Genomic evaluation of crossbred dairy cattle in the United States – an update


  • George Wiggans Animal Improvement Programs Laboratory, ARS-USDA
  • Paul VanRaden U.S. Department of Agriculture, Agricultural Research Service, Animal Genomics and Improvement Laboratory
  • Dan Null U.S. Department of Agriculture, Agricultural Research Service, Animal Genomics and Improvement Laboratory
  • Ezequiel Nicolazzi
  • Gerald Jansen
  • Jay Megonigal, Jr


Genomic evaluations of crossbred dairy cattle were first released in the United States in April 2019 by the Council on Dairy Cattle Breeding (Bowie, MD). Each crossbred animal has an evaluation calculated for each of 5 breeds: Ayrshire, Brown Swiss, Guernsey, Holstein, and Jersey. The evaluations are expressed on an all-breed base and combined by weighting the breed proportions — called breed base representation (BBR) — for each animal. The BBRs are estimated with the same process used for other traits. The breed reference populations are updated annually and are composed of AI progeny-tested bulls (≥10 daughters) with at least 4 generations of same breed ancestry. A blending procedure is applied to animals that have a highest BBR of <90%. The SNP effects are estimated from animals with a BBR of ≥94%, which are considered to be purebreds. Evaluations are reported on individual breed bases. In most cases, the breed with the highest BBR is used. However, to facilitate comparisons within herd, the breed associated with an animal’s identification record is used if the BBR is ≥40%. For weekly evaluations, an initial approximate BBR is calculated based on only the SNPs present on the genotyping chip so that imputation is not required. Currently, 5 groups of chips are defined, and SNP effects for breed determination are estimated for each group. To minimize the portion of animals that are incorrectly designated for blended evaluations based on the approximate BBR, animals with a BBR of ≥85% are processed with purebreds until a BBR based on an imputed genotype is calculated; then animals with a new BBR of <90% are moved to the blended group. A breed error is declared if the highest BBR is ≥85% for a breed other than that from the identification record. The blending process is not applied to type traits, which are not comparable across breeds, or to feed saved and calving traits, which are not calculated for all breeds. Recessive haplotypes are not reported for crossbreds because the largely within-breed nature of those conditions leads to less accuracy in detecting crossbred carriers. Crossbred evaluations have lower reliability than those for purebreds; however, by processing crossbreds, parentage discovery and BBRs are provided. In July 2021, 154,670 animals had blended evaluations. Of those animals, 48% were born in 2019 or later, an indication of recent increases in crossbreeding and genotyping of crossbreds.