Toward predictive control of the gut microbiome: from perturbation to precision intervention

The human gut microbiome is a densely interconnected ecosystem whose responses to perturbation remain difficult to predict, limiting our ability to rationally design therapies. In this talk, I will present an integrated framework for understanding and engineering microbiome dynamics by combining systematic perturbation screens with in situ measurements of microbial communities along the intestinal tract. First, I will describe large-scale drug and small-molecule screens in human-derived microbial communities, revealing that even narrowly targeted perturbations can produce broad, community-wide effects through metabolic competition and ecological release. These experiments uncover recurring response modes – collapse, resilience, and opportunistic expansion – that are not readily predictable from monoculture behavior or genomic content alone. Second, I will introduce a new approach for directly sampling the small intestinal microbiome using an ingestible capsule device, enabling spatially resolved measurements of microbial composition, gene content, and metabolite profiles in regions that have historically been inaccessible. These data reveal that the small intestine harbors distinct, dynamic communities with unique metabolic constraints, and that perturbations can have region-specific effects that are invisible in stool-based measurements.

By integrating these approaches, we begin to link controlled perturbations with their in vivo consequences, providing a path toward predictive, mechanism-based interventions. I will close by discussing how these insights can be leveraged to design targeted strategies – combining selective molecules, ecological principles, and spatial context – to reshape microbial communities for therapeutic benefit.