Selection for a QTL trait in a realistic multiple-trait scenario in a simulated animal population

Responses to different modes of single-trait selection on a trait with identified quantitative trait locus (QTL trait) and multiple-trait selection on a QTL trait and genetically correlated binary traits were studied. The aim was to identify the selection mode which most efficiently lowered the prevalence of a QTL trait or of a QTL trait and correlated binary traits under conditions encountered in animal breeding. Simulated data resembling a multiple-trait scenario including one QTL trait in a realistic setting with respect to the number of individuals with available phenotype, genotype or phenotype and genotype information were used to study selection responses after one generation. Selection responses were defined as prevalence changes of the binary traits in the offspring of selected sires relative to the prevalence in the offspring of all sires. Simulations were performed for two different levels of heritability of the QTL trait (0.2, 0.4), two different proportions of genetic variance of this trait explained by the QTL (0.5, 0.1) and two different frequencies of the favorable QTL allele (0.5, 0.2). Genetic parameters for the simulated populations were estimated in mixed linear-threshold animal models using Gibbs sampling. Single-trait selection reduced the prevalence of the QTL trait relatively by up to 10% when based on phenotype, by up to 41% when based on genotype or genotype and phenotype, and by up to 81% when based on polygenic BV and genotype of the QTL trait. At the same time, the prevalences of the correlated binary traits were almost unchanged. Multiple-trait selection based on polygenic BV, polygenic BV and QTL genotype or overall BV simultaneously reduced the prevalences of all binary traits by 20–55%. Relative prevalence decrease of the QTL trait under multiple-trait selection was largest when selecting only sires with homozygous favorable QTL genotype and above-average total index derived from polygenic BV. Low proportions of individuals homozygous for the favorable QTL allele may limit practicability of most efficient single- and multiple-trait selection schemes, so selection of individuals with above-average BV which are homo- or heterozygous for the favorable QTL allele is recommended to quickly achieve relevant selection response, while keeping sufficient numbers of breeding animals. Advantages of combined use of phenotype and genotype information were obvious even under practical conditions of small numbers of individuals with trait information, low heritability of the QTL trait, small proportion of genetic variance explained by the QTL, and low frequency of the favorable QTL allele.