Trying to plot significance value for multiple boxplots

[Rohit-Satyam]

. The file has been attached here
brain.zip

I was trying to add significance bar for each EP pair in every tissue type using the following code
I used the following code:

library(ggplot2)
library(reshape2)
library(ggsignif)
df<- read.table(brain.csv,header = FALSE, col.names = c("Tissuetype", "E", "P"), sep = "\t")
df.m <- melt(brain)
ggplot(data = df.m, aes(x=value, y=Tissuetype)) + geom_boxplot(aes(fill=variable))+geom_signif(comparisons = list(c("E", "P")),map_signif_level = TRUE)

However it throws the following warnings:

Warning messages:
1: In f(..., self = self) : NAs introduced by coercion
2: Computation failed in `stat_signif()`:
missing value where TRUE/FALSE needed 

Can someone help me understand what's wrong with defining the function?

[const-ae]

Hey Rohit,

unfortunately it is difficult to calculate the exact positions of boxplots with position = "dodge". Please see the second example here.

The function fails because geom_signif() searches for "E" and "P" on the x-axis, but cannot find them.

You could set aes(x = paste0(Tissuetype, "_", variable), y = value) and then making a list with the correct comparison: eg. geom_signif(comparisons = list(c("insua_E", "insula_P")).

Btw. ggsignif does not support rotated boxplots, so you will need to put the continuous variable on the y axis and the discrete ones on the x-axis.

I hope this is helpful. I will close this issue, but feel free to comment or re-open if I missed something.

Best, Constantin