Data and code from: Effects of phylogeny on coexistence in model communities

This repository contains code and data needed for the analyses presented in the paper 'Effects of phylogeny on coexistence in model communities'. In that contribution, by coupling a simple model of trait evolution on a phylogenetic tree with Lotka-Volterra community dynamics, allows us to derive properties of a community of coexisting species as a function of the number of traits, tree topology and the size of the species pool. Our analysis highlights how phylogenies, through traits, affect the coexistence of a set of species. Together, these results provide much-needed baseline expectations for the ways in which evolutionary history, summarized by phylogeny, is reflected in the size and structure of ecological communities. Here we provide de code needed to compute analytic expectations, to simulate community dynamics, as well as the dataset (Biodiversity II experiments, Tilman et al., Science 2001) used to test model predictions.

Methods
Raw data was taken from Biodiversity II experiments (Tilman et al., Science 2001). That biodversity dataset was filtered by Lemos-Costa et al. (https://onlinelibrary.wiley.com/doi/10.1111/ele.14490, 2024). Only datasets satisfying three conditions were taken into account by Lemos-Costa et al.: a) only species that appear in multiple plots are considered to be part of the pool; b) unidentified biomass or non-target species do not dominate the plots; c) there is sufficient replication or diversity of assemblages to confidently fit the model. In the raw dataset, aside from the original target pool of eighteen species, many other species are found in the plots, along with a substantial amount of litter or unidentified biomass. Therefore, a larger number of species was included in the “species pool” , and also plots in which the total biomass for species not in the pool surpassed 10% of the total biomass were excluded. In this way, Lemos-Costa et al. extracted nine data sets (for distinct years) containing between 13 and 20 species and a substantial number of assemblages. These datasets were downloaded from https://github.com/StefanoAllesina/lemos-costa-2023, together with the phylogenetic trees computed using the V.PhyloMaker R package. They contain assemblages in which different combinations of the 16 species were planted in plots.

We took those datasets and, by comparing them with raw data, we extracted only the plots in which all the 16 species in the experiment were planted. This way we obtained diversity data to fit the model to diversity data.