I have two questions:
First: I was just wondering if there is another way to calculate FDR (or other pvalue correction methods) besides using statsmodels.stats.multitest.multipletests?
Especially I am looking for FDR as stand alone (for example NOT frd_bh) which - as far as I have seen - is not given here.
Second: In which cases are the corrected and adjusted p-values always the same among all methods when using stats.multitest? It happened in my case (352 tests, alpha=0.05) where the adjusted p-value is always like the bonferroni correction definition (alpha/number of tests) 0.05/352 - I am wondering what I did wrong.
Thank you for all answers!!
An answer to your first question is
scipy.stats.false_discovery_control.