Variant Interpretation for Cancer Consortium


Group Logo

Specific Aims

  1. Harmonize global efforts for clinical interpretation of cancer variants. Form an open consortium of developers and curators committed to eliminating the interpretation bottlenecks for precision medicine in cancer.
    1. Standardize data model. Establish the minimal data elements and standards needed to describe genotype, clinical phenotype, and evidence for clinically relevant genomic alterations in cancer.
    2. Coordinate curation activities according to the domain expertise of different groups/institutes. Unify efforts and leverage domain-specific expertise to reduce redundant curation effort.
  2. Implement software systems to query across standardized knowledgebases. Given a clinical sequencing assay result, for a patient with a specific cancer type, produce a comprehensive report of clinically relevant associations between genomic alterations and diagnosis, prognosis and treatment options using all publicly available sources of expert-curated interpretations.
    1. Implement federated query. Demonstrate ability to submit clinical genomics queries across disparate knowledgebases through a public interface.
    2. Implement web application. Develop capability to display clinically actionable recommendations based on cross-knowledgebase queries and accept contributions through a shared interface.

Guiding Principles

  • While each institution’s individual knowledgebase will continue to exist to service the specific needs of each institute, there is clear value and necessity in sharing knowledge of cancer type-variant-drug associations. Such sharing will increase confidence where interpretations overlap, fill gaps, reduce redundancy, and leverage disparate domain expertise.
  • A minimal set of data elements for sharing interpretations is required. Each institute will commit to including at least these minimal elements in their curation data model.
  • Each institute will share all interpretations (with at least the minimal required data elements) accumulated by their current curation efforts. This will require releasing content under permissive licenses (free and non-exclusive for at least academic use) and providing documented public APIs for data access.
  • Building on software developed for independent efforts will be critical to create a community resource. Therefore, software will be released in public repositories with open source licenses.
  • To avoid patient data privacy concerns, the project will focus initially on clinical interpretations derived from published findings (literature, conference proceedings, and clinical trial records), not individual patient/variant-level observations.
  • Data sharing will be facilitated by use of the GA4GH genotype-to-phenotype (G2P) schemas, APIs and demonstration implementations.
  • The GA4GH group will make available bulk downloads of all cross-knowledgebase interpretations to facilitate data mining.


  • Obi Griffith Obi Griffith
    • Washington University in St. Louis
    • St. Louis, United States
    • Co-Lead, Group
  • Malachi Griffith Malachi Griffith
    • Washington University in St. Louis
    • St. Louis, United States
    • Co-Lead, Group
  • Nuria Lopez-Bigas Nuria Lopez-Bigas
    • Co-Lead, Group
  • David Tamborero David Tamborero
    • Co-Lead, Group



  • Obi Griffith , Washington University in St. Louis , St. Louis, United States
  • Malachi Griffith , Washington University in St. Louis , St. Louis, United States
  • Nuria Lopez-Bigas , Universitat Pompeu Fabra , Barcelona, Spain
  • David Tamborero , Universitat Pompeu Fabra , Barcelona, Spain