Estimating the rate of quantitative trait evolution in the presence of gene tree discordance by calculating likelihoods across trees

Abstract

Quantitative traits provide insights into how phenotypes evolve across species. However, standard comparative methods often assume a single species tree and overlook the discordant gene tree histories that may underlie complex traits. Here, we develop a model and software (Spinney) that explicitly incorporate gene tree heterogeneity into rate estimation. Spinney finds the optimal rate of evolution and ancestral states by jointly maximizing the likelihoods across a set of gene trees. Using simulated data, we compare rate estimates from Spinney with those using the species-tree alone, test Spinneys ability to distinguish true rate variation from spurious signals caused by gene tree discordance, and evaluate ancestral state reconstruction. Spinney consistently produced more accurate rate estimates and reduced incorrect inferences of rate variation. This method therefore provides a flexible framework to integrate gene tree heterogeneity into comparative methods and to produce reliable inferences of quantitative trait evolution, regardless of the source of discordance.