Meeting Abstract
Phylogenetic comparative methods (PCMs) are used to make comparisons among organisms while considering their shared evolutionary histories. Brownian motion (BM) and the Ornstein-Uhlenbeck (OU) process are the main models of continuous trait evolution utilized in the context of PCMs. However, because PCMs can be applied to a wide range of organisms occupying a number of environments and various spans of evolutionary time, these models may not be able to fully capture all observed patterns of trait evolution. How valid are your biological conclusions if characters of interest do not conform to either mode of evolution? Here we propose a new parametric bootstrapping method using phylogenetic contrast to assess the fit of either BM or OU models to continuous traits. Phylogenetic contrasts were calculated for each trait of interest along the tree and used as a test statistic. We performed simulations using estimated parameters after fitting each trait to either BM or OU. We used effect size to compare phylogenetic contrast between empirical and simulated datasets. We applied this method to data on seven structural, temporal and frequency characteristics of birdsong from the genus Catharus. Surprisingly, comparisons indicate that neither BM or OU models properly characterizes the mode of evolution, as evidenced by high effect size values. Additionally, this method allows for the identification of specific outlier nodes that contrast with what we would expect if the tree where evolving under BM or OU models. To verify our method, we performed a small simulation study to demonstrate that it has a proper Type I error rate. Ultimately, this study highlights the importance of carefully considering other possible modes of evolution.
