phylo2vec: a library for vector-based phylogenetic tree manipulation

2025-06-24

AI Summary

Traditional Newick format incurs substantial memory overhead and inefficient operations when handling large-scale phylogenetic trees, hindering downstream analyses such as machine learning. To address this, we propose an integer-vector-based representation for binary phylogenetic trees, establishing a bijective mapping between tree topology and an (n−1)-dimensional integer vector—enabling lossless, compact encoding. Our method significantly improves storage efficiency, tree sampling speed, and pairwise tree comparison performance. The core library is implemented in Rust for memory safety and performance, with Python and R bindings enabling cross-platform, high-efficiency memory management. Experiments demonstrate substantial acceleration in large-scale genomic and linguistic evolutionary analyses; the framework has been integrated into multiple bioinformatics workflows. This work provides a scalable, foundational representation for phylogeny-driven AI modeling.

Technology Category

• Machine Learning: Representation Learning
• Knowledge Representation and Reasoning: Computational Complexity of Reasoning
• Cognitive Modeling & Cognitive Systems: Symbolic Representations

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding

Abstract

Phylogenetics is a fundamental component of many analysis frameworks in biology as well as linguistics to study the evolutionary relationships of different entities. Recently, the advent of large-scale genomics and the SARS-CoV-2 pandemic has underscored the necessity for phylogenetic software to handle large datasets of genomes or phylogenetic trees. While significant efforts have focused on scaling optimisation algorithms, visualization, and lineage identification, an emerging body of research has been dedicated to efficient representations of data for genomes and phylogenetic trees such as phylo2vec. Compared to traditional tree representations such as the Newick format, which represents trees using strings of nested parentheses, modern representations of phylogenetic trees utilize integer vectors to define the tree topology traversal. This approach offers several advantages, including easier manipulability, increased memory efficiency, and applicability to downstream tasks such as machine learning. Here, we present the latest release of phylo2vec (or Phylo2Vec), a high-performance software package for encoding, manipulating, and analysing binary phylogenetic trees. At its core, the package is based on the phylo2vec representation of binary trees, which defines a bijection from any tree topology with $n$ leaves into an integer vector of size $n-1$. Compared to the traditional Newick format, phylo2vec is designed to enable fast sampling and comparison of binary trees. This release features a core implementation in Rust, providing significant performance improvements and memory efficiency, while remaining available in Python (superseding the release described in the original paper) and R via dedicated wrappers, making it accessible to a broad audience in the bioinformatics community.

Problem

• Efficient representation of phylogenetic tree data
• Handling large datasets for genomes and trees
• Enabling fast tree sampling and comparison

Innovation

• Uses integer vectors for tree topology
• Enables fast sampling and comparison
• Implements Rust for performance efficiency