Spoiler: the experience of publishing a manuscript with PeerJ was very positive. My first paper published with them just arrived online: Fast and accurate semantic annotation of bioassays exploiting a hybrid of machine learning and user confirmation. This is a summary of the work that I was doing in San Francisco for the first half of summer 2014 (with Collaborative Drug Discovery), and it describes a hybrid machine learning/interactive method for marking up bioassay data, which is an optimised compromise between the two extremes of methods for taking plain text and turning it into semantically rich markup.

One of the contemporary issues in science that I grumble about fairly frequently is the journal publishing industry’s reluctance to migrate their business model to the information age. When computers weren’t invented or had yet to be connected to the internet, and the corpus of scientific knowledge was much smaller, journal publishers provided an essential service: centralising the evaluation and printing of manuscripts, and distribution of carefully distilled knowledge around the globe. Every major university had a large room full of shelves of leather-bound tomes that contained pretty much all of the good science journals from all fields from the days of Newton et al. The system worked, and companies like Elsevier, Springer and Wiley, and professional organisations like the American Chemical Society and the Royal Society of Chemistry, added a lot of value, and incurred substantial costs which had to be recovered, mainly from university libraries.

And then two things happened: the exponentially increasing number of papers broke the stagnant-or-declining university library budgets; and aroundabout the same time, internet access became available to most researchers, which meant that printing and shipping the journals became optional. As the submission and manuscript production systems became more heavily networked and automated, the administrative costs (theoretically) went down too. Eventually it got to the point when a humble scientist might ask the question: what do journal publishers actually do? Editors and peer reviewers work for free, the research is done by the authors, as is the manuscript preparation, and the publishers offer no assistance in reformatting according to arcane specifications (e.g. a different wierdo reference format for each journal), and it’s far from clear that the conversion of an incoming Microsoft Word document into the production-ready PDF file requires a large amount of time from highly paid professionals (and one or two of my experiences suggest that this step got outsourced to the lowest bidder). Somebody at the publishing house needs to sit down and monitor the incoming submissions manually, but it seems like one competent full time secretary could easily manage several dozen journals.

But of course that’s naive, and missing the point: the fact is that journal publishing as we know it is a legacy business, and is based on an obsolete cost structure. The companies and non-profit professional organisations have their existing infrastructure, which is based on the 19th & 20th century business model, when publishing was expensive. Which brings me to my next point of grumbling. In order to address the complaints of scientists about unreasonable per-paper viewing costs ($35-50 per view per article) for intellectual property that the sellers acquired for free, the journal publishers came up with something they called open access. This is terminology that makes me quite angry, because it attempts to associate the process with a similarly named social movement, open source. As someone who has been involved with the latter for a fair while (e.g. SketchEl), there is a very important distinction: when you create a software product and make it available to the world for free (as in speech), you can use any number of sites, such as the venerable SourceForge, and it will never cost you a cent. The nominal cost of distributing the software is compensated for by other means (e.g. advertising). Making research available for free (as in speech, also) ought to be the same, but unfortunately so-called open access is disingenuously named. A better term for it would be author pays, since there is typically a fee of $3000-5000 that the writers of the manuscript have to cough up in order to let everyone else read it for free (as in beer). Contrast this to, say, a columnist for a magazine who gets paid for creating the value proposition around which the product is based, and you might be starting to think that this sounds like a bit of a scam. Pragmatically, though, it’s not quite that nefarious, it’s just the way capitalism works during a disruptive transition phase. The publishers are resisting the market pressure to move away from their legacy cost structure, and are trying to prolong the agony by deflecting the costs onto the authors instead of the readers. On a personal level I find this to be particularly offensive, since these days my literature creation-to-consumption ratio is relatively high, due to the nature of my work. Pushing the burden of payment onto authors is also predicated on the notion that all practising scientists are backed up by generous government grants or patron companies with deep pockets, and makes no allowances for researchers in small cash strapped companies. Every dollar that I spend on author pays schemes is for all practical purposes a dollar that doesn’t find its way into my already unimpressive paycheque or an essential business expense.

Enter PeerJ. This is a startup company that provides all of the services that you would expect from a high end journal (i.e. a good reputation and high integrity peer review). Papers are free to read for everyone (open access), and while it is technically author pays, the authors do not pay very much. This is because the great folks at PeerJ do not have the costs associated with a legacy publishing company, and so they are free to keep their processes lean and slim and use as much automation as possible. The only reason it has taken me so long to make use of their services is that they focus on biology, which is unfortunate for me, because I’m a chemist. However, since I worked on a bioinformatics project this summer, this squeaked into the domain range, so I was quite excited about the opportunity to participate.

Their official philosophy of manuscript submission is that it should be easy and painless, and I found this claim to be valid. The submission process is quite a detailed sequence, but the web interface is very good, and the many different questions and categories make a lot of sense (e.g. the section requiring clarifications about which authors did what kinds of things). There are one or two corners that are a bit confusing, but that’s OK: the staff are very helpful, and respond quickly to email inquiries. This is as it should be, because PeerJ is building a system that can be much more highly automated than conventional publishers, and of course it can merely be almost-perfect in its early stages. Being responsive and attentive to the minor glitches is a characteristic that one might expect from a startup company that intends to be in it for the long haul.

All things considered, I’ve been a huge fan of PeerJ and what it seeks to accomplish from the moment I read about it, and now that I have my first paper with them, I’d make the same statement, but with experience to back it up. I very much hope that their experimental business model works out, and that they expand their domain in the general direction if cheminformatics, so I can publish with them more often.