Using divergent selection and genomics to uncover genetic variation underlying larkspur tolerance and susceptibility in cattle

In the Rocky Mountain region of western US, selection for larkspur tolerance would reduce mortality of cattle from larkspur poisoning and increase opportunity to utilize pastures at peak nutrient availability resulting in increased sustainability of beef production. Previous research indicated that there are breed differences for tolerance to toxic larkspur. Our objective was to estimate heritability for larkspur tolerance within breed and evaluate the potential for increasing larkspur tolerance through artificial selection. Larkspur challenge was administered to 141 yearling steers (32 Angus, 13 Brahman, 49 Line 1 Hereford, 33 Holstein, and 14 Jersey) with a standardized dose of dried ground larkspur suspended in water and gavaged directly into the rumen. Larkspur tolerance was measured at 24 h after dosing as the length of time (up to 40 min maximum) in which the animal could sustain walking at 6.44 km/h while being led behind a tractor on a circular track. High-density SNP arrays (770,000 or 30,000 SNP) were used to genotype each steer and genotypes were used to compute the genomic relationship matrix which is a precursor to estimating heritability. Larkspur tolerance heritability estimates were similar whether estimated with REML (0.36 ± 0.30; P = 0.10) or Bayesian Monte Carlo Markov chain (MCMC) (0.42 ± 0.23; MCMC posterior distribution 2.5, 25, 50, 75, and 97.5th percentiles were 0.035, 0.24, 0.40, 0.59, and 0.90). To evaluate the potential for using our larkspur challenge data to calculate EBV of an untested population, we computed genomic relationship coefficients between 190 previously genotyped (but untested for larkspur tolerance and comprising the same 5 breeds in this study) cattle and the 141 steers tested for larkspur tolerance. Because of uncertainty in the heritability estimate, EBV were computed for each iteration of the MCMC to average over all possible values for heritability and weight by the appropriate posterior density. The most extreme EBV were for target animals with the strongest genetic ties to tested animals. Simulations indicated that divergent selection of parents can more than double the power for estimating heritability. Our results indicate that selection for larkspur tolerance should be effective. The rate of selection response will critically depend on challenging and testing animals with strong genetic ties to candidates for selection. Genetic ties can either be estimated from SNP genotypes or computed from common ances