Collaborative Research: THE THEORETICAL FOUNDATIONS OF MICROBIAL EVOLUTION UNDER VARIABLE CONDITIONS

Biological soil crusts (biocrusts) are living skins on soil surfaces that connect the ground with the atmosphere. They bind soil particles, reduce erosion, regulate water movement and atmospheric carbon and nitrogen. This in turn supports the establishment and productivity of plants. Drylands are water-limited ecosystems that support billions of people worldwide, and in these fragile landscapes, biocrusts play a key ecosystem role in maintaining the stability of natural and agricultural systems. However, biocrusts are not static. They are living communities made of microbes, lichens, mosses, and other organisms that change over time as their members interact, gain and lose genes, adapt to stress, and alter how the community functions. Despite their importance, we still know little about how ecological interactions and evolutionary change jointly shape the function, resilience, and recovery of biocrust communities in changing environments. This project will integrate ecological and evolutionary modeling, computer simulations, genomic data, and artificial intelligence to reveal how biocrust communities change across scales, from genes and microbes to whole landscapes. Insights from this project can guide restoration, land management, and protection of fragile dryland ecosystems, supporting healthier soils, more resilient natural and agricultural systems, and the communities that depend on them. The project will also strengthen education by engaging students with how ecology, evolution, genomics, computer modeling, and artificial intelligence can be integrated to address real-world environmental problems. This project develops an integrated theoretical and computational framework to understand how microbiomes evolve and persist under environmental variability. It combines new stochastic computational models to advance eco- evolutionary theory, with a novel machine-learning–assisted statistical framework for theory-driven inference from longitudinal metagenomes. It applies these tools to data from rainfall-manipulation experiments at the Sevilleta Long-Term Ecological Research Program in New Mexico to identify genes and mechanisms that enable microbial persistence under increasingly variable precipitation regimes. Simulations will be used to test how organisms, genes, and species interact, spread, adapt, or disappear under different environmental conditions. Machine-learning tools will then connect these theoretical insights to real biocrust metagenomic data. By linking ecological interactions with evolutionary change, this work can reveal why some biocrust communities remain resilient while others degrade or recover slowly from environmental stress. It can help identify conditions that support soil stabilization, water regulation, nutrient cycling, and plant establishment. More broadly, the theory and AI-assisted tools developed here will advance science by helping explain how complex biological communities change through interactions over time, with applications beyond biocrusts to host-associated microbiomes and other rapidly changing ecological systems. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.