Post date: Aug 29, 2013 11:15:13 PM
I made two mistakes earlier with the simulated data. The first population number needs to be 0, but I started numbering with 1. Also, I did not correctly convert the genotypes to phred-scaled genotype likelihoods (I forgot that phred 0 is the most likely genotype). I fixed these mistakes and I am re-running bgsr on the simulated data. This time I have runs with the beta sd = 0.01 or 0.05 (last time it was just 0.01). The last set of runs took about 5.5 hours; these should be similar.