Generalized mixture models for molecular phylogenetic estimation.

Area of Science:

• Evolutionary biology
• Computational biology
• Genomics

Background:

• Multigene sequence data availability has surged, enabling phylogenomic analyses.
• Handling evolutionary process heterogeneity across genomes presents significant challenges in phylogenetic estimation.

Purpose of the Study:

• To develop a novel mixture model approach for phylogenomic analysis.
• To incorporate reversible jump Markov chain Monte Carlo (MCMC) estimation for flexible model selection.
• To allow for hard polytomies in phylogenetic tree inference.

Main Methods:

• Implemented a mixture model using reversible jump MCMC estimation.
• Allowed for distinct evolutionary models, including general time-reversible models and special cases.
• Expanded proposal mechanisms to include hard polytomies (zero-length internal branches).
• Applied the approach using the Crux software toolkit.

Main Results:

• Successfully demonstrated the feasibility of reversible jump MCMC on mixture models.
• Reanalyzed a 44-taxon mammalian dataset with 22 concatenated genes.
• Reproduced original results under identical assumptions.
• Inferred significantly lower bipartition support values for key bipartitions when allowing for polytomies and/or mixture models.

Conclusions:

• The developed mixture model approach effectively handles evolutionary heterogeneity in phylogenomics.
• Accounting for polytomies and employing data-driven mixture models can lead to more conservative estimates of phylogenetic support.
• This method provides a more robust framework for inferring evolutionary relationships from complex genomic data.