Predicting carcass composition and individual feed requirement in live cattle widely varying in body size

A total of 192 feeder steers of five breed types and body sizes commonly found in the United States cattle population were fed high-energy diets to three endpoints (275-, 300-, and 360-kg carcass weights) to determine their carcass composition. Before slaughter, ultrasound was used to predict fat thickness, longissimus muscle area, and marbling. Individual steer data were used for developing prediction equations, which were validated with three independent data sets. These data were used to develop and validate equations to predict carcass composition and DM requirements for individuals fed in pens and varying in breed type, body weight and size, and ADG. Equations to predict carcass weight during growth accounted for 84, 83, and 88% of the variation in the three data sets with 0, 1, and 3% bias. An equation to predict percentage of carcass fat from fat thickness and equivalent shrunk weight accounted for 96% of the variation in the percentage of carcass fat. An equation to predict yield grade from longissimus muscle area per 100 kg, fat thickness, and equivalent shrunk weight accounted for 93% of the variation. Dry matter requirement predicted by the system for individuals accounted for 48% of the variation in actual DMI with a 3% overprediction bias. The equations allow the user to allocate feed to individual animals in group-feeding environments along with marketing cattle on an individual basis at optimum endpoints given cattle types, feeding costs, and market prices.