A vectorial tree distance measure.

Area of Science:

• Computational Biology
• Bioinformatics
• Data Science

Background:

• Tree distance measures are crucial for analyzing complex hierarchical data structures.
• Existing methods may not fully capture nuanced differences between tree topologies.
• Phylogenetic tree analysis requires robust and informative distance metrics.

Purpose of the Study:

• To introduce a novel vectorial distance measure for trees, termed Tree-Alignment (T-Alignment).
• To develop a recursive algorithm for calculating tree distances based on hierarchical alignment.
• To evaluate the efficacy of the vectorial tree distance (VTD) in clustering and classifying phylogenetic trees.

Main Methods:

• A recursive T-alignment algorithm is proposed, aligning trees from root-branches outward.
• A minimal T-alignment under lexicographic order defines a distance vector based on branch differences at each level.
• The VTD algorithm was compared against existing tree distance measures using simulated phylogenetic trees from TreeSimGM.
• The algorithm was tested for classification using real-world phylogenetic trees from fungi.

Main Results:

• The VTD algorithm successfully differentiates between symmetric and asymmetric trees.
• The VTD algorithm effectively distinguishes between hierarchical and non-hierarchical tree structures.
• The VTD algorithm demonstrated capability in classifying real-world phylogenetic trees from mushrooms and mildews.

Conclusions:

• The developed vectorial tree distance provides a novel and effective approach for quantifying tree dissimilarities.
• The VTD algorithm shows promise for applications in phylogenetic tree analysis, clustering, and classification.
• The algorithm's ability to separate distinct tree types highlights its utility in biological data analysis.