Genome-Wide Association Study and Genomic Prediction for Maximum Litter Size in Sows

Abstract

Total born number (TBN) and number born alive (NBA) are primary selection targets for improving reproduction performance in pigs; however, genetic progress is constrained by the low heritability of single-parity records. To better capture the genetic potential underlying litter size, we evaluated maximum (maxTBN and maxNBA) and mean (meanTBN and meanNBA) litter size traits derived from multiple parities as alternative selection criteria. Using data from a Dongliao Black (DL) pig population, heritability estimates for maxTBN (h² = 0.19), meanTBN (h² = 0.16), maxNBA (h² = 0.24) and meanNBA (h² = 0.17) were substantially higher than those for single-parity TBN (h² = 0.06) and NBA (h² = 0.07). Genome-wide association analyses (GWAS) based on estimated breeding values of maximum and mean traits identified 158 significant single-nucleotide polymorphisms, implicating several candidate genes, including BPI, KCNC2, ELMO1, CHD6, OPCML, and MND1. Validation in an independent Large White population with high litter size (TBN > 16) showed that favorable alleles at representative loci were close to fixation, supporting their effects on reproduction performance. Furthermore, genomic prediction models based on maxTBN and maxNBA exhibited higher cross-validation variance than those using conventional litter size traits, reflecting greater prediction instability. Collectively, these results demonstrate that mean litter size traits constitute robust phenotypes for genetic evaluation of sow fertility, while maximum traits may serve as supplementary indicators for identifying elite individuals, and provide functionally relevant markers that can be exploited to inform balanced selection strategies for reproduction in pig breeding programs.