Skip to content

dbNSFP

Resource Location

/labs/SysMedBio/Q/Resources/Genomics/dbNSFP

/Volumes/lab/SysMedBio/Resources/Genomics/dbNSFP

Q:\SysMedBio\Resources\Genomics\dbNSFP

dbNSFP is a comprehensive, one-stop resource compiling functional annotations and pathogenicity predictions for all possible non-synonymous single-nucleotide variants (nsSNVs) in the human genome.

Scientific Context

Non-synonymous single-nucleotide variants (nsSNVs) are single base-pair substitutions that alter the amino acid sequence of a protein. Because they change the primary protein structure, nsSNVs can affect protein folding, stability, ligand binding, or catalytic activity, representing the primary class of mutations underlying rare Mendelian disorders and genetic susceptibility to complex diseases.

Exome or whole-genome sequencing of patients typically yields thousands of nsSNVs. Discerning which ones are pathologically deleterious versus benign is a major challenge.

dbNSFP resolves this bottleneck by precalculating and consolidating: - Pathogenicity Scores: Consolidating predictions from dozens of algorithms (e.g., SIFT, PolyPhen-2, LRT, MutationTaster, MutationAssessor, FATHMM, PROVEAN, CADD, REVEL, ClinPred, popEVE, MisFit, and AlphaMissense). - Evolutionary Conservation Scores: Including PhyloP, PhastCons, GERP++, and SiPhy. - Population Frequencies: Compiling allele frequencies from 1000 Genomes, gnomAD, and local cohorts. - Gene-Level Annotations: Standardizing gene names, function, tissue expression, and phenotypes.

By providing a unified local database, dbNSFP allows annotation engines (like VEP or ANNOVAR) to perform high-throughput variant annotation offline, avoiding network bottlenecks and protecting clinical data privacy.

Main Publication

  • Title: dbNSFP: A Lightweight Database of Human Nonsynonymous SNPs and Their Functional Predictions
  • Authors: Xiaoming Liu, Xueqiu Jian, and Eric Boerwinkle
  • Journal: Human Mutation (2011, Vol. 32, Issue 8, Pages 894–899)
  • DOI: 10.1002/humu.21517

  • Latest Major Update (dbNSFP v4): Liu, X., et al. (2020). "dbNSFP v4: a comprehensive database of all patterns of non-synonymous variants in the human genome." Genome Biology, 21(1): 290.

  • DOI: 10.1186/s13073-020-00803-9

Available Files

The dbNSFP directory contains precomputed databases (version 5.3.1a):

  • dbNSFP5.3.1a_grch37.gz: Gzipped tab-delimited annotation database mapped to the GRCh37 (hg19) reference genome.
  • dbNSFP5.3.1a_grch37.gz.tbi: Tabix index file enabling fast coordinate queries for the GRCh37 database.
  • dbNSFP5.3.1a_grch38.gz: Gzipped tab-delimited annotation database mapped to the GRCh38 (hg38) reference genome.
  • dbNSFP5.3.1a_grch38.gz.tbi: Tabix index file enabling fast coordinate queries for the GRCh38 database.