A Flexible Phylogenetic Generalized Linear Mixed Model Framework for Complex Trait Evolution Analysis

Phylogenetic Comparative Methods (PCMs) offer a critical approach to analyzing species traits by accounting for evolutionary relationships. Traditionally, these models incorporate phylogenetic signals and control for shared ancestry, relying on Brownian motion (BM) models to capture trait evolution. However, classic PCMs often fall short in flexibility and computational efficiency, particularly when applied to large datasets or when partitioning complex variation, such as phylogenetic mixed models with random intercepts and slopes. Here, an extended phylogenetic generalized linear mixed model (PGLMM) framework is presented, allowing for both Gaussian and non-Gaussian responses while incorporating multiple levels of variability. This approach enables the modeling of complex evolutionary effects. Using R packages like lme4 and glmmTMB, the method is applied to simulated datasets, demonstrating improved flexibility and computational efficiency compared to existing techniques. The framework broadens the applicability of PCMs, offering a robust tool for evolutionary biology studies involving large datasets and complex trait-environment relationships.