Ideogram with cluster marks using ggplot2

This is some notes on using the facet_grid function in ggplot2.

My data looks like this:

chr	window	function id	name	genes	start	end	chrlen	source	label
chr1	206	164a8da3b80	Dynein	7	77424754	81043029	88881357	gfam
chr1	206	PF01221	Dynein_light	14	77424754	81056868	88881357	pfam	yes
chr2	5	PF05019	Coq4	3	2611802	3155461	48130368	pfam
chr2	5	f40b1ae1da7e4	CMTR1	8	2687527	4607079	48130368	gfam	yes

Basically it’s a collection of enriched functional clusters (gene family and pfam) on each chromosome. I would like to plot those gene numbers as bar (at least 4 genes) and label those marked “yes” with the function names. From ggplot2 facet_grid() function seems suitable in categorising the data into each chromosome. And I used geom_bar and geom_label to plot and label clusters. For the chromosome I used geom_rect to draw a rectangle (not ideal though).

Here is the script with notes to myself:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21


library(ggplot2)
library(ggrepel)
# read the data
clstall<-read.delim("combined_clusters.txt", sep=" ", header = F, col.names=c("chr", "block", "func", "name", "genes", "fdr", "start", "end", "chrlen", "source", "label"))
# subset the data
clst<-clstall[which(clstall$genes>=4),]

pp <- ggplot(data=clst, aes(x=start/1000000, y=genes, fill=source)) # bar fill color based on source
	+ geom_bar(stat="identity", position="dodge", width=0.3)  # overlapped bars will not stack
	+ geom_text_repel(data=clst[which(clst$label=="yes"),], mapping=aes(color=source, label=paste0(name, " (", genes, ")")), size=3.3, direction="both", max.iter=6000) # label from "yes"-marked, label color based on source
	+ geom_rect(mapping=aes(xmin=0, xmax=chrlen/1000000, ymin=0, ymax=25), fill="white", color="black", alpha=0, size=0.3) # the outer framebox as chromosome; size is the thickness
	+ facet_grid(rows=vars(chr)) # facet data based on chromosomes


pp <- pp + scale_x_continuous(expand = c(0, 0)) + scale_y_continuous(expand = c(0, 0)) # x, y axis starting from 0
	+ labs(x="Coordinate (Mb)", y="Genes") 
	+ ylim(0, 25) + scale_y_continuous(breaks=c(0, 10, 20)) # y limit and ticks/labels 
	+ scale_fill_manual(labels=c("Gene family", "Pfam"),values=c("blue", "red")) # re-define bar fill color 
	+ scale_color_manual(labels=c("Gene family", "Pfam"),values=c("blue", "red")) # re-define bar label color
	+ theme(legend.position="bottom") + theme(legend.text = element_text(size=9.5, face="bold"))  # legend position and text style
	+ theme_classic() # classic without grid lines

There is still one problem with the plot: some labels are overlapped with bars. Not sure which function to avoid this?