SDM: a fast distance-based approach for (super) tree building in phylogenomics

Area of Science:

• Phylogenomics
• Computational Biology
• Evolutionary Biology

Background:

• Phylogenomic studies utilize large datasets of homologous genes to construct evolutionary trees (phylogenies).
• Existing methods face challenges with computational speed and accurately inferring evolutionary distances due to gene rate heterogeneity.
• Distance-based methods are valuable for initial tree construction but require refinement for accuracy.

Purpose of the Study:

• To propose a novel, efficient method, the super distance matrix (SDM), for phylogenomic analyses.
• To address the limitations of estimating evolutionary distances from concatenated genes with varying evolutionary rates.
• To provide a method that is both fast for exploratory analyses and can improve accuracy in more computationally intensive approaches.

Main Methods:

• The super distance matrix (SDM) method combines evolutionary distances from individual genes into a single supermatrix.
• SDM deforms source distance matrices to align their topological information before averaging, minimizing a least-squares criterion.
• The method involves solving a sparse linear system, offering computational efficiency (O(naka) time complexity).

Main Results:

• SDM provides a relevant alternative to standard methods like matrix representation with parsimony (MRP), particularly when gene datasets have low overlap.
• Simulations demonstrate SDM's effectiveness in handling rate heterogeneity across genes.
• SDM successfully generated an accurate starting tree for maximum likelihood (ML) analysis, reducing computation time and enhancing topological accuracy.

Conclusions:

• The super distance matrix (SDM) method offers a computationally efficient and accurate approach for phylogenomic studies.
• SDM is particularly beneficial for large datasets and situations with missing data or low gene overlap.
• The method's ability to generate improved starting trees for ML analyses highlights its utility in advancing evolutionary inference.