How to turn axis labels into shapes? - r

Question

I'm trying to add boxes to replace the Y axis labels, such as in this figure.
I'm using ggplot but any other R package would work too.
input <- data.frame(
Pre.Subtype=c("Basal sq.", "Basal Sq.", "Basal sq.", "Neuronal", "Luminal", "Luminal", "Luminal"),
Post.Subtype=c("Basal sq.", "No tumor", "Luminal inf.", "Basal sq.", "Luminal", "Luminal pap.", "No tumor"),
Number.of.Patients=c(2,3,4,5,2,4,4), Percentage=c(34,23,23,53,20,34,56))
ggplot(data=input, aes(x=Pre.Subtype, y=Percentage, fill=Post.Subtype)) +
geom_bar(position = 'dodge', stat='identity') +
theme_classic() +
theme( axis.text.y = element_text(face="bold", size=12, colour="black"), axis.title=element_text(size=16,face="bold")) +
theme(legend.text=element_text(size=12), legend.title = element_text(size=14, face="bold")) +
labs(x= "Pre", y="Percentage", title="Post") +
geom_text(aes(label=Number.of.Patients), position=position_dodge(width=0.9), vjust=0.5, hjust=-0.3) +
scale_fill_manual(values=c("#fed976", "#7fcdbb", "#4d004b", "grey"), name="") +
theme(text = element_text(size=20), axis.text.x = element_text(size=13, face="bold", colour="black")) +
theme(plot.title = element_text(hjust = 0.5, face="bold", size=16), axis.ticks.y = element_blank()) +
coord_flip()

Answer 1

You can set clip = "off" in coord_flip as well as ensuring the y limit doesn't drop below zero:
... + coord_flip(ylim = c(0,50), clip = "off") + ...
Now you can draw a geom_rect outside of the plot area:
... + geom_rect(aes(xmin = 0.55, xmax = 0.85, ymin = -10, ymax = -7),
fill = "blue", size = 2, colour = "darkblue")
Which looks something like this:
Or use a geom_label:
geom_label(aes(label = "2", x = 0.55, y = -7), fill = "blue", size = 5, colour = "white")
Like this:
Of course, you'll need to add multiple geoms to get this effect for each category.

Answer 2

For a more automated way to do this, I suggest generating two separate plots, and combining them using ggarrange as shown below.
I have also fixed an issue where different bars were showing up in different widths in your original code.
library(tidyverse)
library(ggpubr)
input <- data.frame(
Pre.Subtype=c("Basal sq.", "Basal Sq.", "Basal sq.", "Neuronal", "Luminal", "Luminal", "Luminal"),
Post.Subtype=c("Basal sq.", "No tumor", "Luminal inf.", "Basal sq.", "Luminal", "Luminal pap.", "No tumor"),
Number.of.Patients=c(2,3,4,5,2,4,4), Percentage=c(34,23,23,53,20,34,56))
p1 <- ggplot(data=input, aes(x=Pre.Subtype, y=Percentage, fill=Post.Subtype)) +
#geom_bar(position = 'dodge', stat='identity') +
geom_col(position = position_dodge2(width = 0.9, preserve = "single")) +
theme_classic() +
theme( axis.text.y = element_blank(), axis.title.y = element_blank(), axis.title.x=element_text(size=16,face="bold")) +
theme(legend.text=element_text(size=12), legend.title = element_text(size=14, face="bold")) +
labs(x= "Pre", y="Percentage", title="Post") +
geom_text(aes(label=Number.of.Patients), position=position_dodge(width=0.9), vjust=0.5, hjust=-0.3) +
scale_fill_manual(values=c("#fed976", "#7fcdbb", "#4d004b", "grey", "red"), name="") +
theme(text = element_text(size=20), axis.text.x = element_text(size=13, face="bold", colour="black")) +
theme(plot.title = element_text(hjust = 0.5, face="bold", size=16), axis.ticks.y = element_blank()) +
scale_y_continuous(expand = c(0,0)) +
coord_flip()
p2 <- ggplot(data=input, aes(x=Pre.Subtype, y=Percentage)) +
theme_void() +
scale_y_continuous(expand = c(0,0)) +
geom_label(data = (input %>% group_by(Pre.Subtype) %>%
summarise_at(c('Number.of.Patients'),sum)), aes(label=Number.of.Patients, y = -0.1), fill = 'blue', color = 'white') +
coord_flip() + labs(title = 'Pre') + theme(plot.title = element_text(hjust = 0.5)) + theme(plot.margin = unit(c(0,-1,0,0), "cm"))
ggarrange(p2,p1, widths = c(0.1,0.9), align = "h")
This will give you the picture below.