Yansong Ge, Ziyao Wang, Xu Sun, Xueyan Lin, Zhonghua Wang
To improve feed evaluation for precision dairy nutrition, this study integrated chemical analysis, CNCPS fractionation, in situ ruminal disappearance, exploratory regression, and amino acid-flow calculations for 19 dairy feedstuffs. Conventional nutrients and CNCPS fractions were determined, and total CP and carbohydrate-fraction disappearance were measured in four ruminally cannulated Holstein cows. Protein-fraction Kd values were inferred by nonlinear fitting with measured CNCPS fractions held constant. In the three silages, non-protein nitrogen represented 93.82-96.80% of soluble CP. Energy feeds had greater total carbohydrate and starch-associated CB1 fractions, whereas cottonseed had the greatest unavailable cell-wall fraction. Exploratory equations for Kd1, Kd2, and Kd3 yielded in-sample R2 values of 0.7422, 0.7365, and 0.3871, respectively. Two CNCPS-based sets of intestinally absorbable essential amino acid estimates were then calculated using either in situ-fitted or regression-estimated Kd values. The regression-assisted estimates were generally comparable with those based on in situ-fitted Kd values. However, neither set represented direct intestinal measurements, agreement between them did not constitute validation, and Kp was derived from NRC equations under a standardized dietary scenario. Thus, the amino acid estimates are scenario-dependent, and the regression equations should be considered exploratory pending external validation.