Establishing a base population for a breeding program in aquaculture, from multiple subpopulations, differentiated by genetic drift: II. Sensitivity to assumptions on the additive genetic relationships of base animals

Abstract: The simulation study examined the effect in optimal contribution selection of specifying an additive genetic relationship matrix (A-matrix) deviating from the true A-matrix of the base population. A base population of 160 males and 160 females was sampled evenly from 4 or 8 randomly chosen, drift-differentiated, wild subpopulations. Wild fish had true inbreeding coefficient of F wild, either 0.05 or 0.20, giving additive genetic relationships of 2F wild within, and 0 between subpopulations. Selection based upon phenotypes was simulated on one continuous trait of heritability, h wild 2, either 0.1 or 0.3. Base animals were mated either within subpopulations, or at random, so that the first round of selection occurred either on purebred or crossbred fish respectively. Optimum contributions was implemented from generation 1 for a desired rate of inbreeding of 0.5% per generation, using an assumed F wild ⁎ which varied between 0.0 and 0.3. Assuming F wild ⁎ =0.0 compared to 0.3, resulted in steeper regressions of the fractional contributions of subpopulations on average breeding value, higher true inbreeding and consequently reduced additive genetic variance, albeit with larger genetic gain from the first round of selection. When the first round of selection followed random mating the regression of the fractional contributions of the subpopulations was also less steep. These effects persisted, but originated from the misspecification of F wild ⁎ and so will not accumulate further in the medium-to long-term. Consequently, assuming F wild ⁎ was less than the true value led to sub-optimal use of subpopulations with an increased risk of loss of alleles of direct and strategic relevance to the breeding program. In conclusion, assuming F wild ⁎ ≈0.0 should be avoided and F wild ⁎ should be estimated with procedures such as marker estimated kinships, Q ST estimates or empirical observations of clines. [Copyright &y& Elsevier]