Hi,
I understand GSEA is typically run with an expression dataset. If I have just a list of genes, I can run compute overlaps with the MSigDB database online:
http://software.broadinstitute.org/gsea/msigdb/annotate.jsp
However, I am wondering if I can compute overlaps in MSigDB datasets locally through a command line for a list of genes. I want to do this because I have tons of such datasets (only genes) which I need to query for overlaps with MSigDB. Has anyone done this?
Thanks.
1 answer
What you need is an implementation of Fisher's exact test and some sort of matching algorithm.
I've never used it in a separate tool, but you probably can apply some of the code used to calculate GO enrichments, or just write it yourself - should be pretty easy.
Log in to answer this question.