An idea whose time has come.

Orac has a post up about MacGyver science — you know, supercolliders made out of toilet rolls and chewing gum, or in this case an electrophoresis rig made out of kitchen stuff. Orac concludes, sadly, that it’s not a practical way to cut lab costs. He’s right, but there are good ways to cut lab costs.
(There are bad ways, too. I’ve done the grow-your-own Taq thing that RPM mentions in comments; it’s not worth it. Too much fiddling and no one else in the lab will trust their experiments to your crappy enzyme anyway.)
For instance, commenter Dave raises a good point about resource pooling. A colleague of mine, lab manager in the last lab I worked in, estimated that he saved the lab about 30% of its running costs just by instituting a central ordering system. Once all orders went through him, he could shop around for best prices and pool orders with other labs to save on shipping. The institute that lab was in also saved itself a ton of money by putting together a central Store, so they could buy in bulk.
(A brief digression. It occurs to me that most of my tens of readers won’t be familiar with what it costs to do biomed research. Quite apart from salaries, on-costs and infrastructure, I’d guess that most labs spend at least $500/month/staff member just on reagents and consumables (such as disposable plasticware). For a medium sized lab of five people, that’s $30K per year. On top of that, costs vary widely from experiment to experiment; for instance, the lab I’m in now probably spends at least a further $20K/year on facilities for transgenic mice. If, like Orac, you do a lot of qRT-PCR, that’s spendy too — I think it goes close to $0.5/reaction and “a lot” is thousands of reactions per month, if not per week. To take a less fine-grained view, the average cost of an NIH grant (1992-96) was $274,710/year. Them’s your dollars, taxpayers, so you should be keeping an eye on us — in fact, there’s a whole nother post — hell, a whole nother career — right there.)
Anyway, the whole point of this post is: there’s another kind of resource pooling that is due for an internet-era upgrade: simple “hey have you got an antibody against X?” sharing. A while back, my current PI came up with the idea of a central database for sharing biological reagents; it’s an idea best illustrated by example. (For non-scientists that is; labrats reading this will already be punching themselves and going “oh man why didn’t I think of that, does it already exist, where is it gimme gimme gimme”. Patience, I’ll get to it.)
I happen to be interested at the moment in a protein called Smad3. We had an antibody to the molecule, but I also wanted to be able to distinguish between the phosphorylated (active) and non-phosphorylated (inactive) forms. You can buy an anti-phospho-Smad3 antibody, but it’ll cost a bundle and you may be buying a lot more than you need. For instance, the one I linked comes in 40 µl lots for $110 (though most antibodies typically aren’t sold in such small lots; the 100 µl/$250 size is much more usual). The company says that’s enough for 4 blots, but I could probably stretch it to 40 — if I wanted to run 40 blots, that is. Until I ran the first experiment, I didn’t know whether I was going to pursue that line of inquiry, so I didn’t want to toss 110 hard-earned taxpayer dollars (plus shipping and handling, and you really get screwed on that believe me) at something that might not pan out. (Plus, I wasn’t too keen on the cross-reactivity with pSmad1, a related molecule, that the linked antibody displays.)
In such cases, and there are MANY, MANY such cases, what you typically do is wander forlornly around the building, asking if anyone has the antibody (or plasmid, or yeast strain, or oligo, or whatever it is) that you want. I did that — even sent a couple of emails to groups elsewhere on campus — but no luck. So I did the next thing you do, which is I ran a few searches and read a few papers, and discovered that there were a couple of antibodies in the literature that fit my requirements nicely. One of these was made in the lab of Prof Ed Leof at the Mayo Clinic; promisingly, it was cited in several papers by other groups (“the anti-pSmad3 antibody was a gift from Ed Leof”). So I sent Prof Leof email, and about 24 hours later someone in his lab sent me enough of his antibody for at least 100 blots (Prof Leof, Dr Edens — if you’re reading this, thanks again, and FYI the Ab can be re-used at least ten times, just put azide in the dilution buffer). All it cost our lab was FedEx shipping for a small container on dry ice.
Now, that’s the way it’s supposed to work — and in fact, in my experience, the majority of such requests are met with similar collegiality and generosity. For myself, I am always pleased when I can help a colleague out. But here’s the thing — there’s probably a lab right here on campus that has an antibody I could have used. I tried the obvious suspects (labs working on systems in which Smad3 might play a role), but even though they didn’t have any I bet there’s someone on campus who does. It’s even likely that they bit the bullet and coughed up for the antibody on spec, and it’s been sitting in their -80°C freezer since that first experiment didn’t go the way they hoped. That shit happens all the time, ask any researcher. I want to emphasize that: this whole example, from me wanting something for just one look-see experiment to the likelihood that it was available on campus but I just couldn’t find it, happens all the time.
Enter the idea whose time has come: an online database into which labs everywhere input the biological reagents they’re willing to share: antibodies, plasmids, viruses, bacteria, yeast, mutant model animals, peptides, oligos, primer sets, cytokines, spendy chemicals — the list of potential shareables is enormous and ever-expanding. Some of this functionality exists — for instance if the mouse you want already exists, Jackson Labs probably has it or knows about it, and you can always do the literature thing like I did — but it’s scattered and inefficient. Think how much easier my quest for an anti-pSmad3 antibody would have been made by such a tool: one search and up comes a list of labs and antibodies, pick an antibody, sort the resulting labs by location, email (or walk over to) the nearest one. Here’s another example: I have a new search going right now — I want some Smad2/3-null mouse embryo fibroblasts and a set of Smad2/3 expression plasmids. I’ve sent out seven or eight emails to colleagues I found in the literature; I’ve had one negative and one positive response, but the positive response depends on permission from someone else from whom I’m still waiting to hear. I’m still not sure I’m bothering the right people, it’s been almost a week (and Thanksgiving’s coming up), and dammit there’s probably someone in Portland, or maybe at the Hutch in Seattle, who has what I want and would share it with me.
See what I mean? Happens all the time. I want that database and I want it now! Peter suggested I make it happen, at least initially on a limited, local scale — start with the six labs in our institute, then expand to include the whole OHSU campus. Great idea, so as a first step I googled around to see whether anyone had already done it — turns out they have:

Welcome to BioRoot Bioinformatics
BioRoot is a non-profit organization dedicated to fostering communication, collaboration, and increased productivity in the biological sciences through information exchange.
We provide centralized databases to collect, store, and disseminate information about commonly used molecular biology reagents: antibodies, plasmids, strains, and oligonucleotides.
Use of these BioReagent databases will cut costs, save time, and accelerate research benefiting the bench scientist, the PI, and the public.

The guy behind it is David Nix, who clearly has the programming chops to go from “an Excel spreadsheet uploaded to my web space somewhere”, which is probably where I was going to start, to a fully-functional database complete with privacy/security measures. Major, major kudos, dude. (What is slightly odd is that I found out about BioRoot by googling, and found David the same way. Why haven’t I heard of this everywhere? It’s the best thing since PubMed, it should be huge. Apart from one subscriber-only article in The Scientist, I couldn’t find anything.)
I’ve sent BioRoot an email, so we’ll see how things work out. If it’s what it looks like (and why wouldn’t it be?), I’m going to become a hardcore BioRoot evangelist.

2 thoughts on “An idea whose time has come.

  1. Jim, I didn’t do it. I looked for a way to ping Orac but couldn’t find one. My guess is that he has a Technorati search set up to find links to each of his posts.

Comments are closed.