RNA-based phylogenetic methods: application to mammalian mitochondrial RNA sequences.

Area of Science:

• Evolutionary Biology
• Bioinformatics
• Genomics

Background:

• Phylogenetic inference for RNA sequences with conserved secondary structure requires specialized evolutionary models.
• Compensatory substitutions in paired RNA regions introduce correlations that impact likelihood calculations.
• Existing models may not fully account for these correlations, potentially biasing phylogenetic results.

Purpose of the Study:

• To construct accurate phylogenetic trees for mammalian mitochondrial genomes using RNA sequences.
• To implement and evaluate evolutionary models that account for correlations in paired RNA regions.
• To resolve problematic phylogenetic placements within mammalian orders.

Main Methods:

• Utilized the PHASE software package for phylogenetic tree construction.
• Employed two simultaneous evolutionary models: a paired-site model for paired regions and a single-site model for unpaired regions.
• Applied Bayesian phylogenetic methods with a Markov chain Monte Carlo algorithm to determine the most probable trees and clade posterior probabilities.

Main Results:

• Achieved well-resolved phylogenetic trees for most major branches of the mammalian evolutionary tree.
• Results support the established four supra-ordinal clades of mammals, consistent with studies using nuclear genes.
• Successfully placed previously problematic groups, such as hedgehogs and murid rodents, in their expected ordinal positions.

Conclusions:

• The developed evolutionary models and gene selection provide a more reliable phylogenetic inference for mammalian mitochondrial genomes.
• The approach mitigates biases associated with base composition variation observed in previous studies.
• This method enhances the accuracy of evolutionary relationships, particularly for challenging taxa.