The industrial production of chicken soup has been a persistent challenge, underscoring the crucial importance of searching for suitable sustainable processing technologies. In this study, the effects of various ultra-high pressure (UHP) times (0, 5, 10, and 15 min) on the flavor of Tibetan chicken soup were investigated. When the processing time reached 10 min, the solubility of myofibrillar protein (MP) and the small-fragmentation index (MFI) increased by 66.41 % and 43.46 %, respectively. There was a significant decrease in α-helix content, while β-turn content showed a marked increase. This demonstrated the full release of macromolecules in the chicken, resulting in a notable rise in the nutrient content, such as proteins, fats, and sugars, in the chicken soup. Consequently, there was a gradual enlargement in the particle size of micro-nanoparticles (MNPs), causing instability phenomena like delamination and flocculation. Additionally, this study integrated six machine learning (ML) models to develop a novel strategy for screening characteristic flavor compounds in Tibetan chicken soup. Among these models, the RF model achieved the highest AUC value (AUC = 0.861) and exhibited a relatively small median residual, showcasing its excellent classification and predictive capabilities. Concurrently, its SHAP (SHapley Additive exPlanations) analysis results revealed that hexanal, 2-pentylfuran, 1-octen-3-ol, and stearic acid were key contributors to the flavor profile of Tibetan chicken soup. 5′-IMP and Glu were consistently identified as key features across multiple models, highlighting their significant role in the flavor profile of Tibetan chicken soup. These components collectively contribute to the distinctive taste of the soup.
扫码关注我们
求助内容:
应助结果提醒方式:
