A model for phylogenetic inference using structural and chemical covariates.

Area of Science:

• Molecular Evolution
• Genomics
• Bioinformatics

Background:

• Evolutionary change in DNA sequences is typically modeled assuming homogeneity.
• Understanding variation in evolutionary rates across different nucleotide sites is crucial for accurate phylogenetic inference.

Purpose of the Study:

• To investigate if evolutionary change in DNA sequence data is homogeneous across different classes of base pairs.
• To identify distinct classes of nucleotide sites that exhibit differential evolutionary behavior.

Main Methods:

• Obtained DNA sequences for eight protein-coding mitochondrial genes from 38 vertebrate taxa.
• Classified nucleotide sites by codon position, genetic code degeneracy, and hydrophobicity.
• Estimated evolutionary transition matrices using parsimony on a phylogenetic tree and analyzed with canonical variates analysis.

Main Results:

• Identified five distinct clusters of transition matrices, primarily defined by codon position and degeneracy.
• This clustering pattern was consistent across all investigated mitochondrial genes.
• A stochastic model incorporating covariates significantly explained more sequence variation than simpler models.

Conclusions:

• Confirms that different classes of base pairs evolve at different rates.
• Suggests that incorporating covariate information into models of sequence evolution can improve accuracy.
• Highlights a common underlying process in the molecular evolution of the mitochondrial genome.