Guide the design of lactic acid bacteria synthesis community through computational metabolic interaction experimental pipeline

Understanding whether ecological assembly principles can inform synthetic microbial community construction remains a critical challenge in fermented food research, hindered by the complexity of natural ecosystems. Here, we integrate top-down metabolic modeling with bottom-up experimental validation to establish a generalizable framework for food microbial community engineering. Genome-scale metabolic modeling of 507 spontaneously assembled lactic acid bacterial (LAB) communities revealed a significant enrichment of cooperative interactions in naturally fermented ecosystems compared to randomly assembled consortia. This cooperative propensity was driven by amino acid auxotrophies shaped by biosynthetic cost trade-offs, which structured cross-feeding networks across phylogenetically distant strains. Leveraging these ecological patterns, we developed a computational model that quantifies metabolic interaction costs to predict optimal strain combinations. We characterized 15 functional lactic acid bacterial species isolated from fermented foods based on their auxotrophic profiles and interaction capacities. Synthetic communities (2-6 members) engineered via this framework exhibited superior stress resilience (e.g., resistance to osmotic pressure and lactate accumulation) and accelerated substrate utilization compared to non-cooperative communities. Resource conversion efficiency increased by 18 %-37 % in cooperative communities, driven by complementary amino acid exchange. These findings establish metabolic interdependencies as a key driver of community assembly in both natural and synthetic contexts. Our approach provides actionable insights for designing robust starter cultures tailored to industrial fermentation challenges, advancing precision control in food microbiome engineering.