GSEA Phenotype Table Input

Question

[DVA]

Hello All,

I am trying use GSEA for gene enrichment analysis, and I was wondering about the phenotype labels. Is there a way to test for the difference between two phenotypes while controlling for other phenotypes in a different category?

For example, if I wanted to compare 5 cancer samples to 5 non-cancer samples, but 4 of the cancer samples are male while 4 of the non-cancer samples are female, is there a way to feed this sex information to GSEA so that it is controlled for (similar to what DEseq2 can do?). Sorry if this is a clueless question, I am very new to the program.

Thank you.

Answer 1

[c.chakraborty]

Hey, I am not sure, if I am answering you properly, but when you do diferential gene expression analysis using DEseq2, you will get hits which are definitely biassed for the sex factor. In such a condition, I would try to look for raw data from Array Express or GeoOmnibus,for male wild-type and female wild-type RNA-Seq, run a DEG analysis, using DEseq2 (keeping all other factors same). This analysis will give use the DEGs between male and female wt. Then you can omit the overlapping genes from your data, and this should probably give you a list of genes unbiased for sex. I am not so sure about this idea, and if someone else has a better solution, I am so ready to know and learn, because this is an interesting problem. Best of luck.