Ok, so there are some press-releases popping up now from the University of Washington, about a new video game that is making some bold claims (or at least has some bold goals), their have made a video game where us mere mortals are allowed to contort proteins into 3D shapes in hopes of creating something that is of actual use (link to game website).
The game is based on the actual physics of proteins, and the hope is that gamers around the world will embrace it and gain a sixth sense of sorts to building proteins in a good way, in fact, David Baker, one of the biochemists working on the project, says that his 13 year old son is better at the in-game protein bending then he is. So i guess they’re hoping that kids will gain an intuitive understanding of the physics involved and contort the proteins accordingly, utilizing the natural 3D puzzle solving skills that we humans have.
You might think that it’d be easier to just have a computer have a go at brute-forcing it’s way through it (try every permutation and see what works best), well they’ve already tried that and the number of different proteins needed to be tested are ridiculous, according to the article, it would still take centuries to solve even if all the computers in the world worked on it. So this is why they are trying to come up with an innovative way of finding new breakthroughs in medical science.
I tried downloading the game and playing it, and it was fairly easy to get started and work through the introduction part that explains to you the basic tools at your disposal (even worked fine with wine on my Linux computer). I do have serious doubts about the hefty amount of buzzwords used in their press-release though, take for example this line from one of the scientists:
Our ultimate goal is to have ordinary people play the game and eventually be candidates for winning the Nobel Prize.
Plus of course the mention of perhaps curing HIV with it. This is obviously just shrewd PR work, as it is something that makes for quite catchy titles (”Will the next noble prize winner be a 12 year old gamer?”, “Will gaming nerds cure HIV?” etc), but it almost always seems that science press-releases with too many buzz-words are too good to be true.
I really wish i knew more about biology/chemistry so I’d be more qualified to comment on this whole thing (being a physicist myself), as i DO find it to be a very interesting idea, regardless of my distaste for their PR campaign. They claim that very soon they will be holding competitions for people to make proteins that might be of actual use in the medical industry and such, where the top proteins in the competition will actually be produced and tested in petri dishes at Dr. Bakers lab (and given credit if they publish an article on it).
I’m curious though, how they actually intend on picking winners out of possibly thousands of protein structures submitted, as they don’t know beforehand exactly what protein will solve whatever problem it is they have at hand, but that might just be skepticism from someone that does not grasp this field well enough.
May 10th, 2008 at 1:24 pm
Well, one thought is that they already have a simulation model for testing whether a given protein has an effect on certain other target proteins. However, the brute force algorithm for generating proteins to test might have been too inefficient — or simply take too long, given the huge search space — to be worth using. So by using humans to generate testable proteins, they’re hoping to get a representative subset of that search space based on pattern-matching inuition of the human brain that can’t really be programmed into computer algorithms yet.
Maybe. ;)
May 10th, 2008 at 3:19 pm
Yeah i think that’s right, my main problem was how they would know which proteins are better then others? How will they judge from the 1000 submissions which 3 are worthy of being produced if they don’t know what the solution is beforehand? This might just be my ignorance in biology/chemistry speaking, but it seems kind of weird to me.
May 11th, 2008 at 10:20 pm
By knowing the exact shape of the protein, one can deduct the way it will interact with other proteins (the hull of a bacterium, for example). It’s well known among chemists which form a proteine consisting of a given set of amino acids will take on: that of the lowest energy. The score you get in the game is based on how low the energy of your protein is. So the guy (or gal) who has the highest score has found a likely shape of the proteine.
May 12th, 2008 at 10:25 pm
Cool Ben, thanks for enlightening me. I figured it was something in that sense, as they wouldn’t have made the game without being able to figure that out in a fairly easy way. I just didn’t really understand how :)