Actually there is a very simple solution without need to search for specific functions. But it did not come to my mind in the past several days, when I tried to use apply and its variants and to set up function or a loop inside the rows and columns. The idea is to set the maxmium value in each row to 1 and the rest values scaled.

Now comes the easiest way.

1. To make a single plot

Assume that you have a matrix called ‘mx’ and you want to make a barplot for a specific row “gene_id_1”

1
2
3

idx<-"gene_id_1"
barplot(mx[idx, ]/max(mx[idx,]), col = colo3, space = 0, 
        main = idx, ylab="Relative expression")

So the keypoint is mx[idx, ]/max(mx[idx,], which divides each row element by the largest one in this row.

2. To plot all rows and make each row a single barplot

A for loop just did the right thing.

1
2
3
4
5
6

i<-1;
for (i in 1:length(mx)) {
  pdf(paste(i,".pdf", sep=""));
  barplot(mx[i, ]/max(mx[i, ]), col = colo3, space = 0, ylab="Relative expression");
  dev.off()
}

3. To define rows from a file (or input) and make batch plotting

Add this to your script:

1
2

idx<-readLines(file("stdin"),1)
GI<-scan(file=idx, what=character());

And use mx[GI[i],]/max(mx[GI[i],]).