732 posters,  2 sessions,  11 topics,  3542 authors,  923 institutions

ePostersLive® by SciGen® Technologies S.A. All rights reserved.

ClinGen Efforts to Improve the Accuracy of Variant Interpretations in ClinVar

Primary tabs


No votes yet


527 reads

ClinGen and ClinVar Collaborate to Improve Variant Interpretations in ClinVar

Heidi Rehm1,2, Steven Harrison1, Danielle R. Azzariti1, Donna Maglott3, Erin Rooney Riggs4, Marina T. DiStefano1, Justin Aronson5, Sarah E. Hemphill1 , Brandon J. Cushman1 , Melissa Landrum3 , Christa Lese Martin4 

1 Laboratory for Molecular Medicine, Partners Healthcare Personalized Medicine, Cambridge, MA2 Department of Pathology, Brigham & Women’s Hospital, Harvard Medical School, Boston, MA3National Center for Biotechnology Information, National Library of Medicine, National Institutes of Health, Bethesda, MD, USA4Geisinger Health System, Danville, PA, USA, 5Pierce Middle School, Brookline, MA, USA

ClinVar (a database) and ClinGen (a program), are two NIH-based efforts, that have formed a critical partnership to improve our knowledge of clinically relevant genomic variation.  ClinGen is an authoritative central resource that defines the clinical relevance of genes and variants for use in precision medicine and research. ClinVar is an archival database that aggregates information about genomic variation and its relationship to human health. This partnership includes significant efforts in data sharing, data archiving, and curation to characterize and disseminate the clinical relevance of genomic variation.

ClinVar contains 438,178 submissions on 285,376 unique variants from 668 submitters in 59 countries

Distribution of variant interpretation comparisons between Ambry, GeneDx, Partners LMM, and University of Chicago.
(a) Interpretation comparison of data in ClinVar before resolution efforts. (b) Interpretation comparison after reassessing 33% (242/724) and resolving 87% (211/242) of variants with differences. (Harrison et al, GIM, in press).

VariantExplorer.org allows submitters to identify all of their interpretation differences as well as compare how often they call a variant one category (e.g. P or LP) versus other labs calling the same variant a different category (e.g. VUS, LB, B). Plotting these ratios allows identification of each submitter’s relative tendency towards pathogenic interpretations versus more benign or uncertain interpretations, which helps identify systematic differences in variant classification.

‘Literature only’ submissions represent significant tendencies towards P/LP calls compared to clinical laboratories. Removal of variant classifications from non-primary source submissions would reduce conflicting interpretations by 19% (from 4,895 to 3,983).

Basis of initial interpretation differences for resolved variants. 17% resolved because the re-interpretation was not yet submitted to ClinVar; 36% resolved by reassessing old interpretations; 17% resolved by sharing non-published internal data; 14% resolved by agreeing on use or weighting of public data. (Harrison et al, GIM, in press).

Data sharing in ClinVar and interlaboratory conflict resolution activities, as well as evolving policies on the representation of data submitted to ClinVar, are improving the accuracy of variant interpretations in ClinVar, which will undoubtedly improve the care of patients who are dependent on accurate variant interpretation for their health.

This work was supported by NHGRI in conjunction with funding from the NICHD under award U41HG006834. ClinVar is supported by the Intramural Research Program of the NIH, National Library of Medicine.

Enter Poster ID (e.gGoNextPreviousCurrent