Stochastic and deterministic processes shape microbial communities and flavor characteristics of strong-flavor daqu during storage.

Daqu is a wheat-based fermentation starter essential for Chinese Baijiu production. The storage of daqu critically influences its quality. However, the assembly mechanisms of microbial communities in daqu during the storage phase and their impact on volatile flavor compounds remain unclear. This study investigated the microbial community assembly process and volatile compound profiles of daqu from three major regions of strong-flavor Baijiu production, Luzhou (LZ), Chengdu (CD), and Deyang (DY), during the storage phase. Fifteen biomarkers were identified across the regions using the random forest algorithm. Co-occurrence network analysis indicated that the LZ network was the most stable, while the CD network had the lowest complexity and stability, revealing 14 keystone species. Additionally, 42 differential volatile flavor compounds were identified. The analysis of microbial assembly mechanisms revealed that both stochastic and deterministic processes influenced the assembly of microbial communities in daqu during storage, as indicated by the modified stochasticity ratio (MST) and the neutral community model. Stochastic processes predominantly shaped the community of samples DY, while deterministic processes were more influential in the communities of samples LZ and CD. Notably, the important taxa (dominant microorganisms, keystone species, and biomarkers) were mainly influenced by deterministic processes, including temperature, dewpoint temperature, and surface air pressure, showing significant correlations with volatile flavor compounds. In conclusion, these findings highlight the effects of regional differences and storage processes on microbial community assembly and volatile flavor compounds in daqu, offering insights to regulate the daqu storage process and thereby optimize production.