CheckV

CheckV is a software tool designed to assess the quality of viral genomes recovered from metagenomic data. It provides standardized quantitative metrics that estimate genome completeness and contamination, enabling researchers to compare genome quality across studies and report results consistently. CheckV is used for sequences assembled from environmental samples, including complete genomes, partial contigs, and prophages integrated in host genomes.

Methodology and features: CheckV uses a curated database of complete viral genomes and a gene-content based approach to infer how much of a genome is present in a given sequence. The tool classifies sequences as viral, distinguishes them from potential host or non-viral contaminants, and estimates genome length, circularization, and terminal repeats where applicable. It computes completeness and contamination percentages and assigns a quality tier to each sequence, such as complete, high quality, medium quality, or low quality. Additional metrics reported include the number of viral hallmark genes and indicators of potential host contamination. The workflow is designed to work with single contigs as well as larger metagenomic assemblies.

Implementation and availability: CheckV is open-source and distributed as a command-line tool. It accepts standard FASTA inputs and can process large datasets common in metagenomic projects. Outputs include machine-readable reports (for example JSON) and human-readable summaries, facilitating integration into existing analysis pipelines and reporting workflows.

History and context: The concept was introduced to standardize quality assessment of viral genomes recovered from metagenomes. Since its initial release, CheckV has undergone updates that expand its reference database with uncultivated viral genomes and improve accuracy and scalability. It is hosted on a public repository and maintained by a community of developers and researchers.

Limitations: Quality estimates depend on the breadth and representativeness of the reference database. Novel or highly divergent viral lineages may yield uncertain completeness or contamination estimates, so results should be interpreted in the context of complementary analyses.