Quality Improvement of Fermented Soybean Meat Flavouring by Targeted Regulation of Flavour Precursors via the O2PLS Model

Fermented meat flavouring, characterized by its natural, layered, and rich sensory profile, addresses the growing demand for premium and differentiated food products, and has become a prominent research area in the flavour industry. This study elucidated the precursor‐flavour relationships in dual‐strain fermentation using O2PLS modelling, enabling targeted process optimization and substantial quality improvement. Meat flavouring was first prepared through multi‐strain fermentation of soybean protein, followed by the Maillard reaction. HPLC was subsequently employed to quantitatively track the dynamic changes of free amino acids, while key volatile flavour compounds were identified through HS‐SPME‐GC‐O‐MS combined with AEDA. The O2PLS model was then applied to elucidate the correlation between free amino acids and volatile flavour compounds, revealing alanine, arginine, histidine, and the < 1000 Da peptide content as critical precursors for target optimization. Under optimal conditions (40 h fermentation, 10% inoculum, 1:1 inoculation ratio, the total amino acid content of the resulting sample was 142.2 ± 2.69 mg/100 mL, and the peptide components with molecular weights < 1000 Da was 84.9% ± 1.11%, p < 0.05), sensory evaluation confirmed significantly enhanced intensity of meaty (3.2 ± 0.2, p < 0.05), roasted (1.62 ± 0.01, p < 0.05), and caramel (1.6 ± 0.47, p < 0.05) notes with reduced sulfur (1.0 ± 0.22, p < 0.05) off‐odour. This study provides theoretical insights and technical guidance for flavour quality modulation of fermented meat flavourings, contributing to the advancement of flavour‐oriented regulation strategies in fermented seasoning systems.