Robot does 12 human-years of trial testing in one week.

November 14, 2013 11:18 PM

Robot scientists! A Pretty cool video about research automation from the Wall Street Journal. [slyt]
posted by Lutoslawski (10 comments total)

A few years ago a friend of mine who worked in this sort of lab explained his job to me and I was amazed to realize that this phase of pharmaceutical research is basically just mixing a whole bunch of stuff together to see if anything happens. Nine-year-old me might have been emboldened to a career in science if I'd known that.
posted by XMLicious at 11:42 PM on November 14, 2013

this phase of pharmaceutical research is basically just mixing a whole bunch of stuff together to see if anything happens.

True. But this is not an attempt to discover new drugs. It's an attempt to discover new uses for existing drugs. And you can't do that unless you develop those drugs in the first place. It's only possible because someone else--a lot of someones, really--did the hard work of coming up with the 450k drugs that are currently being used in these tests. Not to mention the hard work of coming up with the robotic system in the first place.
posted by valkyryn at 3:18 AM on November 15, 2013

First they came for the automobile assembly line jobs, but I didn't complain because I wasn't an automobile assembly line worker...
posted by The 10th Regiment of Foot at 4:58 AM on November 15, 2013 [1 favorite]

Here is a great example of what government can do for you. What private entity would invest the money in a system like this that could potentially give an edge to a competitor by finding another application for their drug? What private company would willingly give their drug formulations to another private company to do this kind of testing if it were possible that the testing company could find a new use for it and steal it away from them? The government funds these efforts because others will not, and it ends up helping the public (new cancer drug keeps you from dying) and the industry (new application for drug opens new markets).

Tax dollars, bitches.
posted by backseatpilot at 5:23 AM on November 15, 2013 [3 favorites]

But this is not an attempt to discover new drugs.

Not in the instance of this video, but it's a similar approach to a lot of drug discovery: seeking to inhibit (or promote) a particular enzyme, gene, or protein, and using robots to pipette thousands of different compounds to test for an effective one.
posted by rocketman at 5:48 AM on November 15, 2013

At Janelia Farm (a HHMI research campus), they have a robot that does complete maintenance of their mind-boggling number of fruit flies. (Video is low-quality, sorry, but robot is super cool!)
posted by Maecenas at 6:59 AM on November 15, 2013

I can't begin to describe how incredible this seems to me, and it is the perfect example of automation done right. There is no conceivable way that these new drug applications would be discovered, aside from brute force luck, and it enables scientists to analyze results and propose test protocols instead of using pipettes all day long.

I would have liked for them to elaborate a little on what they mean toward the end about the self learning aspects of the robotics. They threw it out there but I can't imagine offhand some way that the system learns from failure. Anyone have any idea or reasonable conjecture?
posted by dgran at 7:16 AM on November 15, 2013

I spend a lot of time in screening labs. 10-15 years ago you'd see maybe 30 gainfully employed middle class research associates with B.S. degrees working in each one. These days there's 1-2 middle or upper middle class Ph.Ds supervising the machines. Good for making new medicines (I guess, the pipelines at most pharmas are running dry and new new drugs are increasingly biologics not discovered through compound screens), but bad for the many people on my LinkedIn contact list who've been struggling to find long-term employment for the past few years.
posted by Thoughtcrime at 9:28 AM on November 15, 2013

Ah, I see the "robot apocalypse" thread posted further down the front page now. I'll leave this thread to all of you welcoming our new robotic (or the increasingly few capitalists who own them) overlords.
posted by Thoughtcrime at 9:33 AM on November 15, 2013

But this is not an attempt to discover new drugs. It's an attempt to discover new uses for existing drugs.

Watch the video again - he spends most of the time talking about "potential drugs". It's only at the end that he says that if one of the people with these untreatable diseases is lucky, the substance that shows a promising reaction will be an existing drug already approved by the FDA because they can skip a bunch of the steps and go directly to testing in humans.
posted by XMLicious at 4:10 PM on November 15, 2013