Dear all,

I generated pooled RADSeq data for two populations (20 individuals per pool) of a South African bee species with the aim to identify loci under selection using an FST-outlier approach. I first used Stacks to call loci and then used the consensus sequences of each stack as reference contig against which I mapped all my RADSeq reads with BWA to finally call SNPs with Popoolations. Now I used Lositan to identify genes under selection which resulted in app. 300 outlier from my overall 19,000 marker. However, when I run the same data in Bayescan I do not find any outlier at all (neither with a prior of 100 nor 10). What could cause this large discrepancy between the two programs? Does Bayescan have problems in identifying outlier between only two populations? Are there any other programs that are suitable to detect selection when you have pooled RADSeq data? Below is the plot for my Bayescan analyses and I would very much appreciate any comments and help with that issue. Thank you very.

Best wishes,

Belinda