Submitted to: Journal of Animal Science Supplement
Publication Type: Abstract only
Publication Acceptance Date: 3/10/2003
Publication Date: 7/1/2003
Citation: YAMKA, R.M., MCLEOD, K.R., HARMON, D.L., FREETLY, H.C., SCHOENHERR, W.D. ACCOUNTING FOR THE PROPORTION OF ALPHA-AMINO NITROGEN IN CRUDE PROTEIN IMPROVES METABOLIZABLE ENERGY PREDICTION IN DRY EXTRUDED DOG FOODS. JOURNAL OF ANIMAL SCIENCE. 2003. v. 81(SUPP.1): ABSTRACT P. 260. Interpretive Summary:
Technical Abstract: The modified Atwater's equation [metabolizable energy (ME) kcal/kg = 3.5*crude protein (CP) + 8.5*ether extract (EE) + 3.5*nitrogen free extract (NFE)] is currently used to predict the metabolizable energy (ME) content of dog foods. However, we found that the equation consistently under predicted ME compared to the observed ME in 55 balance trials. It was our objective to use these balance trials to develop an equation based on chemical composition of the diet to predict ME content of the diets. Eight diets that varied in ME content (3,463 to 4,233 kcal/kg) were fed at maintenance and used in the analysis. The diets varied in protein source, with the major protein sources being low-oligosaccharide whole soybeans, low-oligosaccharide low-phytate whole soybeans (2 sources), conventional soybean meal (2 sources), low-ash poultry meal, low-oligosaccharide low-phytate soybean meal or conventional whole soybeans. A mutlivariate regression analysis was used to predict ME content based on chemical composition. Two initial models were fit to the data. Model 1 included CP, EE and CF. Model 2 replaced the CF term with acid detergent fiber (ADF) and neutral detergent fiber (NDF), which resulted in a model that contained CP, EE, NDF, and ADF. Because the diets varied in protein sources, the ratio of alpha-amino N (AAN) to non-alpha-amino N (NAAN) ranged from 3.5 to 14.4; therefore, we hypothesized that accounting for the proportion of AAN in CP would improve the fit of the models. Model 1 had an r2 of 0.46 and when AAN and NAAN were substituted for CP, the model had an r2 of 0.79. Similarly, Model 2 had an r2 of 0.43 and when AAN and NAAN were substituted for CP, the model had an r2 of 0.82. Residual analysis suggests that by replacing the CF term in Model 1 with ADF and NDF in Model 2 there was an improvement in prediction of ME content. By splitting CP into an AAN and NAAN fractions we have further defined the chemical composition of the diet. These data suggest that defining protein quality improves the ability to predict ME content of dog foods.