kdetrees

kdetrees estimates distributions of phylogenetic trees and identifies trees that significantly deviate from a sample to reveal gene trees with distinct evolutionary histories.

Key Features:

• Non-Parametric Estimation: Employs a non-parametric approach to estimate the distribution of phylogenetic trees without assuming a specific parametric form.
• Outlier Detection: Identifies phylogenetic trees that are significantly different from others in the sample, enabling detection of events such as horizontal gene transfer and gene duplication.
• Computational Efficiency: Achieves quadratic time complexity relative to the number of trees, enabling analysis of larger tree sets.
• Accuracy and Reliability: Demonstrates high classification accuracy in simulation studies while maintaining improved computational efficiency.

Scientific Applications:

• Apicomplexa Genes: Detected unreliable sequence alignments and identified a gene suspected of horizontal gene transfer in Apicomplexa datasets.
• Epichloë Genes: Identified instances of paralogy in Epichloë gene datasets, revealing complex evolutionary relationships.

Methodology:

Implemented as an R package, kdetrees uses a non-parametric estimator of phylogenetic tree distributions and operates with quadratic time complexity relative to the number of trees.