Fermentation vessels affect the characteristics of food fermentation; however, we lack an approach to identify the biomarkers indicating fermentation. In this study, we applied metabolomics and high-throughput sequencing analysis to reveal the dynamic of metabolites and microbial communities in age-gradient fermentation vessels for baijiu production. Furthermore, we identified 64 metabolites during fermentation, and 19 metabolites significantly varied among the three vessels (p < 0.05). Moreover, the formation of these 19 metabolites were positively correlated with the core microbiota (including Aspergillus, Saccharomyces, Lactobacillus, and Bacillus). In addition, ethyl lactate or ethyl acetate were identified as the biomarkers for indicating the metabolism among age-gradient fermentation vessels by BP-ANN (R2 > 0.40). Therefore, this study combined the biological analysis and predictive model to identify the biomarkers indicating metabolism in different fermentation vessels, and it also provides a potential approach to assess the profiling of food fermentations.
Keywords: alcoholic beverages; baijiu; fermentation vessels; machine learning; metabolite shift; microbial succession.