EvoLunch Seminar: Martin Polz (University of Vienna)

Microbial communities contain enormous diversity, but the taxonomic units commonly used to describe that diversity do not necessarily correspond to ecologically meaningful populations. As a result, adaptive differences can remain hidden. We present a reverse-ecology framework that uses genomic variation—especially signatures of natural selection—to identify populations with distinct ecological adaptations. Across human and animal microbiomes, we find that genome-wide selective sweeps frequently structure bacterial populations. These sweeps happen under a wide range of recombination rates, and can show gene specific sweeps as signatures of secondary adaptations. In the human microbiome, these sweep-defined populations can occur across geographically and ethnically diverse hosts, resembling global epidemics. Several such populations may coexist within a single traditionally defined species, and divergence-time estimates suggest that some can spread worldwide within decades. Importantly, sweep-defined populations show associations with human phenotypes that are not apparent at the species level. Using an example from the cow microbiome, we further show that differential adaptation among sweep clusters can promote coexistence through the evolution of metabolic tradeoffs.Together, these results show how genomes can be used not only to catalogue microbial diversity, but also to identify ecologically cohesive populations and generate testable hypotheses about the processes that structure them.