Dear Matthieu,
I'm trying to compare BAYESCAN outlier results with those of BAYPASS. The latest calculates and integrates a covariance matrix which is useful when assessing populations with a "complicated" hierarchical structure as the ones I'm working with. However, unlike BAYESCAN, BAYPASS does not correct for multiple testing. I tried to correct the p-values using the R function q-value, but that only applies to p-value from one side test. Therefore, I was wondering if you could share how does BAYESCAN estimate the q-values for p-values from two side test (i.e loci with FST lowest than expected will potentially be under balancing selection, while loci with FST highest than expected will potentially be under positive selection). I think that applying the same BAYESCAN multiple test correction to BAYPASS will provide a fair comparison.
Thanks in advance,
Dear Matthieu,
I'm trying to compare BAYESCAN outlier results with those of BAYPASS. The latest calculates and integrates a covariance matrix which is useful when assessing populations with a "complicated" hierarchical structure as the ones I'm working with. However, unlike BAYESCAN, BAYPASS does not correct for multiple testing. I tried to correct the p-values using the R function q-value, but that only applies to p-value from one side test. Therefore, I was wondering if you could share how does BAYESCAN estimate the q-values for p-values from two side test (i.e loci with FST lowest than expected will potentially be under balancing selection, while loci with FST highest than expected will potentially be under positive selection). I think that applying the same BAYESCAN multiple test correction to BAYPASS will provide a fair comparison.
Thanks in advance,