Development of prediction equations for apparent metabolizable energy corrected for nitrogen of fish meal and poultry by-product meal in broiler chickens

This experiment was conducted to determine the AME_n content of fish meal (FM) and poultry by-product meal (PBM) samples from different origins to develop prediction equations based on the chemical composition. A total of 300 one-day-old male broiler chicks were randomly assigned to fifteen treatments (fourteen experimental diets and 1 reference diet) with 4 replicates of each treatment and 5 birds per replicate. In order to determine AME_n using a replacement method, the test diets consisted of 60% reference diet (based on corn and soybean meal), 38% FM or PBM samples, and 2% micro-ingredients. The broilers were fed the experimental diets for 4 d (24–27), and feces were collected on d 28. The AME_n values of the FM samples varied from 2,886 to 3,157 kcal/kg, and the range for the PBM was from 2,186 to 4,114 kcal/kg. The regression equation prediction of AME_n was determined by SPSS software. Finally, the equations AME_n = 35.658 × CP + 61.975 × EE (standard error of prediction [SEP]: 98.79; Adjusted R²: 0.89) and AME_n = 84.396 × EE + 21.221 × CP (SEP: 269.83; Adjusted R²: 0.93) were suggested to predict the AME_n for FM and PBM, respectively. These equations can save time and cost for poultry nutritionists to formulate the diets for broilers and serve as a reference to develop calibration equations for the predicted values of AME_n of FM and PBM for broiler chickens using near-infrared reflectance spectroscopy.