Boar taint in entire male pigs: a genomewide association study for direct and indirect genetic effects on androstenone.

Abstract

Androstenone is one of the compounds causing boar taint of pork and is highly heritable (approximately 0.6). Recently, indirect genetic effects (IGE; also known as associative effects or social genetic effects) were found for androstenone, meaning that pen mates (boars) affect each other's androstenone level genetically. Similar to estimating variance components with a direct-indirect animal model, direct and indirect genetic SNP effects can be estimated for androstenone. This study aims to detect SNP with significant direct genetic effects and IGE on androstenone. The dataset consisted of 1,282 noncastrated boars (993 boars genotyped) from 184 groups of pen members. After quality control, 46,421 SNP were included in the analysis. One model for single-SNP regression was fitted, where both the direct SNP effect of the individual itself and the indirect SNP effects of its pen mates were included. None of the SNP (direct or indirect) were found genomewide significant. One QTL on SSC6 was chromosome-wide significant for the direct effect. A single SNP on SSC9 and 2 regions and a single SNP on SSC14 were found for the indirect effect. A backwards elimination method and haplotype analysis were used to quantify the variance explained by the SNP. The backwards elimination method identified 4 independent regions affecting androstenone. The QTL on SSC6 explained 2.1 and 2.6% of the phenotypic variance using the backwards elimination method or the haplotype analysis. The QTL on SSC14 explained 3.4 and 2.7% of the phenotypic variance using the backwards elimination method or the haplotype analysis. The single association on SSC9 explained 2.2% of the phenotypic variance. All significant QTL together explained 7 to 8% of phenotypic variance and 40 to 44% of the total genetic variance available for response to selection. Besides the newly discovered QTL and the confirmation of known QTL, this study also presents a methodology to model SNP for IGE.