Estimation of genetic parameters and genetic trends for ewe longevity indicators in U.S. Katahdin sheep

Direct genetic selection for increased ewe longevity can improve flock profitability and animal welfare. However, longevity indicator traits are not presently evaluated by the National Sheep Improvement Program (NSIP). The primary objective of this study was, therefore, to estimate genetic parameters for 8 longevity indicator traits using data collected in NSIP Katahdin flocks. Ewes (n = 12,734) were born between 1989 and 2020 in 58 flocks across the U.S. and were daughters of 1,245 sires and 6,325 dams. Traits evaluated were age at the last lambing (ALL), length of productive life (PL; number of days between the first and last lambing), total number of litters (TNL), total number of lambs born (TNB) and weaned (TNW) over ewe lifetime, total lamb birth weight (TLB) and total lamb weight at weaning (TLW) over ewe lifetime, and TLW divided by the ewe's 120 d adjusted body weight (TLWadj). Variance components were estimated using the AIREML method based on fitting an animal model using either a pedigree (A) or blended pedigree and genomic (H) relationship matrix. Genomic information of 10,032 animals genotyped with a 50K SNP chip was included in the analyses based on H matrix. Age at first lambing and birth-rearing type of the ewe were fitted as fixed effects, while the contemporary group (CG: flock-year-season of ewe's birth) was fitted as either a fixed (contemporary group fitted as a fixed effect) or random (contemporary group fitted as a random effect) effect. Breeding values and their accuracies were obtained for 127,535 animals in the pedigree using either best linear unbiased prediction (BLUP) or single-step genomic BLUP. Genetic trends were evaluated based on all combinations of CG type and method for predicting breeding values. The averages of ALL, PL, TNL, TNB, TNW, TLB, TLW, and TLWadj were 1100 d, 890 d, 2.7 litters, 4.6 lambs, 4.3 lambs, 18 kg, 70 kg, and 2.8 kg/kg of ewe weight, respectively. The H matrix performed better than the A matrix, based on akaike information criterion and estimates of breeding value accuracy. Higher average accuracy values were observed when fitting CG as a random effect. The heritability estimates ranged from 0.06 ± 0.02 (TLWadj) to 0.15 ± 0.02 (TLB). All genetic and phenotypic correlations between longevity traits were greater than 0.80. Genetic trends were significant and positive for all traits, but no substantial genetic gains were observed. Considering the observed average values and the estimated genetic parameters, we recommend that longevity becomes part of the selection objectives for U.S. Katahdin sheep.