Detecting Adaptive Evolution in Phylogenetic Comparative Analysis Using the Ornstein-Uhlenbeck Model.

Area of Science:

• Evolutionary biology
• Phylogenetics
• Quantitative genetics

Background:

• Phylogenetic comparative analysis integrates phylogenetic and phenotypic data to infer evolutionary processes.
• Sophisticated models are increasingly used for detailed evolutionary hypotheses, but their statistical performance is under-examined.
• Ornstein-Uhlenbeck processes are commonly used models for phenotypic evolution.

Purpose of the Study:

• To systematically evaluate the statistical performance of Ornstein-Uhlenbeck models in phylogenetic comparative analysis.
• To identify the conditions under which these models perform well or break down.
• To provide guidance for users on the strengths and weaknesses of these analytical methods.

Main Methods:

• Conducted an extensive simulation study.
• Focused on models of phenotypic evolution using Ornstein-Uhlenbeck processes.
• Analyzed the impact of discriminability ratio, signal-to-noise ratio, and taxon number on model performance.

Main Results:

• Identified three key determinants of model performance: discriminability ratio, signal-to-noise ratio, and number of taxa.
• Found high model-selection power even with small tree sizes, contrary to previous assumptions.
• Observed that model parameters are often difficult to estimate accurately due to limited data information.
• Demonstrated that accurate model selection is frequently possible even when parameters are weakly identified.

Conclusions:

• Model performance in phylogenetic comparative analysis is governed by specific ratios and sample size.
• Despite challenges in parameter estimation, accurate model selection is achievable under certain conditions.
• Findings have implications for the application and development of more complex phylogenetic models.