I am using GOseq for go term enrichment over 2 strains and 2 conditions. I noticed that the results have something peculiar - or something I am misunderstanding. The results returned from GOseq contains a numInCat column, which I take as number of genes for the given GO category. I expected this to be the same over different comparisons I performed over strains and conditions since gene2cat is the same. However, it differs from comparison to comparison. Could anyone help me understand?