From what I can tell, the way to process Nimblegen data is use limma. The latter does not read Nimblegen ".pair" files, so you sort of glue together your own limma-friendly data structures and then let limma take over. Is that about it? Thanks, Tom