Wrong cluster for downregulated genes

[Cristian Fiorucci]

Hi there,

I'm trying to running an analysis of my DEGs but when i enter it i notice that genes in the downregulated bar are, instead, upregulated. In fact the score is positive. I tried running Log2FC and FC but the result is still the same.
I attach the result html down there, awaing for your response

[John Elizarraras]

Hello,

I believe this is a result of the functionality of GSEA. It doesn't necessarily care about the sign of the score. The score of GSEA is used to sort the list from most positive to most negative. Essentially, negative enrichment scores from GSEA indicate that the genes in the pathway were more commonly found near the end of the list. For your case, those genes still have positive scores, but when you sort your input, they are found near the end of your list. To avoid this confusion, ideally your ranking metric has a median of 0, but this may not always be possible depending on your conditions. A popular ranking metric for GSEA is -log10(p)*sign(FC) which you can try to see if that produces better results.

Please let me know if you have any questions.

Best,

John