Figure 1 The Deafness Variation Database
(A) Kafeen, a custom internal pipeline, gathers data for the DVD by collecting variants and annotations from multiple data sources. Deleteriousness predictions collected from dbNSFP and MAF data are extracted from our local database, EVS, 1000 Genomes, and ExAC to inform the DVD classification. A comparison between DVD versus ClinVar and HGMD classifications captures all changes that result in medically significant differences (defined as up-grading a variant to P/LP or down-grading a variant from P/LP), each of which is manually curated to ensure that the DVD reclassification is appropriate.
(B) Venn diagram showing number of variants collated from major population-scale MAF databases and the count of variants that are shared among them and those that are database specific.
(C) Decision tree for Kafeen classification.
(D) Decision matrix detailing Kafeen logic regarding variants classified in ClinVar and HGMD.