Preprints are beginning to arrive in biology, will medicine be next?

Amy Williams

Preprints, long used in physics, astronomy and mathematics, finally seem to be catching on in the life sciences. A recent online seminar, hosted by internet reference linking organization Crossref, showcased the benefits that preprints can lend to biology researchers.

Preprints serve to bridge the often lengthy period between submission to a peer-reviewed journal and official publication. The final draft of the article is hosted in a repository and made accessible online ahead of or upon submission to a journal, allowing it to be read while the peer-review process is underway. Posting preprints has other key benefits: ideas can be shared and discussed faster, and ‘invisible’ research such as doctoral theses can be easily and publicly disclosed. A timestamp on the preprint can also defend against ‘scooping’.

The Crossref meeting focused on how preprint platforms, which have traditionally catered to other fields of study, are expanding into biology. Meeting Chair, Rachel Lammy, began with an introduction to Crossref’s platform, which aims to link each item related to an article together and curate associated metadata. The topic of preprints was selected on account of the new services introduced by Crossref in November 2016, which can now link articles to their preprinted drafts.

The advantages of preprints to biologists were outlined by ASAPbio representative Jessica Polka, with specific emphasis on the progress made by funders of biological research. Last year saw several major funding bodies, including the Wellcome Trust, the Helmsley Charitable Trust, the National Institutes of Health (NIH), EMBO, the Simons foundation, the UK’s Medical Research Council and the Chan Zuckerburg Initiative announcing pro-preprint policies. The Chan Zuckerberg Initiative preprint policy is particularly radical, requiring that each piece of research funded by the Initiative be posted to a preprint server upon journal submission without exception, thereby effectively preventing researchers from submitting to journals that do not accept preprinted articles.

Also highlighted was the 4D Nucleome research consortium, which is funded by the NIH. The consortium encourages preprints by ensuring that all research it conducts is posted on one unified preprint server. This was the first major biology research programme to mandate preprints for research outputs. However, the policy has not received unanimous approval even within the Nucleome project itself, with 30% of investigators at the Nucleome consortium opposing the policy. The founder of ASAPbio, Ron Vale, is supportive of the consortium’s policy, and hopes that “over time, biologists grow to see the value and want to do this… being forced to do something is not the best way to win over someone’s heart”.

The final speaker was Richard Sever, Assistant Director of bioRxiv, who presented on bioRxiv as a platform and the behavioural changes surrounding preprint submissions seen in recent years. In September 2016, 1564 life science preprints were posted to eight of the largest preprint servers available to life scientists. This is a five-fold increase from September 2011, when only 300 preprints were posted, and only three of the platforms examined, if they existed at all at that point, hosted any life science preprints. The increase in submissions has prompted journal policy changes and attitude changes among funders and research institutions.

But what does this mean for medicine? Biomedical sciences still only constitute approximately 22% of preprints submitted to bioRxiv, with genetics research accounting for almost half of this figure. Clinical trials in particular are rarely posted, and account for less than 1%. The restrictions placed upon medical researchers by journals have been a leading cause of this. Some prominent medical publishers still abide by conservative policies: for instance, the American Heart Association has stated in correspondence as recently as September 2016 that it will not review preprinted submissions. A similar policy was reported in communications from the American Association for Cancer research in November 2015.

Concerns also exist surrounding the sharing of medical research before peer review. This is understandable as poorly conducted research, particularly in medicine, can certainly be damaging. For this reason pharmaceutical companies, which are major funders of medical research, have been cautious of using preprint platforms. This is particularly true for clinical trial results, and stems from fears that sharing research publicly ahead of peer review could violate regulations regarding off-label or direct-to-consumer promotion.

The meeting concluded with the benefits of preprints seen so far, and encouraging wider uptake in all fields of the life sciences. It is clear from the figures presented that many significant shortcomings of journal publishing can be ameliorated through the hosting of preprints upon submission. Engagement with research is facilitated, with 10% of the preprints posted on bioRxiv receiving comments from other users on the site; 90% of submissions to the site are made publicly available and all that pass the basic editorial check are shared in less than 24 hours from submission. Whether or not preprint platforms become widely adopted in biomedical research is yet to be seen, but they have great potential if author behaviour and funder attitudes continue on their present trajectory.

By Amy Williams

Disclosure: Amy Williams is an employee of Oxford PharmaGenesis, and attended the meeting as part of her employment at the company. Oxford PharmaGenesis organized another meeting earlier in the year focusing on publishing innovations with relation to the pharmaceutical industry. For more information on this meeting, find a summary by Richard Smith here.