Document sharing speeds up research

Science, perhaps more than any other human endeavour, depends on the sharing of information and ideas. This takes place by a variety of means, from informal chats and letters, to large, well-orchestrated conferences and peer-reviewed journals. The internet and the web have left their mark on all of these activities. One-to-one discussions of the kind previously pursued verbally or through posted letters are now more often conducted by email. The one-to-many dissemination of information by conferences and journals has been greatly facilitated by the downloadable PDF and the webcast. Such developments, particularly over the last 10-15 years, have created a cumulative effect in science that is nothing short of revolutionary.

But to see this as an endpoint is to miss the web’s most important characteristic of all: it is an environment ideally suited to facilitating many-to-many discussions. Which scientist has not dreamt of overcoming the constraints of geography and time by getting just the right group of brains into a room to discuss a pressing challenge, or debate a counter-intuitive result? Well, now we can do exactly that, except that this group no longer needs to come together at the same location and instant in the physical world, they need only to frequent the same website. In addition, broad scientific discourse no longer needs to take place over timescales of months and years, but rather hours and days.

Nature Publishing Group (NPG) is working on a number of initiatives to exploit this feature for the benefit of science. One of our most recent projects is Nature Precedings (precedings.nature.com), an open documentsharing website for scientists in the biological, chemical and earth sciences.

Nature Precedings is perhaps most easily understood as a preprint server that allows scientists to upload non-peer-reviewed (or pre-peer-review) documents so that they can be discovered, downloaded, read, and cited by other researchers. In this way, it is similar to arXiv.org, a website that has been providing an indispensable service to physicists, mathematicians, and computer scientists for over 15 years. In contrast to arXiv.org, however, Nature Precedings serves scientific fields in which the open and early exchange of research findings has yet to take root. Running a preprint server might seem like a bold, or even quixotic, goal for a publishing company, but we strongly believe that such open and early exchange of information between researchers is in the best interests of science as a whole, and NPG exists to enable such activities.

NPG has developed Nature Precedings with a series of partner institutions, including the British Library, the European Bioinformatics Institute (EBI), Science Commons and the Wellcome Trust. The service is provided free-of-charge to authors and readers alike, and NPG claims no special rights over the content (which is released under the Creative Commons Attribution License). We have committed to keeping the content openly available and our partners will help to guarantee this by ‘mirroring’ Nature Precedings documents on their servers, where they will also be available free-of-charge.

Early results in biology, chemistry and the earth sciences are often described in posters and presentations so Nature Precedings accepts a range of document types in addition to manuscripts. All contributions are screened by curators, who check for new submissions regularly throughout the day. They ensure that anything accepted is from a suitablyqualified scientist, that it coherently reports findings in an appropriate subject area (e.g., not in clinical medicine or physics), and that there are no obvious technical errors or omissions in the document or its associated metadata. These criteria are obviously much less stringent than those applied to NPG’s peer-reviewed journals.

Also unlike our journals, Nature Precedings does not aim to highlight the very best research, but to provide a snapshot of current activity in the fields that it covers and help balance the situations where delays of several months between writing-up and publishing really do impede progress. The service should be most useful for researchers reading within their own areas of expertise so that they are well equipped to judge for themselves the merits (or otherwise) of non-peer-reviewed work.

Readers can leave comments (which are moderated) and can vote for contributions they consider particularly noteworthy. When work appearing in Nature Precedings is later published elsewhere, especially in peerreviewed form, there is a mechanism for authors and readers to provide links to the new version. Contributing authors can also upload new versions of previously-released documents. The older versions remain online and are separately citable, but readers are advised of the availability of the newer version.

Addressing author concerns

Even if open information sharing is in the interests of science as a whole, individual researchers do not necessarily have an incentive to participate. We recognise that modern science is such a competitive environment that the service needs to appeal not only to researchers’ altruism but also their self-interest. One worry is that early dissemination of findings might enable another research group to publish similar results in a peer-reviewed journal.

Nature Precedings attempts to address such concerns by making early findings citable. Documents are cited not merely by referencing a URL, but by using Digital Object Identifiers (DOIs) and Handles, the same global system used to provide online citability for the peerreviewed literature. Nature Precedings can therefore be viewed as a system for claiming priority rather than surrendering it. Any future peer-reviewed papers based on work released on Nature Precedings can cite the original work – and in our opinion ought to do so.

This notion of citability and comparison with journals leads to another potential concern: that some readers will mistake nonpeer-reviewed content on Nature Precedings with manuscripts in peer-reviewed journals. Although Nature Precedings contains manuscripts similar to those found in peerreviewed journals, they are not ‘reviewed’ by ‘editors’ or ‘published’, but ‘screened’ by ‘curators’ and ‘posted’. Although these are only language differences, they help to set readers’ expectations. Furthermore, as people generally become familiar with services such as discussion boards, social networking sites and wikis, this distinction will become more intuitive. The site also contains many references to the fact that the content is not peer-reviewed. To our knowledge, no one has yet been seriously misled. But we are not complacent, not least because our own reputation is at risk, and we are constantly on the lookout for ways to clarify this issue further. For this reason, too, Nature Precedings does not currently accept documents focusing on clinical or therapeutic claims as the stakes arising from misinformation are obviously much higher there than in basic research.

Meeting potential threats

Some people have wondered why a highly successful journal publisher would embark on a project that seems to undermine the traditional publishing model. Aren’t preprint servers and open commenting systems a potent long-term threat to journals and the traditional models of peer-review on which they depend? In our view, they are not. In physics, a strong preprint culture and arXiv. org have co-existed peacefully for many years with peer-reviewed journals, including some of NPG’s own titles such as Nature and Nature Physics. We don’t believe that a competent publisher that is confident in the value it adds through editing and peer-review has anything to fear from the exchange of preprints between scientists.

But suppose we are wrong and that the sharing of preprints is the first step to eroding journal publications in their current form. That would be an even more compelling reason to create a service like Nature Precedings. If the foundations of our business are going to be torn up then to stick our heads in the sand would be folly. Better to reinvent the world ourselves –and understand quickly how we can add value in new ways – than leave this to others. Even if our ultimate destination is out of sight, those leading the charge get a much better view of the road ahead than those reluctantly bringing up the rear.

Timo Hannay is NPG’s publishing director for Nature.com and is based in London. Hilary Spencer is product development manager for Nature Precedings and works in New York.