Genomic Data Sharing: From Utopian Principle to Practical Reality

“Publicly available data should be treated as open data, a shared resource with unrestricted use for analysis, interpretation, and publication.”

–The journal Science report on genomic data sharing

As a principle, it’s hard to take issue with Science‘s recommendation; but as the article itself notes, actually achieving genomic data sharing remains highly difficult due to the lack of clear guiding rules for data usage. The good news is that recent progress in standards and software development has made it achievable.

Criteria for Data Sharing
A number of organizations are working to facilitate the process of and establish standards for data sharing. Among them is ClinGen, a National Institutes of Health (NIH)-funded resource whose mission is to build a genomic knowledge base to improve patient care. ClinGen has established a minimum standard of data sharing for clinical labs that meet the following criteria:

• The lab’s submissions are registered in ClinVar (a free government resource for the medical community) as “Single Submitter, Assertion criteria provided;”

• A complete list of the lab’s test offerings is publicly available either through a registry such as Genetic Test Registry (GTR) or Orphanet, or on the lab’s website (Requirement modified February 2019 to any public display of the lab’s complete list of test offerings, not limited to GTR);

• The lab submits at least once per year adding new variants and updating reclassified variants as necessary;

• The lab submits all categories of variants reported to patients (labs are also encouraged to share B/LB variants and VUS if not reported);

• The lab has attested to submitting at least 75% of all sequence and/or copy number variants reported in the past year;

• The lab has submitted at least 100 variants; and

• The lab is CLIA certified (USA) or meets an equivalent standard in another country.

Tools for Data Sharing
Another barrier to achieving data sharing is the lack of available tools to facilitate the process. However, medical software providers are working to solve that problem.

For example, ApolloLIMS has developed a product called eXchange that allows a lab to send and receive data using common protocols and file formats. While this solution in and of itself does not create the “open data” for which some, including the authors of the Science article advocate, it does highlight an important component. The technology to share data is readily available.

Takeaway: While much work remains to be done, genomic data sharing has become achievable thanks to standards like the ClinGen criteria and software like eXchange.