23
Accuracies of genomic and pedigree based predictions for swine litter size traits in Large White and Landrace breeds

Tuesday, March 17, 2015: 2:30 PM
302-303 (Community Choice Credit Union Convention Center)
Austin M Putz , North Carolina State University, Raleigh, NC
Kent A. Gray , Smithfield Premium Genetics, Rose Hill, NC
Christian Maltecca , North Carolina State University, Raleigh, NC
Mark Knauer , North Carolina State University, Raleigh, NC
Francesco Tiezzi , North Carolina State University, Raleigh, NC
Abstract Text:

Breeding value accuracy of young animals is a key determinant of achievable genetic gain. The objective of this study was to compare the accuracy of prediction for young animals using GEBV from single-step GBLUP versus traditional pedigree EBV for litter size traits. Four years of sow performance data were provided by Smithfield Premium Genetics (Rose Hill, NC) on 4810 Large White and 4418 Landrace sows. Traits included: total number born (TNB), number born alive (NBA), litter size at day 5 (LS5), litter size at weaning (LSW), and number weaned (NW). Litter size at day 5 and LSW were calculated by subtracting preweaning mortalities with age less than 5 d or total deaths from TNB, respectively, according to the biological litter. Number weaned was recorded as the number of piglets that were removed from the sow at the time of weaning. Genotyping was completed with the Illumina Porcine 60k BeadChip. After editing, the total number of effective SNP was 32,719 and 31,350 for Large White and Landrace, respectively. The number of animals genotyped for Large White and Landrace was 3264 and 1976, respectively. Accuracy of prediction was defined as the correlation between adjusted phenotype with GEBV and EBV for dams with genotypes, after accounting for the effect of year-season, farm, and parity. Full datasets included 16 year-seasons over the four years. Validation datasets included the last three year-seasons of the full dataset; 29 and 27% for Large White and Landrace, respectively. Heritability estimates for TNB, NBA, LS5, LSW, and NW were 0.05, 0.06, 0.06, 0.06, and 0.05 for Large White and 0.07, 0.07, 0.08, 0.09, and 0.07 for Landrace, respectively. GEBV (EBV) accuracies for Large White were 0.082 (0.073), 0.104 (0.085), 0.101 (0.096), 0.099 (0.107), and 0.058 (0.088) for TNB, NBA, LS5, LSW, and NW, respectively. GEBV (EBV) accuracies for Landrace were 0.101 (0.144), 0.118 (0.149), 0.210 (0.185), 0.269 (0.236), and 0.223 (0.192) for TNB, NBA, LS5, LSW, and NW, respectively. Accuracies were greater for Landrace than Large White and inconsistent between breeds for each trait. Results showed that the GEBV had a greater correlation with adjusted phenotype than EBV in 6 out of the 10 comparisons. Further research will need to be completed to understand accuracy of prediction in litter size traits for young animals. 

Keywords: accuracy, litter size, piglet mortality