Adding xbreaks to ggplot - r

Question

Hello I have generated the following:
Using the following code:
library(ggplot2,quietly=TRUE)
args <- commandArgs(TRUE)
data <-read.table(args[1],header=T,sep="\t")
pdf(args[2])
p1 <- ggplot( data,aes(x=Mutation,y=difference) ) +
geom_bar( width=.75,aes(fill=as.factor(position),stat="identity"),position="dodge") +
geom_errorbar(aes(y=difference, ymin=difference-difference_std, ymax=difference+difference_std )) +
theme(legend.key = element_blank()) +
ylab(expression(Delta*"Energy (Design - Wild Type)" )) +
xlab( "Mutation" ) +
ylim(-2.5,2.5) +
theme_bw() +
theme(axis.text.x = element_text(angle = 90,size=9,hjust=1))+
opts(legend.position="none")+
ggtitle(expression("PG9 Mutational Analysis"))
p1
dev.off()
As you can see I've sort of grouped the positions using the fill layer so you can group them together. Ideally though I would like to put a break in the xbreak when the position changes instead of grouping them by the color.
So it would go like orange bar, orange bar, break, green bar, break, cyan bar, break, etc...
*EDIT:
Here is what the data from an input table looks like:
position Mutation difference difference_std
97 R97L -0.3174848488298787 0.2867477631591484
97 R97N 0.5333956571765566 0.35232170408577224
99 A99H -0.2294999853769939 0.24017957128601522
99 A99S -0.45171049187057877 0.0013816966425459903
101 G101S 0.5315110947026147 0.08483810927415139
102 P102S -0.04872141488960846 0.02890820273131048
103 D103Y 0.6692000007629395 0.07312815307204293
....
So all the positions would be grouped together with a break on either side.
I hope I'm making sense. Is there an easy way to do this?
J

Answer 1

I think the easiest way would be to use faceting (used sample data from question). To get new variable for the faceting, first, order your data frame according to Mutation. Then add new column pos2 where using cumsum() and diff() according to column position sequence of numbers is added (idea of #agstudy).
df2<-df[order(df$Mutation),]
df2$pos2<-cumsum(c(0,diff(df2$position)) != 0)
df2
position Mutation difference difference_std pos2
3 99 A99H -0.22949999 0.240179571 0
4 99 A99S -0.45171049 0.001381697 0
7 103 D103Y 0.66920000 0.073128153 1
5 101 G101S 0.53151109 0.084838109 2
6 102 P102S -0.04872141 0.028908203 3
1 97 R97L -0.31748485 0.286747763 4
2 97 R97N 0.53339566 0.352321704 4
Now use new column pos2 for facetting. With theme() and strip.text= and strip.background= you can remove strip texts and grey background.
ggplot(df2,aes(x=Mutation,y=difference) ) +
geom_bar( width=.75,aes(fill=as.factor(position),
stat="identity"),position="dodge") +
geom_errorbar(aes(y=difference, ymin=difference-difference_std,
ymax=difference+difference_std )) +
theme(legend.key = element_blank()) +
ylim(-2.5,2.5) +
theme_bw() +
theme(axis.text.x = element_text(angle = 90,size=9,hjust=1))+
theme(legend.position="none")+
theme(strip.text=element_blank(),
strip.background=element_blank())+
facet_grid(.~pos2,scales="free_x",space="free_x")
UPDATE - empty levels
Other possibility is to use scale_x_discrete() and argument limits= and add empty levels in places where you need space between bars (actual levels). Problem with this approach is that you need to add those levels manually.
For example used the same data as in previous example (ordered question data).
ggplot(df2,aes(x=Mutation,y=difference) ) +
geom_bar( width=.75,aes(fill=as.factor(position),
stat="identity"),position="dodge") +
geom_errorbar(aes(y=difference, ymin=difference-difference_std,
ymax=difference+difference_std )) +
theme(legend.key = element_blank()) +
ylim(-2.5,2.5) +
theme_bw() +
theme(axis.text.x = element_text(angle = 90,size=9,hjust=1))+
theme(legend.position="none")+
scale_x_discrete(limits=c("A99H", "A99S","", "D103Y","","G101S",
"","P102S","","R97L", "R97N"))

Answer 2

I think you need first to create a new variable that group the labels.
dat$group <- cumsum(c(0,diff(dat$position)) != 0)
Mutation difference position group
20 0wYpO 0.93746859 4 0
17 63L00 -0.57833783 4 0
3 6hfEp -1.01620448 3 1
1 FvAz4 0.09496127 2 2
8 ghNTj -1.10180158 3 3
14 GxYzD 0.41997924 3 3
Then I don't think it will be an easy way to add a gap to your plot. But You can create a barplot by group and add the genes names with geom_text. here a first version, I hope that someone more experienced with ggplot2 can help to adjust the texts in the middle
ggplot( dat,aes(x=factor(group),y=difference) ) +
geom_bar( width=.75,aes(fill=as.factor(Mutation)),stat="identity",
position="dodge") +
geom_text(aes(label=Mutation),position=position_dodge(height=0.9),angle=90)+
theme_bw() +
opts(legend.position="none")
PS: The code below can be used to generate data:
N <- 21
Mutation <- replicate(N,paste(sample(c(0:9, letters, LETTERS),
5, replace=TRUE),
collapse=""))
difference <- rnorm(N)
position <- c(4, 4, 3, 2, 3, 3, 2, 3, 2,
5, 2, 2, 2, 5, 5,
5, 2, 4, 5, 4, 2)
difference_std <- sd(difference)
dat <- data.frame(Mutation,difference)
dat <- dat[order(dat$Mutation),]
dat$position <- position
dat$group <- cumsum(c(0,diff(dat$position)) != 0)