Saturday, May 24, 2008

I am so a scientician, and so can you.

So my brother turned me onto FoldIt not too long ago. It's a game that allows the player to fold real proteins and scores them based on stability. A number of different things go into protein stability and how the protein is folded effects all these. At its core, it's a whole bunch of complex chemistry that I haven't thought about since High School. In my high school chemistry class, we didn't cover anything more complicated than crystalline structures based on a single type of molecule, so the whole protein thing is way more complicated than what I've studied.

Anyway, there are big, powerful computers that are devoted to this sort of thing, but it's still a time-consuming and expensive process. The project that released FoldIt has a theory that if they can better understand human problem-solving strategies, which are intuitive for us, they can program computers to emulate some of them and solve protein-folding problems more effectively. They also think humans might beat computers to the solutions of some proteins using this program.

Most of the users, at least those that talk about themselves, are in math or science fields. I think there aren't too many biologists, because it's too much like what a lot of them spent a summer doing at some point during their education, but the point is that most of the people playing with this thing have some background. My background is theatre. The program represents degrees of stability and other pieces of information mainly with color-coding, and with a couple of other graphics. Their web site explains it better.

So now for the ego portion of the blog. I was up way too late last night playing with one of the proteins. Something to do with lipids. I managed to get all the way to 7th place, out of a few hundred players. I think that 98th percentile, give or take, is pretty good for not having a clue what I'm doing. I pretty much just have visual cues to go on, and some recollection of the material that this is supposed to represent, from the tiny bit of organic chemistry covered at the tail end of my class in high school. As of now, I'm in 11th, with 9,117 points. For context, most puzzles will go to scores in the mid-8000s just by telling the computer to shake the sidechains and then wiggle the backbone. Basically, telling it to automatically find the lowest-energy state of the sidechains, which is time-consuming but relatively easy, and then mess with the configuration of the backbone. While it sounds like this is allowing the computer to solve the whole puzzle automatically, it doesn't really make any positive or creative choices.

I think it's interesting because it puts a complex scientific pursuit in the context of being a puzzle game, with competitive scoring on the internet. While I don't think it would hurt to include a bit more of the science, it's interesting to me that such a simulation could be represented in such an accessible way.

No comments: