Mean and Variance of Phylogenetic Trees

Area of Science:

• Phylogenetics
• Computational Biology
• Statistics

Background:

• Phylogenetic trees represent evolutionary relationships.
• Summarizing and exploring the diversity within a set of phylogenetic trees is crucial for understanding evolutionary processes.
• Existing methods for summarizing tree diversity have limitations in terms of theoretical justification, robustness, and efficiency.

Purpose of the Study:

• To introduce and evaluate the Fréchet mean and variance as novel summary statistics for phylogenetic trees within the Billera-Holmes-Vogtmann (BHV) treespace.
• To compare the performance of the Fréchet mean and variance against established methods for summarizing tree diversity.
• To establish a foundation for advanced statistical methods in phylogenetics, including hypothesis testing and outlier detection.

Main Methods:

• Utilizing the Fréchet mean and variance within the Billera-Holmes-Vogtmann (BHV) treespace.
• Comparing the Fréchet mean to majority-rule consensus trees regarding binary properties.
• Assessing the speed and precision of the Fréchet variance against common variance measures.

Main Results:

• The Fréchet mean is comparable to other summary methods and more likely to be binary than majority-rule consensus trees.
• The Fréchet variance demonstrates superior speed and precision compared to commonly used variance measures.
• The Fréchet mean and variance are theoretically justified, robust, and efficiently estimable.

Conclusions:

• The Fréchet mean and variance provide a theoretically sound and robust framework for summarizing and exploring phylogenetic tree diversity.
• These methods offer advantages in efficiency and precision over existing techniques.
• The Fréchet mean and variance lay the groundwork for developing advanced statistical applications in phylogenetics, such as hypothesis testing and outlier detection.