Genetic analysis and genomic selection of stayability and productive life in New Zealand ewes1

Genetic parameters for ewes were estimated for stayability (STAY) and productive life (Prodlife). Records from 658,871 animals from 241 seed stock or ram breeder flocks and one research flock were used to estimate parameters for ram breeder flock STAY and Prodlife. These flocks would have a culling policy based on criteria typical of a commercial enterprise, but including culls based on knowledge such as estimated breeding values. A subset of 35,688 of these animals from 4 ram breeder and one research flock was used to infer the animals that might be culled based on the perspective of a commercial flock as these flocks recorded reasons for culling. These data were used to estimate parameters for commercial flock STAY and Prodlife. Heritability for ram breeder flock STAY until age 3 to 6 yr ranged from 0.07 to 0.09. The heritability of commercial flock STAY was similar in yr 5 and 6, but lower at 0.05 (age 3) and 0.07 (age 4). Heritability of Prodlife for ram breeder flock and commercial flock was respectively 0.10 and 0.13. The genetic correlations between ram breeder flock STAY and commercial flock STAY ranged from 0.20 to 0.99. Both STAY and Prodlife were most correlated genetically with the trait number of lambs born, at about 0.4, and to a lesser degree, live weight up to the age of 12 mo for the traits analyzed with lower genetic correlations estimated for the commercial flocks. Live weight at 18 mo and fleece weight had little or no correlation with STAY and Prodlife. From the data set used to estimate genetic parameters from a ram breeder flock, 4,330 animals were genotyped with the Illumina OvineSNP50 BeadChip and the genotypes used to develop genomic predictions via genomic best linear unbiased prediction. The animals used in the training data set were multibreed, but were heavily Romney breed based. The genomic prediction accuracy for ram breeder flock STAY, according to breed, was dependent on sample size and ranged from 0.4 to 0.44 for Romney and 0.26 to 0.4 for Coopworth. Accuracies for Perendale and Composites, for which sample sizes were less, were more variable. This study has derived genetic parameters for STAY and Prodlife and investigated genomic predictions for these traits. As these traits are of low heritability, sex limited and expressed late in life genomic selection will be useful.