VICAST: An Integrated Toolkit for Viral Genome Annotation Curation and Low-Frequency Variant Analysis in Passage Studies

Abstract

Cultured virus passage studies are fundamental to understanding viral evolution, attenuation, and host adaptation, yet analyzing genomic changes across passages requires both accurate functional annotation of viral genomes and sensitive detection of low-frequency variants. Existing tools address these needs separately: automated annotation pipelines such as VADR and VIGOR4 perform well for well-characterized virus families but struggle with poorly-annotated or novel genomes, while variant calling pipelines designed for clinical diagnostics focus on consensus sequences rather than the low-frequency variants (3-50% frequency) that are biologically meaningful in passage studies. Here we present VICAST (Viral Cultured-virus Annotation and SnpEff Toolkit), an integrated software suite that combines semi-automated genome annotation with manual curation checkpoints and low-frequency variant calling optimized for viral populations. VICAST provides four annotation pathways to accommodate diverse genome annotation quality, including polyproteins, unannotated and multi-segmented genomes. It integrates with SnpEff for functional variant annotation and includes a BAM-level read co-occurrence module for haplotype validation. We validated VICAST using publicly available datasets from three virus families representing distinct analytical challenges: SARS-CoV-2 for polyprotein cleavage-aware annotation, Dengue virus 2 for standard flavivirus annotation and low-frequency variant detection, and Influenza A H1N1 for multi-segmented genome handling. Additionally, VICASTs annotation curation workflow has produced validated annotations not available from NCBI, including protein-level annotations for Chikungunya virus (NC_004162.2). These curated annotations have been incorporated into a custom SnpEff database distributed with VICAST, enabling immediate functional variant annotation for Chikungunya without requiring users to build the database from scratch. Benchmark comparisons with VADR demonstrate VICASTs advantages for passage study workflows, including 5.6-8.1 times faster processing and integrated contamination screening. VICAST is freely available at https://github.com/mihinduk/VICAST and distributed as both Docker containers and conda-based installations.