VariSNP 2 database for benign variations from dbSNP

Abstract

To develop and evaluate methods for predicting the effects of genetic variations, robust benchmark datasets are essential. While some existing datasets serve this purpose, there has been a notable lack of newer benchmark sets specifically for benign variants. The VariSNP 2 database addresses this gap by selecting subsets from dbSNP and filtering them against disease-related variants in the ClinVar and UniProtKB/Swiss-Prot databases to identify neutral or nonpathogenic cases. Each variant description in these datasets includes detailed mappings. To ensure their relevance and accuracy, the datasets can be periodically updated using scripts. This approach provides a comprehensive and up-to-date resource for the accurate prediction of the effects of genetic variations, including essential details such as variant ID, location, variant type, gene, molecular consequences, most severe clinical significance, alleles, population allele frequencies (1000G MAF, ExAC MAF), position, HGVS descriptions, positions in cDNA, CDS, and protein, amino acid and codon changes, co-located variations, and distance from relevant genomic features. This enriched dataset enhances the ability to accurately predict and understand the effects of genetic variations.