Hello,

I'd like to do an enrichment using romer/roast/camera and my input are HGNC official human gene symbols.

For the symbols2indices() function I need a list of vectors for the reference gene sets, either from the rdata files provided by you or generated from the gmt-files from Broad institute.

As discussed previously in this post:

Romer and symbols2indices query

you had to curate the gene symbols in the gmt-files heavily to make them consistent, which is why you nicely provide these downloads. Does this all apply to the entrez-id gmts only or also to the gene-symbol gmts downloadable from Broad?

I wonder what is the better option, to use the gene symbol gmts or to map my ids to entrez first and use the rdata files. To my experience enrichments can be pretty vulnerable to how good the mappings work.

Do you have any experience with the gene-symbol version of the gmt-files? Or what would you recommend to do starting from HGNC-symbols (or ensembl gene ids)?

Best,

Maik

Either way is fine.

Mapping official human symbols to Entrez Gene IDs is easy and pretty much one-to-one. That's what I would do.

Or working from the Broad Institute gene symbol gmts is also fine. It was different 6 years ago, but these days the symbol gmts from MSigDb contain official human symbols in a consistent format.