Hi all,

I have a quirk in relatedness estimates I'm trying to figure out. Any help would be much appreciated! See the TL;DR at the bottom.

I have genotype data from 11 microsatellites for African buffalo from different populations. I estimated relatedness for each population (individually) using the TrioML estimator in Coancestry. I noticed just by looking at the values there seemed to be quite a few at exactly 0.5, which I thought was odd. So I plotted the estimates in a density curve (like a histogram, but with a line) and saw most populations had a blip at exactly 0.5. See attached PDF "Relatedness_by_pop.pdf".

For one population (AENP), I also did this estimation while accounting for inbreeding (this is an option in Coancestry), because this population had a known bottleneck and has low genetic diversity (see attached PDF entitled "AENP_Relatedness.pdf"). The blip at 0.5 went away and the overall distribution was more even. However, in his 2011 paper (Wang J. (2011). Unbiased relatedness estimation in structured populations. Genetics, 187, 887-901.), Jinliang Wang (who developed the TrioML method and Coancestry) states the following: "The results imply that, in practical situations where