Phylogenetic invariants for genome rearrangements

Area of Science:

• Computational Biology
• Bioinformatics
• Evolutionary Biology

Background:

• Combinatorial optimization problems are central to calculating genomic edit distances and phylogenetic inference.
• Existing tree-building methods can incur high computational costs and suffer from the "long branches attract" artifact.

Purpose of the Study:

• To explore probabilization of genome rearrangement models to mitigate computational expense and artifacts.
• To develop a novel methodology for phylogenetic inference based on branch-length invariants.

Main Methods:

• Probabilistic characterization of gene adjacency set evolution for reversals on unsigned and signed circular genomes.
• Development of linear invariants using concepts from the theory of invariants and an extended Jukes-Cantor semigroup.
• Application of invariants to analyze mitochondrial genomes of invertebrate animals.

Main Results:

• A complete set of linear invariants was derived for unsigned reversals and a mixed rearrangement model for signed genomes.
• The methodology provides an alternative to computationally intensive methods and addresses the "long branches attract" artifact.
• The derived invariants were successfully used to relate mitochondrial genomes across various invertebrate species.

Conclusions:

• The developed branch-length invariant methodology offers a computationally efficient and accurate approach to phylogenetic inference based on gene order changes.
• This work advances the understanding of genome evolution and provides a powerful tool for comparative genomics.
• The study highlights the utility of invariant theory in addressing complex problems in bioinformatics and evolutionary biology.