Phylogenetic tree statistics: A systematic overview using the new R package 'treestats'.

Area of Science:

• Phylogenetics
• Computational Biology
• Evolutionary Biology

Background:

• Phylogenetic trees offer insights into diversification but are high-dimensional and difficult to compare.
• Numerous summary statistics exist, but their interrelationships and information coverage are poorly understood.
• Existing studies often focus on extreme tree balance (fully balanced vs. imbalanced), neglecting intermediate states.

Purpose of the Study:

• To investigate the correlations among diverse phylogenetic summary statistics.
• To assess the information content of summary statistics in explaining variation across phylogenies.
• To explore tree balance properties using a new algorithm and identify key balance statistics.

Main Methods:

• Developed the R package 'treestats' for efficient computation of 70 summary statistics.
• Analyzed correlations between statistics using empirical and simulated phylogenetic trees.
• Introduced an algorithm to generate trees along a balance gradient to study statistic variation.

Main Results:

• Most summary statistics correlate with tree size, posing challenges for size correction without knowing the generating model.
• Identified at least three major clusters of correlated statistics based on topology or branching times.
• Tree balance statistics exhibit complex, non-linear, and non-monotonic relationships with tree balance.

Conclusions:

• The 'treestats' package provides a valuable tool for phylogenetic analysis.
• Tree balance is a more intricate characteristic than previously assumed.
• Recommendations are provided for selecting appropriate summary statistics for phylogenetic tree comparison and analysis.