Community-Level Selection Evolves Group Assemblage Phenotypes Inaccessible Through Individual-Level Selection in a Simple Agent-Based Model with Emergent Behavior
61 Pages Posted: 12 Apr 2025
Abstract
How do communities of diverse organisms maintain relative frequencies of members, even as the community evolves, grows, and replicates? How do communities that contain mutually dependent species, such as bacterial biofilms, self-organize and maintain these niches as individuals compete for resources? Are there community-wide characteristics, such as a distinct flavor of cheese caused by a diverse assemblage of microbes, that are only possible through the careful curation and selection of communities by humans? While evidence of selection at levels lower than the individual (such as at the gene-level) are well-documented, the possible outcomes of selection at higher levels (such as communities and community-level properties) are unclear. We used a simple, agent-based simulation to investigate the degree and the predictability of the impacts of community-level selection. In this model, we independently toggled individual-level selection and community-level selection. We found an unexpectedly complex fitness landscape and an emergent behavior — discrete partitioning of individuals into groups — connecting community composition and reproductive success in this simulation. Community-level selection maximized success of communities even with this complex fitness landscape, but the addition of individual-level selection accelerated this process if, and only if, both levels of selection were selecting for complementary properties. Community-level selection also increased the final diversity in the community compared to individual-level selection alone. While our model is simple compared to natural systems, we provide guidance for experimenters curating and selecting communities for diversity or for properties inaccessible by individual species alone.
Keywords: evolution, selection, levels of selection, community-level selection, artificial ecosystem selection
Suggested Citation: Suggested Citation