Balancing Gain and Relatedness in Selection

Abstract

Summary A population merit criterion, B ω , for a set of genotypes, ω, is formulated as where, c is a weighting constant, g w is the average of their breeding values, and Θω is the average coancestry of the considered genotypes, which is a measure of their relatedness. The breeding objective studied here is selecting the set ω that maximises B ω . An iterative search algorithm is proposed for finding this maximum under a given breeding-population size. This algorithm was applied to an example using simulation techniques. Results were presented as graphs where the gain was plotted against the status effective number, which was used to quantify the degree of relatedness as an inverse function of average coancestry. For all except extreme c values the algorithm gave markedly better combinations of gain and average coancestry when compared with a conventional method to control relatedness by restricting contributions from individual parents.