Abstract Industrial upgrading is widely applied to improve hygiene and consistency in traditional solid-state fermentations, yet its implications for production-environment microbiomes remain unclear. Here, we compared two Shaoxing wheat Qu production modes: an enclosed, automated line (ZN) and a semi-open, semi-mechanized line (JX). Time-resolved Qu samples and environmental matrices along each line were profiled by 16S rRNA and ITS amplicon sequencing, Bayesian source tracking, and ecological null-model analysis. The two modes showed distinct successional trajectories and yielded finished Qu with different quality traits (finished Qu: Day 14 for ZN; Day 60 for JX). ZN finished Qu showed higher saccharification power (1566±36 U/g) and acidic protease activity (26.40±2.70 U/g) than JX (874±19 U/g and 11.6±1.30 U/g), whereas JX retained a broader volatile profile including higher ketone-associated signals (e.g., 2-octanone). Source tracking indicated a stage-dependent change in ZN from raw wheat dominance at inoculation (fungi: 96.88%) to increased recruitment from indoor air at maturity (bacteria: 15.65%; fungi: 46.12%). In contrast, JX showed stronger contact-mediated signatures, with major fungal contributions from the grinder surface at Day 4 (77.42%) and from cultivation tools at Day 16 (88.47%). βNTI-based inference suggested that stochastic processes dominated in both modes, but the deterministic fraction was higher in ZN (bacteria: 17.39%; fungi: 6.88%) than in JX (bacteria: 1.90%; fungi: 0.10%) and was mainly attributable to heterogeneous selection. Distance–decay analysis further suggested a more spatially confined external microbial signal around the enclosed ZN workshop than around the semi-open JX workshop. Together, these results support an environmental terroir perspective for wheat Qu and suggest a potential stability–complexity trade-off that accompanies industrial upgrading.
Zhao et al. (Sat,) studied this question.