Linear and nonlinear models for measuring protein nutritional quality

Area of Science:

• Nutritional Science
• Biochemistry
• Animal Nutrition

Background:

• Traditional linear methods (PER, NPR, NPU) provide a single protein quality estimate.
• Assessing protein nutritional quality is crucial for diet formulation and health.
• Existing linear models may not capture the full complexity of protein utilization across varying intake levels.

Purpose of the Study:

• To compare the efficacy of widely used linear methods for assessing protein nutritional quality against a nonlinear saturated kinetics (SK) model.
• To evaluate how different protein sources (casein, peanut, wheat gluten) are ranked by linear versus nonlinear models at various intake levels.
• To determine the strengths and limitations of the SK model in protein quality assessment.

Main Methods:

• Male Sprague-Dawley rats were fed diets with 2-31% casein, peanut, or wheat gluten protein for 4 weeks.
• Nitrogen intake, body weight gain, and body nitrogen retention were measured.
• Protein Efficiency Ratio (PER), Net Protein Ratio (NPR), Net Protein Utilization (NPU), linear regression (LR) slopes, and saturated kinetics (SK) parameters were calculated.

Main Results:

• The SK model ranked peanut and gluten proteins higher than casein at maintenance intake levels compared to linear methods.
• At higher protein intakes, the SK model provided lower relative rankings for peanut and gluten proteins.
• The SK model demonstrated a close fit to experimental data across the intake range but failed to converge with insufficient protein.
• Linear models provided single estimates, while SK allowed assessment across the intake-response curve.

Conclusions:

• The nonlinear SK model offers a more nuanced evaluation of protein nutritional quality, particularly at maintenance, compared to linear methods.
• The SK model's performance is dependent on adequate protein quantity and quality in the diet.
• While the SK model provides dynamic insights, linear methods may be more robust in distinguishing statistically significant differences between protein qualities in certain scenarios.