Genomic Data Sharing Spotlight

From The Genomics Landscape
Originally published: February 3, 2022

The Genotype-Tissue Expression (GTEx) project generated NHGRI’s most frequently requested controlled-access datasets. It is one of the most widely used resources for studying the relationship between genomic variation and gene expression. Because GTEx data can be shared for general research use, they are being used for many different research purposes. For example, GTEx data are used as controls to study COVID-19 as well as many other diseases and conditions. These datasets are also used to develop computational methods. The GTEx Consortium consists of over 200 investigators from over 40 institutions in the United States and abroad.

The latest version of the GTEx dataset, v8, is now available on the NHGRI cloud repository: NHGRI Analysis Visualization and Informatics Lab-space (AnVIL). Having the data reside in the cloud and running analyses on them there lowers the barrier to performing large-scale computational research by eliminating the need to download data to local platforms. Cloud storage also enhances and enables collaborative efforts. The net effect is an impressive demand for accessing the data; for example, NHGRI received 1237 requests for GTEx controlled-access data in 2020, with 1068 (86%) of those requests approved. Disapprovals were most often related to administrative issues, with most of those requests approved upon resubmission.

From The Genomics Landscape
Originally published: April 7, 2022

When pursuing clinical genomic studies, it is important to be transparent with both research participants and patients about how their genomic data will be used and shared. In the clinical setting, where patients are not necessarily enrolled in a research study when first seeking clinical care, conversations about how genomic data may later be shared in databases can sometimes be an afterthought, making it difficult to determine whether and how data can be shared down the line. To address this issue, the ClinGen Education, Communication, and Training Working Group developed a helpful one page consent form and training materials, which could be used as a model by others in point-of-care settings when collecting specimens that may later be used to generate genomic data that will be submitted to a database for future clinical and/or research use. Available in four different languages, the document is now accessible to a wide range of patients.

As the lines between patient care and research continue to evolve and blur, these sorts of mechanisms for patient consent are critical to maintaining the trust of the public in genomics research. Did you know that the top contributors of clinically relevant genomic variants to ClinVar are commercial laboratories? These groups are doing their part to ensure that future patients, no matter where they have their genomic tests performed, can have a better picture of the clinical significance of their results.

From The Genomics Landscape
Originally published: June 2, 2022

Finding a needle in a haystack is never an easy task. In the field of genomics, this analogy often applies when trying to identify genomic variants associated with phenotypes in large-scale sequence datasets. NIH is increasingly hosting genome-sequence datasets and other genomic datasets in cloud environments. Cloud-based computing helps researchers more efficiently sift through the haystack to find that needle. Pooling datasets from multiple studies enhances the statistical power of the resulting analyses, thereby increasing the likelihood that scientists will find both rare and common genomic variants that contribute to health and disease.

Enabling scientists to seamlessly combine datasets that were generated in different studies and that are housed in different repositories requires well-defined standards for interoperability. To this end, NIH is supporting efforts to improve the interoperability of the emerging set of NIH-supported cloud-based platforms and resources. When researchers access data from one site, there is no guarantee that those data will be readily usable along with data from a different site. By focusing on interoperability, the new NIH effort is improving the ability of researchers to integrate and analyze the data derived from multiple different datasets.

Are you interested in what it takes to build interoperable genomic resources in the cloud? The NIH Cloud Platforms Interoperability (NCPI) Effort will be hosting its first open meeting on June 22 and 23. Did you know that the existing NCPI sites currently house a whopping 11 petabytes of data? As the various genomic resources continue to grow and interoperate, so too will the power of human genomic studies.

From The Genomics Landscape
Originally published: August 4, 2022

Did you know that the goal of researchers contributing to the Human Genome Project (HGP) was to submit their sequence data into a public database within 24 hours of being generated? This target was documented during a strategy meeting in Bermuda in 1997, which is why this expectation is often referred to as the “Bermuda Principles.” The legacy of rapid data sharing within the HGP, as well as other large collaborative genomics projects, remains today. Open, quick, and quality data sharing—though not without its challenges—is an ideal that scientists, funders, journals, and many other stakeholders strive for across the biomedical research landscape. To read more about the origins of the Bermuda Principles, see here.

The NIH, inspired by the success of the HGP and other data-sharing efforts, has been advancing its data-sharing policies over the last two decades. NHGRI, as a leader in genomics, has also developed policies and practices that go beyond the baseline set by NIH. For instance, NHGRI expects researchers to leverage data standards and provide comprehensive metadata and phenotypic data for data shared per NIH data-sharing policies. The upcoming NIH Data Management and Sharing (DMS) Policy makes it more important than ever to ensure that shared data are maximally useful to the broader scientific community. NIH will be holding a webinar series in August and September focused implementing this policy. To learn more about the upcoming policy, see sharing.nih.gov and register now for the two-part interactive series to learn what the DMS policy means for you!

From The Genomics Landscape
Originally published: October 6, 2022

HL7 Fast Healthcare Interoperability Resources (FHIR) is a clinical data standard and an application programming interface to exchange Electronic Health Records (EHRs). FHIR provides an opportunity for the genomic medicine research community to develop ways to consistently report genomics and genomics-related data among different healthcare systems. However, additional developments are needed to better understand how these data should be represented in FHIR in order to best serve both the clinical and basic research communities. In 2019, NIH issued a notice encouraging researchers to explore the use of FHIR to capture, integrate, and exchange clinical data for research purposes and to enhance capabilities to share research data. Did you know that NHGRI-supported programs are involved in improving FHIR’s ability to exchange genomics and genomics-related information among discrete healthcare systems? For example, the Electronic Medical Records and Genomics (eMERGE) Network has been involved in expanding the capabilities of FHIR to represent clinical genomics results, and, more recently, FHIR is being incorporated into NHGRI resource programs such as the Genomic Data Science Analysis, Visualization, and Informatics Lab-space (AnVIL).

Efforts to standardize genomic medicine research are not finished. In February 2021, the NHGRI Genomic Medicine Working Group hosted a virtual meeting titled “Genomic Medicine XIII: Developing a Clinical Genomic Informatics Research Agenda.” The meeting called out “identifying and addressing semantic and syntactic gaps related to the representation of genomic information in existing clinical data standards and models” as a short-term need for the clinical informatics field. Over the coming years, NHGRI will be focusing on this issue, so as to pave the way for genomics to be responsibly incorporated into the clinical setting.

From The Genomics Landscape
Originally published: December 8, 2022

Over the past year, various aspects of genomic data sharing (GDS) have been highlighted via this spotlight series. There are many stakeholders that are part of this data-sharing ecosystem – patients and research participants, those that manage data repositories, data-access committees, policymakers, standards-generating organizations, and many more. The initial inspiration for this GDS Spotlight series was a conversation in 2020 that NHGRI hosted with scientific journal editors, another important group of stakeholders. Scientific journals share in the responsibility to ensure that the data underlying published research findings are made public and meet appropriate expectations for scientific rigor and reproducibility. That conversation highlighted several challenges that can complicate data sharing, including consent, data ownership, privacy, and international regulations. By increasing awareness about the importance and nuances of data sharing, the research community will hopefully be proactive in facilitating data sharing from the very beginning of each new study. Did you know that NHGRI maintains an Informed Consent Resource to provide guidance and sample language for genomics research, including how to describe data sharing to research participants?

The entries for this year’s GDS Spotlight series are now available on this page for future reference. Hopefully, others will get involved in this campaign to inform the community about the resources available to both share data and make those data maximally useful down the line!