There has been a revolution in how scientific discoveries are being reported. More and more scientists and researchers are now uploading early drafts of their research papers to online repositories called preprint servers. Traditionally, a preprint is a version of a research article that is made available to other researchers before formal peer review takes place. Preprints have been commonly available in certain disciplines for decades; for example, the arXiv preprint server was started in 1991 and has served as a popular repository for preprints in physics, astronomy, mathematics, and related fields.
In the past several years, there has been a proliferation of preprint servers across many disciplines, including chemistry (ChemRxiv), earth science (EarthArXiv), and biomedicine (bioRxiv and medRxiv). There has also been an increase in multidisciplinary preprint servers hosted by commercial and nonprofit third parties, such as Research Square and the Center for Open Science. This explosion in preprint servers has had various effects, some positive and some controversial, with convincing arguments on both sides. However, one effect that is not being argued is the impact preprints are having on scientific journal workflows. At first a disrupting force, preprints are assimilating into journal workflows while also evolving those workflows in interesting ways. HighWire, a subsidiary of MPS, has been able to participate in this evolution through our relationship with Cold Spring Harbor Laboratory, the host of bioRxiv and medRxiv. These preprint servers use HighWire’s BenchPress submission system to receive and screen preprints and HighWire’s hosting platform to display the preprints and associated comments.
It has long been established that the scientific record is made up of research articles that have been submitted to peer review journals, that those articles have been through a rigorous peer review process, and that an editor or editorial board has “accepted” the research paper, resulting in formal publication in an established scientific periodical. These research articles reference other research articles that have been through the same process, and, in turn, these articles, once published, can be cited. There are stringent rules governing research journal workflow. A research paper cannot be submitted for consideration to more than one journal at a time, which means that a researcher may have to go through the peer review process multiple times at multiple journals before their research gets published. This means that it can take months or even years before a piece of research becomes part of the scientific record, which frustrates scientists.
Preprints have upset this process by introducing research into the scientific record before peer review takes place. They have grown in popularity as an accepted way to expose research early in the research cycle and to establish provenance over a scholarly concept or scientific advancement. Over the past ten years, many journals have altered their policies to allow researchers to submit articles that were originally posted as preprints, which previously many forbade. In fact, processes and protocols have been developed to automate the transfer of manuscripts from preprint servers to journal submission systems. The Manuscript Exchange Common Approach (MECA) protocol is one such mechanism. Developed by a group of publishing workflow vendors, including HighWire, MECA is used to transfer research articles, including commentary and peer review, from system to system, in an automated, machine-readable way.
The Covid-19 pandemic drew more attention to the use of the preprint. As the world recognized the severity of the pandemic, many scholarly publishers made virus-related research freely available to readers. At the same time, preprint servers, such as bioRxiv and medRxiv, began to receive thousands of preprints from scientists eager to help battle the pandemic by getting their findings distributed quickly, without waiting for the prolonged journal process. The resulting avalanche of Covid-related studies included many that were helpful and some that were not. To help readers sift through these, various groups created preprint peer review services. Among the first was a collaboration between MIT Press and the University of California, Berkeley called Rapid Reviews: COVID-19 (RR:C19), which sought to rapidly evaluate Covid research once it was made public. The editorial team at RR:C19 would identify preprints on Covid, organize peer review of those papers, and also direct authors to appropriate publishing opportunities, including their own overlay journal.
This model has now expanded in different ways, producing different types of innovative publishing workflows. There are several services available that perform independent peer review of preprints. Review Commons offers peer review to preprint authors and will then send the reviews, along with the authors’ comments, to a preprint server or on to a peer review journal. Peer Community In (PCI) builds communities of researchers in various disciplines who review each other’s research papers, resulting in a valid, citable article or in a research paper that can be sent along to a traditional journal. These are just two of many examples of preprint peer review that have emerged to help legitimize preprints. A comprehensive list, called ReimagineReview, has been compiled by ASAPbio, a nonprofit organization promoting the evolution of the peer review process.
Another interesting innovation being used in preprint peer review is an XML framework called DocMaps, which captures the editorial process used in evaluating a piece of research. DocMaps was developed by a consortium of publishers, representatives from preprint servers, service providers, and transparent peer review advocates, with the aim of capturing editorial processes and events in a machine-readable way so that the workflow details can travel with the article and be displayed for the purpose of making the underlying editorial process transparent. The DocMaps XML is available alongside the article XML and can be read by any compatible system. This means that as an article moves through a publishing workflow, each system it encounters can add information to the DocMap, building an editorial history that can be later exposed to the reader.
A great end-to-end example of this innovative workflow using third-party peer review services, DocMaps, and MECA has been implemented by Cold Spring Harbor Laboratory on their bioRxiv and medRxiv preprint servers. A researcher who posts their research on one of these preprint servers can choose to have the preprint reviewed by a preprint peer review service, like Review Commons, or it can be selected and reviewed by a service like PCI. Once reviewed, the information about the review is added to the DocMap, which is accessible at the appropriate preprint server. Once the reviews are received, they are posted alongside the preprint. In addition, the preprint and associated peer reviews can be sent along to a traditional journal. As more traditional journals are exposed to this workflow, preprint peer review will likely be seen as a useful augmentation, and may even displace the current peer review workflow, shortening the time between submission and publication.
At HighWire Press, it has been fascinating to watch and be part of this preprint revolution. Working with partners like Cold Spring Harbor Laboratory, we have been able to transform our workflow systems to support evolving forms of peer review and to accommodate innovation in the editorial process. The need for flexibility has driven innovation in our workflow platforms in other ways as well. In anticipation of even more disruption in publishing workflow, we are now modernizing our systems so that they are more modular and microservice-based, which means they are capable of handling new processes without extensive development. This will enable us to respond even more rapidly to the inevitable changes in both technology and workflow and allow us to bring innovation to our partner publishers.
Tony Alves, senior v-p of product management at HighWire Press
