Doctor of Philosophy (PhD)
Animal Science (Animal, Dairy, and Poultry Sciences)
Generalized mixed model methodology and MCMC simulations were used to estimate genetic parameters for calving rate and calf survival with the normal, probit, and logistic models. Calving rate and calf survival were defined as 0 each time a cow failed to calf or a calf failed to survive to weaning age, otherwise they were set to 1. Data were available on 1,458 cows and on 5,015 calves. Cows produced a total of 4,808 records over 4 discrete generations of rotational crosses between Angus, Brahman, Charolais, and Hereford from 1977 to 1995. The heritability of calving rate and calf survival, the EPDs of sires, and mean performance for calving rate and calf survival for various rotational crossbreeding systems were computed. The probit model and the logistic model each failed a lack of fit test based on the scaled deviance for calf survival. Spearmen correlations measured potential change in the ranking of bull EPDs across models. The normal model estimate of heritability for calving rate and calf survival was 0.062 ± 0.023 and 0.038 ± 0.019, respectively. Heritability estimates from the other models were slightly larger when adjusted, but smaller than 20%. Spearman rank correlations were larger than 0.98 indicating a minimal change in the ranking of bull EPDs. The H-B two-breed rotation cows had a higher calving rate than A-B or C-B two-breed rotation cows. The best mating system for calving rate was the A-H two-breed rotation system (0.93 ± 0.07), and the best system for calf survival was the A-B-H three-breed rotation system (0.98 ± 0.03). Three- and four-breed rotation systems were similar to two-breed rotation cows for calving rate. The differences between three-breed and four-breed rotation systems were minimal. Heritability estimates found in this study for calving rate and calf survival were similar to the literature estimates. Sire EPD range varied among models but was less for the normal model. Predicted performance for mating systems is possible with estimates of genetic effects.
Document Availability at the Time of Submission
Release the entire work immediately for access worldwide.
Guerra, Jose Lucio, "Statistical models and genetic evaluation of binomial traits" (2004). LSU Doctoral Dissertations. 3032.