Efficient Recycled Algorithms for Quantitative Trait Models on Phylogenies.

Area of Science:

• Evolutionary Biology
• Phylogenetics
• Quantitative Genetics

Background:

• Calculating likelihoods of phenotypic traits on phylogenetic trees is crucial for understanding evolutionary processes.
• Existing methods, often relying on Monte Carlo simulations, can be computationally intensive and limited in scope.
• There is a need for efficient and flexible computational frameworks that can handle complex trait models and uncertainty.

Purpose of the Study:

• To present an efficient and flexible computational method for calculating trait likelihoods on phylogenies.
• To integrate Felsenstein's pruning algorithm with numerical quadrature for enhanced computational performance.
• To adapt the framework for non-Gaussian models and incorporate uncertainty in observed trait data.

Main Methods:

• Developed a novel computational framework combining Felsenstein's discrete character pruning algorithm with numerical quadrature.
• Implemented efficient algorithms for likelihood calculation and ancestral state reconstruction.
• Applied the methods to Wright's threshold model for discrete phenotypic traits.

Main Results:

• The proposed method provides efficient and flexible likelihood computations for phenotypic traits on phylogenies.
• The framework avoids computationally expensive Monte Carlo simulations.
• Demonstrated successful application to a dataset of extrafloral nectary traits in 839 Fabales species.

Conclusions:

• The new computational approach offers a significant advancement for phylogenetic trait evolution studies.
• The method's flexibility accommodates various models and data uncertainties.
• This framework facilitates more robust ancestral state reconstruction and evolutionary analysis.