Comparison of methods for prediction of genetic trend in beef cattle selection programs

The purpose of this study was to (1) evaluate the interaction between preweaning nutritional management and selection response in an open herd breeding program and (2) compare methods to estimate or predict genetic trend in both a simulated and a live animal selection experiment.;A selection experiment with Hereford cattle was conducted from 1978 to 1985. Selection line females were mated to top sires chosen from the American Hereford Association's National Cattle Evaluation (NCE). Sire selection was based solely on yearling weight expected progeny difference. An unselected control line was maintained to monitor environmental trend.;In the last two years of the experiment, one-half of calves from both lines were creep fed during the preweaning period. During this period, direct response to yearling weight selection amounted to 28 kg. Significant correlated response was found in preweaning average daily gain (.057 kg/d), weaning weight (14 kg), postweaning average daily gain (.085 kg/d), yearling hip height (4.6 cm) and yearling pelvic area (11.2 cm;Data was simulated to mimic all 8 years of this selection project. Least-squares and reduced animal model (RAM) analyses were conducted to compare genetic trend results. The RAM incorporating Bayesian techniques yielded the most accurate predictions of genetic trend. Following in accuracy was the RAM with grouping based on portion of selected sire's genes an animal possessed. The least-square analyses estimated weaning weight very closely but over-estimated postweaning gain markedly. Other analysis methods utilized were less accurate.;In the live animal project, results from NCE and RAM with Bayesian methodology yielded the largest genetic trend predictions. Next in magnitude were RAM analyses with grouping strategies. Least-square and the RAM with no grouping resulted in the smallest genetic trend predictions.;In this selection experiment, no genotype X creep feeding interaction was found. Analysis methods yielded large differences in results when estimating or predicting genetic trend.