I have hundreds of arrays for which I would like to obtain NUSE and RLE plots. I would like to make one call to affyPLM, but then plot a subset (boxplot), perhaps a few random samples that cover the entire set. Otherwise, there will be too many samples to show effectively in one boxplot. What is a way to go about this?  Example below taken from affyPLM vignette.

require(affydata)
library("affyPLM")
data(Dilution)
Dilution = updateObject(Dilution)
Pset <- fitPLM(Dilution)
boxplot(Pset)

Unfortunately the PLMset class never had a [ function, so it's not as simple as

boxplot(Pset[,1:3])

In addition, if you want NUSE or RLE values, you should use either NUSE or RLE rather than boxplot, because you can return the values.

Let's say you have 200 samples, and you want to do a NUSE plot on a random subset of 25:

nuse.dat <- NUSE(Pset, type = "values")

boxplot(nuse.dat[,sample(1:200, 25)])

You can do the same for RLE.