Two stationary nonhomogeneous Markov models of nucleotide sequence evolution

Area of Science:

• Evolutionary Biology
• Phylogenetics
• Computational Biology

Background:

• The general Markov model (GMM) is a foundational tool for nucleotide substitution analysis.
• Existing models like the general time-reversible (GTR) model impose assumptions of stationarity, reversibility, and homogeneity.
• Compositional homogeneity in phylogenetic datasets is crucial for applying stationary models.

Purpose of the Study:

• To develop and evaluate new stationary and nonhomogeneous models for nucleotide substitution.
• To test the assumptions of reversibility and homogeneity within stationary evolutionary processes.
• To improve phylogenetic inference by offering more flexible evolutionary models.

Main Methods:

• Proposed two novel stationary, nonhomogeneous models, one reversible and one not.
• Integrated these models with the GTR model to create a nested set for assumption testing.
• Extended models to include invariable sites and applied them to a hominoid dataset.

Main Results:

• The proposed nonhomogeneous models provided a superior fit to the hominoid data compared to the GTR model within the stationary class.
• The study demonstrated the utility of the nested model set for evaluating evolutionary process assumptions.
• Non-stationary models, while not fully explored, showed potential for even better data fit.

Conclusions:

• Stationary, nonhomogeneous models offer significant improvements over traditional models like GTR for certain datasets.
• Model selection is critical for accurate phylogenetic reconstruction, especially when evolutionary processes vary.
• Further research is needed to develop methods for handling complex nonstationary evolutionary models.