Searching for virus phylotypes.

Area of Science:

• Computational Biology
• Virology
• Bioinformatics

Background:

• Large-scale phylogenies are crucial for studying virus evolution, epidemic origins, transmission, and drug resistance.
• Existing tools lack the ability to efficiently inspect phylogenies and integrate extrinsic traits for targeted strain group identification and surveillance.

Purpose of the Study:

• To develop a novel computational method for identifying biologically relevant strain groups (phylotypes) from annotated phylogenies.
• To provide a tool that aids in the exploration and analysis of large viral datasets for epidemiological and evolutionary insights.

Main Methods:

• The phylotype method combines ancestral trait reconstruction (parsimony) with criteria for tree shape, diversity, and separation.
• Statistical significance of identified phylotypes is assessed using a shuffling procedure.
• Algorithms exhibit linear time complexity, enabling rapid analysis of large datasets.

Main Results:

• The method successfully identifies phylotypes, defined as taxa with close phylogenetic relationships and shared trait values.
• Analysis of HIV-1 subtype C data (>3000 strains) demonstrated the ability to recover known clusters and transmission routes, and to detect novel ones.
• The approach offers low computing times, typically minutes for large datasets.

Conclusions:

• The developed phylotype method provides an efficient and statistically robust approach for analyzing complex viral phylogenies.
• The interactive web interface (www.phylotype.org) facilitates exploratory data analysis and surveillance of viral strains.
• This tool has significant implications for public health, aiding in the tracking and management of infectious diseases.