blog

We are particularly excited to see the first preprints being indexed in a major bibliographic database, namely Europe PubMed Central (ePMC). This is a major step forward in moving us beyond the peer reviewed article being seen as the only valuable research output for many fields and enabling new findings to be made more visible much more quickly.

The growth of preprinting

This expansion of ePMC’s coverage recognises the trend across many fields of research that now encourage the early sharing of work through preprints. This of course started back in 1991 in physics with arXiv but has spread much more recently to biology, geosciences, social sciences and beyond with servers like bioRxiv, EarthArXiv and SocArXiv.

This trend has been boosted by the support that major funders have shown in recognising preprints as a valid representative of work that has been carried out by their researchers. Chan Zuckerberg Initiative (CZI) have developed one of the strongest policies where they ‘strongly encourage, and in some cases, may require’ researchers to deposit manuscripts as preprints before peer review. The NIH, Wellcome and others also encourage their researchers to submit preprints and allow their citation in grant applications.

Publication of a preprint ahead of peer review enables researchers/authors to retain control of their work; it is them that decide what they want to communicate and when, removing the typical gatekeeping role of the journal Editor. This increases access to a much broader range of research findings (including null results, data notes, software tools, etc), helps to reduce publication bias and research waste, and enables much more rapid communication of results – from months/years to days.

Building on preprints – from editorial checks to peer review

As ePMC comment in their announcement, there is work to do to bring some consistency to the basic screening process that many preprint servers conduct prior to public posting of a submission.  We and others believe that systems need to be developed to provide readers with simple information on the level of trust they can place in a preprint, e.g. have plagiarism checks been done, has someone checked the article meets ethical or reporting requirements for the relevant field, has anyone checked any associated images for image manipulation etc. Interesting approaches are being developed such as the recent PLOS-bioRxiv collaboration where PLOS submissions that pass their initial screening checks can be immediately posted to bioRxiv without further checks. The second ASAPbio meeting earlier this year also helped to trigger the development of several new community review experiments that will be interesting to watch as they grow.

Of course, peer review of new discoveries by relevant experts is still very important in assessing whether new findings are technically sound. By conducting this review directly on the preprint and completely in the open, together with the ability for the authors to revise and version their article, removes unnecessary confusion between article versions on different sites (preprints and journal publications). It also makes it clear to a reader the level of review an article has been through, by whom, and the outcome of that review.

This approach, which circumvents the need to go through a separate (and often protracted) journal system, has seen a significant growth in uptake, despite the stranglehold that the journal brand has on funding and promotion/tenure decisions for most of the submitting researchers.  We initially developed this ‘preprint–peer review’ approach on our own platform F1000Research.

The significant growth in researchers using this approach instead of the traditional journal system to publish their work has been bolstered by the crucial support of major research funders around the world through white-labelled platforms such as Wellcome Open Research, Gates Open Research, HRB Open Research. These open research publishing platforms are controlled by the funders but run by us as a service provider.

New evaluation approaches needed

With these platforms, the venue of publication is no longer relevant as a proxy of article quality, and new approaches to evaluation are needed that better recognise the broader range of outputs that this ‘preprint–peer review’ route enables, as called for by DORA. New approaches are emerging that provide valuable output-level metrics beyond traditional indicators of reach, public engagement, use and reuse such as views, downloads, tweets, reference manager downloads and citations.

For example, preLights is a service in biology where early career researchers select preprints that they think are particularly important. Similarly, F1000Prime uses a large virtual Faculty of about 8000 experts to identify life science articles published anywhere – whether because they believe the article is a major scientific breakthrough, a robust data or resource, has implications for policy or practice, etc. This service also highlights many preprints, irrespective of where (or even if) they end up being ‘published’ elsewhere. Preprinting and giving control of the publication process back to the authors opens up a whole new world of opportunities towards a better and fairer system of publication, evaluation and recognition for researchers.

Maximising preprints

We are keen to work with ePMC and other providers of preprint-related services to develop the metadata standards that they rightly call for to maximise the benefits that indexing this content can provide to the research community. Of course this is part of a larger requirement for the development of better metadata standards across scholarly communication to reduce needless duplication of effort and admin and to maximise the benefits of truly open research. We hope that the leadership that ePMC have shown here with preprints will be followed by other major bibliographic indexers.