Database Deadlock

After months of high-level correspondence, the National Institutes of Health and the American Chemical Society do not appear to be any closer to finding a mutually acceptable agreement regarding the scope of NIH's chemical database PubChem, which ACS argues will duplicate its Chemical Abstracts Service (CAS) Registry.

In an exchange of letters last month, ACS President William F. Carroll and NIH Director Elias A. Zerhouni proposed possible options for an amenable solution that would allow NIH to retain the intent of the database while leveraging existing private-sector chemical databases. Neither side has yet to agree to the other's proposal, although both sides tell C&EN they remain committed to working out a solution.

PubChem was launched last fall under the molecular libraries initiative, which is part of NIH's Roadmap for Biomedical Research. The purpose of this free database is to compile information on small molecules with potential medical applications--including compounds yielded by the agency's screening center network and publicly available ones--and provide links from these compounds to biomedical data (C&EN, June 13, page 23). It's this broad inclusion of compounds that ACS worries will duplicate its own comprehensive chemical structure database.

To avoid this duplication, Carroll proposed in an Aug. 9 letter to NIH that CAS be allowed to set up, run, house, and maintain PubChem for NIH at no cost to the agency. As part of the plan, ACS would commit up to 15 members of its staff for a minimum of five years at a cost to ACS of more than $10 million. ACS and NIH would work together to disseminate the data in a way that takes advantage of existing databases, Carroll explained in his letter.

As Carroll notes, NIH hasn't yet agreed to ACS's proposal. "We are confident that we [CAS] can create a resource to advance the NIH Roadmap that would be fully interoperable with biomedical databases and serve broad scientific needs," Carroll tells C&EN. "So far, NIH and ACS haven't agreed on the approach, but we're still working toward an acceptable solution," he notes.

In an Aug. 22 letter from Zerhouni, the agency cites a combination of federal regulations--which bar the government from entering into an exclusive relationship without giving other private-sector entities a chance to compete--and concerns that allowing ACS to take over PubChem would result in the loss of some of the most critical aspects of the database. Specifically of concern to NIH is how decisions on what compounds would be added to the CAS-operated version of PubChem would be made and whether CAS would be able to integrate the database with other NIH databases.

"NIH's position is that any molecule of potential biomedical relevance is appropriate for inclusion in the database," Zerhouni wrote in a letter to Carroll. He added that unfettered access to this broad range of compounds is necessary for the database to have a significant impact on biomedical research.

But ACS is concerned that because all small molecules can have potential relevance, PubChem will not be a more limited database than the CAS Registry, as the agency has argued, according to Brian Dougherty, senior adviser to the chief strategy officer at ACS. It's this unlimited nature that ACS contends will reproduce its registry.

"This is a public policy decision," says Dougherty. He questions whether there is sufficient public need for a federal government agency to step in and create a competitive database rather than leveraging existing databases--in other words, asking "where the federal role ends and the private sector begins."

Neither side has yet to agree to the other's proposal, although they remain committed to working out a solution.

In a push to have this public policy issue addressed, ACS has taken its case to Congress. And as Carroll notes, both the House and Senate appropriations bills include language that urges NIH to avoid unnecessary duplication and competition; he adds that "we believe our offer constitutes a good platform for doing so."

For its part, NIH has countered the ACS proposal with one of its own. The six-part plan includes an opportunity for CAS to work with NIH to "validate or assign registry numbers for all PubChem structures." In return, NIH would provide "reasonable financial compensation."

Also in the proposed plan is the formation of a new working group of outside chemical information providers and users to advise NIH on PubChem's development. NIH has already moved to set up such a group by issuing a Federal Register notice earlier this month (C&EN, Sept. 12, page 27).

One of the tasks of this working group will be to develop a retrospective process for judging and removing compounds from PubChem that are deemed not to be biomedically relevant. The group will also be charged with providing input on ways to ensure that data are being properly deposited into PubChem.

Finally, the plan proposes to have a representative from the private sector serve on the existing PubChem Advisory Board and expresses NIH's continued openness to having an independent, outside group study the impacts of PubChem on science progress and on private-sector databases.

As NIH awaits ACS's response to this proposal, both sides remain hopeful that a resolution can be reached. "We are determined to move forward in our efforts to work with the private sector to ensure that PubChem develops in a way that enhances all of biomedical research without having a negative impact on the private sector," Zerhouni said in his letter to ACS.