Pathway enrichment with z-scores

Question

[glady]

Hello,

Can I perform a pathway enrichment study on z-scores by fgsea/GSVA package?

Say, if I have 4 samples from the same condition, but collected from different locations. And I want to check for the pathways enriched in each of these 4 samples.

Converting read_counts into z-scores, and then performing a pathway enrichment study on these z-scores, will be right?

Or we need log2FC values only for performing an enrichment study?

Answer 1

[jared.andrews07]

I guess you could do this, but I'd consider the statistics behind it meaningless given that you're violating a whole bunch of assumptions of GSEA analysis, e.g. replicates, a proper ranking statistic, etc. I'd just use ssGSEA instead, that at least gives you relative scores within a sample that are kind of comparable.