Entering edit mode
7.9 years ago
I'd like to correct for a biased gene enrichment analysis test.
Consider the following usual gene expression experiment: Consider a database D with N pathways having each {K1,....,Kn} genes. I did a gene expression experiment where the organism considered has L < N pathways having each {M1,...,Ml} < {K1,....,Kn} genes, and the relative proportion (representation) between pathways in L is different from N. e.g. Bateria B has 50 genes of a given pathway and 2 genes in the L-1 others. Database D has 100 genes in all pathways.
Doesn't that call for a correction? Isn't that most organisms analysed in gene expression assays have a different representation of pathways compared to the queried DB? If so how can I correct for it?
You shouldn't need any correction because the DB has to be species specific (otherwise you wouldn't have gene to pathway mappings). Your gene enrichment testing is with respect to L and {M1, .., Ml} not N and {K1, ... , Kn}.