On Preprints, and Beer and Tacos

A meeting (ASAPbio) was held on February 16–17 to discuss the role of preprints in biology, which sparked a New York Times Op-Ed by Amy Harmon under the provocative title “Handful of Biologists Went Rogue and Published Directly to Internet”¹ and an article in the Economist entitled “Taking the Online Medicine.” ² These articles expose to the public a need to fix a slow and antiquated publishing system that holds our community in a stranglehold, and the potential for scientists to regain control. Ron Vale, who held the position of ASCB President in 2012, writes about one tangible effort that shows how we could move forward.

—Peter Walter,
ASCB President

Ron Vale

In 2008, your ASCB President Peter Walter (together with Martin Raff and Alexander Johnson) published a Commentary in Science entitled “Painful Publishing.”³ The callout quote from the article was “The stress associated with publishing experimental results—a process that can take as long as obtaining the results in the first place—can drain much of the joy from practicing science.” The situation does not appear to have improved in the past eight years and arguably has gotten worse. Scientists feel pressure to publish in elite journals (which accept the same number or even fewer papers now than they did a couple of decades ago), spend their time dealing with journal rejections and often resubmit even their best work to multiple journals, deal with lengthy resubmissions to satisfy three or four referees, and shorten their increasingly massive studies to comply with the haiku-like formats of some journals. A recent poll conducted by ASAPbio (a grass roots science organization described below) revealed that 91% of 392 respondents were dissatisfied with the current state of publishing in biology (http://asapbio.org/survey).

As Walter and colleagues point out in their article, the current publishing system is affecting the culture of biology and the joy of science, making it less attractive as a career. But beyond “job satisfaction,” there are deeper issues at stake.4 First and foremost, scientists are funded by the public (e.g., the National Institutes of Health [NIH]) and private foundations to create knowledge for the public good. If new knowledge (the most powerful mechanism advancing science) is not being delivered in a timely manner, then scientific progress and potential downstream public good are delayed. Second, career advancement is linked to journal publications, and a slow system of publication means potential delays in career transitions. Over the past few decades, it has been taking longer for students and postdocs to complete their training and longer for junior faculty to receive their first NIH grants. While not the sole reason, increasing time to publication is likely contributing to the lengthening time to independence.

Addressing Problems

Scientists and funding agencies are beginning to reexamine our system of scientific communication, particularly with respect to: 1) open access (the ability to view scientific results by scientists worldwide without paywalls), 2) improving the accuracy and transparency of peer review, 3) improving metrics for evaluating scientific work, and 4) minimizing delays in scientific publication. Progress has been made in the first two issues, particularly by creating new journals with new models of access and review. ASCB has been a leader in these issues, particularly in addressing issue 3 through its role in drafting the San Francisco Declaration on Research Assessment (DORA).⁵ But overall, transformation has been slow because the ecosystem of biology journals is large and enormously complex (with each journal having its own policies, governance, and business model). Furthermore, scientists themselves have been complacent in improving their own standards and behaviors of evaluation and peer review. Improvements in evaluation and review must continue. But open access (topic 1) and minimizing publication delays (topic 4) could be more under the immediate control of scientists than we might think, as described below.

Preprints and ASAPbio

Physicists, mathematicians, and computer scientists use a mechanism called a “preprint” to communicate their findings prior to publication in a journal. Simply put, a preprint is a complete scientific manuscript (usually the same one being submitted to a journal) that is uploaded to a server. After a brief inspection to screen for “pseudoscience,” the manuscript is posted without cost to the author and can be viewed without charge by anyone in the world. Thus, transmission of a completed scientific study is fast, open, and free, although it is not validated through peer review.

The physicist Paul Ginsparg developed the server called arXiv 25 years ago, a time before the World Wide Web even existed in its present-day form. Housed at Cornell University, arXiv accepts 100,000 scientific communications per year (at a cost of ~$10 per communication). Virtually all physicists also submit their papers to a peer-reviewed journal; the journal publication provides validation of the work and helps in evaluations by promotion and grant committees. However, in their daily routine, physicists wake up in the morning and first check arXiv for the latest work coming out in their field. Work posted on arXiv also is universally acknowledged as establishing priority for a new discovery and is cited as such. With its 25 years of experience, arXiv provides a strong proof of principle of the impact of preprints as well as experience and data on issues that worry many biologists (e.g., the potential for scooping, posting of low-quality work, and dissemination of irreproducible work or pseudoscience, all of which will be discussed later).

Preprint servers exist in biology (e.g., bioRxiv, as well as several others); their current usage is relatively low but growing. A recent meeting called ASAPbio (Accelerating Science and Publication in biology) was held to explore the future of preprints in biology, discussing possible benefits and concerns. The meeting was held at the Howard Hughes Medical Institute in Chevy Chase, MD, on February 16 and17 and co-organized by Jessica Polka (whose meeting report appeared in the ASCB Newsletter⁶), Daniel Colón Ramos, Harold Varmus, and myself. Participants included a diverse group of junior and senior scientists, representatives of various funding agencies (public and private), senior editors of scientific journals, and thought leaders on data management. All of the talks and breakout sessions were live streamed to encourage broader community participation. A full record of the meeting is available on the ASAPbio website (http://asapbio.org/meeting-information). Meeting reports also have been written by Needhi Bhalla⁷ and Gary McDowell,⁸ two junior scientists, and another report that reflects the views of scientists, funders, and journals that emerged from the meeting will be submitted shortly to a journal. The meeting also was recently covered in the popular press, including the Economist² and the New York Times1 and in science journals such as Nature.⁹  The consensus of the meeting was that the wider adoption of preprints could have significant benefits for life sciences with relatively little downside, especially if combined with peer review through a journal. In this column, I discuss benefits and concerns regarding preprints and how the ASCB can participate in further developments and discussions.

Frame capture from a video on how to modernize publication of research results (produced by Youreka Science in collaboration with ASAPbio and iBiology; http://www.ibiology.org/ibioeducation/what-are-preprints.html)

The Roles of Preprints and Journals

“Beer and tacos—why not both?” exclaims my University of California, San Francisco, faculty colleague James Fraser to describe his attitude toward using both preprints and journals for communicating his scientific work. Preprints get his work disseminated rapidly (perhaps like the beer, which is served first and puts you in a good mood) and journals provide peer review (the sustenance of the tacos, which might take longer to get out the kitchen). If this California analogy does not work for the international ASCB community, then you can think of saki and sushi or perhaps Bordeaux and baguettes. Amusement aside, physicists have been using preprints and journals for years, with the two providing different services, as described here.

Preprints:

1. Your work is visible to the science community at the moment when you think it is ready to be released. The scientist is empowered with the ability and responsibility of dissemination of the work to her or his fellow scientists. There are no unpredictable delays due to serial journal rejections, prolonged efforts to satisfy referee requests, or very late stage journal rejection. If your work contains valuable information, the scientific enterprise benefits from it immediately.
2. Your science is accessible to everyone (including those in the developing world and at small colleges).
3. Your publicly released preprint establishes a time stamp and your evidence for a discovery (see later discussion). In contrast to other established modes of communicating science, such as a conference talk, a preprint constitutes a permanent, freely accessible record of your work. Therefore, it can be cited in the literature.
4. You can receive feedback on your work from a broader audience than just the two or three scientists who referee your work anonymously after journal submission. The added input can help you to improve your work for your final journal publication. You can also revise and update your preprint along the way, although the original and subsequent preprint versions are retained as permanent records of the work.
5. Your work is visible. If it is good, you can be recognized and invited to meetings, as organizers are always looking for new, exciting work. You can send your preprint link to a future employer or grant study section so that they can see and assess your most recent work. Such practices already happen in the physics community.
6. Scientists in your field might read your work. Yes, wouldn’t it be nice if biologists started actually reading scientific work in the form of preprints rather than just scanning for trophy journal names on a CV? Biologists will be forced to read preprints to keep up with their fields and perform their jobs as evaluators. Again, such practices happen in the physics community.

Journals:

1. Journals provide an infrastructure for peer review and quality control. While some question the benefits of the current peer-review mechanism, journals still provide the best-tested mechanism for assessing and improving scientific work. Journal-based peer review also remains the present-day gold standard of validation, which is required by other scientists (particularly those outside of one’s immediate field), universities and granting agencies, and the popular press. For these reasons, arXiv has not put journals out of business in physics, but rather the two have co-existed.
2. Your work can become more visible through a broad journal readership or added journal features (highlights, perspectives, etc). Preprint servers, by their nature, are democratic and do not highlight or provide higher visibility.
3. A paper can improve through revisions and editorial corrections.
4. A journal can provide assurance that the authors have complied with standards of the field for database depositions, conflict of interest disclosures, and other issues.
5. A journal can reformat the paper for easier human reading as a PDF and and facilitate increased automated data-mining (e.g., by adding metadata to figures or formatting citations).

Okay, beer and tacos sound good. But can I, a biologist, order both just like the physicists? Many life science journals, such as Molecular Biology of the Cell, the Journal of Cell Biology, eLife, and the Proceedings of the National Academy of Sciences, have preprint-friendly policies. Nature and Science, which have been working with physicists for years, also allow preprints. Some life science journals (e.g., Cell and its sister journals) have complex policies where editors ultimately decide whether the prospective author can submit a preprint or not. ASAPbio hopes to work with many journals over the coming year so that life scientists who want to use preprints will have more choices and more control. Furthermore, some journals are starting to work with bioRxiv to facilitate journal submission right from the preprint server itself. But what about the scientist who just wants tacos but no beer? That is fine—many scientists will choose to just submit to journals; preprints are a choice and not a mandate.

Concerns That Biologists Have about Preprints

Since preprints are new in biology, there are several concerns about possible unintended consequences:

1. If I post a preprint, will I get scooped by a competitor who then rushes his or her work to a journal? This is probably the fear most commonly articulated by biologists, since we are in an era where many scientists are working on similar problems. Interestingly, physicists have the 180-degree opposite view; they accept the preprint as a public disclosure that protects their discovery rather than making it more vulnerable to being scooped. I asked Paul Ginsparg about other scientists rushing out a journal article after seeing a preprint and then not citing it. He said that he was unaware of any such case and that physicists simply would not tolerate such behavior. It is simply too hard to ignore a manuscript that appears on arXiv. Biologists will likely come to the same view once they become more familiar with preprints and preprints become more widely adopted.
2. If someone posts a preprint before me on a similar work, am I scooped? Let’s consider a related question—if someone publishes a paper in a journal on a similar work, am I scooped? The situation is comparable and the answer could be yes or no, depending upon similarity and also the quality of the work. Both are disclosures of the work. You as the “second scientist” could potentially post a preprint relatively quickly if your work was complete. The community will then see that both studies appeared in a similar period of time, and thus were independently conducted work, and judge them on quality (and differences that are usually present and important for scientific progress). If the only option is a journal, the second work might be very difficult to get past a journal editor and, even if that is possible, there will be a long delay until it appears. It also might be possible for scientists who know of one another’s work to establish co-priority through an easily coordinated submission to a preprint server, which is more predictable than  coordinating publication in journals.
3. Will the low barrier of posting preprints result in the proliferation of incomplete and low-quality work? This has not been the case with arXiv. Scientists are motivated primarily by their reputation within the community. Publishing low-quality work tarnishes a scientist’s reputation. Posting a preprint reveals your scientific standards to your colleagues, without the filtration of a journal. A poor or rushed preprint is a public record. In contrast, a poor work transmitted to a journal is not public and there are no negative consequences of rejection. Ask yourself what type of work you would want to submit to a preprint server knowing that it will be broadly visible to all of your colleagues?
4. There is too much literature already and I cannot cope with the extra burden of preprints. Indeed, we are in an era of information overload. But the solution lies in developing better discovery tools rather than limiting information. If preprints become more widely used, we will need to develop better search mechanisms and easier ways to link together various elements of the same piece of scientific work (different versions of preprints, a final journal article, data existing on other servers).
5. A preprint server might become a haven for reports of “pseudoscience” by the anti-vaccine or other anti-science movements. This concern also exists in the physical sciences where climate change is challenged by anti-science forces. However, most pseudoscience papers have been blocked by moderators of arXiv; the few that have made it through have had minimal effect. Furthermore, preprints are acknowledged to be non-vetted due to lack of peer review. In contrast, journal publications can be wrongly construed as “fact” by the public and thus an inaccurate article can have more damaging consequences. For example, a connection between autism and vaccination was published in a prestigious journal, not on a preprint server. Indeed, preprints might even serve a useful role in uncovering obviously flawed studies before they reach the final stage of a publication in a high-profile journal.

In summary, as is true of many useful inventions, preprints come with both benefits and risks. But the experiences of arXiv have shown that benefits outweigh the risks and that the risks, while not absent, are manageable.

What’s Next?

Changing the practice and culture of scientific communication in biology will require several steps:

1. Identify a preprint mechanism that will best serve the life science community and that will be supported in a sustainable manner by funding agencies. ASAPbio will hold a one-day meeting in late spring of 2016 at the NIH that will include several public and private foundations (U.S. and international) to discuss these issues. News about the meeting will be relayed through the ASAPbio.org website.
2. Engage and inform the entire life sciences community—those conducting biomedical as well as non-biomedical research (e.g., ecology). Scientists, both junior and senior, will have to decide whether or not to use preprints as a form of communication. Currently, many biologists have either never heard of preprints or have incomplete or inaccurate information. Even among the ASAPbio meeting attendees who were primed with advance materials, we witnessed how concerns with preprints changed during the course of the breakout sessions after the exchange of information and discussion. The challenge now will be to expand the success of the ASAPbio meeting conversations (with 73 people in attendance) to hundreds of thousands of scientists worldwide. Town hall meetings and the ASAPbio website will help to tackle this problem. If you are interested in participating, please visit the ASAPbio website and become an Ambassador.
3. Work with funding agencies and universities worldwide to accept and use preprints. A first important goal will be for funding agencies to accept preprints as interim evidence of productivity in a grant application process. Similarly, university hiring and promotion committees should consider and encourage applicants to include preprints on their CVs, as evidence of recent achievement. These practices are common in physics, math, and computer science but not yet in biology.
4. Create a viable partnership between preprints and journals, as has occurred in the physics and mathematics communities. This effort will require working with journals to allow a paper that has been posted as a preprint to be given fair review and treatment at the journal.

Involvement of ASCB and Other Scientific Societies in Preprints

Scientific societies such as ASCB could play very important roles in the development of preprints in biology. Many societies publish scientific journals, and ASAPbio hopes that all journals in the life sciences will adopt a clear policy allowing scientists to post preprints of submitted manuscripts. Of note, ASCB’s journal Molecular Biology of the Cell embraces the use of preprints. Furthermore, preprints open up new opportunities for societies to participate in science communication, for example by engaging their membership in evaluating or highlighting work that appears on a preprint server.

Scientific societies also serve their members by communicating accurate information. Most biologists do not know about preprints or receive fragmented or incorrect information. Conveying accurate information and discussions on preprints through society communications (e.g., newsletters) and at an annual meeting are important actions. Societies also serve their members by crafting and advocating changes in science policy. As an example, DORA has been very influential. We hope that ASCB will catalyze its own discussions of preprints within its leadership and membership and work with ASAPbio in improving scientific communication.

In 2018, we will mark the 10th anniversary of Peter Walter’s “Painful Publishing” article. Let’s hope to have a faster, fairer, and more joyful system of research communication, especially for the many young scientists who deserve something better than what is being offered right now.

References

¹Harmon A (March 15, 2015). Handful of biologists
went rogue and published directly to Internet. New York
Times. http://nyti.ms/1XsYa6t.
²Anonymous (March 19, 2016). Taking the online
medicine: Old-fashioned ways of reporting new
discoveries are holding back medical research. Some
scientists are pushing for change. Economist.
http://econ.st/1XxSSa6.
³Raff M, Johnson A, Walter P (2008). Painful
publishing. Science 321, 36–36.
⁴Vale RD (2015). Accelerating scientific publication in
biology. Proc. Natl. Acad. Sci. USA 112, 13439–13446.
⁵Bertuzzi S, Drubin DG (2013). No shortcuts for
research assessment. Mol. Biol. Cell 24, 1505–1506.
⁶Polka J (2016). Consensus grows on preprints in
biology. ASCB Newsletter 39(2), 22–24. www.ascb.org/
consensus-grows-preprints-biology-2.
⁷Bhalla N (2016). Has the time come for preprints in
biology? bioRxiv 10.1101/045799.
⁸McDowell G (2016). Junior biomedical scientists
and preprints. F1000Research 5, 294 (doi: 10.12688/
f1000research.8216.1).
⁹Calloway E, Powell K (February 16, 2016). Biologists
urged to hug a preprint: ASAPbio meeting discusses the
ins and outs of posting work online before peer review.
Nature. http://bit.ly/24asJ69.

Questions and comments are welcome and
should be sent to president@ascb.org.

“ASCB,” “The American Society for Cell Biology,” “iBioSeminars,” “DORA,” and “Molecular Biology of the Cell”
are registered trademarks of The American Society for Cell Biology. “The Cell: An Image Library” is a common law
trademark of The American Society for Cell Biology.