Enhancing beef tallow flavor through enzymatic hydrolysis: Unveiling key aroma precursors and volatile compounds using machine learning

Lipids are critical precursors of aroma compounds in beef tallow. This study investigated how enzymatic hydrolysis treatment affected the aroma precursors and flavor of beef tallow during the manufacturing process. Using gas chromatography-mass spectrometry and high-resolution gas chromatography-ion mobility spectrometry, 100, 111, and 122 aroma compounds were identified in beef tallow at three processing stages namely, raw beef fat, enzymatic hydrolysates, and enzymatic beef tallow. Employing machine learning methods, including fold change analysis, partial least squares-discriminant analysis, and random forest algorithms, we identified 26 potential aroma biomarkers strongly associated with the manufacturing process. Furthermore, debiased sparse partial correlation analysis revealed the significant contribution of C22:2, C22:6 n3, C20:3 n3, C20:1, C18:1 n9t, C18:1 n6t, and C18:2 n6t to aroma compound formation, thereby enhancing our fundamental understanding of flavor precursors and their enzymatic transformation.