Bioanalysis of Flavor Features in Fermented Vegetables Based on SERS Spectrum and AI algorithm

The unique flavor and nutritional value of fermented vegetables are dominated by the microbial activities during fermentation. This study aimed to investigate the flavor formation mechanisms of fermented vegetables by employing a multidimensional approach, incorporating both surface-enhanced raman spectroscopy (SERS) and artificial intelligence (AI). With the help of sensory evaluations, the flavor of fermented vegetables was identified by SERS spectra with high accuracy at 91%. Feature importance analysis was performed to identify the key Raman peaks associated with amino acids and organic acids that significantly affect flavor. Microbial diversity and composition were analyzed at different fermentation temperatures. The correlation analysis revealed that microorganisms such as Bacillus, Psychrobacter, and Pseudomonas were highly correlated with the production of key amino acids and organic acids. Their metabolic pathways were further elucidated, demonstrating the influence of microbial activity on the flavor profile of fermented vegetables.