dbNSFP v5: A Comprehensive Resource for Functional Prediction and Annotation of Human Nonsynonymous and Splice-Site SNVs with Enhanced User Services

Le résumé fourni par la source

The increasing use of whole-exome and whole-genome sequencing requires efficient tools for prioritizing potentially disease-associated variants. dbNSFP v5 provides a comprehensive and frequently updated resource for annotating human nonsynonymous single-nucleotide variants (nsSNVs) and splice-site SNVs (ssSNVs). The latest release, dbNSFP v5.3, is based on the hg38 reference genome and GENCODE v49. It includes 83,049,507 potential nsSNVs and 2,446,464 potential ssSNVs—approximately 1.6 million more variants than v5.2, primarily because of expanded protein-coding transcript annotations informed by long-read sequencing. dbNSFP v5.3 integrates prediction scores from 35 algorithms, nine conservation scores, and population allele frequencies from major resources, including the 1000 Genomes Project, gnomAD, TOPMed, All of Us, the RGC Million Exome dataset, and ALFA. It also provides gene identifiers, functions, expression profiles, and disease associations from resources such as the Human Protein Atlas, UniProt, OMIM, and ClinGen. User services have been substantially enhanced in dbNSFP v5. A redesigned web interface enables users to search and explore variants without downloading the full database. A new DuckDB-based architecture improves data-processing speed and can annotate a whole-genome sequencing VCF file with dbNSFP resources in under 50 seconds on a desktop computer. dbNSFP v5 provides a fast, comprehensive, and current platform for prioritizing and interpreting human genetic variants.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.