The Metabolizable Energy Evaluation of Rapeseed-Containing Diets in Broiler Chickens Using the Regression Method

Abstract

The present research aims to determine the chemical composition, apparent metabolizable energy (AME), Nitrogen-corrected apparent metabolizable energy (AMEn), true metabolizable energy (TME), and nitrogen-corrected true metabolizable energy (TMEn) in rapeseed specimens, and to predict their energy content using regression prediction equations. 30 rapeseed specimens were provided from different animal and poultry feed factories across Iran. The different energy content of each rapeseed specimen was measured in 4 replicates using the force-feeding method. The results of the laboratory analysis of the specimens were as follows: average raw dry matter (89.91%), crude protein (9.25%), crude fat (3.18%), crude fiber (3.83%), ash (1.35%), starch (65.23%), and glucose (9.69%). The results of field experiments were as follows: average raw energy (4365 kcal/kg), AME (3274), AMEn (3209), TME (3863), and TMEn (3771). Using the obtained data, regression equations were fitted to predict AME, AMEn, TME, and TMEn based on chemical compositions. The results showed that rapeseed specimens from different origins were different in raw energy, crude protein, crude fiber, and ash, but no difference was observed between them in crude fat and TMEn. The GPR model showed a favorable performance based on the evaluation criteria of R²=0.91 and RMSE= 33.67 kcal/kg). The GPR model may improve the ability and capacity to accurately predict the energy of diets to achieve optimal performance in poultry nutrition.

Keywords: Metabolizable energy, Rapeseed, Model prediction, Regression equation