The Howard Hughes Medical Institute (HHMI), a non-profit research organization that funds more than 300 US researchers, is considering a plan to pressure its investigators into making their published papers freely accessible.

The plan, if approved, would dictate that publications must be deposited in a public database within six months of publication in order to count towards an investigator's application for reappointment. HHMI investigators apply for reappointment every five or seven years.

The proposal embodies the latest stage of the open-access movement: enforcement. After years of requesting voluntary compliance, several funding agencies are considering tougher stances. In 2005, the US National Institutes of Health (NIH) asked grantees to voluntarily deposit articles in a public database such as PubMed Central within 12 months of publication. A year after the request, only 4% of NIH grantees had done so, prompting Congress to propose legislation mandating compliance.

Meanwhile, on 1 October, Britain's Medical Research Council and the Wellcome Trust medical charity began requiring grantees to deposit final, peer-reviewed manuscripts in public databases as soon as possible, but no later than within six months of publication. Failure to do so, says Wellcome Trust director Mark Walport, would be a breach of granting conditions.

HHMI president Thomas Cech says a decision on the proposed policy will probably be made in early 2007. Although HHMI officials say they will not legislate where their investigators publish, several researchers say the threat of weakening their reappointment application represents significant pressure.

Cech says the proposed policy is simply an extension of HHMI guidelines about sharing published reagents and other research material. “The publications are the most useful product of our investigators' research,” says Cech.

The HHMI is still negotiating with publishers, but Walport says most major journals, including Science and Nature, have complied with the Wellcome Trust's guidelines.

Many publishers let authors pay to make their articles available immediately. For example, Wellcome Trust grantees can make their papers in most Elsevier journals publicly accessible for $3,000 per article. Both the HHMI and the Wellcome Trust already provide funds for publication in open-access journals. “We see payment to the publisher as part of the cost of research,” says Walport. “And we're prepared to pay appropriately.”

But some publishers continue to struggle with the Wellcome Trust's requirement. If a paper is available in a public database, this reduces the number of visits to subscription sites, says Martin Frank, executive director of the American Physiological Society.

Frank says this is why his society refuses to comply with the Wellcome Trust's guidelines. But such decisions may be more difficult with the proposed expansion of open-access requirements — Wellcome Trust-funded research accounts for only 2% of the papers in the society's 14 journals, but the NIH supports almost half.

Kathleen Case, publisher of the American Association of Cancer Research's five peer-reviewed journals, also considers the requirements too strict. “The Wellcome Trust asked us to change our policy and we said no,” says Case. “And now we've been blacklisted.”