One thing people might not realize is that the pdInfo packages for the Gene ST arrays also contain a parsed version of Affy's transcript and probeset csv file, which can be used as well. This is sort of a pain because it's not that easy to parse out the relevant information, and since the order of the probesets in the csv is different from the ExpressionSet of summarized data, you have to remember to reorder (which to my dismay I forgot in a recent analysis... Tsk tsk).
Anyway, since I am apparently incapable of remembering such things, I added some functionality to my affycoretools package to automate this, and ensure that I don't bork it up like usual. This is in the devel version, so you need a devel installation of R/BioC. But it's a one-liner. Say you do something like this:
dat <- read.celfiles(filenames = list.celfiles())
eset <- rma(dat)
So now you have an ExpressionSet called 'eset', which has 'pd.porgene.1.1.st' in its annotation slot. You can now annotate the ExpressionSet like this:
eset.annot <- annotateEset(eset, annotation(eset))
Which will then parse out the information from the Affy csv file that comes with the pd.porgene.1.1.st package. The nice thing about this is that the limma package will use this annotation data by default, so if you use limma to make comparisons, you don't have to do anything else to include the annotation data.
If you follow Guido's suggestion and use the MBNI packages, you can install the corresponding ChipDb package from MBNI rather than using the OrgDb package directly, in which case you can then do (assuming now that your ExpressionSet is based on the MBNI package instead):
eset.annot <- annotateEset(eset, porgene11stssentrezg.db, columns = c("ENTREZID","SYMBOL","GENENAME","ALIAS"))
And you will get the same exact annotation that Guido got, with less work on your part, a double-check that your data and annotations are in correct 1:1 correspondence, and as above these annotations will propagate down through limma.