Genotyping-By-Sequencing for Genomic Evaluation in Pigs

Genotyping-by-sequencing (GBS) technology has a capacity for discovering and genotyping a high density of SNP markers with potentially lower cost than SNP chips. In this study, 2,836 Duroc boars were genotyped by a modified GBS pipeline we previously reported, and on average 3.5 million high quality reads were produced for each individual. With the requirement of minor allele frequency greater than 0.01, missing rate less than 40%, and Hardy-Weinberg equilibrium test with p>10^-6, 103,173 SNPs were used for estimating genomic heritabilities and breeding values (GBVs). Eight traits including average daily gain (ADG), back fat thickness (BF), lion muscle area (LMA), lion muscle depth (LMD), intra-muscular fat content (IMF), feed conversion ratio (FCR), residual feed intake (RFI), average daily feed intake (ADFI) were analyzed. Genomic heritability estimates by GREML were 0.23±0.03 for ADG, 0.41±0.04 for BF, 0.51±0.03 for LMA, 0.50±0.04 for LMD, 0.19±0.03 for IMF, 0.38±0.05 for FCR, 0.38±0.05 for RFI and 0.43±0.05 for ADFI, which were close to the heritability estimates using pedigree relationships. The accuracies of GBV (measured as correlations between true breeding values and GBVs) calculated by GBLUP and SSBLUP were greater than the accuracies of EBVs predicted by traditional BLUP for genotyped animals. The average of accuracy for the eight traits was 0.46 (ranging 0.40-0.51) using traditional BLUP, increased to 0.56 (0.52-0.63) using GBLUP, and further increased to 0.66 (0.63- 0.70) using SSBLUP.