GRAViTy-V2: a grounded viral taxonomy application

Area of Science:

• Virology
• Bioinformatics
• Genomics
• Taxonomy

Background:

• Accurate virus classification is crucial for understanding viral evolution and disease.
• Current methods rely on amino acid motifs for higher ranks and nucleotide identities for lower ranks.
• Existing frameworks require optimization for broader applicability and efficiency.

Purpose of the Study:

• To introduce Genome Relationships Applied to Viral Taxonomy Version 2 (GRAViTy-V2), an enhanced computational tool for virus classification.
• To evaluate GRAViTy-V2's performance against expert-curated viral taxonomies.
• To provide a general-purpose, optimized application for genomic virus classification.

Main Methods:

• Development of GRAViTy-V2, incorporating expanded features and optimizations.
• Analysis of 28 datasets based on ICTV 2022 taxonomy proposals.
• Comparison of GRAViTy-V2 output with human expert-curated classifications from the ICTV 2023 round.

Main Results:

• GRAViTy-V2 generated taxonomies equivalent to manual classifications at family, genus, and species levels.
• Discrepancies were primarily attributed to coding sequence annotation errors or incomplete genome data.
• Analysis times were efficient, with a median of 3.59 minutes across diverse datasets.

Conclusions:

• GRAViTy-V2 is a robust and efficient tool for automated viral taxonomic classification.
• The tool demonstrates high concordance with expert-driven classifications, supporting its use in routine taxonomy.
• Addressing data quality issues in sequence records will further enhance classification accuracy.