so a bit of a mouthful of a title, but best to be specific about this. I am trying NMF subtyping of tumor RNAseq data using the NMF package. All was fine until I tried to create metagenes (subset of genes indicative of particular subtypes). The NMF method (nmf package in R) uses ExpressionSet data-type, which is microarray-centric. I am using RNAseq data counted based on Ensembl IDs but believe NMF wants Entrez IDs, based on output of their example. So it appears the only way to annotate an ExpressionSet object is using the AnnotationDBi package. I have looked at making my own annotation, but surely someone out there has either Ensembl to Entrez annotation strategy for ExpressionSet, or has done exactly this. Or at least knows a little work-around to the ExpressionSet annotation issue(?)
I have read all the effing manuals, there are lots, and most I cannot get any use from, so I have reverted to pleading with the internet gods. Hope someone has some ideas for me!