Bonferroni correction actually makes results more significant

[josiahseaman]

When I used goenrich.enrich.analyze(O, query_genes, 'Background', method='bonferroni') I was surprised to find that the q values were actually lower than the p values. Is it possible the math in your bonferroni correction is backwards?

if method == 'bonferroni':
        q[mask] = p / len(p)

where it should probably be q[mask] = min(1, p * len(p))

[jdrudolph]

Well spotted! You are definitely correct. I'll try to find some time to apply the fix, or you can go ahead and make a pull request. Thank you for reporting this.

[jdrudolph]

Fixed by 68ee3f7. A fixed version 1.10.2 was already released and goenrich can be updated with pip.
Thanks again @josiahseaman