Quantitative Characterization of Nutritional Composition and Biochemical Variability in Alternative Protein Ingredients for Sustainable Aquaculture Feed Development Using Proximate Analysis and High-Performance Liquid Chromatography

The search for nutritionally adequate and environmentally sustainable protein sources remains a central challenge in aquaculture feed development. This study develops a quantitative framework for characterizing alternative protein ingredients using proximate composition and high-performance liquid chromatography (HPLC)-derived amino-acid profiles, and then evaluates whether multivariate laboratory information can support evidence-based ingredient ranking for aquafeed design. A reproducible synthetic experiment was constructed to emulate laboratory measurements from 216 batches representing nine candidate protein ingredients: soy protein concentrate, corn gluten meal, wheat gluten, black soldier fly meal, mealworm meal, spirulina meal, chlorella meal, yeast single-cell protein, and pea protein concentrate. Measured attributes comprised crude protein, lipid, ash, fiber, moisture, digestibility proxy, biochemical stability proxy, and ten essential amino-acid variables. A composite nutritional suitability score was defined from protein density, total essential amino acids, lysine, methionine, digestibility, biochemical stability, fiber, ash, and lipid balance. Logistic regression, random forest, and histogram gradient boosting were compared for classification of high-suitability batches. Logistic regression achieved the highest ROC-AUC (0.994), PR-AUC (0.992), and the lowest Brier score (0.036), while soy protein concentrate, wheat gluten, and spirulina meal achieved the highest average suitability scores. The study offers a transparent, simulation-based analytical template for integrating bench-top nutritional characterization with ingredient prioritization.