Adsa Logo White Adsa Title White

Genomic predictions for crossbred animals with single-step genomic BLUP.

D. Gonzalez Pena



Join D. Gonzalez Pena on this page for a live text chat!
6:00 PM - 8:00 PM GMT


Genomic predictions for crossbred animals with single-step genomic BLUP.
D. Gonzalez Pena*1, N. Vukasinovic1, J. Brooker1, C. Przybyla1, Y. Steyn2, S. DeNise1. 1Zoetis Portage, MI, 2University of Georgia Athens, GA.

Crossbreeding in US dairy herds is increasing, motivating research on breed-specific parameters. This study investigates the use of ssGBLUP for generating genomic estimated breeding values (gEBV) of crossbreed dairy animals for milk (MY), fat (FY), and protein (PY) yield, daughter pregnancy rate (DPR), somatic cell score (SCS), and productive life (PL) under different scenarios. Genotypes from 375K Holsteins (H), 60K Jerseys (J), and 21K HxJ crosses (C) were available. The C genotypes were combined in 6 scenarios: S1) only C; S2) C + H and J connected through pedigree; S3) C + sires connected through pedigree and phenotypes; S4) C + sires connected through pedigree and phenotypes + all J; S5) C + all H; S6) C + all H + all J. All scenarios used the same phenotypes and a pedigree depth of 20 generations. Number of records was 13,636,724 for MY, FY, PY, and SCC; 9,531,611 for DPR, and 4,100,343 for PL. The performance of each scenario was validated using 1,987 C animals with phenotypes and genotypes, with phenotypes being removed from the analysis. The predictivity was defined as the correlation (r) between gEBV and phenotypes adjusted for the fixed effects in the model. The inflation of gEBV was measured by the regression coefficient (b) of the adjusted phenotype on the gEBV. For MY, S2 had the highest (r = 0.42) and S6 had the lowest predictivity (r = 0.30). For all other traits, S4 performed best with r = 0.31, 0.27, 0.14, and 0.22 for FY, PY, DPR, and SCS, respectively. Both S4 and S6 had r = 0.12 for PL. The worst scenarios for FY, PY, DPR, PL, and SCS were S6 (r = 0.21), S6 (r = 0.15), S5 (r = 0.09), S3 (r = 0.02), and both S1 and S6 (r = 0.17), respectively. Regarding gEBV inflation, S2 was the best for MY (b = 0.89), whereas S4 was the best for all other traits (b = 0.87, 0.73, 1.04, 1.0, 0.78 for FY, PY, DPR, SCS, and PL, respectively). Limited predictivity of only C genotypes and the benefits of including related purebred animals are visible. The number of purebred genotypes in the analysis had an impact on the crossbred gEBV under ssGBLUP, which could be associated with the proportion of chromosomal segments shared among crossbred and purebred animals.

Keywords: crossbred, genomic.