SUMMARY Each adult human harbors hundreds of bacterial species in their intestine. However, the networks of microbe- microbe interactions that underly the stable co-existence of resident species, and exclude additional species, are not well defined. The intestinal lumen is a turbulent, semi-fluid landscape where microbial cells and dietary plant cell wall fragments are distributed with high heterogeneity, and redistributed on the time scale of seconds. We propose that bacteria selectively adhere to dietary particles in the gut lumen and that interactions with their co-adherent microbial partners dictate whether they persist. We created multiplex libraries of artificial food particles (consisting of glycan-coated magnetic beads) to measure gut bacterial adhesion in vivo, and discovered that many members of the phylum Bacteroidetes adhere to dietary glycan particles in a strain- specific and glycan-specific manner. We will first identify families of adhesion proteins required for these binding phenotypes using transposon mutagenesis and insertion site sequencing. Next, we will identify networks of interacting strains that co-adhere to dietary particles. Using orally administered libraries of fluorescently labeled beads, we will map co-adhesion networks in gnotobiotic mice colonized with strains that have evolved together in a single donor host. Finally, we will establish a mutational selection strategy that permits the simultaneous generation of different binding specificities in genetically intractable gut microbes. Analysis of these mutations will reveal the potential origins of adhesion-dependent interspecies relationships. These studies will shed light on the poorly defined spatial structure of the gut microbiota. The technologies we develop in this proposal hold promise as a means to intentionally position the members of microbial communities in physical configurations that prevent or ameliorate metabolic, immunologic, and infectious diseases.