Energy represents a major cost in feed formulation. Arkansas Net Energy (ArkNE) has shown soybean meal (SBM) provides more energy than reported with other energy systems. This study intended to develop a model to predict Ark NE of SBM based on its content of digestible nutrients, including the essential to non-essential digestible amino acid ratio (AAR), compared to apparent metabolizable (AME, AMEn) and classic net (CNE) energy. Four experiments (E1 to E4) were conducted 1-56 d in floor pens with diets formulated following Cobb 500 guidelines. Pelleted treatment diets (five-phase feeding program) included SBM dilutions with cellulose (E1) or corn (E2) and graded levels of oil (E3) or total digestible amino acids (E4). Body composition (Dual Energy X-Ray Absorptiometry), heat production (calorimetry chambers), performance, and digestible nutrients of 60 test diets were assessed. The ArkNE value of SBM was determined by regression (E1), and the digestible nutrient profiles of 4 SBM were characterized (E1 to E4). A database was built up (n=646), and data were analyzed in JMP using mixed models. A multiple linear model was obtained (E1 to E4) applying a stratified random data-splitting with a refitting approach to predict SBM ArkNE considering digestible crude protein (dCP, %), digestible fat (dEE, %), digestible starch (dST, %), and age (d) as fixed predictors: ArkNE (kcal/kg) = –134.7 + 81.09 dCP + 70.65 dEE + 36.7 dST – 13.76 age – 0.45 (age – 31.1548)2 (every predictor P<0.0001; adjR2 = 0.62). The model was validated. A coefficient to adjust the ArkNE for AAR was inferred from the model above and validated with a multiple linear mixed model that included AAR, dCP, and dEE as fixed predictors (AAR P=0.0178; E1 to E3; 96% accuracy). An AAR-optimized ArkNE model (ArkNEAAR; >92% accuracy) was obtained by combining the prediction model above and the correction factor: ArkNEAAR (kcal/kg) = ArkNE (0.95 AART/AARR + 0.05), being AART and AARR the AAR of the tested and a reference (AARR = 0.82) SBM, respectively. dCP% in SBM explained 95% of its ArkNE value. AME and AMEn were insensitive to dCP%, while CNE was less influenced by dCP% than ArkNE. Only ArkNE was consistently and positively correlated with AAR (P<0.0001), while AME and AMEn showed no consistent relationships. In contrast, CNE was negatively correlated with AAR (P=0.003). FCR and feed intake were correlated more with ArkNE than AME or CNE (>adjR2; < P value).
In conclusion, the ArkNE value of SBM can be predicted based on its digestible nutrient content, credits its AAR, and is a better predictor of performance.