Graph Neural Networks approximate phylogenetic tree distances

Researchers have developed a Graph Neural Network (GNN) to approximate the Subtree Prune and Regraft (SPR) distance between phylogenetic trees. This method aims to provide a near-constant time computation after training, addressing the NP-hard nature of exact SPR distance calculations for large datasets. The study released a dataset of 864 phylogenetic trees and validated a heuristic that closely correlates with exact SPR distances, achieving approximately 87-90% variance explanation on in-distribution data. AI

IMPACT This research could enable faster analysis of large-scale phylogenetic data, potentially accelerating research in areas like epidemic dynamics.

RANK_REASON This is a research paper detailing a novel method for approximating a complex biological computation using machine learning. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• Graph Neural Network
• neighbor joining
• phangorn
• Siamese Graph Isomorphism Network
• UPGMA

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →