High-Order Interactions between Species Strongly Influence the Activity of Microbial Communities

Abstract

We are surrounded by complex communities of microbes, many that play a fundamental role in our everyday lives. The activity of a biological community is the outcome of complex interactions between community members, and it is often unclear how those interactions work. Developing a quantitative understanding of how ecosystems of microbes interact will be essential to predicting how microbial networks respond to environmental or biological changes and aid in designing synthetic communities with tailored functionality. Here we combine modeling and experimental data to analyze the interaction networks within a multispecies community of microbes.The community consists of four species isolated from fresh water. We measured the overall metabolic rate of all subsets of this community to quantify interactions between community members, including high-order interactions between three or four species. For two-species interactions, we found a range of interactions, from no interaction to strongly positive. Accounting for interactions between species greatly improved the ability to predict ecosystem function. When not taking any interactions into account, the predictions of the overall metabolic rate were off by more than 30%. Further measurements to include 3 and 4 species interactions revealed a complex set of interactions between these four species. The 3 and 4 species interactions coefficients on average resulted in 20-40% changes in metabolic rate of individual species. We compare how model predictions of the overall metabolic rate improve when incorporating pairwise and higher-order interactions between species.We incorporated these findings into a theoretical model which examined how the overall metabolic rate of a community depended on the distribution of interaction strengths and the distribution of cells in space. We found that the distribution of cells in space was particularly important for networks containing strong high-order interactions.