Marker-assisted selection can reduce true as well as pedigree-estimated inbreeding

Abstract

This study investigated whether selection using genotype information reduced the rate and level of true inbreeding, that is, identity by descent, at a selectively neutral locus as well as a locus under selection compared with traditional BLUP selection. In addition, the founder representation at these loci and the within-family selection at the nonneutral locus were studied. The study was carried out using stochastic simulation of a population resembling the breeding nucleus of a dairy cattle population for 25 yr. Each year, 10 proven bulls were selected across herds along with 100 dams from within each of 40 herds. Selection was performed using BLUP, marker-assisted, or gene-assisted selection for a trait with low heritability (h2=0.04) only expressed in females, mimicking a health trait. The simulated genome consisted of 2 chromosomes. One biallelic quantitative trait loci (QTL) with an initial frequency of the favorable allele of 0.1, and initially explaining 25% of the genetic variance as well as 4 markers were simulated in linkage disequilibrium, all positioned at chromosome 1. Chromosome 2 was selectively neutral, and consisted of a single neutral locus. The results showed that in addition to reducing pedigree-estimated inbreeding, the incorporation of genotype information in the selection criteria also reduced the level and rate of true inbreeding. In general, true inbreeding in the QTL was greater than pedigree-estimated inbreeding with respect to both the level and rate of inbreeding, as expected. Also as expected, true and pedigree-estimated inbreeding in the neutral locus were the same. Furthermore, after 25 yr, or approximately 5 generations, the pedigree-estimated level of inbreeding was reduced by 11 and 24% compared with BLUP in gene- and marker-assisted selection, respectively, and the level of true inbreeding in the QTL was reduced by 22 and 13%, respectively. The difference between selection scenarios was found to be caused by a larger number of founders being represented at the QTL when using genotype information in the selection criteria. This in turn was caused by an increased selection of individuals sharing the favorable QTL allele rather than individuals sharing genes on average, which was shown by a higher Mendelian selection differential in the QTL. Hence, even though the selection pressure was increased at the QTL, more variation was retained. The results suggest that marker-assisted selection is a useful selection strategy.