Stochastic Forces in Microbial Community Assembly: Founding Community Size Governs Divergent Ecological Trajectories

Biological community dynamics are organized by the interplay of deterministic and stochastic processes. While species responses to biotic and abiotic environments determine the attractors of community compositions, stochastic processes, especially those in early stages of community assembly, can drastically influence the direction of temporal ecological dynamics. Nonetheless, establishing quantitative insights into the roles of stochastic processes has remained a major challenge. By developing a multi-replicated experimental system for tracking community assembly, we quantitatively evaluated the extent to which stochasticity could cause the divergence of community dynamics. We constructed soil- and freshwater-derived experimental microbial assemblages (microbiomes) with varying foundation community size, thereby controlling variation in initial community compositions among replicate communities (i.e., initial stochasticity). An analysis of > 3,000 community samples across four time points then indicated that higher levels of initial stochasticity could result in greater divergence of population- and community-level consequences. From the quantitative perspective, conspicuous differentiation into alternative trajectories of community assembly occurred when the absolute number of founding prokaryotic cells was less than the order of 104. Furthermore, the replicate microbiomes differed in dominant Pseudomonas species, suggesting divergence into alternative transient/stable states through competitive exclusion between species occupying similar niches. Overall, our experiment showed that quantitative differences in stochasticity could result in qualitative differences in the fate of ecological communities. Further quantitative insights into stochasticity in community assembly will reorganize our views on biological invasions, agroecosystem microbiome management, and therapeutics of human-associated microbiomes.