Abstract #T100

# T100
Exploring methodology for application of genomic information in South African dairy breeds.
Bernice E. Mostert1, Robert R. Van der Westhuizen2, Este Van Marle-Köster*1, Brian Van Doormaal3, 1Department of Animal and Wildlife Sciences, University of Pretoria, Pretoria, South Africa, 2SA Stud Book and Animal Improvement Association, Bloemfontein, South Africa, 3Canadian Dairy Network, Guelph, ON, Canada.

The establishment of reference populations for genomic selection for Holstein and Jersey breeds in South Africa (SA) is underway. SA breeders have been using direct genomic values (DGVs) based on foreign reference populations, depending on base year definitions and units of measurement of foreign countries to make within herd selection decisions. Due to extensive linkage with international dairy populations, potential exists for utilizing DGVs originating from foreign reference populations to provide SA breeders with genomic information, comparable to SA breeding values. International collaboration was sought for foreign genomic profiles. Genomic profiles of SA Holstein and Jersey animals were generated based on either the GGPHD78K or GGPLD23K chips and DGV on the Canadian scale were estimated, using the North American Consortium’s reference population. These were converted to the SA scale using Interbull conversion equations and blended with the SA Parent Averages (PAs), based on performance of SA cows. This blending was done according to weightings determined by the reliabilities of the DGV and SA PAs. To account for genotype × environmental interactions between Canada and SA, the reliability of the DGV was multiplied by the correlation coefficient between SA and Canada for the specific trait, thereby ensuring that DGV from traits with lower correlations received less weight in the blending with local estimated breeding values (EBV). The weighting for the blending was determined as: EBV weight = EBV reliability/(EBV reliability + DGV reliability). These blended EBVs have been released as genomically enhanced breeding values (GEBVs) and serves as the official breeding values for these animals in the genetic evaluation, which are directly comparable to EBVs of the same breed in SA. Significant increases in the reliability with implementation of this methodology have been observed, with an average increase of 27% for milk yield and 23% for functional herd life. This approach has resulted in the release of GEBVs for production, udder health, fertility, longevity and conformation traits for SA Holstein and Jersey breeders.

Key Words: dairy, genomic information, reliability