Reducing haystacks to needles - ViralClust: A Nextflow pipeline to cluster viral sequences.

ViralClust reduces viral genome datasets by approximately 95% across six virus types while maintaining genetic diversity and mitigating sampling biases.

• Why it matters: Handling large viral genome datasets poses challenges for downstream analysis due to computational limits and sampling biases from outbreaks. Effective clustering can address these issues, but selecting appropriate strategies is complex and context-dependent.
• What they did: ViralClust is a modular Nextflow pipeline that integrates five clustering algorithms—CD-HIT-EST, SUMACLUST, VSEARCH, MMSeqs2, and HDBSCAN—to enable bias-aware, scalable representative selection from datasets ranging from 632 to 156,586 sequences and genome lengths from 890 to 197,185 nucleotides.
• The result: Clustering with ViralClust preserves genetic diversity across taxonomic levels and effectively reduces dataset size, facilitating efficient, reproducible downstream analyses and allowing users to empirically evaluate and select optimal clustering strategies.