VADR

VADR validates and annotates viral nucleotide sequences submitted to GenBank using curated RefSeq-derived models to perform sequence classification, feature mapping, protein validation, and deterministic alerting.

Key Features:

• Validation and classification: Uses curated RefSeqs and Hidden Markov Models to classify input nucleotide sequences by similarity to reference sequences.
• Feature mapping: Maps features from the closest matching RefSeq using nucleotide alignment against a covariance model.
• Protein validation: Validates predicted proteins via nucleotide-to-protein alignments using BLAST.
• Alert system: Identifies 43 distinct alert types that provide deterministic feedback on unexpected sequence characteristics.
• Integration with GenBank processing: Integrates into GenBank's submission processing to enable automatic acceptance and annotation of submissions that pass all tests.
• Non-influenza viral focus: Targets validation and annotation for non-influenza viral sequence submissions to GenBank.

Scientific Applications:

• Norovirus submissions: Employed for Norovirus sequence submissions beginning in May 2018.
• Dengue virus submissions: Applied to Dengue virus sequence submissions beginning in January 2019.
• SARS-CoV-2 validation: Used to validate SARS-CoV-2 sequences by March 2020.
• High-throughput viral submission processing: Applied to processing and annotating large volumes of non-influenza viral submissions to GenBank.

Methodology:

VADR uses curated RefSeqs, Hidden Markov Models for sequence classification, nucleotide alignment to covariance models for feature mapping, nucleotide-to-protein BLAST alignments for protein validation, and reports 43 alert types.