Evaluation of carcass, live, and real-time ultrasound measures in feedlot cattle: I. Assessment of sex and breed effects.

Carcass and live-animal measures from 1,029 cattle were collected at the Iowa State University Rhodes and McNay research farms over a 6-yr period. Data were from bull, heifer, and steer progeny of composite, Angus, and Simmental sires mated to three composite lines of dams. The objectives of this study were to estimate genetic parameters for carcass traits, to evaluate effects of sex and breed of sire on growth models (curves), and to suggest a strategy to adjust serially measured data to a constant age end point. Estimation of genetic parameters using a three-trait mixed model showed differences between bulls and steers in estimates of h2 and genetic correlations. Heritability for carcass weight, percentage of retail product, retail product weight, fat thickness, and longissimus muscle area from bull data were .43, .04, .46, .05, and .21, respectively. The corresponding values for steer data were in order of .32, .24, .40, .42, and .07, respectively. Analysis of serially measured fat thickness, longissimus muscle area, body weight, hip height, and ultrasound percentage of intramuscular fat using a repeated measures model showed a limitation in the use of growth models based on pooled data. In further evaluation of regression parameters using a linear mixed model analysis, sex and breed of sire showed an important (P < .05) effect on intercept and slope values. Regression of serially measured traits on age within animal showed a relatively larger R2 (62 to 98%) and a smaller root mean square error (RMSE, .09 to 8.85) as compared with R2 (0 to 58%) and RMSE (.31 to 67.9) values when the same model was used on pooled data. We concluded that regression parameters from a within-animal regression of a serially measured trait on age, averaged by sex and breed, are the best choice in describing growth and adjusting data to a constant age end point.