castor

castor performs high-performance comparative phylogenetic analyses on large phylogenetic trees (e.g., the Open Tree of Life with over 1.4 million metazoan and 300,000 bacterial taxa) to enable efficient tree manipulations, ancestral state reconstruction, and trait-evolution inference.

Key Features:

• Efficient tree manipulations: Performs pruning, rerooting, calculation of most-recent common ancestors (MRCA), root-to-tip distances, and pairwise distances on large phylogenies.
• Phylogenetic signal and trait analysis: Computes phylogenetic signal metrics and mean trait depths to assess trait conservatism across species.
• Ancestral state reconstruction and hidden character prediction: Reconstructs ancestral states and predicts hidden discrete character states across nodes.
• Performance optimization: Implements optimized algorithms that achieve approximately 100–1,000× speedups on trees comprising millions of tips compared to existing tools.

Scientific Applications:

• Large-scale trait evolution studies: Enables analysis of trait distribution and evolution across very large phylogenies such as those from the Open Tree of Life.
• Ancestral biodiversity reconstruction: Supports inference of ancestral character states to reconstruct historical biodiversity patterns.
• Comparative phylogenetics at scale: Facilitates comparative analyses and hypothesis testing on datasets containing millions of taxa.

Methodology:

Implemented as an R package that leverages optimized algorithms for efficient tree manipulation, ancestral state reconstruction, and phylogenetic signal and trait computations.