Genetic prediction of mature weight and mature cow maintenance energy requirements in Red Angus cattle

Abstract

The objectives of the multivariate mature weight analysis were 1) to determine if mature weight was a heritable trait in Red Angus cattle, 2) to determine if the assumptions of a repeated measures model were valid for prediction of mature weight, and 3) to evaluate the feasibility of developing a mature weight prediction using random regression models. I obtained cow weight data on animals 2 years of age and older from the Red Angus Association of America, Denton, TX. The data included 111,460 weight observations. After editing the data for incomplete fixed effects and missing contemporary group information, the reduced data set included 74,124 observations. A subset of 29,380 records representing four Red Angus herds was used in the multivariate analysis. Results of the first analysis showed that weight in Red Angus cattle 2 to 10 yr of age is a heritable trait with estimates of .50 to .78. Genetic correlations between different ages remained high with a range of .86 to .98; residual correlations ranged from -.04 to .50 and were inconsistent in magnitude; and, phenotypic correlations were .42 to .71. The estimates of genetic, residual, and phenotypic correlations satisfied some of the assumptions for a repeated measures model, but the genetic, residual, and phenotypic covariances exhibited a scaling pattern with increasing variation as age increased. According to this study, the assumptions of a repeated measures model are not appropriate for analysis of mature cow weight. Modifications were made to existing Method R variance components software to estimate variance components using random regression models. Validation of the techniques showed that it was feasible to obtain variance components using random regression models. Simulated data results showed that three to four records per animal were needed to obtain reliable variance components estimates. Mature weight predictions were obtained using random regression models and BLUP procedures. The variance components for this analysis included those from the repeated measures model analysis, which were adapted for use with random regression models. Previous attempts to estimate variance - components using a random regression model resulted in inflated variance component estimates. Predictions from the random regression analysis were compared with the repeated measures model results. There was an improved fit with the random regression model compared to the repeated measures model (R2 = .940 vs. .804). The correlations between estimated breeding values from different models were .95 to .98, for ages 2 to 6 yr. The objectives of the weaning weight and postweaning gain with mature weight analysis were 1) to determine if there was a genetic relationship between mature weight and weaning weight or postweaning gain in Red Angus cattle; and 2) to determine if the genetic relationship changed across cow weight age classes. Genetic correlation estimates between weaning weight (direct) and weight at 2, 3, and 4 to 9 yr of age were .61, .53, and .81, respectively. Genetic correlation estimates between postweaning gain and weight at 2, 4, 6, 8, and 2 to 9 yr of age were .65, .60, .67, .64, and .68, respectively. An additive genetic relationship was present between mature weight and weaning weight, and postweaning gain. Including immature growth traits in a multivariate analysis with mature weight may serve to increase accuracy of mature weight genetic predictions. I analyzed Red Angus cow weight with stayability using a multivariate continuous and categorical model. The objective was to evaluate the additive genetic relationship between traits. For the RAAA, stayability is defined as the probability of daughters staying in the herd to six years of age. Genetic correlation estimates between 2, 3, and 5 yr old weight with stayability were .28, .03, and .04, respectively. Results of the multivariate analysis showed a genetic relationship between weight at 2 yr of age with stayability, but this went to zero at older ages. These results were inconclusive and further study with different analytical procedures was needed. Using additive genetic groups models, I analyzed stayability with cow weight genetic groups. Results showed there was a nonlinear genetic relationship between traits. Additional genetic groups models were fit to assist the interpretation, because results of the first analysis were inconclusive. Using alternative grouping strategies, results showed that there was still a nonlinear relationship between traits. The results showed that animals with different EBV for yearling growth were present in the same mature weight group. A mature cow maintenance energy EPD was developed using genetic predictions for mature cow weight and maternal weaning weight. The mean, SD, and range of EPD were 22.37, 102.14, and -427.95 to 581.88 Meal / yr. The genetic trend for years 1970 to present was increasing and followed the same pattern as the genetic trend for mature weight. Using the results of the mature cow maintenance energy study, it is feasible to develop a genetic prediction for a new economically relevant trait. Further work is needed to determine the best model for prediction of mature cow maintenance energy requirements.