YuLab-SMU / ReactomePA

Reactome Pathway Analysis
https://yulab-smu.top/biomedical-knowledge-mining-book/
37 stars 10 forks source link

P-value correction for multiple testing necessary? #40

Open L3ft2di3 opened 1 year ago

L3ft2di3 commented 1 year ago

Hey,

first: thank you for the package along with the data comprehensive book/vignette, it's great!

I was digging a bit deeper in the original Reactome analysis provided by the web-tool and there p-values are corrected by benjamini-hochberg and can be reported and subsequently filtered. Is this correction already implemented in the reactomePA::enrichPathway() function and hence pAdjustMethod = "BH" would be 'double correction'? I am asking, because setting pAdjustMehthod = "none" give very comparable results to the webtool result, whereas leaving it to "BH" seems to be much more stringent.

What would you suggest?

Kind regards Thorben