GSEA

[Thomas]

Hi folks.

I tested GSEA and re-run the same rnk file three times and three times i get different results, which might be interpreted differently. How can i get a more comparable or reproducible approach when running GSEA?

Best.

[Anirban Sinha]

@Thomas... thanks for pointing this out. I experienced the same issue. Running GSEA with the same set of ranked genes/proteins with same other-parameters settings returns different results. Any thoughts on this by the authors could be helpful.

[John Elizarraras]

Sorry for the delayed response.

GSEA creates random permutations to create a sample population. Depending on your lists and other factors, the effect of this randomness can vary. For example, a small gene sets and rank lists will typically have different results each run. You can experiment with increasing the minimum number of genes in a category in the advanced parameter section, which should decrease the randomness effect, but this will result in some gene sets being skipped.

Let me know if you have any more questions.

Best,

John