Hi,
I would be grateful for your opinion.
I performed a DGE analysis and received a list of differentially expressed genes. Next, we focus on specific list of pre-defined genes for further analysis, since their biological functions are relevant for our scientific question (and all other differentially expressed genes will not be further included).
Since the list of genes-of-interest was selected by previous research I thought to run a functional enrichment analysis on those pre-defined genes to reassure that they are truly represent the biological functions we assume they are.
That is different from functional enrichment analyses I performed in the past, which were on all the differentially expressed genes.
Do you think our approach is reasonable?
Thanks!
To quote a really excellent paper on the topic:
While I suppose it may depend on the specific tool you're using, I would generally think you want to use the entire set of genes with non-zero expression in your RNA-seq data set for your enrichment analysis, not just a pre-determined subset or even the set of differentially expressed genes.
Thank you, The quote refers to the background gene list while I am asking about restricting the gene list of interest (which I query against the background gene list). Furthermore, in ORA analysis (over-representation analysis) only the differentially expressed genes are being used, as the quoted paper also describes, but again I wonder if I can limit this list of differentially expressed genes and using only pert of them.
your background gene list for ORA should be all the genes that underwent differential expression testing.
Yes, fully agree, but my dilemma is not about the background gene list but about the genes of interest that I test against the background.