如何在 ggplot2 中定义 y 轴段
How to define y axis segments in gglpot2
我有一个数据框：
df <- structure(list(Sample = c("1: FL_643", "2: FL_645", "3: FL_647","4: FL_656", "5: FL_658", "6: cKO_644", "7: cKO_646", "8: cKO_654","9: cKO_655", "10: cKO_657", "1: FL_643", "2: FL_645", "3: FL_647", "4: FL_656", "5: FL_658", "6: cKO_644", "7: cKO_646", "8: cKO_654", "9: cKO_655", "10: cKO_657"), Genotype = structure(c(1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L), .Label = c("miR-15/16 FL", "miR-15/16 cKO"), class = "factor"),
Tissue = structure(c(2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L,
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L), .Label = c("Thymus",
"iLN", "Spleen", "Skin", "Colon"), class = "factor"), variable = structure(c(1L,
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L,
2L, 2L, 2L, 2L), .Label = c("Cells/SC/Live/CD8—,, CD4+,Freq. of Parent",
"Cells/SC/Live/CD8—,, CD4+/Foxp3+,Freq. of Parent", "Cells/SC/Live/CD8—,, CD4+/Foxp3+,Median,<BV421-A>,CD127",
"Cells/SC/Live/CD8—,, CD4+/Foxp3+/CD25+,Freq. of Parent",
"Cells/SC/Live/CD8—,, CD4+/Foxp3+/CD25-,Freq. of Parent",
"Cells/SC/Live/CD8—,, CD4+/Foxp3-,Freq. of Parent", "Cells/SC/Live/CD8—,, CD4+/Foxp3-,Median,<BV421-A>,CD127",
"Cells/SC/Live/CD8—,, CD4+/Foxp3-/CD62L—,, CD44—,Freq. of Parent",
"Cells/SC/Live/CD8—,, CD4+/Foxp3-/CD62L—,, CD44+,Freq. of Parent",
"Cells/SC/Live/CD8—,, CD4+/Foxp3-/CD62L+,, CD44—,Freq. of Parent",
"Cells/SC/Live/CD8—,, CD4+/Foxp3-/CD62L+,, CD44+,Freq. of Parent",
"Cells/SC/Live/CD8—,, CD4+/Foxp3-/CD44+,Freq. of Parent",
"Cells/SC/Live/CD8+,, CD4—,Freq. of Parent", "Cells/SC/Live/CD8+,, CD4—,Median,<BV421-A>,CD127",
"Cells/SC/Live/CD8+,, CD4—/CD62L—,, CD44—,Freq. of Parent",
"Cells/SC/Live/CD8+,, CD4—/CD62L—,, CD44+,Freq. of Parent",
"Cells/SC/Live/CD8+,, CD4—/CD62L+,, CD44—,Freq. of Parent",
"Cells/SC/Live/CD8+,, CD4—/CD62L+,, CD44+,Freq. of Parent",
"Cells/SC/Live/CD8+,, CD4—/CD62L+,, CD44+,Freq. of Parent_1",
"Cells/SC/Live,Count", "Cells/SC/Live/CD8—,, CD4+,Count",
"Cells/SC/Live/CD8—,, CD4+/Foxp3+,Count", "Cells/SC/Live/CD8—,, CD4+/Foxp3+/CD25+,Count",
"Cells/SC/Live/CD8—,, CD4+/Foxp3+/CD25-,Count", "Cells/SC/Live/CD8—,, CD4+/Foxp3-,Count",
"Cells/SC/Live/CD8—,, CD4+/Foxp3-/CD62L—,, CD44—,Count",
"Cells/SC/Live/CD8—,, CD4+/Foxp3-/CD62L—,, CD44+,Count",
"Cells/SC/Live/CD8—,, CD4+/Foxp3-/CD62L+,, CD44—,Count",
"Cells/SC/Live/CD8—,, CD4+/Foxp3-/CD62L+,, CD44+,Count",
"Cells/SC/Live/CD8+,, CD4—,Count", "Cells/SC/Live/CD8+,, CD4—/CD62L—,, CD44—,Count",
"Cells/SC/Live/CD8+,, CD4—/CD62L—,, CD44+,Count", "Cells/SC/Live/CD8+,, CD4—/CD62L+,, CD44—,Count",
"Cells/SC/Live/CD8+,, CD4—/CD62L+,, CD44+,Count"), class = "factor"),
value = c(41.2, 35.5, 39.5, 33.2, 39.1, 35.5, 35.7, 33.9,
39.7, 42.4, 10.9, 12.1, 10.9, 12.5, 12.3, 12.8, 14.1, 15.8,
14.6, 12.5)), .Names = c("Sample", "Genotype", "Tissue", "variable", "value"), row.names = c(NA, -20L), class = "data.frame")
我正在使用以下函数绘制数据的各种组合
library(ggplot2)
library(ggpubr)
plot_it <- function(Tissue,
row_add = (1:nrow(temp)),
y.lab = "Did you forget to add a label?",
font_choice = "Helvetica",
font_size = 12,
stat_test = "t.test",
p_display = "p.signif",
legend_position = c("right")) {
# Subset data frame based on row_add
rownames(temp) <- NULL
df <- droplevels(temp[c(row_add),])
rownames(df) <- NULL
View(temp)
# Define color and shape of variables
color.groups <- c("black","red")
names(color.groups) <- unique(df$Genotype)
shape.groups <- c(16, 1)
names(shape.groups) <- unique(df$Genotype)
# Generate data frame of reference y-values for p-value labels and bracket positions
dmax = df %>% group_by(variable) %>%
summarise(value=max(value, na.rm=TRUE),
Genotype=NA)
# For tweaking position of brackets
e = max(dmax$value)*0.1
r = 0.6
w = 0.19
bcol = "black"
# Define y axis and wrap label
y.axis <- df$value
y.lab <- str_wrap(y.lab, width = 40)
ggplot(df, aes(x = variable, y = value, color = Genotype, shape = Genotype)) +
# geom_violin(position = position_dodge(width = 0.75)) +
geom_boxplot(position = position_dodge(width = 0.75), outlier.shape = NULL) +
geom_point(position=position_dodge(width=0.75), size = 2) +
ylim(0,1.2*max(y.axis, na.rm = TRUE)) + ylab(y.lab) + xlab(df$Tissue) +
scale_color_manual(values=color.groups) +
scale_shape_manual(values=shape.groups) +
scale_x_discrete(labels = function(x) str_wrap(x, width = 20)) +
theme_bw() + theme(panel.border = element_blank(), panel.grid.major = element_blank(),
panel.grid.minor = element_blank(), axis.line = element_line(colour = "black"),
aspect.ratio = 1, text = element_text(family=font_choice, size = font_size),
legend.position = legend_position) +
stat_compare_means(show.legend = FALSE, label = p_display, method = stat_test,
label.y = e + dmax$value, family = font_choice) +
geom_segment(data=dmax,
aes(x=as.numeric(variable)-w, xend=as.numeric(variable)+w,
y=value + r*e, yend=value + r*e), size=0.3, color=bcol, inherit.aes=FALSE)
}
通过调用以下函数来绘制：
plot_it(Tissue = "Thymus", row_add = c(c(1:30), c(141:150)))
这会生成此图：
我想让该函数创建一个 facet wrap，它可以有效地将 y 轴转换为两个部分，以便两个部分的比例允许更好的数据可视化。事实上，当我用具有高度不同值的变量绘制图表时，比例尺不适合所有变量。
如果这是不可能的，那么有没有一种简单的方法可以在每个绘图的基础上手动引入小平面环绕来分解 y 轴？
您可以使用聚类将具有相似手段的组放在一起。但是，由于比例不同，这些图可能会产生误导。
在下面的示例中，我使用了假数据，因为样本数据只有两组。
library(tidyverse)
# Fake data with five groups
set.seed(2)
dat = data.frame(group=rep(LETTERS[1:5], each=20),
sub=rep(rep(letters[1:2], each=10), 5),
value=rnorm(100, rep(c(20, 17, 27, 56, 80), each=20), 5))
# Add facet groups using kmeans clustering
dat = dat %>%
group_by(group) %>%
mutate(mean=mean(value)) %>%
ungroup %>%
mutate(facet_group = kmeans(mean, 2)$cluster)
ggplot(dat, aes(group, value, colour=sub)) +
geom_boxplot() +
facet_wrap(~ facet_group, scales="free", ncol=2) +
expand_limits(y=0) +
theme_classic() +
theme(strip.background=element_blank(),
strip.text=element_blank())
或更复杂的布局选项：
library(gridExtra)
# Create the two plots separately and store them in a list
plots = unique(dat$facet_group) %>%
map(
~ ggplot(dat[dat$facet_group==.x, ], aes(group, value, colour=sub)) +
geom_boxplot() +
expand_limits(y=0) +
labs(y="", x="") +
theme_bw() +
theme(strip.background=element_blank(),
strip.text=element_blank())
)
# Extract Legend
g_legend <- function(a.gplot) {
tmp <- ggplot_gtable(ggplot_build(a.gplot))
leg <- which(sapply(tmp$grobs, function(x) x$name) == "guide-box")
legend <- tmp$grobs[[leg]]
return(legend)
}
leg = g_legend(plots[[1]])
# Lay out the plots and the legend
grid.arrange(
arrangeGrob(grobs=map(plots, function(x) x + guides(colour=FALSE)), widths=c(3,2)),
leg, widths=c(10,1), left="Value", bottom="Group"
)
我有一个数据框：
df <- structure(list(Sample = c("1: FL_643", "2: FL_645", "3: FL_647","4: FL_656", "5: FL_658", "6: cKO_644", "7: cKO_646", "8: cKO_654","9: cKO_655", "10: cKO_657", "1: FL_643", "2: FL_645", "3: FL_647", "4: FL_656", "5: FL_658", "6: cKO_644", "7: cKO_646", "8: cKO_654", "9: cKO_655", "10: cKO_657"), Genotype = structure(c(1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L), .Label = c("miR-15/16 FL", "miR-15/16 cKO"), class = "factor"),
Tissue = structure(c(2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L,
2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L), .Label = c("Thymus",
"iLN", "Spleen", "Skin", "Colon"), class = "factor"), variable = structure(c(1L,
1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L,
2L, 2L, 2L, 2L), .Label = c("Cells/SC/Live/CD8—,, CD4+,Freq. of Parent",
"Cells/SC/Live/CD8—,, CD4+/Foxp3+,Freq. of Parent", "Cells/SC/Live/CD8—,, CD4+/Foxp3+,Median,<BV421-A>,CD127",
"Cells/SC/Live/CD8—,, CD4+/Foxp3+/CD25+,Freq. of Parent",
"Cells/SC/Live/CD8—,, CD4+/Foxp3+/CD25-,Freq. of Parent",
"Cells/SC/Live/CD8—,, CD4+/Foxp3-,Freq. of Parent", "Cells/SC/Live/CD8—,, CD4+/Foxp3-,Median,<BV421-A>,CD127",
"Cells/SC/Live/CD8—,, CD4+/Foxp3-/CD62L—,, CD44—,Freq. of Parent",
"Cells/SC/Live/CD8—,, CD4+/Foxp3-/CD62L—,, CD44+,Freq. of Parent",
"Cells/SC/Live/CD8—,, CD4+/Foxp3-/CD62L+,, CD44—,Freq. of Parent",
"Cells/SC/Live/CD8—,, CD4+/Foxp3-/CD62L+,, CD44+,Freq. of Parent",
"Cells/SC/Live/CD8—,, CD4+/Foxp3-/CD44+,Freq. of Parent",
"Cells/SC/Live/CD8+,, CD4—,Freq. of Parent", "Cells/SC/Live/CD8+,, CD4—,Median,<BV421-A>,CD127",
"Cells/SC/Live/CD8+,, CD4—/CD62L—,, CD44—,Freq. of Parent",
"Cells/SC/Live/CD8+,, CD4—/CD62L—,, CD44+,Freq. of Parent",
"Cells/SC/Live/CD8+,, CD4—/CD62L+,, CD44—,Freq. of Parent",
"Cells/SC/Live/CD8+,, CD4—/CD62L+,, CD44+,Freq. of Parent",
"Cells/SC/Live/CD8+,, CD4—/CD62L+,, CD44+,Freq. of Parent_1",
"Cells/SC/Live,Count", "Cells/SC/Live/CD8—,, CD4+,Count",
"Cells/SC/Live/CD8—,, CD4+/Foxp3+,Count", "Cells/SC/Live/CD8—,, CD4+/Foxp3+/CD25+,Count",
"Cells/SC/Live/CD8—,, CD4+/Foxp3+/CD25-,Count", "Cells/SC/Live/CD8—,, CD4+/Foxp3-,Count",
"Cells/SC/Live/CD8—,, CD4+/Foxp3-/CD62L—,, CD44—,Count",
"Cells/SC/Live/CD8—,, CD4+/Foxp3-/CD62L—,, CD44+,Count",
"Cells/SC/Live/CD8—,, CD4+/Foxp3-/CD62L+,, CD44—,Count",
"Cells/SC/Live/CD8—,, CD4+/Foxp3-/CD62L+,, CD44+,Count",
"Cells/SC/Live/CD8+,, CD4—,Count", "Cells/SC/Live/CD8+,, CD4—/CD62L—,, CD44—,Count",
"Cells/SC/Live/CD8+,, CD4—/CD62L—,, CD44+,Count", "Cells/SC/Live/CD8+,, CD4—/CD62L+,, CD44—,Count",
"Cells/SC/Live/CD8+,, CD4—/CD62L+,, CD44+,Count"), class = "factor"),
value = c(41.2, 35.5, 39.5, 33.2, 39.1, 35.5, 35.7, 33.9,
39.7, 42.4, 10.9, 12.1, 10.9, 12.5, 12.3, 12.8, 14.1, 15.8,
14.6, 12.5)), .Names = c("Sample", "Genotype", "Tissue", "variable", "value"), row.names = c(NA, -20L), class = "data.frame")
我正在使用以下函数绘制数据的各种组合
library(ggplot2)
library(ggpubr)
plot_it <- function(Tissue,
row_add = (1:nrow(temp)),
y.lab = "Did you forget to add a label?",
font_choice = "Helvetica",
font_size = 12,
stat_test = "t.test",
p_display = "p.signif",
legend_position = c("right")) {
# Subset data frame based on row_add
rownames(temp) <- NULL
df <- droplevels(temp[c(row_add),])
rownames(df) <- NULL
View(temp)
# Define color and shape of variables
color.groups <- c("black","red")
names(color.groups) <- unique(df$Genotype)
shape.groups <- c(16, 1)
names(shape.groups) <- unique(df$Genotype)
# Generate data frame of reference y-values for p-value labels and bracket positions
dmax = df %>% group_by(variable) %>%
summarise(value=max(value, na.rm=TRUE),
Genotype=NA)
# For tweaking position of brackets
e = max(dmax$value)*0.1
r = 0.6
w = 0.19
bcol = "black"
# Define y axis and wrap label
y.axis <- df$value
y.lab <- str_wrap(y.lab, width = 40)
ggplot(df, aes(x = variable, y = value, color = Genotype, shape = Genotype)) +
# geom_violin(position = position_dodge(width = 0.75)) +
geom_boxplot(position = position_dodge(width = 0.75), outlier.shape = NULL) +
geom_point(position=position_dodge(width=0.75), size = 2) +
ylim(0,1.2*max(y.axis, na.rm = TRUE)) + ylab(y.lab) + xlab(df$Tissue) +
scale_color_manual(values=color.groups) +
scale_shape_manual(values=shape.groups) +
scale_x_discrete(labels = function(x) str_wrap(x, width = 20)) +
theme_bw() + theme(panel.border = element_blank(), panel.grid.major = element_blank(),
panel.grid.minor = element_blank(), axis.line = element_line(colour = "black"),
aspect.ratio = 1, text = element_text(family=font_choice, size = font_size),
legend.position = legend_position) +
stat_compare_means(show.legend = FALSE, label = p_display, method = stat_test,
label.y = e + dmax$value, family = font_choice) +
geom_segment(data=dmax,
aes(x=as.numeric(variable)-w, xend=as.numeric(variable)+w,
y=value + r*e, yend=value + r*e), size=0.3, color=bcol, inherit.aes=FALSE)
}
通过调用以下函数来绘制：
plot_it(Tissue = "Thymus", row_add = c(c(1:30), c(141:150)))
这会生成此图：
我想让该函数创建一个 facet wrap，它可以有效地将 y 轴转换为两个部分，以便两个部分的比例允许更好的数据可视化。事实上，当我用具有高度不同值的变量绘制图表时，比例尺不适合所有变量。
如果这是不可能的，那么有没有一种简单的方法可以在每个绘图的基础上手动引入小平面环绕来分解 y 轴？
您可以使用聚类将具有相似手段的组放在一起。但是，由于比例不同，这些图可能会产生误导。
在下面的示例中，我使用了假数据，因为样本数据只有两组。
library(tidyverse)
# Fake data with five groups
set.seed(2)
dat = data.frame(group=rep(LETTERS[1:5], each=20),
sub=rep(rep(letters[1:2], each=10), 5),
value=rnorm(100, rep(c(20, 17, 27, 56, 80), each=20), 5))
# Add facet groups using kmeans clustering
dat = dat %>%
group_by(group) %>%
mutate(mean=mean(value)) %>%
ungroup %>%
mutate(facet_group = kmeans(mean, 2)$cluster)
ggplot(dat, aes(group, value, colour=sub)) +
geom_boxplot() +
facet_wrap(~ facet_group, scales="free", ncol=2) +
expand_limits(y=0) +
theme_classic() +
theme(strip.background=element_blank(),
strip.text=element_blank())
或更复杂的布局选项：
library(gridExtra)
# Create the two plots separately and store them in a list
plots = unique(dat$facet_group) %>%
map(
~ ggplot(dat[dat$facet_group==.x, ], aes(group, value, colour=sub)) +
geom_boxplot() +
expand_limits(y=0) +
labs(y="", x="") +
theme_bw() +
theme(strip.background=element_blank(),
strip.text=element_blank())
)
# Extract Legend
g_legend <- function(a.gplot) {
tmp <- ggplot_gtable(ggplot_build(a.gplot))
leg <- which(sapply(tmp$grobs, function(x) x$name) == "guide-box")
legend <- tmp$grobs[[leg]]
return(legend)
}
leg = g_legend(plots[[1]])
# Lay out the plots and the legend
grid.arrange(
arrangeGrob(grobs=map(plots, function(x) x + guides(colour=FALSE)), widths=c(3,2)),
leg, widths=c(10,1), left="Value", bottom="Group"
)