Simulation-based likelihood approach for evolutionary models of phenotypic traits on phylogeny.

Area of Science:

• Evolutionary biology
• Phylogenetics
• Population genetics

Background:

• Phylogenetic comparative methods (PCMs) traditionally use Brownian motion and Ornstein-Uhlenbeck models.
• Complex evolutionary models, like branch-specific directional selection, are challenging for PCMs due to computational difficulties.
• Existing methods struggle with accurate likelihood and parameter estimation for intricate evolutionary scenarios.

Purpose of the Study:

• To develop a flexible and comprehensive framework for estimating evolutionary parameters using simulation-based likelihood computations.
• To overcome the limitations of analytical calculations in complex phylogenetic comparative models.
• To integrate population genetics principles into PCMs for more robust evolutionary analyses.

Main Methods:

• Introduced a population genetics framework into phylogenetic comparative methods.
• Utilized simulation-based likelihood computations, avoiding the need for analytical solutions.
• Developed a method applicable to any evolutionary model that allows for simulation.

Main Results:

• Presented a novel framework for estimating evolutionary parameters that accommodates diverse evolutionary modes and intraspecific variation.
• Demonstrated the method's capability to evaluate full likelihood and estimate ancestral traits.
• Successfully applied the framework to analyze primate brain size evolution.

Conclusions:

• The new simulation-based PCM framework offers a flexible and comprehensive approach to studying phenotypic evolution.
• This method enhances the incorporation of complex evolutionary models and intraspecific variation into phylogenetic analyses.
• The framework's compatibility with computational methods like approximate Bayesian computation (ABC) will advance the field of evolutionary biology.