The effect of heterogeneity for yearling weights measured in different test phases of the South African National Beef Cattle Improvement Scheme

Abstract

In animal breeding, animals originating from different environments are often considered in selection decisions. Production data in these groups may differ both in mean and variance. In South Africa, yearling weights of beef cattle are recorded in two different phases of the National Beef Cattle Improvement Scheme (NBCIS). The aim of this study was to assess changes in the ranking of animals in the selection process where animals are selected on estimated breeding values (EBVs) for yearling weight before and after correcting for heterogeneity between the two phases. Yearling weights are measured in Phases B and D of the NBCIS. Yearling weights recorded in Phase D were adjusted according to the standard deviation and heritabilities of yearling weights in Phase B. To determine whether the adjustment was justified, two VCE REML analyses were performed on combined Phase B and D data to estimate breeding values. The first analysis involved unadjusted Phase D data, while the second run involved Phase D yearling weights, adjusted for heterogeneity. The ranking correlation between EBVs derived from the two runs, was 99% indicating almost no re-ranked after adjustment. It is concluded that Afrikaner yearling weights in Phase B and D can be analysed jointly without any pre-adjustments when estimating breeding values.