Harold Varmus is still waiting for the revolution. In a 1999 proposal he later described as a “manifesto,” the then director of the US National Institutes of Health (NIH) suggested creating an electronic repository to host freely accessible research papers, including manuscripts prior to their formal publication. The next year, a version of Varmus's proposal spawned PubMed Central, a digital archive where all papers stemming from NIH-funded research must be submitted within 12 months of publication. However, the agency never created one key aspect of Varmus' ultimate vision: a preprint portal.

A decade later, the practice of prepublication archiving, now routine in physics, is looking for another chance in the biomedical and life sciences. On 3 April, the journal PeerJ launched a preprint server aimed at life scientists, while Cold Spring Harbor Laboratory Press, the publishing arm of the Long Island, New York–based research institute, is hoping to lure in biologists with its own preprint website later this year. Both are modeled on the wildly successful physical sciences preprint server, arXiv.org, which is gaining a fast following among some quantitative-minded biologists.

ArXiv's founder Paul Ginsparg, a theoretical physicist at Cornell University in Ithaca, New York, believes it is a matter of when and how—not if—biomedical scientists embrace preprints. “It's just going to happen at some point,” he says, citing the speed they bring to scientific communication. Ginsparg has watched a number of other scientific communities, from astronomy to mathematics to computer science, flock to arXiv over its 20 years and never look back. “The interesting question is: what the hell's wrong with biology?” says the University of California–Berkeley's Michael Eisen, who co-founded the open-access publisher Public Library of Science (PLoS) in 2000.

Like many leading population geneticists, Eisen has recently begun posting many of his lab's papers on arXiv prior to publication (see Nature 488, 19, 2012). Last month, for example, Eisen posted a paper exploring how the parasite Toxoplasma gondii makes infected mice less fearful of cats, an important host in the parasite's life cycle (http://arxiv.org/abs/1304.0479). Researchers who study emerging infectious diseases have also found arXiv useful. Chinese scientists on 7 April posted a genomic analysis of the country's ongoing H7N9 influenza virus outbreak (http://arxiv.org/abs/1304.1985).

Peering into the future: Time for prepublication? Credit: PeerJ

Some have suggested that the fear of being scooped or the potential for commercial applications have kept biomedical scientists from embracing preprint publication. Others point to the potential for unvetted research to influence public health practices. But Eisen blames the importance placed on journal titles as a proxy for quality. “I think the real reason biology is different is that, unlike most other fields, journals are way more powerful,” he says.

Landing a paper in Nature, Science or Cell can make or break careers in biomedicine, more so than in the physical sciences. Many scientists are afraid that posting papers to a preprint server will keep them from being accepted in such journals. (Nature journals allow preprint publication, Science permits it in “many cases” and Cell journals expressly prohibit prepublication.)

PeerJ pressure

Not everyone will be a convert, but biomedical researchers looking to release preprints while still ensuring they have a home for their final peer-reviewed manuscripts might want to use PeerJ Preprints and its companion journal PeerJ. With the single click of a mouse, researchers can simultaneously share a paper on the recently launched PeerJ Preprints site and submit that manuscript to the open-access journal. “It's almost a seamless experience,” says Jason Hoyt, a geneticist by training who co-founded PeerJ last year and now serves as the journal's chief executive.

Just because a manuscript ends up on PeerJ Preprints doesn't mean it has to be published in PeerJ, though. According to Hoyt, researchers are free to use the preprint service and submit their papers elsewhere. “We're happy either way,” he says.

Martin Fenner, a hematologist at the Hanover Medical School in Germany who consults for PLoS, worries that the service looks too much like Nature Precedings, a preprint site launched by Nature Publishing Group in 2007 that never really achieved community buy-in. (The site was shuttered last year.) Instead of launching another copycat portal, Fenner would like to see services that make it easier to share outputs of research other than manuscripts, such as conference talks, posters and seminars. Faculty of 1000, a London-based platform best known for providing post-publication review, has websites to host both preprint articles and conference presentations, but they have yet to truly take off.

Meanwhile, Cold Spring Harbor Laboratory Press is set to test the waters in preprint publishing before the end of the year. The service, called bioRxiv, will be largely modeled after arXiv, with a few additional features to entice life scientists. These include public commenting, room for supplementary information and links to established databases such as GenBank. According to Executive Director John Inglis, bioRxiv isn't designed to compete for life scientists with ArXiv. Rather, “we are hoping to attract a new population of scientists,” particularly from those scientific communities whose members have yet to fully embrace preprinting, he says.

Varmus, now the director of the NIH's National Cancer Institute, thinks the biomedical community finally may be ready to adopt preprinting—if they include tools to evaluate preprints and a way of apportioning academic credit. “If the virtues of speedy dissemination and vigorous exchange can be realized and appreciated, the practice could become widespread,” he told Nature Medicine.

But it's by no means certain that Varmus's 1999 manifesto will finally become a reality, notes Eisen. “Biology had an opportunity to be out in front on this in 1999,” he says. “We're 14 years later, and we still haven't done it.”