The popular model plant Arabidopsis thaliana. Credit: J. BURGESS/SPL

The world's most valued plant database faces extinction because its funding is being phased out by the US National Science Foundation (NSF), and no alternative source is on the horizon.

"This is the wrong way to go," says genomics researcher Ernest Retzel of the National Center for Genome Resources in Santa Fe, New Mexico. "I believe it will set the field back."

The NSF says that it does not have a policy to support long-term, established research-infrastructure projects such as the Arabidopsis Information Resource (TAIR), which maintains a free, open-access database of genetic and molecular-biology data for Arabidopsis thaliana, or thale cress, the widely used model plant. "We didn't approach this decision in isolation, we considered our whole portfolio," says Peter Arzberger, director of the Division of Biological Infrastructure at the NSF. "We rely on peer review in setting our priorities." The NSF has suggested that TAIR develop its own self-supporting funding model, based on user subscriptions and other sources of income.

But TAIR director Eva Huala told an international meeting on database and bioresource sustainability, held in Rome on 11–12 November, that introducing a subscription system would destroy, not save, TAIR.

Huala, a member of the Department of Plant Biology at the Carnegie Institution for Science in Stanford, California, presented preliminary results of a survey among TAIR users, which revealed that many would be reluctant to submit data to TAIR if these were not freely shared.

Established ten years ago, TAIR integrates data submitted by the community with data extracted from the literature, and it has evolved into the plant community's foremost authority on matters relating to plant genomics, regulating nomenclature and developing curation standards. It is much more widely used than other plant databases because of its all-inclusive nature and the quality of its curation.

TAIR also feeds information into other specialist databases, such as those of the National Center for Biotechnology Information in Bethesda, Maryland, and the international protein database UniProt. In addition, it links to the Arabidopsis Biological Resource Center in Columbus, Ohio, which provides seed and DNA resources to researchers.

Credit: TAIR

TAIR has been supported by two consecutive five-year NSF grants, the second of which came to an end on 31 August. The NSF is planning to maintain the current budget of $1.6 million for 2010, and then to phase out funding over the following three years (see graph).

Huala's survey, sent on 4 November to more than 900 TAIR users, asked respondents which of TAIR's features are most important and which could feasibly be sacrificed in the hunt for alternative funding mechanisms. The majority of the 250 or so responses she has received so far say that there should be no log-in requirement, that everyone should have equal, free access to the data, and that data should continue to flow freely into other databases. Around two-thirds said they would be less likely to submit data to TAIR if these were not then freely shared with all researchers.

Respondents said that they would accept a situation in which publicly funded institutions and individuals had free access, but companies were required to buy subscriptions. They also said that they would be happy for advertising to appear on the website to raise revenue. But Huala says that neither measure would raise enough money to sustain the database.

"As soon as we introduce any form of subscription, we would not be able to export information to other free, open-access databases as we do now," she says. "The whole system would break down."

If TAIR were lost, another free database would inevitably spring up to take its place, thereby fragmenting the community, she adds. "The self-supporting models proposed by the NSF are more like a track to extinction."

Huala has put out feelers to other potential funding sources in the United States, and also to other countries, because only around a quarter of TAIR users are US-based. "But these are all long shots," she admits. "I am not optimistic."

Along with the Multinational Arabidopsis Steering Committee, which represents Arabidopsis researchers, the NSF is planning workshops for 2010 to gather input from members of the Arabidopsis community on their database and informatics needs.

The plight of TAIR is the most recent and most drastic example of funding crises now facing many databases and bioresources, says Paul Schofield, a molecular geneticist at the University of Cambridge, UK, who coordinated the Rome meeting. "There is a disparity between what science needs and available funding instruments for infrastructure," he says. "National research agencies need to get together to design new strategies."