Abstract Microbes in host microbiomes, human infections and the natural environment often live in spatially structured aggregates and interact antagonistically with each other. Toxin-mediated antagonistic interactions are widespread in the gut, skin, and other human microbiomes, and protect these communities against external invasion. Recent results suggest that spatial structure can strongly affect the evolutionary dynamics of microbial populations, and, in turn, microbial interactions can feedback on the formation of spatial structure. For example, we found that mechanical interactions among dividing cells in growing yeast colonies reduce the power of natural selection by reducing the rates at which lower fitness strains go extinct and fitter ones expand in these populations. Despite spatial structure and microbial interactions have a strong impact on the evolutionary dynamics of microbes relevant for human health, most of what we know about microbial evolutionary dynamics comes from experiments with well-mixed liquid cultures with limited interactions among cells. To fill this gap, my group is interested in understanding quantitatively how spatial structure, mechanics and biological interactions impact the adaptive evolutionary dynamics of microbial populations. We approach this question via experimental evolution, synthetic biology, and mathematical modeling. In preliminary experiments, we found that evolving yeast colonies selecting for faster expansion on agar surfaces results in notable changes in cell shape: cells evolved from an ellipsoidally shaped ancestor to being elongated and almost rod- like, changing the way cells interact mechanically when growing and dividing. We hypothesize that an elongated cell shape is advantageous for faster expansion because it reduces cell packing, and that this adaptive change is associated with changes in the way genotypes cluster in space leading to increased genetic drift, the temporal change in allele frequencies due to chance events. Recently, we showed that a toxin- producing microbe can only invade a landscape occupied by a weaker toxin-producer if its inoculum is larger than a critical size, and that adaptive evolution can alter the dynamics of antagonism. We will experimentally investigate the dependence of the critical inoculum size on the strength of the interaction, and we will study how spatial structure controls the fate of mutations that confer resistance to the toxin produced by either the invader or resident strain. Finally, we will investigate how antagonistic interactions among microbes affect the dynamics of invasion in the gut of the nematode Caenorhabditis elegans: these experiments will help us translate results obtained in simple laboratory settings to the more complicated but more realistic dynamics of invasion of a host microbiome.