Artificial intelligence-assisted optimization of extraction process, characterization, and functional analysis of globulin from safflower seed meal.

Safflower seed meal, a protein-rich byproduct of oil extraction, is often discarded as waste, resulting in both resource inefficiency and environmental concerns. In this study, albumin, globulin, prolamin, and glutelin were sequentially extracted from safflower seed meal via Osborne extraction method, accounting for 22.69%, 27.69%, 37.33%, and 12.28% of the total protein in the meal, respectively. Physicochemical and functional characterizations revealed distinct functional advantages among the protein fractions. Specifically, the globulin fraction demonstrated high foaming property, favorable emulsifying capacity, and strong surface hydrophobicity, whereas albumin and glutelin exhibited good oil-holding capacity and water-holding capacity, respectively. Given globulin's outstanding performance, its extraction process was further optimized using artificial intelligence-assisted approaches. The suitable extraction conditions for globulin were determined as follows: extraction time of 110 min, solid-liquid ratio of 1:47 g/mL, extraction temperature of 37 °C, and NaCl concentration of 1.24 mol/L. Under these conditions, the globulin yield reached 7.33 ± 0.10%. SDS-PAGE analysis indicated that the molecular weight of globulin was characterized by small molecular weights (13-53 kDa). FTIR spectra revealed β-sheet (30%) was the dominant secondary structure of globulin, while the α-helix content was the lowest (18%); this structural feature may contribute to the globulin's high foaming and emulsifying capabilities. Amino acid analysis identified 17 amino acids in globulin, including eight essential amino acids, with hydrophobic amino acids accounting for 46.88%. Collectively, these results demonstrate that safflower seed meal-derived globulin is a nutritionally balanced and functionally potent plant protein, exhibiting great application potential in the food field.