Estimation of evolutionary parameters with phylogenetic trees

Area of Science:

• Computational Biology
• Evolutionary Biology
• Bioinformatics

Background:

• Accurate phylogenetic analysis of nucleotide sequences relies heavily on the chosen evolutionary model.
• Maximum likelihood (ML) methods are widely used but require robust parameter estimation and variance assessment.

Purpose of the Study:

• To develop and evaluate novel methods for simultaneously estimating phylogenetic tree topology and evolutionary model parameters.
• To improve the estimation of standard errors for these parameters, crucial for assessing phylogenetic uncertainty.

Main Methods:

• Proposed a new conditional variance estimate using Monte Carlo sampling (MCS) to approximate expected information.
• Compared the MCS-based variance estimate with the traditional observed information method via simulations.
• Developed a bootstrapping approach combined with MCS for estimating unconditional standard errors when tree topology and parameters are estimated simultaneously.

Main Results:

• The MCS method provides a reliable estimate of conditional variance for evolutionary model parameters.
• Simulations demonstrated the performance of the MCS method under various conditions.
• The bootstrapping and MCS combination effectively estimates unconditional standard errors.

Conclusions:

• The proposed MCS method enhances variance estimation in phylogenetic analysis.
• This approach offers a robust way to assess uncertainty in evolutionary model parameters and tree topology.
• The methods were successfully applied to a real dataset of papillomavirus sequences, demonstrating practical utility.