, 0.67, 0.67, 0.78, 0.78, 0.55, 0.3, 0.3, 0.3, 0.26, 0.26, 0.26)
Y <- c(75.45, 75.45, 86.66, 86.66, 86.66, 103.36, 103.36, 103.36, NA, NA, 107.53, NA, NA, NA, 128.49, 128.49, 128.49)
AA <- c(0.003437318, 0.005842468, 0.005573348, 0.006074338, 0.002537367, 0.006583666, 0.006015314, 0.010983784, 0.009116288, 0.010872489, 0.010924257, 0.009359167, 0.009068434, 0.00601658, 0.017616501, 0.014813675, 0.018048576)
BB <- c(0.007614672, 0.007632451, 0.007066506, 0.007524053, 0.008337992, 0.012520277, 0.012249, 0.011351902, 0.01263021, 0.009969673, 0.008850031, 0.007290232, 0.00724349, 0.007161781, 0.004299581, 0.004896156, 0.005970637)
CC <- c(0.002133046, 0.00168291, 0.001580502, 0.001491037, 0.001295399, 0.001644785, 0.001738881, 0.001496376, 0.00140218, 0.001247361, 0.001364975, 0.001209774, 0.000933038, 0.002034014, 0.000665552, 0.000855588, 0.000878233)
我有三个 geom_point 图表，绘制了 3 个不同的变量，这些变量与 ggarrange 排列在一起。最终输出显示了相互叠加的图例。尝试 common_legend = TRUE 时，它只显示第一个的图例。是否可以安排图例，以便我为每个图表（右侧）提供三种颜色比例，然后在每个图表的某处也有变量名称。
这是一个可重现的例子： 数据集：
Samples <- c(1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17)
X = c(1.16, 1.16, 0.96, 0.96, 0.96, 0.67, 0.67, 0.67, 0.78, 0.78, 0.55, 0.3, 0.3, 0.3, 0.26, 0.26, 0.26)
Y = c(75.45, 75.45, 86.66, 86.66, 86.66, 103.36, 103.36, 103.36, NA, NA, 107.53, NA, NA, NA, 128.49, 128.49, 128.49)
AA = c(0.003437318, 0.005842468, 0.005573348, 0.006074338, 0.002537367, 0.006583666, 0.006015314, 0.010983784, 0.009116288, 0.010872489, 0.010924257, 0.009359167, 0.009068434, 0.00601658, 0.017616501, 0.014813675, 0.018048576)
BB = c(0.007614672, 0.007632451, 0.007066506, 0.007524053, 0.008337992, 0.012520277, 0.012249, 0.011351902, 0.01263021, 0.009969673, 0.008850031, 0.007290232, 0.00724349, 0.007161781, 0.004299581, 0.004896156, 0.005970637)
CC = c(0.002133046, 0.00168291, 0.001580502, 0.001491037, 0.001295399, 0.001644785, 0.001738881, 0.001496376, 0.00140218, 0.001247361, 0.001364975, 0.001209774, 0.000933038, 0.002034014, 0.000665552, 0.000855588, 0.000878233)
这是用于创建单个数据框并使用三个合并的 gig-ggplots 绘图的代码：
library(ggplot2)
library(ggpubr)
coex1 = data.frame(Samples, X, Y, AA)
coex1 <- data.frame(X,Y,value = c(AA), letters = rep(c("AA"), each = length(AA)))
coex2 = data.frame(Samples, X, Y, BB)
coex2 <- data.frame(X,Y,value = c(BB), letters = rep(c("BB"), each = length(BB)))
coex3 = data.frame(Samples, X, Y, CC)
coex3 <- data.frame(X,Y,value = c(CC), letters = rep(c("CC"), each = length(CC)))
p1 <- ggplot(coex1,aes(x=X,y=Y,shape=letters,col=value, col=value))+geom_jitter(width=0.05) + scale_color_gradient(low="blue", high="red")
p2 <- ggplot(coex2,aes(x=X,y=Y,shape=letters,col=value, size=value))+geom_jitter(width=0.05) + scale_color_gradient(low="blue", high="red")
p3 <- ggplot(coex3,aes(x=X,y=Y,shape=letters,col=value, size=value))+geom_jitter(width=0.05) + scale_color_gradient(low="blue", high="red")
ggarrange(p1, p2, p3, rremove("x.text"),
labels = c("a", "b", "c"),
ncol = 1, nrow = 3, legend = "right")
输出：
我们可以添加 legend.box = "horizontal 并为每个 ggplot 上的图例设置正确的顺序。
然后，在 ggarrange
width 和 align = "v"
p1 <- ggplot(coex1,aes(x=X, y=Y, shape=letters, col=value, col=value)) +
geom_jitter(width=0.05) +
scale_color_gradient(low="blue", high="red") +
theme(legend.box = "horizontal")+
guides(color = guide_legend(order=1),
size = guide_legend(order=2),
shape = guide_legend(order=3))
p2 <- ggplot(coex2,aes(x=X, y=Y, shape=letters, col=value, size=value)) +
geom_jitter(width=0.05) +
scale_color_gradient(low="orange", high="yellow") +
theme(legend.box = "horizontal")+
guides(color = guide_legend(order=1),
size = guide_legend(order=2),
shape = guide_legend(order=3))
p3 <- ggplot(coex3,aes(x=X, y=Y, shape=letters, col=value, size=value)) +
geom_jitter(width=0.05) +
scale_color_gradient(low="green", high="cyan") +
theme(legend.box = "horizontal")+
guides(color = guide_legend(order=1),
size = guide_legend(order=2),
shape = guide_legend(order=3))
ggarrange(p1, p2, p3 + rremove("x.text"),
labels = c("a", "b", "c"),
ncol = 1, nrow = 3,
legend = "right",
widths = c(2, 2, 3),
align = "v"
)
你用这种方法让自己的生活变得很困难。使用 ggplot 出色的分组功能，然后就不需要绘图组合包了。
您可以按美学 and/or 方面进行分组。见下文。关键是让你的数据变长，另见下文。了解它如何显着减少您的代码。
我还介绍了如何将颜色和尺寸组合成一个图例。
library(tidyverse)
mydat <- data.frame(Samples, X, Y, AA, BB, CC) %>%
pivot_longer(names_to = "letters", values_to = "value", cols = AA:CC) %>%
group_by(letters) %>%
mutate(scaled_val = scale(value)) %>%
ungroup()
ggplot(mydat, aes(X, Y, col = scaled_val, size = scaled_val)) +
geom_jitter(width = 0.05) +
scale_color_gradient(low = "blue", high = "red") +
facet_grid(~letters) +
guides(color=guide_legend(), size = guide_legend())
#> Warning: Removed 15 rows containing missing values (geom_point).
由 reprex package (v0.3.0)
于 2020-03-30 创建数据
Samples <- c(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17)
X <- c(1.16, 1.16, 0.96, 0.96, 0.96, 0.67, 0.67, 0.67, 0.78, 0.78, 0.55, 0.3, 0.3, 0.3, 0.26, 0.26, 0.26)
Y <- c(75.45, 75.45, 86.66, 86.66, 86.66, 103.36, 103.36, 103.36, NA, NA, 107.53, NA, NA, NA, 128.49, 128.49, 128.49)
AA <- c(0.003437318, 0.005842468, 0.005573348, 0.006074338, 0.002537367, 0.006583666, 0.006015314, 0.010983784, 0.009116288, 0.010872489, 0.010924257, 0.009359167, 0.009068434, 0.00601658, 0.017616501, 0.014813675, 0.018048576)
BB <- c(0.007614672, 0.007632451, 0.007066506, 0.007524053, 0.008337992, 0.012520277, 0.012249, 0.011351902, 0.01263021, 0.009969673, 0.008850031, 0.007290232, 0.00724349, 0.007161781, 0.004299581, 0.004896156, 0.005970637)
CC <- c(0.002133046, 0.00168291, 0.001580502, 0.001491037, 0.001295399, 0.001644785, 0.001738881, 0.001496376, 0.00140218, 0.001247361, 0.001364975, 0.001209774, 0.000933038, 0.002034014, 0.000665552, 0.000855588, 0.000878233)