I'm analyzing enrichment scores of MSigDB gene list in the subsets of my single cell rnaseq data. I'm using S2N metric implemented in GSEA (I just wrote a function to perform this for me and feed it into fgsea). I got everything to work but I'm trying to ensure what I see is actually real. Please see the image below as a representative output for top enrichment scores for a subset of my between two conditions. As you can see, most of the significant top hits have the same exact `pval` and `padj` values. Is there a chance this can be just an artifact altogether and p-values may not be trustworthy? Have you experienced something like this before?