Site-heterogeneous mutation-selection models within the PhyloBayes-MPI package.

Area of Science:

• Evolutionary biology
• Computational biology
• Genomics

Background:

• Codon substitution models are increasingly important for biological applications.
• Previous models faced computational challenges, leading to simplifications and limitations with large datasets.

Purpose of the Study:

• To develop a scalable, site-heterogeneous codon model using a Bayesian approach.
• To address computational demands and improve accuracy in evolutionary inference.

Main Methods:

• Implemented a message-passing-interface-based Bayesian framework for codon models.
• Jointly inferred nucleotide-level mutational parameters, tree branch lengths, and amino acid-level variation using a Dirichlet process.
• Integrated the models into the PhyloBayes-MPI package.

Main Results:

• Demonstrated a scalable implementation of site-heterogeneous codon models.
• Successfully estimated selection coefficient distributions from influenza PB2 gene sequences.
• Highlighted the capability for site-specific evolutionary characterization.

Conclusions:

• The developed software offers a powerful tool for evolutionary inferences.
• Enables more robust analyses of large-scale sequence data.
• Opens avenues for future applications in evolutionary studies.