|WIGGANS, GEORGE - Council On Dairy Cattle Breeding|
|NICOLAZZI, EZEQUIEL - Council On Dairy Cattle Breeding|
|TOOKER, MELVIN - Retired ARS Employee|
|MEGONIGAL, JR, JOEL - Council On Dairy Cattle Breeding|
|WALTON, LEIGH - Council On Dairy Cattle Breeding|
Submitted to: Interbull Annual Meeting Proceedings
Publication Type: Proceedings
Publication Acceptance Date: 11/13/2019
Publication Date: 11/13/2019
Citation: Wiggans, G.R., Van Raden, P.M., Nicolazzi, E., Tooker, M.E., Megonigal, Jr, J.H., Walton, L.M. 2019. Extending genomic evaluation to crossbred dairy cattle – US implementation. Interbull Bulletin. 55:46-49.
Technical Abstract: In April 2019, the Council on Dairy Cattle Breeding (CDCB) extended its genomic evaluation system to crossbred dairy cattle by weighting estimates for effects of individual-breed single nucleotide polymorphisms (SNPs) by breed proportions. Previously, animals detected as crossbreds were excluded because SNP effect estimates differ by breed and imputation relies on breed-specific haplotype frequencies if parental genotypes are missing. Breed proportions for all animals were estimated using a combined reference population of purebred bulls from each of the five evaluated breeds. Their phenotypic values were set to zero or 100 to estimate SNP effects for breed base representation (BBR). Genotypes for animals that appeared to be crossbred were imputed using the same multi-breed haplotype library as for estimation of BBR. Non-crossbred genotyped parents were included to improve accuracy. Because combination of SNP effect estimates across breeds requires that genetic effects be on the same multi-breed base, CDCB modified the genomic evaluation procedure in April 2018 to use traditional evaluations on multi-breed base before adjusting to individual breed bases. Animals with a breed percentage of at least 90 were not included in the blending, and a minimum breed percentage of 94 was imposed for an animal to contribute to SNP effect estimation. This limitation ensured that SNP effect estimates were not affected by animals of other breeds but reduced the size of the reference population. For animals with a breed percentage below 90, weighting evaluations by BBR was not possible for type traits because they are not on a common base or for calving and health traits because they are not evaluated for all breeds. In those cases, the single-breed evaluation for the breed with the highest percentage is reported. For two-breed crosses, some allowance was made to use the breed reported in an animal’s identification as the evaluation breed. For a March 2019 test evaluation, 68,691 animals received evaluations that were weighted by BBR, and 31,521 of those animals had not previously received an official evaluation. Implementation of the weighting process extends genomic evaluation to more animals and provides more complete information for herds that genotype all females. Even for animals that received evaluations in the past, the evaluations weighted by BBR more accurately reflect an animal’s multi-breed origin.