Branch Length Transforms using Optimal Tree Metric Matching

Area of Science:

• Phylogenetics
• Computational Biology
• Evolutionary Genomics

Background:

• Phylogenomic analyses require comparing evolutionary trees, but branch length heterogeneity poses challenges.
• Existing methods struggle with diverse branch length units and evolutionary rate variations across genomes.
• Accurate branch lengths are crucial for downstream applications, yet tools for weighted tree comparison are limited.

Purpose of the Study:

• To develop a computational method for matching phylogenetic trees based on branch lengths.
• To address the challenge of transforming gene tree branch lengths to align with a reference species tree.
• To introduce Topology-Constrained Metric Matching (TCMM) for phylogenomic data integration.

Main Methods:

• Defined Topology-Constrained Metric Matching (TCMM) problems to adjust query tree branch lengths using a reference tree.
• Developed an efficient solution using linear algebra and dynamic programming.
• Applied TCMM to embed gene tree leaves in Euclidean space and summarize gene tree branch lengths onto the species tree.

Main Results:

• TCMM problems are solvable efficiently through a linear algebraic formulation and dynamic programming.
• The method successfully identifies potential outliers in gene tree estimation.
• Summarizing gene tree branch lengths onto the species tree improves accuracy of existing methods.

Conclusions:

• TCMM provides an efficient framework for comparing and combining phylogenetic trees based on branch lengths.
• This approach enhances the accuracy of phylogenomic analyses, including outlier detection and branch length summarization.
• The method offers significant improvements at a limited computational cost.