Reputation: 167
I have a working solution but am. I've been using the mtcars data set and trying to color by the "disp" variable as my reproducible example.
> library(gplots)
> m<-cbind(mtcars[,3],mtcars[,3])
> rownames(m)<-rownames(mtcars)
> heatmap.2(x=m,dendrogram="none",trace="none",Colv=FALSE,Rowv=FALSE,cellnote=cbind(rownames(m),rownames(m)),notecol="black")
I can always cut out the extra row of the pdf, replace with a representation of the p-value in my actual data set (plotting ratios and p-value is fisher's exact difference from whole population), however that would add a large amount of processing on the full table of 500 values (right now just printing a giant length pdf). I may color the p-value column separately and paste them together for the final figure but assuming this will look sloppy. Suggestions at any step are appreciated.
Upvotes: 1
Views: 967
Reputation: 1127
A solution with ggplot2:
d=as.data.frame.table(m[,1,drop=FALSE])
ggplot(d,aes(x=Var2,y=Var1,fill=Freq)) + geom_tile()
Upvotes: 1