I have CEL files from GeneChip miRNA v4.0. I would like to generate normalized data using a Loess curve subset with only spike-in probesets. Package affy has a function normalize.loess(mat, subset=...) which can accomplish the subseting portion, however after importing the CEL data using the oligo package I end up with ~300,000 annotations corresponding to PM sequences. How do I determine which PM sequencing in the annotation are associated with the spike-ins? I've checked the CSV annotation provided online, and they are all N-masked.