Strain-specific flavor signatures in fermented Chinese spicy cabbage: An integrated framework of kinetic modeling and machine learning
Year: 2025
Type: article
Abstract: This study investigated the effects of Lactiplantibacillus plantarum , Leuconostoc mesenteroides , and Levilactobacillus brevis starters on the fermentation and flavor profile of Chinese spicy cabbage (CSC). The starters exhibited distinct fermentation kinetics, accurately described by Logistic modeling (R 2 > 0.91), which directly impacted processing time. These kinetic differences resulted in unique metabolic signatures that defined the final flavor. While Leu . Mesenteroides fermentation produced higher sulfides (pungent notes), L . plantarum fermentation yielded higher aldehydes (undesirable fatty odors). A Random Forest model successfully classified the starter groups and identified ( E,E )-2,4-heptadienal as a key differentiating volatile organic compound. In contrast, Lb. brevis fermentation resulted in higher mannitol, balanced acids, lower off-notes, and enhanced physical attributes. Sensory analysis confirmed Lb . brevis -fermented CSC achieved the highest overall acceptability due to its balanced, refreshing flavor. These findings demonstrate that starter selection is a critical control point for CSC quality. The innovative integration of kinetic modeling and machine learning provided novel insights into starter-specific flavor development, offering a valuable strategy for the rational design of industrial fermentations. • Logistic modeling (R 2 > 0.91) quantified distinct starter kinetic profiles in CSC. • LP showed slow kinetics (homo-fermentative), high tolerance, higher aldehydes. • LM showed fast kinetics, poor tolerance, high mannitol & sulfides, pungent notes. • Lb. brevis , identified as optimal, yielded balanced flavor, lower off-notes. • RF modeling identified ( E,E )-2,4-heptadienal as a crucial differentiating VOC.
Authors: Yunjing Gu, Yun Cen, Lufeng Zhang, Weiye Cheng, Qingyang Zhang, Xuan Wang, Yaxin Li, Hui Guan, Kanghee Ko, Hansheng Gong, Wenli Liu, Xiaoping Liu, Huamin Li
Language: English
Subfield: Food Science
Domain: Life Sciences
SDG: Zero hunger
Open Access status: gold
Estimated APC paid: $2,530
Funder: Ludong University
Award: G197662784
OpenAlex ID: W7106247237