Genetic studies of assortative mating - a simulation study .3. Assortative mating in selected populations

Effects of 25 generations of positive assortative mating, random mating and negative assortative mating in simulated selected populations of various effective size (N-e = 40, 100, 200) were compared. The trait under consideration was controlled by either 100 or 2500 loci. Positive assortative mating produced the highest cumulative selection response (11.61-13.24 sigma(P)), followed by random mating (11.00-12.48 sigma(P)) and negative assortative mating (10.88-11.98 sigma(P)). The differences between the various mating systems were highly significant (P < 0.001) after the second generation and depended on the covariance due to linkage disequilibrium (C-l) and varying rates of fixation of the favourable alleles, leading to different genic variance (V-a). Positive assortative mating first caused C-l to increase to maximum values equal to 10.8% of the base population genetic variance (V-A) when the trait was controlled by 100 loci and then to decline gradually. A faster rate of change in gene frequency caused V-a to decrease initially, but to return towards random mating levels later (after generations 13-17 when the trait was controlled by 100 loci), when fixation of the favourable alleles approached completion. Inbreeding coefficient and degree of homozygosity increased slightly under positive assortative mating. Negative assortative mating had similar, but smaller, effects on C-l and V-a in the opposite direction of positive assortative mating.