Modeling microbial communities as a network of chemically mediated interactions
This event is part of the Biophysics/Condensed Matter Seminar Series.
Microbial communities are often modeled as a network of pairwise fitness (NPF) influences between species. Such pairwise models have limitations in capturing common interaction types among microbes. Yet, pairwise models are popular, because they have received empirical support in some cases and, not incorporating interaction mechanisms, are easy to construct. Here, we analyze the validity of NPF modeling for microbial communities. For this analysis, we build a reference model as a network of chemically mediated interactions (NMI) by explicitly incorporating interaction mediators. Comparing the predictions between this reference model and a pairwise model derived from it reveals the conditions under which NPF is a valid approximation. We find that NPF modeling may fail to capture community dynamics in many commonly encountered cases: for example, when both consumable and reusable mediators are present, when one species affects another via multiple mediators, or when a mediator is produced or consumed by multiple species. We further use NMI modeling to explore the network properties that promote coexistence of species in microbial communities.