search

MannLabs / alphapeptstats

Python Package for the downstream analysis of mass-spectrometry-based proteomics data
https://alphapeptstats.readthedocs.io/en/latest/
Apache License 2.0
56 stars 12 forks source link

Type of FDR correction #257

Open pejota66 opened 11 months ago

pejota66 commented 11 months ago

What are the Type of FDR corrections availables at the Alphapeptstats? Benjamini Hochberg?

Thank you

steph-robinson commented 6 months ago

Hi, I had a similar question coming here. After going through some of the source code, I was able to see that the perform_ttest_analysis and perform_ttest functions found in alphapeptstats/alphastats/multicova/multicova.py will return the unadjusted and adjusted t and p-values according to the definition for s0 from Tusher et al 2001 PNAS. Later, the only options for FDR references 'BH' in some other calculations. However, how the qvalue and final significance is called in annotate_fdr_significance seems to be different than Benjamini Hochberg's method.

I'm curious if this is a permutation-based FDR correction like what's used in Perseus (Tyanova et al 2016), or is this another q-value method? I've tried passing an old dataset worked up in Perseus, but I get fairly different numbers of significant hits despite using the same s0 and kind of ttest. I do get the correct unadjusted p-values and correct adjusted t-values. I'm quite excited for this package, so thank you for making this code available!