Modelling ecological communities when composition is manipulated experimentally

Abstract

In an experimental setting, the composition of ecological communities can be manipulated directly. Starting from a pool of n species, it is possible to co‐culture species in different combinations, ranging from monocultures, to pairs, and all the way up to the full species pool. Leveraging datasets with this experimental design, we advance methods to infer species interactions using density measurements taken at a single time point across a variety of distinct community compositions. First, we introduce a fast and robust algorithm to estimate parameters for simple statistical models describing these data, which can be combined with likelihood maximization approaches. Second, we derive from consumer–resource dynamics a family of statistical models with few parameters, which can be applied to study systems where only a small fraction of the potential community compositions have been observed. Third, we show how a Weighted Least Squares framework can be used to account for the fact that species abundances often display a strong relationship between means and variances. To illustrate our approach, we analyse datasets spanning plant, bacteria and phytoplankton communities, as well as simulations, consistently recovering a good fit to the data and demonstrating the ability of our methods to predict equilibrium densities in out‐of‐sample communities. By combining more robust model structures and fitting procedures along with a more flexible error model, we greatly extend the applicability of recently proposed methods to model community composition from experimental data, opening the door for the analysis of larger pools of species using sparser and noisier datasets than was previously possible.

Phylogeny structures species’ interactions in experimental ecological communities

“…This approach is grounded in a simple framework to infer interactions and predict biomasses from community “endpoints” — a framework that is gaining traction in the ecological literature (Xiao et al 2017; Fort 2018; Maynard, Miller, and Allesina 2020; Ansari et al 2021; Skwara et al 2023). One of the main advantages of this framework is that it makes predictions that are compatible with models of population dynamics: if we were to simulate communities using the fitted interaction strengths as parameters of the Generalized Lotka-Volterra model, we would find the same predicted abundances as equilibria of the dynamical system.…”

“…Versions of this basic framework have been proposed numerous times in the ecological literature (Xiao et al 2017; Fort 2018; Maynard, Miller, and Allesina 2020; Ansari et al 2021), and here we closely follow the approach of Maynard et al . (Maynard, Miller, and Allesina 2020; Skwara et al 2023).…”

Sparsity of higher-order landscape interactions enables learning and prediction for microbiomes

“…If we can write down and parameterize a model that accurately represents interactions, but using only a limited amount of experimental data, we might be able to use such a model to predict outcomes beyond those experiments. A longstanding approach to modeling interactions has been to use pairwise interactions among species, with the strength of this pairwise dependence encoded in a community interaction matrix [16,17]. But there is a recent realization that microbial communities may be infused with higher-order interactions, where the influence of one species on another is dependent on the presence or absence of a third, fourth, or fifth species [18][19][20][21][22].…”

“…But there is a recent realization that microbial communities may be infused with higher-order interactions, where the influence of one species on another is dependent on the presence or absence of a third, fourth, or fifth species [18][19][20][21][22]. With the potential for such higher-order interactions, fitting mechanistic models could just bring us back to a similar combinatorial problem to the one we started with [17,18,23,24]. Finally, deep learning has been deployed to address this question [23,25], at the expense of making results and predictions challenging to interpret.…”

“…If we can write down and parameterize a model that accurately represents interactions, but using only a limited amount of experimental data, we might be able to use such a model to predict outcomes beyond those experiments. A longstanding approach to modeling interspecific interactions, inspired by Robert May's seminal work [16] on complex systems, has been to use pairwise interactions among species, with the strength of this pairwise dependence encoded in a community interaction matrix [17,18]. There is certainly no conceptual obstacle to pairwise interactions providing a potential description of microbial communities-theoretical work demonstrates that pairwise interactions can lead to diverse, realistic communities [19][20][21][22][23][24][25][26].…”