Designing fiber–gut microbiome interactions with active learning

Abstract

Identifying synergies between dietary fibers and beneficial bacteria holds promise for precision interventions that optimize gut health, yet these interactions remain largely unexplored. Here we integrate machine learning, Bayesian optimization and high-throughput community construction to investigate how dietary fibers shape health-relevant functions of human gut microbial communities. To efficiently navigate the landscape of fiber–microbiome interactions, we implemented a design–test–learn cycle to identify fiber–species combinations that maximize a multiobjective function capturing beneficial community properties. Our model-guided approach revealed a highly butyrogenic and robust ecological motif characterized by the copresence of inulin, Bacteroides uniformis and Anaerostipes caccae and a higher-order interaction with Prevotella copri. Human fecal communities invaded with model-designed species–fiber combinations displayed predictable gut-beneficial outputs. In sum, we demonstrate a framework for designing synthetic microbial communities with desired functions in response to key nutrients. Dietary fibers are a primary energy source for gut bacteria and shape the composition and function of the microbiome. Now, an active learning approach integrating explainable machine learning, Bayesian optimization and high-throughput community construction designs microbial communities with desired functions in response to key nutrients.