ggboxplot - How can I inrease distance between adjacent boxplots? - r

Question

I am using the ggpubr package to do boxplots with ggboxplot. Any suggestions on how to increase the distance between the adjacent boxplots?
I have been using R for a couple of weeks and I am aware that my script might be written better.
My code:
flowdata <- read.csv("flowdata.csv", header = TRUE, sep = ";")
flowdata$Haplotype = factor(flowdata$Haplotype,levels(flowdata$Haplotype)
[c(5,1,2,3,4,6)])
library(ggpubr)
p<-ggboxplot(flowdata, x="TP", y="Treg", add = "jitter",width = 0.5, shape
= "Treatment", fill = "Haplotype", palette = c("#0092d1","#62b232","#b23a32","#b232a3","#99cccc","#132a64"))+scale_shape_manual(values = c(21,23))
p1<-p+theme(legend.title = element_blank(), legend.text = element_text(size=8), text = element_text(family = "Calibri"), axis.text.x = element_text(angle = 45, hjust = 1))+ labs(x = expression(paste("")),
y = expression(paste(CD4^+{}, CD25^+{}, "cells/µL")))
p1

Answer 1

The parameter 'width' specifies the width of the boxes, so a simple solution would be to reduce that value (from 0.5). This would not increase spacing of the boxes, but increase spacing between them and therefore make the boxes narrower.
However, it seems to me like your boxplots are well spaced, but your points (jitter) are overlapping, making the graph look messy. A simpler solution would be to remove them, or change them to points instead of jitter. Alternatively you could use a violin plot.
For finer control, 'standard' ggplot2 can be used, perhaps with cowplot which can give you formatting:
p <- ggplot(data = flowdata, mapping = aes(x = TP, y = Treg, fill = Haplotype)) +
geom_boxplot(position = position_dodge(0.5)) +
geom_jitter(aes(shape = Treatment)) +
scale_shape_manual(c(21, 23)) +
scale_fill_manual(c("#0092d1","#62b232","#b23a32","#b232a3","#99cccc","#132a64")) +
theme(legend.title = element_blank(), legend.text = element_text(size=8), text = element_text(family = "Calibri"), axis.text.x = element_text(angle = 45, hjust = 1)) +
labs(x = expression(paste("")), y = expression(paste(CD4^+{}, CD25^+{}, "cells/µL")))

Answer 2

Building on Adams answer, I would furthermore suggest you get the boxplots into several facets, so that the plot is actually readable. I suppose the haplotype might be the interesting facet. Also, you could reduce the size of the jitter points or get some transparency alpha so that they are less present, but with the facet the boxplots are larger already, so this might solve the problem of readability by itself.
p <- ggplot(flowdata, aes(x = TP, y = Treg)) +
geom_boxplot(position = position_dodge(0.5)) +
geom_jitter(aes(shape = Treatment), size = 0.5, alpha = 0.8) +
facet_wrap(~Haplotype, ncol = 3) +
scale_shape_manual(c(21, 23)) +
theme(legend.title = element_blank(),
legend.text = element_text(size=8),
text = element_text(family = "Calibri"),
axis.text.x = element_text(angle = 45, hjust = 1)) +
labs(x = "",
y = expression(paste(CD4^+{}, CD25^+{}, "cells/µL")))