Informative Regions In Viral Genomes.

Area of Science:

• Virology
• Bioinformatics
• Microbial Ecology

Background:

• Viruses, especially bacteriophages, are crucial in microbial ecosystems but largely uncharacterized, with most viral sequences lacking annotation.
• Existing annotation methods struggle to identify the function and taxonomic placement of a vast majority of viral genetic material.

Purpose of the Study:

• To address the significant gap in viral genome annotation and classification.
• To develop a robust system for viral taxonomy using conserved viral protein domains.
• To identify reliable viral markers for taxonomic classification.

Main Methods:

• Compilation and analysis of non-redundant viral diversity from public databases.
• Prediction of proteins and identification of protein domains within viral genomes.
• Clustering of domains and unannotated regions into 31,150 Eukaryotic Viruses and Phages Orthologous Groups (ViPhOGs) using cogSoft.
• Application of a random forest model for taxonomic classification based on ViPhOG presence/absence.

Main Results:

• Successfully established a comprehensive set of 31,150 ViPhOGs.
• Demonstrated that ViPhOGs are significantly associated with viral taxonomy.
• Identified 1457 specific ViPhOGs as key markers for classifying viruses at the order, family, and genus levels.
• Significantly improved the annotation and classification of previously uncharacterized viral sequences.

Conclusions:

• ViPhOGs represent a powerful new resource for exploring viral genetic dark matter and understanding viral diversity.
• The identified ViPhOG markers provide a foundation for refining viral taxonomy and annotation pipelines.
• This approach enhances our ability to classify and understand the roles of viruses in various ecosystems.