“Oh BioRxiv, where have you been when I needed you in my scientific youth?”

Anonymous biologist, aged 41

The following is a blog rant in praise of preprints. I personally think preprint evolution is the best thing which happened to sciences, as opposed to Open Access (OA), which I meanwhile view as a major cock-up, a pig’s ear and a dog’s breakfast. Open Access movement may have started with the best of intentions, to break the injustice of subscription paywalls to publicly-funded research. Yet the Gold rush led to greediest and most unscrupulous of publishers profiteering most, and to scientific excrement being turned to Gold (both puns intended). Not just in “classical” predatory journals: my readers know what kind of “scientifically correct” peer-reviewed bunk science OA publisher Frontiers releases in exchange of a fee of up to $3k, applauded by the politicians, funders and certain kind of researchers. Scientific Reports is not much better, not even non-profit journals like PLOS One or those published by learned societies like RSC Advances are clean: in order to sustain the Gold OA income model, journals must publish as much as possible, fast, and on a minimal editorial team. This is where poorly done, short on data and even blatantly fake science gets a foot in the door. Paying customer is king, sometimes even if this customer is a fraudulent idiot.

Universities cannot unsubscribe from bad OA journals, they can’t do much to prevent authors from publishing there on public funds. Of course subscription-based publishing is also just as full of garbage and fraud, but this usual whataboutism from OA advocates is not really helpful. The problem is actually that thanks to internet and digital bundle subscriptions, bad journals cannot be unsubscribed anymore. Universities have to buy even lunatic quackery of Medical Hypotheses and Explore if they want to get access to more respectable of Elsevier journals. Universities lost all leverage against publishers on scientific quality, and a push for Gold OA only makes matters worse.

Open Access vs Open Science

Preprints are exactly the opposite of author-pays Gold OA publishing: they cost nothing to authors, in fact someone must subsidize the preprint platform so it can exist in the first place, and rather heavily if it is to have a minimum of quality control. This means full-time salaries for scientifically-trained personnel and remunerations for academic board members, because, here comes the inconvenient truth most journal editors will recognise: many of those “manuscripts” certain scientists try to get published should be thrown back at them with a brick wrapped inside.

This someone paying for preprint servers is usually a non-profit organisation or a charity, but there are commercial preprint platforms also, even Elsevier is in the business with SSRN (which makes one wonder what the secret game is here?: preprint metrics? “Premium” preprints for a fee?). If research funders were to mandate preprints, they would have to finance the preprint servers, for everyone in the world to use, which however will still be magnitudes cheaper than paying for Gold OA publishing.

Preprints are a probably the most important part of the larger Open Science transformation, where things previously encrypted and hidden in drawers come into the open: research data, codes, protocols, lab books, peer review reports, and with preprints, manuscript drafts. Everyone benefits here, even if they don’t see those benefits immediately. I will list those of preprints later in this blog post.

Open Access on the other hand is, when one thinks about it, not really part of Open Science at all. What OA is about, is making already published final papers free to read for everyone. There is no extra openness coming with it, which also becomes clearer given how the biggest OA operation yet, the European Plan S is designed: it only seeks to make papers free to read, while data sharing or preprints are explicitly not part of Plan S, and not planned to be.

There are of course OA journals which have open peer review and data sharing policies, like eLife. But then again, other OA journals offer little or even none of it, and Frontiers’ constant declamation that OA equals Open Science does not really make it true. Open Access per se does not give you access to research data or peer review reports, if an OA journal does provide that, it is because of its explicit support for Open Science.

The Costs of Knowledge

Meanwhile, the scientific community figured out that the costs to publish in OA might be problematic for their research budgets. Recently, a surprisingly naive idea was launched by the US ecologist Andrew Townsend Peterson and a long list of his peers, who issued this letter in a Wiley journal Biodiversity & Distributions. The occasion was that journal flipping from subscription-based model to Gold OA, which meant that now the authors will have to pay an article processing charge (APC) of $2200. You probably can’t wait to hear the ingenious solution Peterson et al proposed, right? Well, here it comes: Wiley should instead offer free-to-publish OA (called Platinum or Diamond) because the company earns so much money in scholarly publishing.

I am not sure how this should work. Should Wiley switch to charitable philanthropy just for this ecology journal, or for all their journals? What kind of economics do the learned authors of their letter envision for Wiley and other commercial publishers? I asked for clarifications, but Peterson did not reply to my email.

In any case, I never understood why scientists complain to be doing unpaid peer review and editorial tasks for commercial publishers like Wiley, Elsevier or Springer Nature. First of all, as I discussed here, peer review is actually part and parcel of academic job, and not a charitable activity professors do in their spare time. But one doesn’t have to peer review for commercial publishers indeed, there are plenty of non-profit publishers, some foundations, but mostly learned societies, where all income is reinvested into the relevant research field. Instead of begging Wiley for free stuff, go publish, peer review, sit on editorial board or become a member with one of your own relevant scholarly societies.

You can also peer review preprints, in the open as comments, or confidentially as emails to authors. Criticisms can prevent a bad preprint from being ever published as a proper paper (example here), or get a fraudulent preprint withdrawn (as happened here). There are several preprint servers in life sciences alone, though preprints are not listed in the standard search engine NCBI PubMed, being not peer reviewed.

BioRxiv is the best known preprint server in biology, it is also the strictest: only actual research is admissible, be it original research, re-analysis of published data, or reproducibility studies. Unsolicited literature reviews, opinion pieces, theories, hypotheses, unhinged rants against people or inanimate objects (yes, BioRxiv gets all kinds) are not admissible, through there were cases where publisher executives could preprint a white-paper (I discuss one example here). Other preprint servers are less strict, but then again, it is still less of a problem as when blog posts or advertisements pass peer review in “proper” journals.

How BioRxiv works

(according to John Inglis’ talk at the APE 2019 conference in Berlin).

When submitting your manuscript to BioRxiv, you must declare that it is not published anywhere else, and that you have the right and and permission to submit, from all contributors involved. This might create a problem if your principal investigator (PI) or other collaborators do not agree. But unless they have a valid point and your manuscript is indeed embarrassingly bad, do not get discouraged. Use all routes available to twist their arm to concede, in most cases you will succeed. Be constructive and always make the benefits of preprint clear, while stressing that it not at all intended to be the very final manuscript version. If your preprint contains clinical data, you must provide the clinical trial registration with ICMJE-approved registry.

Your preprint will appear online 24-48 hours later, but it must first pass some internal quality controls. Step 1 checks if submission requirements are met, and leads to rejection of writings which are utterly unrelated to life sciences or contain “non-science or pseudoscience”, or “images of human subjects”, due to privacy rights of research subjects. It also it cannot contain “Obscenity, defamation, plagiarism”.

The second step is when actual real scientists (“Affiliate principal investigators”) either green-light your preprint or send it back for further quality checks. It is not just about the minimal level of science, it is also about the information inside the preprint which can be abused if published: for pseudoscientific quackery, for criminal and terrorist activities, or, in case of ecology studies, for poaching and black market of wild life. This is why the following is sanctioned at BioRxiv:

• Dual-use research
• Articles about vaccine safety or infectious disease transmission
• Articles promoting or disputing specific drug regimens
• Articles about the toxicity/carcinogenicity of common substances

Now that you know how preprinting works, I will tell you why I think why preprints are so great.

Benefits of preprints:

1. Open Access. If mandated properly, preprints will provide immediate OA to all science, both to original manuscript draft as well as to final peer-reviewed author’s version, with only publisher’s copy editing and type-setting lacking. Every single paywalled paper will be available online verbatim as preprint, i.e., if everyone preprints. Elsevier denies access to machine-readable manuscript content, or references (because it sells Scopus services)? Not a problem, preprints contain exactly same text, data and references.
2. Research reliability. It is extremely difficult to retract a published paper, most journals just avoid the hustle and do nothing when faced with evidence of irreproducibility or even fraud. But then again, it is extremely easy to not publish a paper in the first place. If editors and peer reviewers are already aware of serious problems with a certain preprint, the paper will sure never make it into a decent journal (example here).
3. Scoop and rejection protection. Preprints establish priority, a journal should disregard them to allow a scoop will act unethically and can be taken to task for that. Also, the unsavoury game some scientists play in preventing the papers from peers they don’t like from seeing the light of the day, does not work when the preprint is out (example here).
4. Early career independence. PhD students and postdocs are utterly dependent on the goodwill of their PIs to publish their research works. Many publications are delayed or even prevented because PI want to maintain their lab’s impactful reputation, or develop the project for some more years (with a new first author), or simply punish the “trainee”, for personal reasons. Preprints allow a way out, where work can be published and used for CVs, authorship claims ascertained, and in most cases, PIs eventually do agree to preprinting, even if through gnashing teeth.
5. Stand to your science. No more hiding behind impact factor and “it passed peer review”. A preprint stands and falls with its objective quality. There are scientists who try to game their CVs or h-index with self-citing preprints (example here), but nobody will be impressed by that. Quite the opposite. Some preprints never make it to peer reviewed paper, which is a read-out in itself.
6. Medical advance. Preprints are actually vilified in clinical and preclinical research, because they allegedly pose mortal danger to patients by not being peer reviewed. As if peer reviewed, yet p-hacked, rigged or fraudulent papers in elite medical journals do not routinely lead to death and suffering, never mind the outright quackery published in predatory journals with their pretence of peer review. Patients are well capable to understand the concept of “not peer reviewed”, in fact it will make them extra cautious, or seek an external opinion from another doctor, or cross check in which journal the preprint was eventually published, or close the case if the preprint never made into any medical journal at all. On the other hand, much of bad medicine will be exposed at the preprint stage, before it makes in into a journal.

Plan U

While the world is debating whether Plan Smits, pardon, Plan S, is a self-imposed sabotage on own science, a conspiracy of commercial Gold OA publishers, a power dominance fantasy of a deluded former EU Commissioner with a Messiah complex, or all of that, the team of BioRxiv already proposed a solution, an alternative to Plan S:

Plan U is indeed simple: funders mandate preprints, and that’s it. Details can be added, for example that also the final accepted peer-reviewed author’s version must be posted as preprint, but otherwise no further interventions are required. Scholarly publishing landscape will simply not be able to remain the same when everybody preprints their papers before submitting them to journals for peer review. An evolution instead of misguided OA revolution will take place, subscription and OA might end up in peaceful and productive coexistence, ideally where no author pays any fees to publish. All the problems scholarly publishing is mired with, might even get resolved, for example:

• Elite journals: they operate as media outlets, with novelty and impact, whatever those are perceived to be, being the key selection criteria. Hot sensationalist papers are pushed through peer review, treated as top secret prior to publication, and trumpeted in all media worldwide the moment they go online. The novelty is lost with a preprint, and journals won’t be even able to dissuade authors from preprinting anymore, like eg Cell or Nature do now. As for impact: if a preprint was already plucked apart, journals will be careful with its sensationalist claims and have a pass.
• Predatory journals. The DORA Assessment‘s central argument that it doesn’t matter where a manuscript is published, but only its objective scientific quality, is being routinely abused to justify predatory publishing behaviour (example here). But with a preprint mandate, to publish same paper in a predatory journal is a direct admittance that it was so awful that no serious journal wanted it.

Otherwise, with Plan U many things become possible, and nothing will remain the same. This is exactly why a preprint mandate must be implemented, and only research funders can do that. One funder already started: the Chan-Zuckerberg Initiative, let’s hope more will join Plan U once Plan S is exposed for what it is: an anti-science cock-up which only Frontiers and some power-hungry Twitter bigmouths can love. Until then: preprint all your science voluntarily, you won’t regret it.

If you liked this blog post, consider donating. I will then reveal to you the secret Impact Factor of BioRxiv.

Update: incidentally, right after this blog post was published, BioRxiv released their full proposal for Plan U.

---