Genotype by environment interaction for age at slaughter in Irish dairy and beef crossbreds using a genomic reaction norm model


  • Birgit Gredler-Grandl
  • Jeremie Vandenplas
  • Alan Twomey
  • Mario Calus


In beef cattle production, the reduction of the number of days from birth until the target weight at slaughter is reached, represents a sustainable option to increase efficiency and reduce the environmental impact. In the presence of genotype by environment interaction (GxE), selection of resilient animals is important. We estimated GxE for age at slaughter in an Irish dairy and beef crossbred population using an analysis protocol modelling either homogeneous or heterogeneous SNP-(co)variances across the genome based on readily available BLUP software packages. We allowed for heterogeneous SNP (co)variances by using different weights across SNPs. In our approach, we divide the data set of interest in two subsets and follow a 2-step approach: (1) derive SNP (co)variances from SNP effects estimated in the first data set, and (2) weight the SNP genotypes using estimated SNP (co)variances from (1) to re-estimate SNP effects in the second data set. The data set consisted of 14,193 genotyped crossbred heifers, steers and bulls in 2,041 herds. Phenotypes used in the genomic analysis were yield deviations for age at slaughter. We estimated contemporary group (CG) effects for age at slaughter in a univariate BLUP analysis to be subsequently used as continuous environmental descriptor in the genomic reaction norm models. Results show large genetic variations for age at slaughter. The average heritability estimated across all CG was 0.24. The genetic parameters for age at slaughter estimated along the environmental gradient support the existence of GxE in extreme environments. Nevertheless, the genetic correlation between the majority of environments was greater than 0.89. Higher accuracy for young selection candidates were achieved when using genomic information instead of using pedigree information only. However, modeling homo- or heterogeneous SNP (co)variances across the genome resulted in similar accuracy of genomic breeding values for age at slaughter.