Monday, May 24, 2010

arXiv your paper!

In its 20-year history, arXiv has gone from a small physics pre-print repository to a giant archive widely used in many disciplines, and so trusted that it sometimes almost acts as a de facto journal [*]. Its growing body of literature on statistical phylogenetics is sure to boom anytime now.

The way I see it, placement of pre-prints on arXiv is a terrific idea. It (a) provides a good way to 'air out' a manuscript, and obtain feedback in case something is wrong, (b) you can cite your permanent arXiv article ID from the time of submission, (c) a version of the manuscript, unadulterated by (what may have been in your opinion) the unfair mauling it got in review, while still being able to correct errors. All the while, it does not interfere with peer-reviewed journal publication, and your manuscript is out instantly.

Most publishers' policies allow the archiving of pre-review manuscripts (including Nature, Science, PLoS, PNAS, PRSoc, Evolution, SystBiol, AmNat, etc.). Some allow post-review manuscripts to posted, as well. Individual journal policies can be checked at SHERPA/ROMEO, which also contains policies for personal/lab website posting and compliance data for funding agency requirements. My next paper is going here.

Try searching arXiv directly, go to their Populations and Evolution collection within Quantitative Biology, or click around to see and arbitrarily chosen sample author. The submission process is reasonable, and one can even submit PDFs generated from MSWord docs (as well as TeX files, and a couple of other formats).

* Perhaps most famously, it is the only place that hosts Grigori Perelman's three-part proof of the Poincaré conjecture [1,2,3; references listed for those who, unlike me, may be both interested and able to understand algebraic topology], which sits on arXiv without a formal peer-review process. He was eventually awarded the Fields Medal that he famously refused.

7 comments:

Dan Warren said...

Thanks for posting this! I knew arXiv was frequently used in math and physics, but didn't know that biologists were using it. I'm about to submit a paper, and I think I might actually arXiv it while I'm at it.

Roderic Page said...

Another nice pre-print server is Nature Precedings, which has a more biological focus, and contains pre-prints, presentations, and even theses (e.g., Brian O'Meara's doi:10.1038/npre.2008.2261.1). Precedings gives you either a Handle or a DOI for your submission (Handle for pre-prints that will be published in the peer-reviewed literature, DOI for anything else). You can also have document collections for papers/presentations from a conference. We'll be using this feature for iEvoBio.

Poletarac said...

Rod, that looks like an equally terrific idea, but I'd strongly prefer to have the repository be in public hands (either a society or a solid non-profit).

I personally dislike Nature's submission protocols (especially the Word template) and their Privacy Policy is far from ideal. I suspect that arXiv will never sink to this level: "We may share the information you provide with affiliates of Nature (i.e. companies in our group) and our joint marketing partners."

Do you think there is a chance that some society (SSB/SSE/ASN) or maybe PLoS could provide an alternative? An automated archiving site that enables pre-print posting and moderated commentary will both improve peer review and have a great shot at pre-eminence.

I really like arXiv's clear no-spam and cross-platform approach (it provides DOIs as well). It's something to be emulated.

Roderic Page said...

@Poletarac As with everything, the issue is "who pays?", or perhaps "how is it sustainable?" Nature is a publishing company, and they've invested money in Precedings, so I think it will be around for a while. A society would have to decide to fund something similar, and I guess they'd ask "why?" when resources already exist.

Note that Nature precedings will take PDFs, which is what I upload (usually produced via LaTeX). There's no Word template required that I'm away of.

Poletarac said...

All good points, Rod. There's no doubt money is a concern and Nature has the deep pockets, motivation, and thus staying power.

One thing that the societies and funding agencies may wish to consider is the fact that arXiv's cost effectiveness is amazing. Each year, the arXiv takes 60,000 submissions, which are downloaded 30,000,000 times. This works out to about US$7 per posting, or US¢1.4 per download [1].

The funding model is about to shift from all-Cornell support to more diffuse institutional funding. I suspect it would cost very little to us join in and ask for a few more rubrics within biology. Then again, I don't know, maybe expansion would somehow ruin it?

Whatever the case, I think that a public concern for dissemination of science is strongly preferable to a for-profit private one, especially when the cost/benefit works out so nicely.

Along these lines, I cannot imagine the scenario with GenBank/EMBL/DDBJ (with its many flaws) privately held. Why have literature go this way?

I am not at all convinced that my points are well-supported, save for the cost of arXiv. These are just some initial thoughts. I'd obviously appreciate further comments.

Anonymous said...

Based on what I've heard from people who were at a recent PLoS mtg, I think PLoS is looking into ways to accept papers in an arXiv fashion. They want to have the extra layers include how to read papers through the archive if they get rated/cited/promoted in some ways via a social media as well as peer reviewed mechanisms.

Not sure how far this is beyond planning discussions or just dreaming but seems like it could be useful.

PhDP said...

I'm a little late but anyway: arXiv is really great, it has done a lot for physics, and could do a lot for the theoretical/mathematical/statistical branches of biology, especially if it was used systematically.

The problem is; some journals consider arXiving a form of pre-publication. I was considering Ecol. Lett. for my first paper, and when I asked them if I could arXiv my paper, their answer was: "Under the terms of the copyright transfer agreement [arXiving] is not possible." Am. Nat. doesn't seem very open to the idea either.

We should write an open letter and arXiv it :P