VirClust

VirClust performs hierarchical clustering and protein-level analyses to support viral taxonomy by computing intergenomic distances from shared protein cluster content and identifying core proteins across viral groups within the megataxonomy of viruses spanning five realms defined by hallmark genes.

Key Features:

• Hierarchical clustering: Computes hierarchical trees of viral genomes using intergenomic distances derived from protein cluster content to organize viruses by genetic similarity.
• Core gene detection: Identifies core proteins shared among defined virus groups to highlight conserved functional elements.
• Protein annotation via clustering: Groups proteins into clusters using BLASTP sequence similarity and Hidden Markov Models (HMMs) to detect both closely and distantly related proteins.
• Integrated visualization: Displays the hierarchical clustering tree alongside the distribution of protein content to relate genomic features to clustering patterns.
• Flexible taxonomic resolution: Generates hierarchical trees that can be split into genome clusters corresponding to different taxonomic ranks by applying alternative intergenomic distance thresholds.

Scientific Applications:

• Viral taxonomy and classification: Organizes viruses into hierarchical clusters that can inform taxonomic ranks within the megataxonomy of viruses.
• Prokaryotic virus genomics: Supports classification and comparative analysis of prokaryotic viruses based on shared protein content.
• Core gene analysis and functional inference: Identifies conserved proteins to aid interpretation of essential viral functions and evolutionary conservation.
• Comparative genomics and evolutionary studies: Reveals genomic features driving clustering to inform studies of viral diversity and evolution and to aid downstream development of targeted antiviral strategies.

Methodology:

VirClust derives intergenomic distances from protein cluster content, performs hierarchical clustering, detects core proteins shared among virus groups, clusters proteins using BLASTP and HMMs, splits trees into genome clusters using distance thresholds, and visualizes trees with protein content distributions.