I use GOseq quite often for RNAseq analyses, including the length bias correction. My question is how to use it properly for data without length bias, e.g., for microarrays.
Normally, with RNAseq data, I would use something like:
pwf <- nullp(gene.vector, "hg19", "geneSymbol", bias.data = lengthhg19) GO.wall <- goseq(pwf, "hg19", "geneSymbol", use_genes_without_cat = TRUE)
But what to change for microarray data? Would it be enough to add
method = "Hypergeometric" to the
goseq function? Or do I also have to feed the
nullp function with dummy values (let's say 100 for each gene)?
Thanks for your advice!