UTILIZATION OF DOMINANCE VARIANCE THROUGH MATE ALLOCATION STRATEGIES

Simulation was used to evaluate the increase in progeny performance from three mating strategies, based on predicted specific combining abilities among sires and maternal grandsires over random mating that do not use specific combining ability but avoid inbreeding. Specific combining abilities were equal to the sum of combination effects from dominance plus the effect of inbreeding depression. Mates were allocated by linear programming with two approximations. Genetic parameters were h2 in the narrow sense equal to .05, .15, or .25 and the ratio of dominance to phenotypic variance equal to .05, .10, or .15. A population of 400 bulls were grouped by .99, .85, and .70 PTA reliability; the first group was sires and maternal grandsires of others. Recurrence equations for combination effects that were due to dominance, not including inbreeding depression, were used to create a matrix of true combination effects among bulls. Predicted specific combining abilities were computed from true combination effects using low, intermediate, and high levels of accuracy plus the effect of inbreeding. Twenty herds of cows were generated for each bull population. Within a herd, four mating groups of 123 cows were mated to 10 bulls from all bull groups. The three mating strategies yielded progeny merits slightly but significantly higher than random mating. Scaled by standard deviation of milk yield, increases with linear programming were 12.3 to 40.1 kg for low, 16.4 to 46.4 kg for intermediate, and 31.0 to 80.3 for high accuracy.