Fast fitting of phylogenetic mixed-effects models

TL;DR

This paper introduces a fast, flexible method for fitting phylogenetic mixed-effects models that incorporate residual covariation and phylogenetic information, significantly reducing computation time compared to existing MCMC-based approaches.

Contribution

The authors develop a novel combination of Variational approximations, sparse precision matrices, and parallel computing to efficiently fit complex phylogenetic mixed-effects models.

Findings

Methods are faster than current state-of-the-art with high accuracy.

Simulation studies confirm the efficiency and precision of the approach.

Application to real data demonstrates practical utility and sensitivity analysis.

Abstract

Mixed-effects models are among the most commonly used statistical methods for the exploration of multispecies data. In recent years, also Joint Species Distribution Models and Generalized Linear Latent Variale Models have gained in popularity when the goal is to incorporate residual covariation between species that cannot be explained due to measured environmental covariates. Few software implementations of such models exist that can additionally incorporate phylogenetic information, and those that exist tend to utilize Markov chain Monte Carlo methods for estimation, so that model fitting takes a long time. In this article we develop new methods for quickly and flexibly fitting phylogenetic mixed-effects models, potentially incorporating residual covariation between species using latent variables, with the possibility to estimate the strength of phylogenetic structuring in species responses per environmental covariate, and while incorporating correlation between different covariate effects. By combining Variational approximations, a sparse approximation to the phylogenetic precision matrix, and parallel computation, phylogenetic mixed-effects models can be fitted much more quickly than the current state-of-the-art. Two simulation studies demonstrate that the proposed combination of approximations is fast and enjoys high accuracy. We explore sensitivity of the approximation to the ordering of species with a real world dataset of wood-decaying fungi.