Hello there, I have a question concerning usage of GOseq for testing custom gene list (approx. 1k genes) enrichment. I am not sure about the use_genes_without_cat argument value. It is referred by https://support.bioconductor.org/p/67003/ and the manual that when this option is set to FALSE as default, p-values are calculated based on only those genes assigned to any GO term. In my case, i have only one category, so it is expected that there are genes not assigned to any of "my" categories. Is their usage (therefore setting use_genes_without_cat=TRUE) in this case justifiable?

Regards, John K.

Yes, you have to set that argument TRUE. goseq will test whether genes in your set are more enriched than genes not in your set. If you drop genes not in your set then there is nothing to test.

Although you can use goseq on a single gene list, the package was not designed with that purpose in mind. I would suggest fry() in the edgeR package for testing a singe gene set or category. See the section called Fry gene set tests in the edgeR QL workflow.