Ancestry-assortative mating in admixed human populations

Non-random mating is an important source of genetic structure in natural populations. In a variety of species, positive correlations have been observed in trait values between mates, a phenonenon termed "assortative mating". In admixed human populations, which derive recent ancestry from multiple source populations, a number of studies have observed positive correlations in ancestry proportion between spouses. However, what this means about how humans choose partners remains unclear.

Population genetic models of ancestry-assortative mating typically consider ancestry proportion as a continuous variable and assume that individuals are directly biased in selecting mates by similarity in ancestry proportion. In my ongoing work, we consider an alternative sociologically-motivated model in which individuals instead choose their partners based on shared group identity. Using forward-in-time simulations written in the SLiM framework, we compare the impacts of different strategies for defining mate choice on observed ancestry tracts.

This work was recently published in Genetics.