EditorialPolicy

Something in Common

See allHide authors and affiliations

Science Translational Medicine  14 Jul 2010:
Vol. 2, Issue 40, pp. 40ed6
DOI: 10.1126/scitranslmed.3001280
CREDIT: MERCK & CO., INC.

VISION

Imagine a world where massive aggregates of clinical, phenotypical, and genomic information drive the discovery of novel drug targets and the development of effective treatments. Imagine that such dynamic and evolving data sets, combined with computational disease models, could provide patients with more potent and personalized therapeutic options than are available today.

To make this vision a reality, Sage Bionetworks (1) and Creative Commons (2) organized the first Sage Commons Congress, which took place in San Francisco on 23 and 24 April 2010. The Congress generated and tested ideas and approaches needed to transition genomics research to a patient-centered focus. The 220 contributor-participants included scientists, disease modelers, patients, policy-makers, funders, and members of the media in similar proportions. The participants came from the full range of stakeholders, including the U.S. National Institutes of Health (NIH), the British Library, Harvard University, RIKEN, the Bill & Melinda Gates Foundation, the Cure Huntington’s Disease Foundation, Nature, WIRED, and the Public Library of Science, to name a few.

HOW WE GOT HERE

Traditional human disease research models are now archaic. The academic NIH RO1 process is choked by favorite gene efforts that result primarily in “impactful” journal articles. This paradigm often assumes a level of simple linear pathways that is more appropriate to the limited data available in the last century. Many companies cling to the concept that they must invest in proprietary, and therefore inherently redundant, disease biology even as a few insightful industry collaboration efforts to share currently sequestered samples and data emerge.

And the patients? They’re getting more and more frustrated.

THE CONGRESS

The program featured research vignettes of network modeling, keynote addresses, and working groups who together had invested thousands of volunteer hours before the event to address barriers and incentives to making disease biology more accessible to the entire research community. Videos and presentations are available on the Congress Web site (3).

Andrew Kasarskis (Sage Bionetworks) and Ilya Kupershmidt (NextBio), with leading network modeling labs, presented a pilot that combined, modeled, and queried available human, mouse, and yeast data sets. Jessie Tenenbaum (Duke University) led the effort to define standards and ontologies for integration, analysis, and exchange of data, soon to result in a public set of draft recommendations. Carole Goble (University of Manchester), Alex Pico (University of California, San Francisco), and Ted Liefeld (Broad Institute) led an effort to integrate existing data sets with open tools such as Taverna, Cytoscape, and GenePattern. Their group is now building broader tool interoperability for network models of disease. Carolina Rossini (Berkman Center for Internet and Society, Harvard Law School) led the group focused on the thorny issues of international data regulations, transfer, and integration. Liz Lyon [the UK Open Library Network (UKOLN)] (4) led a group that studied data citation and proposed an approach for enabling disease models themselves to become citable. This was a predictably emotive issue, because it touches on the metrics used in hiring, funding, and promoting researchers, as well as encroaching on traditional concepts of “turf.” Mechanisms of citation and ownership that are acceptable to researchers, institutions, and funders will be critical if we are to increase data sharing and access.

What was obvious to those present, yet hard to translate externally, was the sense of urgency to tear down the cultural as well as technical barriers that prevent open data and model sharing. There was full agreement that it is essential to build reward and recognition structures to encourage full participation and benefit.

By the end, there was consensus that existing data-sharing and human disease model–building processes need an extreme makeover. The immediate objectives are clear: (i) Develop “shining examples” of progressive data sharing, leading to new insights not currently feasible; (ii) develop specific approaches for data standards, annotation, and tool interoperability; (iii) implement effective mechanisms for citing data sets and models; and (iv) champion patient advocacy and visibility for various issues, including privacy and intellectual property rights.

The Congress vision is achievable but requires a transition to a world wherein large genomics-powered clinical trials are not isolated efforts whose outputs are publications, but rather are used to generate building blocks of data and models for future experiments. The barriers are no longer simply technical, scientific, or legal; they are hurdles tied to scientific culture, funding, and incentives. The barriers are surmountable, but experience suggests that the former obstacles (technical, scientific, or legal) can be directly addressed by constructive engagement, whereas the latter ones (scientific culture, funding, and incentives) may require widespread grassroots endorsement to “win hearts and minds.”

Progress will occur only when like-minded individuals join in efforts such as the energized projects that resulted from the first Sage Commons Congress and that are already progressing toward the second Congress in 2011.

Footnotes

  • Citation S. Friend, Something in common. Sci. Transl. Med. 2, 40ed6 (2010).

References

Stay Connected to Science Translational Medicine

Navigate This Article