Thursday, January 17, 2013

Blog Syn: B.S. to cut through the B.S.

Something I find very exciting happened today: the inaugural post of online synthetic peer-review community Blog Syn. Head over and check it out; remember also to subscribe to it (you'll note no exorbitant subscription fees).

The website is a rapid, international, and open source of peer-review of the synthetic literature focused especially on reproducibility. Essentially, it's a curated set of data: chemists point out interesting or provocative procedures, other chemists try the procedures and document the process and results, and then the efforts are cataloged and published for broader review and comment by the chemically-inclined portion of the internet.

The amount of "junk" in the literature is a well-established problem (if every procedure you've ever tried has worked as advertised, congrats, though). Sometimes you can get difficult literature procedures to work if you discover a variable that wasn't included in the original source (e.g. source of the material, trace impurities, level of rigor in keeping things dry, heating method, etc.) -- but then, since you're merely reproducing a literature procedure, this detail often doesn't make it into your subsequent paper ("Material X was prepared according as described by Author Z."). Moreover, since the publication timescale is slow and synthetic corrections/retractions are rare, lots of money might be wasted on trying to repeat useless procedures (counterpoint: some procedures work really well, and it's in everyone's best interest to hear of those, too!). 

Blog Syn is a really exciting idea, and I think it fills a niche neglected by the current traditional literature. I'll elaborate.

Blog Syn's closest analog is Org Syn, of course (hence the quasi-tongue-in-cheek name). Org Syn is a great journal if you ignore the clunky online interface. It's based on reproducibility. Proposals that are accepted are checked at least twice in a reviewer's lab using meticulous detail supplied by the authors about reagent quality, purification method, etc. The procedure must be reproducible to within 5% yield as written. A tall order--how many chemical steps in the literature would survive those strict guidelines? Accordingly, Org Syn has a good reputation: if you can't reproduce a procedure, it's likely an issue on your end, not on the literature side.

However, Org Syn isn't a panacea for problematic protocols. Procedures are submitted intentionally by authors; Org Syn doesn't review existing publications. Hence, there's self-selection. Authors choose to send things they know will (likely) work. Sketchy stuff is submitted elsewhere.

That's where Blog Syn has an advantage. It's democratic--any interested chemist with the time and materials can potentially contribute. It's an actual dialogue--those without access to the means to carry out experiments can still suggest techniques and comment on the data. And it's rapid. It's very rapid. Take the first post for example, which was compiled less than a month after the article was made available as an ASAP (many of the experiments were done before the article was even assigned a page number). Hence, before the broader chemical community became aware of the article, it had been vetted, including a discussion with the authors.

That's cool. 

There's been a lot of support from bloggers, including inaugural contributors B.R.S.M., Organometallica, and  Matt Katcher, project starter See Arr Oh, and chemblogging king Derek Lowe. There's been some doubt in the comments of the blogs, as well: chiefly about who verifies that the checking was done correctly, and how the whole thing will operate organizationally. I think these concerns are fairly minor hurdles; with sufficient checkers, the number of trials is way over the standard n = 1 for publication. The idea, as I understand it, is for a small group of volunteers to check a reaction that has sufficient interest and feasibility.

I think it would be helpful and important to get some PIs on board. There's two reasons: (1) a PI might become irate if they discover their student burning time and reagents behind their back (after all, reagents and NMR time cost grant money); however, a PI who gives their blessing to a student to conduct trials can contribute resources; and (2) there's a lot of chemists who still doubt that non-traditional publication venues (read: the Internet) offer anything of value; while I think Blog Syn will prove itself, having some PIs contribute might be quite transformative in shaping the face of peer review--a combination of open access and acknowledgement of social media. Attaching "established" names could mollify the perceived connotation of anonymity/sketchiness that lies with blogging in the eyes of many academics.

Anyhow, it'll be quite fascinating to see how it shapes up. Maybe it won't go much further. Or maybe it'll force authors to be more accountable for the science they preach.