Would someone have an explanation for the following?
I've run multiple meta analysis and successfully did ORA at p < 0.05
However, whenever I try doing GSEA (using ReactomePA or clusterProfiler) it fails at p < 0.05 and only works at p < 1 (I'm using the Fisher's p val ranked genes).
If I run GSEA within the individual data sets (not meta analysed) it works fine at 0.05.
The above issue is the case for a total of 6 meta analysis which seriously makes me doubt whether that should be the case...
Any suggestions how I can run GSEA on meta analysed genes or justify the p < 1 threshold?
Much appreciated. Cheers!