Hi all,
I have a question that may be obvious.
With the package fgsea, that uses a pre-ranked gene list, are the permutations carried out are gene set and not phenotype permutations: is that correct?
Thank you!
The gene list itself is permuted; so the comparison is the input ranking versus random rankings.
A phenotype permutation would be repeating (permute case/control labels --> differential expression --> top table --> fGSEA); and in such a case you could disable the internal fGSEA permutations entirely.