Comparing virus classification using genomic materials according to different taxonomic levels

Area of Science:

• Virology
• Bioinformatics
• Genomics

Background:

• Traditional virus classification relies on morphology, which can be subjective and limited.
• Genomic data (DNA, protein sequences) offer objective markers for classification.
• Identifying conserved regions within genomic data can enhance classification accuracy.

Purpose of the Study:

• To compare the effectiveness of DNA sequences, protein sequences, and region names for virus classification.
• To introduce a novel, molecular biology-based approach to virus classification.
• To evaluate classification performance across different taxonomic levels (orders, families, genera).

Main Methods:

• Utilized three types of genomic data: DNA sequences, protein sequences, and region (domain) names.
• Created datasets for 6 orders, 42 families, and 33 genera, ensuring at least 10 instances per class.
• Selected instances containing at least one region name for robust analysis.

Main Results:

• The classification approach using region names achieved superior accuracy: 99.9% for orders, 97.3% for families, and 99.0% for genera.
• Region-based classification outperformed methods relying solely on DNA or protein sequences.
• Exhaustive experiments validated the efficacy of different genomic materials for virus classification.

Conclusions:

• Region names derived from genomic data provide a highly accurate and robust method for virus classification.
• This molecular approach offers a significant advancement over traditional morphology-based methods.
• The findings support the use of genomic region analysis for improved viral taxonomy and understanding.