9 min readOpen Science Revolution – New Ways of Publishing Research in The Digital Age
London, UK — A massive increase in the power of digital technology over the past decade allows us today to publish any article, blog post or tweet in a matter of seconds.
Much of the information on the web is also free – newspapers are embracing open access to their articles and many websites are copyrighting their content under the Creative Commons licenses, most of which allow the re-use and sharing of the original work at no cost.
As opposed to this openness, science publishing is still lagging behind. Most of the scientific knowledge generated in the past two centuries is hidden behind a paywall, requiring an average reader to pay tens to hundreds of euros to access an original study report written by scientists.
Can we not do things differently?
An answer to this question led to the creation of a number of new concepts that emerged over the past few years. A range of innovative open online science platforms are now trying “to do things differently”, offering researchers alternative ways of publishing their discoveries, making the publishing process faster and more transparent.
Here is a handful of examples, implemented by three companies – a recently launched open access journal Research Ideas and Outcomes (RIO), an open publishing platform F1000Research from The Faculty of 1000 and a research and publishing network ScienceOpen. Each has something different to offer, yet all of them seem to agree that science research should be open and accessible to everyone.
New concept – publish all research outputs
While the two-centuries-old tradition of science publishing lives and dies on exposing only the final outcomes of a research project, the RIO journal suggests a different approach. If we can follow new stories online step by step as they unfold (something that journalists have figured out and use in live reporting), they say, why not apply similar principles to research projects?
“RIO is the first journal that aims at publishing the whole research cycle and definitely the first one, to my knowledge, that tries to do that across all science branches – all of humanities, social sciences, engineering and so on,” says a co-founder of the RIO journal, Prof. Lyubomir Penev, in an interview to Scicasts.
From the original project outline, to datasets, software and methodology, each part of the project can be published separately. “The writing platform ARPHA, which underpins RIO, handles the whole workflow – from the stage when you write the first letter, to the end,” explains Prof. Penev.
At an early stage, the writing process is closed from public view and researchers may invite their collaborators and peers to view their project, add data and contribute to its development. Scientists can choose to publish any part of their project as it progresses – they can submit to the open platform their research idea, hypothesis or a newly developed experimental protocol, alongside future datasets and whole final manuscripts.
Some intermediate research stages and preliminary results can also be submitted to the platform F1000Research, which developed their own online authoring tool F1000Workspace, similar to ARPHA. To date, they accept a range of different formats, such as preliminary data and software descriptions, as well as large datasets before the analysis has been completed. F1000Research also publishes clinical trial reports and medical case studies. Each of the above will have to be formatted according to internal guidelines and in its final shape will resemble a traditional research manuscript, with abstract, references and the conflict of interest statement.
Considering that most of science research is not freely accessible even in its final form, both platforms suggest that publishing early and intermediate stages in an open-access format could allow fellow researchers to find out what has been done and suggest the authors what could be improved.
Moreover, argues Managing Director of F1000Research Rebecca Lawrence, access to additional data generated within the project would make each study more substantial and transparent. “One of the big issues with science publishing today,” she says, “is the lack of data supporting the most important findings.”
However, the skeptical members of academic community respond to this proposal with one word: “copycats”. Many researchers are concerned about exposing their ideas prematurely, arguing that the current pace and competition among scientists makes them wary of exposing research ideas and preliminary results.
“What some people are not aware of,” responds to this a co-founder of ScienceOpen, Prof. Alexander Grossmann, “is that the tradition of posting research results prior to the publication of the manuscript in a journal already exists in the field of physics research. I’m talking about preprints. Literally these posts shouldn’t be called “preprints” because they rely on research that had been made public and citeable which means that it has been – published.”
The so-called preprints, or unpublished versions of complete academic manuscripts, have been a common practice among physicists for the past two decades. The largest archive of academic preprints in physics, mathematics, engineering and computer science, arXiv.org, was founded by Paul Ginsparg in 1991 and reached a milestone of 1 million papers in January 2015.
A similar concept also exists in the field of biological sciences and is implemented by such journals as PeerJ. The latter was established in 2013 and has published over 1300 preprints to date, about half of which have been submitted in the past nine months.
“Some researchers in other disciplines than Math or Physics still feel that their studies may be duplicated or misused if they post it on a publishing platform or repository prior to a journal publication,” continues Prof. Grossmann. “This attitude is surprising to me.”
“It’s the same as going to a conference – you can imagine someone seeing a poster, taking some notes, returning home and saying “Now I will do a couple more experiments and it will be accepted by Nature.” I think there is a 1 to 1000 chance or less that it will happen in a way which will harm the researcher who presented the original findings. There has been 1 million preprints published on arXiv for the past 24 years and I have never heard of such a story. The principle of posting or publishing new results immediately on a publishing platform could support the researcher to get priority for that work – if the publication has received a unique identifier, which makes it citeable.”
While preprints reflect completed studies and it may be difficult to perform all experiments within a short time frame, smaller-scale datasets and intermediate outputs are in more danger. To protect scientists from the risk of being “scooped”, RIO, F1000 and ScienceOpen assign each published piece with a Digital Object Identifier (DOI), an official citable reference that allows to prove the authorship of a research idea, method or final manuscript.
“When you publish your idea, you register your priority on it,” says the founder of RIO Prof. Lyubomir Penev. “Once you have published your idea, people cannot steal it so easily.”
So what can be published on these new platforms?
The scheme below is a brief round-up of items accepted for publication by RIO, F1000 and ScienceOpen, based on the official information presented in their article guidelines online. The list is not complete and more information can be found on the publishers’ individual websites.
F1000 commented in an interview to Scicasts that they can also accept grant proposals within their “Research Notes” format. F1000 also publishes software and method description articles in a format of “Method paper”. The “Method protocol” for RIO corresponds to a pure lab protocol, not necessarily supported by experimental findings or datasets. The “Software” published by RIO would be a program package or script, made available for download and/or installation through their platform.
All content at RIO, F1000 and ScienceOpen is published under open access Creative Commons licenses, which allow copying and redistribution of the material in any medium or format, given the appropriate credits.
Post-publication versus pre-publication peer review
The standard practice of anonymous and closed peer reviews has long been an issue of contempt for many researchers. With the arrival of social media and open blogs, a growing number of cases – from retraction scandals and scooping to the ‘sexist’ peer review that stormed the online community this spring – received wide public attention.
“The bias in the pre-publication peer-review process is another issue facing science publishing,” says Lawrence.
In an attempt to improve the evaluation system, many publishers and cross-publishing platforms now offer open commenting forums, such as PubMed Commons, and some decided to make peer reviews visible, once the paper has been accepted and published.
The new generation of publishers, however, take an entirely different approach.
“The fundamental principle of RIO is that all peer-reviews are open and public,” says Prof. Penev to Scicasts. He further explains: “We have two different peer-review stages – pre-submission and post-publication.
“The first stage happens during the authoring process – authors invite reviewers while they are still writing their article. Which saves a lot of time – reviewers can point out things that the author needs to correct so that the article is submitted to the journal already in a good shape.”
Immediate post-publication and pre-submission peer reviews are based on the new principle of authoring and writing manuscripts. Digital technology behind RIO, F1000Research and ScienceOpen allows scientists to create their project within the system, invite collaborators and collectively contribute to adding data or developing the final manuscript.
The initial stages of writing are closed from public view until the authors decide to submit their research to the open platform.
Rebecca Lawrence talks about F1000Research: “Publication of the first version of the article triggers the peer-review process and authors can suggest reviewers. The authors suggest their referees because we believe that the researchers know best who is fit to review their work. In addition to that, other published researchers who would provide their name and official email address can submit a peer-review for any study.”
The RIO journal has two stages of the review process – pre-submission and post-publication. Both F1000Research and ScienceOpen only allow post-publication peer-reviews. “We have an internal team of editors who would do a rapid simple check,” says Lawrence. “They will check if the article is from a reputable institution, that it’s not plagiarized and that it’s readable.”
With all three platforms, after a paper or project part has been published, the authors can take the reviewers’ comments on board and update their study with new results and figures. The new version would then be published as “Version #2”, followed by “Version #3” and so on. Each version receives a different DOI, linked to the original study, and the progress of the article can be traced through all stages.
Reviews are assigned their own DOI and can be cited as official published responses for each study. “What we find is that the referees within this setting tend to provide much more constructive criticism,” says Lawrence.
“Both the reviewers and the authors would probably be afraid to provide a bad review or a bad manuscript if they know that it will all be open and public,” adds Prof. Penev.
A possibility to speed-up the open peer-review process by inviting referees and the ability to update articles in real time could also prove useful when developing a research idea or a grant proposal, suggest new publishers. Dr. Jason McDermott, a Senior Research Scientist in computational biology and bioinformatics at Pacific Northwest National Laboratory in the US, in his blog post describes the experience of publishing preliminary results for his NIH grant proposal at F1000Research.
“I’ve been intrigued by this emerging publishing model and wanted to try it anyway, as an alternative to the usual journals,” he comments on the blog. “But this now also allowed me to reference the online version very soon after I uploaded it, and include it as a bona fide citation for my grant.”
The deadline for his grant application did not allow Dr. McDermott to implement all feedback from the F1000Research reviewers. Moreover, members of the NIH review team did not seem to be familiar with the new concept.
“A recently ‘published’ paper (the quotes are due to the fact that the paper was submitted to a journal that accepts and posts submissions even after peer review – F1000 Research),” one NIH reviewer named the citation of the openly published part of Dr. McDermott’s proposal.
“I think they saw the first version of the paper, read the paper comments, and figured that there were holes in the whole approach,” commented Dr. McDermott on his blog. “I will probably try this, or something like this, again.”
“I think that the views of biologists on preprints, post-publication review, and other ‘alternative’ publishing options are changing. Hopefully more biologists will start using these methods – because, frankly, in a lot of cases they make a lot more sense than the traditional closed-access, non-transparent peer review processes,” concludes Dr. McDermott.
How much does it cost?
Compared to paywall-protected research manuscripts, publishing open access would probably be more expensive, if we assume that the income from submissions is the only revenue (apart from advertising) that an open access journal can generate. Journals like eLife are still exceptions as they are connected to and financed by the funding agencies.
To conclude this case study, presented below are comparative costs of submitting an open access paper to publishing houses that still have subscription fees versus the costs of publishing at open-access-only platforms. For more detailed information, refer to individual publishers’ websites.
Please note that all these journals are published online-only.
The costs of publishing in RIO are only estimates. RIO will release its Article Publication Costs in the end of October and Prof. Penev commented that prices for typical manuscripts will prove lower than the approximate average indicated in the graph.
Is this the way into the future?
“Since the results of investigations financed by public money should be open to all interested parties, the evolution towards Open Access is inevitable,” says Prof. Guido Guidotti from Harvard University, in an interview to ScienceOpen. He recently published his study on ecto-enzymes on the platform and also supports the idea of post-publication peer reviews.
He continues: “In my opinion, reviews should not be anonymous because a reviewer should be prepared to support the remarks made about the paper in a straightforward and candid fashion, and not hide behind the shield of anonymity.
“However, there is the view that a paper should be vetted for accuracy before publication and it will take time to convince authors that transparent discussion after publication is preferable to anonymous pre-publication review. The experiment being done by ScienceOpen is essential in this endeavour.”
Whether these new concepts can uproot the traditional way of publishing research, is perhaps too early to decide. However, Rebecca Lawrence and Prof. Penev also shared with us some positive feedback from the academic community.
“The response [to RIO’s launch] exceeded all our expectations,” says Prof. Penev. “We expected a lot of researchers to answer the call but we didn’t expect so many positive and stimulating responses.”
“Researchers are also often faced with the dilemma,” adds Lawrence, “when they have a really good study, do they send it to a high impact journal and risk getting scooped while the paper goes through all stages of the pre-publication process, or do they just get it out there immediately and publish and receive post-publication peer reviews?
“A lot of prominent researchers told us that once they’ve tried this new way of publishing their research, they find it quite hard to go back to the traditional system.”
UPD This article was updated on September 25, to include additional comments from Prof. Alexander Grossman on the topic of preprints.
UPD2 This article was updated on September 25, to correct the spelling of F1000Workspace tool, which was originally mentioned as ‘Workplace’.
What’s your opinion? Scicasts is currently accepting comments and responses to this story. Please email us at firstname.lastname@example.org for more information on how to contribute. Our public commenting system is currently down so we will be updating this page manually, including all responses.