Genetic evaluation with finite locus models

Abstract

The availability of genotypic data in recent years has resulted in increased interest in the use of marker assisted genetic evaluation (MAGE) in livestock species. Under additive inheritance, Henderson's mixed model equations (HMME) provide an efficient approach to obtain genetic evaluations by marker assisted best linear unbiased prediction (MABLUP) given pedigree relationships, trait, and marker data. For large pedigrees with many missing markers, however, it is not feasible to calculate the exact gametic variance covariance matrix required to construct HMME, and thus, approximations are used. By computer simulation we observed that the use of exact matrices would increase response to selection by 2.2% up to 11.7%. Marker assisted selection (MAS) is efficient especially for traits that have low heritability and non-additive gene action. BLUP methodology under non-additive gene action is not feasible for large inbred or crossbred pedigrees. It is easy to incorporate non-additive gene action in a finite locus model. Under such a model, the unobservable genotypic values can be predicted using the conditional mean of the genotypic values given the available data, which is also known as the best predictor (BP). The potential of alternative methods to compute BP under finite locus models was studied, and it was shown that Markov chain Monte Carlo (MCMC) methods that sample blocks of genotypes jointly hold most promise for such computations. The efficiency of MCMC methods for genetic evaluation by BP under finite locus models, depends on the number of loci considered in the model. Thus, the effect of the number of loci used in the finite locus model used for genetic evaluation by BP was studied by computer simulation. In our study, models with two to six loci yielded accurate BP evaluations for traits determined by 100 loci. Finally, we proposed a strategy to improve the computational efficiency of MAGE under finite locus models.