Partition of metabolizable energy, and prediction of growth performance and lipid deposition in broiler chickens.

The study presented here consisted of the calculation of cross relationships between growth performance parameters, body growth composition, and feed characteristics, using data from an experiment reported in 2 previous publications. In the previous experiment, 30 pelleted diets were given to broiler chickens (8/diet) (21 to 35 d) for in vivo measurement and prediction of AMEn and net energy (NE) values of diets, using 3 trials with 10 diets/trial. In the course of NE determination, individual values for growth, feed intake, and deposition of lipid and protein were measured. Average energy deposited as lipid and protein represented 25.4 and 19.1% AME intake, respectively. Using a multiple regression predicting AME intake, the partial efficiencies of AME for energy deposition as lipid and protein were calculated to be 91.6 and 67.3%, respectively, and the daily amount of AME required for maintenance was evaluated at 0.683 MJ/kg BW0.7. The mean diet NE/AMEn ratios were predicted by an equation combining the lipid content of body growth (positive coefficient) and the apparent digestible protein (ADP) to AMEn ratio (ADP/AMEn), with a quadratic expression for the latter variable. This quadratic response expressed a positive asymptotic relationship, with a plateau for ADP/AMEn values above 1.45 [%/(MJ/kg)]. The equations predicting growth always included either the dietary percentage of water-insoluble cell wall or the AMEn value. The other major parameters predicting growth were either the lipid content of body growth or the CP/AMEn ratio. In many cases, quadratic responses were observed in growth prediction equations. Regressions predicting feed efficiency showed only linear responses. Feed efficiency was predicted precisely by multiple linear regressions based only on AMEn and a dietary protein parameter. According to these regressions, 1% CP was equivalent to 0.247 MJ/kg AMEn in terms of feed efficiency. The most efficient regression predicting the individual lipid content of body growth combined the protein efficiency value (negative coefficient), the CP/AMEn ratio (negative coefficient), AMEn (positive coefficient), and the feather content of body growth (positive coefficient).