3 April 2013 by William

open access  Is the time right for a preprint server for life science?Academics in physics, economics, or math often think that life scientists (like myself) are weird because life science doesn’t have a preprint server. Life science is a fast-paced discipline, but there’s no place where the latest research can be found, discussed, and where the primacy of results can be established. There’s a lot of value in life science research (the reproducible subset, that is) but instead of staking your claim to a finding shortly after you get the data, many researchers feel like they have to write a polished paper, submit it to a prestigious journal, and wait nerve-wracking months to years for the process of review, rejection, resubmission to finally make their results available to a subset of others in their field.As submission-to-publication times grow, fears of someone else getting there first grow and there are often accusations of “anonymous” reviewers asking for more experiments, just to delay the publication of a paper from a competing lab. What can be done about this? As frustrating as these problems are, it’s worth remembering that they’re relatively recent in origin. Pre-publication peer review as a critical part of the publication has only become standard in the last half of the 20th century. As Jason Hoyt notes in a post at Scientific American, Watson & Crick’s paper in Nature wasn’t peer reviewed.

So it might seem like Peerj’s announcement today of a preprint server for life science is a good idea. Indeed, some ecologists and bioinformatics folks already use Arxiv, but as some of you will surely point out, we’ve tried preprint servers before. Nature had Nature Precedings back in 2007, and that didn’t really go anywhere. PeerJ (see our founder interview) would like to convince you that this time it’s different, because Nature weren’t really fully behind Precedings and because it didn’t really seem like it fit anywhere. Although it got thousands of submissions, perhaps the time was just a bit too early as well. Now there are millions more scientists online and we have ways of tracking the attention that a pre-print gets in addition to only citations of final published versions.

So I think that now might be the right time, and PeerJ might be the right people to do it. What do you think? Below I’ve listed some interesting points for discussion.

Benefits of a pre-print server:

It gets your work out immediately, drastically reducing the chance of being scooped unless the other lab legitimately gets the result first. It weakens the role of underhanded tactics.

It improves the final published work. Instead of having work improved by being bounced around by reviewers and editors from journal to journal until it finds a home, the reviews can happen one place, and this also means that the reviews, the “dark matter” of scholarly output, can finally become part of the scholarly record.

It’s just the right thing to do. It’s part of the social contract of doing research in academia that you disseminate your work, and a preprint server allows you to do this as broadly and cheaply as possible.

Risks:

There’s a lot of crazy stuff on Arxiv. Let’s be honest, there are thousands of people who claim to have discovered flaws in the theory of relativity or the secret to perpetual motion. This preprint server will no doubt attract a similar level of attention from people with cures for cancer and herbal remedies that will end all disease and bring about world peace. If they can’t figure out a way to keep that stuff out, or at least moderated appropriately, it’s going to be hard to build broader acceptance.

There may also be concerns about intellectual property. Maybe the slower process of publication works better for discoveries that have commercial value? PeerJ has said that they are working on a more granular set of privacy controls which may address some of these issues.

Some journals won’t publish a paper if there’s a pre-print. This is changing in biology, but still is an issue for some.

Why would a researcher spend time preparing a document for the preprint server or commenting on articles there if they could spend time working on a submission that’s more likely to help them get their next grant or tenure? Altmetrics can help, but can PeerJ get enough of the right articles in front of the right people at the right time to elicit feedback that’s useful to both the author and the reader? Will there be some way for a reader to know that a critical comment was left by a competitor or a glowing comment was left by a co-worker? How will the reputation that scientists have in the real world be made apparent to readers?

Tags: , ,

10 Responses to “Is the time right for a preprint server for life science?”

  1. Mike Taylor Says:

    “Nature had Nature Precedings back in 2007, and that didn’t really go anywhere.”

    Actually, I dispute that. It was working just fine — it was averaging 1000 preprints a year, or three a day, which is very respectable. Nature never explained why they shut it down — all we ever got was “unsustainable as it was originally conceived” which is literally almost devoid of meaning.

    I think NPG let the community down very badly there — not just by closing up Preceedings but by not even having the basic decency to explain why they made that choice and what had changed in the five years it was running. (Though to give credit while it’s due, it is still there in read-only mode.)

  2. Mike Taylor Says:

    BTW — “reviews, the “dark matter” of scholarly output” — very nicely put!

  3. Peter Binfield Says:

    Thanks William,

    To address some of your points:

    We do have a staff moderation step to provide a basic level of vetting (i.e. to ensure that submissions meet our various policies and meet our scope etc – see: https://peerj.com/about/preprints/scope-and-instructions/ and https://peerj.com/about/preprints/policies-and-procedures/). We wont be allowing anything which makes therapeutic claims for example, and we do expect people to adhere to certain ‘discipline norms’. However, I think that only time (and experience) will show what kinds of submissions we receive and whether additional checks are needed.

    As to IP concerns – you are right that we will be introducing some levels of access control, so hopefully people will feel reassured by those access controls being in place.

    As to whether or not other journals will accept articles which have previously appeared as a preprint. Clearly there are a large number of journals which *do* accept preprints (including, of course, PeerJ) and presumably, if an author was sufficiently interested to submit a preprint then they probably wouldnt be interested in submitting it to the type of journal which doesnt accept articles which previously appeared as a preprint…

    And finally, we will certainly be developing functionality that facilitates the kind of evaluation issues you raise.

    Pete

  4. Dr. Allison L. Stelling Says:

    Love the XKCD!

    OK so I have done a bunch of photophysics research through the ACS system (JACS, Biochemistry, and J.Chem.Phys.B.), although nowadays I do optical methods for brain tumor diagnostics.

    I have seen things in bio peer review that any decent Chemistry Editor would not abide by.

    I’m Working On It. Keep it up! This is good stuff.

  5. William Says:

    Pete, thanks so much for dropping in to clarify those points. I wish you the best with this.

    Mike – I’m glad you like my bon mot. Pete and Jason say the same about Precedings in their SciAm post that I linked to, so I didn’t mean to slam it, just to say it isn’t currently a service researchers can use.

    @DrStelling – Keep fighting the good fight.

  6. Dr. Allison L. Stelling Says:

    @ William

    One is doing one’s best. Think balance must be reached b/w applied journals that need quick, subscription access review (like Optics Letters etc.), and general audience stuff like Nature or Science.

    Not every story is a Nature/Science story! And that’s OK! :)

  7. Joel J. Adamson Says:

    “There’s a lot of crazy stuff on Arxiv. Let’s be honest…” Really? Where is it? I don’t doubt that the stuff you list is there, however I don’t see it. I check arXiv qbio regularly and I don’t see anything “crazy.” In other words, the crazy is easily detected and easily ignored. It doesn’t taint the quality of everything else there.

    Sounds like you’re concerned that having “crazy stuff” in a preprint server would diminish the credibility of the preprint server, its associated journal (PeerJ) and the other work there. I don’t see credibility as a main draw for the server itself, since I can assess credibility myself using the standard means (looking up the authors and checking the list of citations is what I usually do). The main reason for using a preprint server is just to get the work out there.

    Joel

  8. Tim Says:

    We make this point in a paper to appear in PLoS Biol: http://figshare.com/articles/The_case_for_open_preprints_in_biology/655710

  9. William Says:

    The blog post I linked to shows one example. I’m not saying crazy stuff doesn’t make it into Pubmed, just that there is crazy stuff in Arxiv. Troll around hep-ph a little while and you’ll see what I mean. However, whereas a paper in Arxiv saying Einstein was wrong about relativity because fractals is different from a paper saying the flu shot causes cancer or bathing in used motor oil is healthy. It has the potential to directly impact a lot more people’s lives. As Peter said, PeerJ has policies about things like this, so they’re on the right track, but they are going to have to watch it closely.

  10. Dr. Allison L. Stelling Says:

    Heh so according to the theoretical physicist I’m dating, astrophysicists are … are bit different in how they are trained and in how they do the Maths.

    I am not allowed to say anymore. People– please for to let your scientists speak!