David Baker's lab at the University of Washington has been working on several approaches to protein structure problems. I mentioned Rosetta@home here, and now the team has published an interesting paper on another one of their efforts, FoldIt.
That one, instead of being a large-scale passive computation effort, is more of an active process - in fact, it's active enough that it's designed as a game:
We hypothesized that human spatial reasoning could improve both the sampling of conformational space and the determination of when to pursue suboptimal conformations if the stochastic elements of the search were replaced with human decision making while retaining the deterministic Rosetta algorithms as user tools. We developed a multiplayer online game, Foldit, with the goal of producing accurate protein structure models through gameplay. Improperly folded protein conformations are posted online as puzzles for a fixed amount of time, during which players interactively reshape them in the direction they believe will lead to the highest score (the negative of the Rosetta energy). The player’s current status is shown, along with a leader board of other players, and groups of players working together, competing in the same puzzle.
So how's it working out? Pretty well, actually. It turns out that human players are willing to do more extensive rearrangements to the protein chains in the quest for lower energies than computational algorithms are. They're also better at evaluating which positions to start from. Both of these remind me of the differences between human chess play and machine play, as I understand them, and probably for quite similar reasons. Baker's team is trying to adapt the automated software to use some of the human-style approaches, when feasible.
There are several dozen participants who clearly seem to have done better in finding low-energy structures than the rest of the crowd. Interestingly, they're mostly not employed in the field, with "Business/Financial/Legal" making up the largest self-declared group in a wide range of fairly even-distributed categories. Compared to the "everyone who's played" set, the biggest difference is that there are far fewer students in the high-end group, proportionally. That group of better problem solvers also tends to be slightly more female (although both groups are still mostly men), definitely older (that loss of students again), and less well-stocked with college graduates and PhDs. Make of that what you will.
Their conclusion is worth thinking about, too:
The solution of challenging structure prediction problems by Foldit players demonstrates the considerable potential of a hybrid human–computer optimization framework in the form of a massively multiplayer game. The approach should be readily extendable to related problems, such as protein design and other scientific domains where human three-dimensional structural problem solving can be used. Our results indicate that scientific advancement is possible if even a small fraction of the energy that goes into playing computer games can be channelled into scientific discovery.
That's crossed my mind, too. In my more pessimistic moments, I've imagined the human race gradually entertaining itself to death, or at least to stasis, as our options for doing so become more and more compelling. (Reading Infinite Jest
a few years ago probably exacerbated such thinking). Perhaps this is one way out of that problem. I'm not sure that it's possible to make a game compelling enough when it's hooked up to some sort of useful gear train, but it's worth a try.
1. Data Police on August 9, 2010 10:51 AM writes...
Interesting! This relates to an experience with one of the quiz questions I gave when I was teaching Advanced Biochemistry lab many years ago.
Permalink to CommentOne of my bonus questions was to derive radioligand binding equation based on law of mass action. I intentionally changed some assumptions. The task was to derive. All, biology, pre-med, biochem and chemistry majors got it wrong. One got it right. That person was a lawyer! He was training to be a patent attorney.
2. A Nonny Mouse on August 9, 2010 11:22 AM writes...
How did you manage to get through Infinite Jest? Gravity's Rainbow (10 times read) and V (8 times), no problem, but this was too much!
Permalink to Comment3. Derek Lowe on August 9, 2010 12:12 PM writes...
I took it one chunk at a time, ANM, and I was never completely sure that I was going to make it through the whole thing. I will admit to skipping a few pages of the agonizing history of "le Jeu du Prochain Train", and the Eschaton game wasn't the easiest thing to get through, either.
But in retrospect, the massive early footnote on the Incandenza filmography is fair warning: if you can get through that, word for word, you can get through the book, I think. If you can't, you surely won't.
I found the book as a whole remarkably consistent in a confined, nightmarish way. When I first read Dante's "Inferno" (in translation, of course!) I had the feeling of the entire poem taking place inside an enormous stone-roofed room, cut off from the rest of the world. All of Infinite Jest took on some of that feeling for me, too.
Permalink to Comment4. RTW on August 9, 2010 12:57 PM writes...
Interesting concept. Might be on our way towards a world described by Sam Landstrom in his SiFi Novel MetaGame. This would be what he describes as a Grinder Game. In this futuristic world, all socio-economic activity revolves around game play. Your work is actual team based game play where solving of problems directly impacts your economic livelyhood. You gain and loose points as you do yuor job. Loose points based on purchase/rent of resources you utilize, greatest gain by solving the problems most economicaly.
The author is was a technical writer at Microsoft, taught software development, and programmed robots for DNA Sequencing. Additionally - its a free E-Book from Barns and Noble.
Permalink to Comment5. Proteinfreak on August 9, 2010 2:01 PM writes...
I wonder if the concept could also be extended to docking of small molecules to protein structures, not only with respect to changes in the compound conformation, but also to possible differences between the crystal structure and the protein conformation in solution? - Maybe it would then be possible to anticipate 'induced fit' scenarios, if the "players" get allowed to introduce movements in the protein backbone?
Another application could, in my opinion, be to improve on the success with clustering of molecules from hit series, i.e., clustering by motives rather than - as often happens - being "blinded" by functional groups that play no or only minor roles in the binding mode.
Permalink to Comment6. Wavefunction on August 9, 2010 2:12 PM writes...
The computational chemistry company OpenEye once organized a mini-contest where they asked us to manually dock kinase inhibitors into a crystal structure. Many of us did quite well and better than a couple of docking programs (although in this case everyone had a PhD and a med chem background of some kind). Would be interesting to package the same contest for laymen as a game and see how they perform.
Permalink to Comment7. LAM on August 9, 2010 3:18 PM writes...
Much of computational chemistry is already a game. So, what is new? Just a different set of guidelines of how to use the cursor and keyboard.
Permalink to Comment8. Aspirin on August 9, 2010 3:21 PM writes...
All of science is a great game. The stakes are high- an understanding of the universe itself!
Permalink to Comment9. ProteinCHemist on August 9, 2010 4:21 PM writes...
From a couple decades ago - "The Challenge" by Timothy Zahn from his Cascade Point Anthology. Briefly, Earth is deemed hopeless to ever achieve space due to its fractal society and penchant for gaming. In the end, a prominent gamer discovers the 'fun' of rocket propulsion theory. Gotta love it...
Permalink to Comment10. RM on August 9, 2010 4:29 PM writes...
Proteinfreak@5
Foldit isn't just limited to structure prediction. They've done some protein-ligand interaction puzzles too. This is with full backbone & sidechain flexibility, although control of the ligand conformation is a little clunky (although no more clunky than control of the sidechain conformation).
However, as I remember, the protein-ligand puzzles were primarily "design" ones, where they were looking for mutants which would improve binding, as opposed to binding predictions. But there's doesn't appear to be a reason why they couldn't do ligand docking predictions too.
The key limitation to all of this, of course, is the scoring function. Players are out to improve their score, and if the in-game score doesn't match reality (e.g. if the lowest energy structure isn't the highest scoring one), there's a limit to what the game can produce.
Permalink to Comment11. Jose on August 9, 2010 7:08 PM writes...
OT, but "Infinite Jest" and "Cryptonomicon" are simply derivative doggerel- both novels should pay royalties to Pynchon for (poor) style infringement.
Permalink to Comment12. A Nonny Mouse on August 10, 2010 3:24 AM writes...
I'm about page 350 of IJ and have been stuck there for some years; I take it on trips with the intention of doing something, but can't force myself to pick it up. Having said that, I still haven't finished Vineland which I have had since it came out (haven't even started Mason & Dixon).
As for the subject in hand, I haven't been able to get any of this to work with the ATI 5770 which is in my son's computer; pity as it's a damn good card.
Permalink to Comment13. chris on August 10, 2010 6:43 AM writes...
Nicola Zonta is developing an environment for human steered computation approaches to drug design. Zodiac, http://www.zeden.org/ To date they have implemented haptic device support so you can feel when the molecule bumps into the receptor, and wiimote head tracking for 3D visulisation.
Permalink to Comment14. Anon 2 on August 10, 2010 9:20 AM writes...
chris #13:
Sounds like a standard docking program adapted to be a Wii-like platform.
Permalink to CommentGame indeed.
Scientific progress---not.
15. PAL on August 10, 2010 9:52 AM writes...
#13, #14: Sounds like modeling for those who are mathematically challenged or disinclined.
Permalink to Comment16. andy on August 10, 2010 2:44 PM writes...
It's really not too surprising that the lawyers and accountants are up there. The last thing I want to do after a day of protein modelling is come home and play a protein modelling game which, although quite breathtakingly well designed, is a lower resolution, zoomed-out format of what I did during the day.
Give me a good lawyer sim, however, and I may happily potter away at it for a few hours.
Permalink to Comment