Folding@Home Reports Success

msheppard writes "This Article describes how the folding@home distributed computing project is reporting that they used the data processed on client machines to "predict the folding rate and trajectory of the average molecule." Too bad Seti@Home hasn't had a hit yet."

Comments

Its just a shame

[dan dan the dna man]

that the sister project genome@home was so monumentally badly mismanaged that it effectively merged with folding@home a distinct project. I lost complete faith in the Pande group at that point along with a lot of other genome crunchers and switched all my CPU's back to SETI...

Re:Its just a shame

[Vijay Pande]

sorry to hear you say that. Our emphasis has always been on the science (hence we have important results whereas the other "science" d.c. projects don't). What I would say to you is the results speak for themselves.

Re:Its just a shame

[Vijay Pande]

Implicit solvation is a good question and deserves a long discourse. The bottom line in my mind is that while it's a pretty harsh approximation, it's still not too bad. Actually, our work shows that. However, we are moving to explicit solvation anyway to compare for sure.

I'm not sure how predictive the Nussinov work was considering the methodology.

Finally, I agree that we need to mine the trajectories more. We're doing that, but perhaps more importantly, we will release the data for others to mine too!

V

So many to choose from!

[TheGreenGoogler]

I hadn't realized how many distributed (grid) computing programs were out there... Check out Google Directory's list of links to distributed computing pages/projects here... Distributed Chess sounds very interesting!

Re:So many to choose from!

[De Lemming]

You can find even more projects here. And they're ordered in categories, including science, mathematics, puzzles and even art (I really like the Electric Sheep Project).

Bottomquark has reviewed a number of projects.

And here is a community site of people participating in such projects.

Re:So many to choose from!

[whereiswaldo]

I almost signed up for a distributed computing project until I read the terms very carefully.

What bothered me is they weren't responsible for damage to your computer, your data, they wouldn't guarantee the non-existence of virii, and they wouldn't tell you what kinds of things were being computed on your computer (it was processing power as a service - you'd get paid a measly amount to let them use your computer).

How was I to know they wouldn't be using MY computer to do things I'd consider unethical or illegal?

Whatever service you join, be SURE to read the fine print carefully beforehand.

Re:So many to choose from!

[das_cookie]

How about United Devices? They have a project that's helping find a cancer cure. Plus, they have a really cool screen saver of a model of the current protein you're working on floating around your screen.

Re:Other uses for Distributed Computing

[Suppafly]

I'm surprised google hasn't come out with a spider at home client which goes out and searches the web caching sites as it goes. Sure distributed computing could help their venture and who doesn't love google?

Article is incorrect...

[motardo]

in saying that it's the first distributed computing success. Look at the success of the distributed.net project, they just recently cracked rc5-64, and have cracked several other ciphers before.

Good CS, bad chemistry

[the gnat]

Okay, it's nice to see that distributed computing is finally becoming a useful tool. Nonetheless, I don't think there's anything particularly impressive about the biological results. The proteins they're folding are so small that most factors that affect the folding and conformation of the vast majority of proteins simply don't exist. When someone accurately predicts the structure of a normal globular protein at atomic resolution, I'll be impressed. When they can predict the structure of the F1F0 ATPase, then we can throw out crystallography- but it's not going to happen. (Ignoring for the moment that crystallography has it's own issues. . . at least it can show active sites and quaternary structure)

Don't get me wrong, the geek half of me thinks that what they're doing is very cool (and far more interesting/useful than Seti@Home). But I don't think it's very relevant to biology, and I doubt it'll ever replace traditional methods. Computers have almost unlimited potential as an aid to experimental structural biology, but in silico protein folding is still a pipe dream and a hand-waving exercise. The theory is really cool, the practical applications are nearly zero.

(Disclaimer: I don't have a PhD so I'm not very qualified in this field, but I do have a BS in biology and a fair amount of experience in programming and some knowledge of molecular simulation.)

Re:Good CS, bad chemistry

[the gnat]

They also have some prior knowledge of the tertiary structure of the protein they're simulating, by the way. So, all they've proved is that they can simulate the folding of a small protein and get results close to what experiment shows. This is a long way off from "we can predict the structure of proteins."

Besides, it tells you nothing about enzymatic activity, interaction with other proteins, interaction with small molecules and ions. . . these are not easily simulated, and certainly aren't covered by their current method. The software they're actually using is fairly standard (I don't mean this in a bad way), and isn't really the type of thing that would be useful for, say, docking experiments.

Frankly, I find David Baker's work at the U of Wasington to be far more impressive from the perspective of biology. (But still not good enough to replace Xray and NMR methods anytime before I retire.) He's not doing distributed computing though, so I guess it doesn't rise to Slashdot levels of sexiness.

Why do people download these blah@home clients?

[zaqattack911]

If SETI can get 100s of thousands of people do lend cpu cycles, and folding@home (which is MUCH more obscure than SETI) can get 30k people.

Then this phenomina should really be looked at by marketing people. It's amazing they can start a project, and just assume people will want to download their little client to use up cpu power ala screen saver.

What drives you people to use these clients? Why bother? And don't tell me it's cuz you want to do your part to find aliens :)

--Me

Seti?

[den_erpel]

Hm,

In the "Space" documentary series (hosted by Sam Neill), one researcher mentioned something about "except that one time".

Apparently they had some signal, but it was gone before they could reallign the dishes to get a confirmation.
I guess they ruled out possible "domestic" signals...

If anyone could guide me to a more elaborate source except that remark of one of the researchers, I would like to read it :)

Re:And so what if SETI did get a hit?

[msheppard]

If we get one Seti hit and we may be able to replace ALL of our current problems with new ones.

Earth: How do we cure cancer/every disease we know of?
ET: Use *this*, but now that you're living forever, you have to worry about massive overpopulation.
Earth: How do we get off this planet?
ET: Use *this*, but now you have to worry about war between your planets.
Earth: How do we achieve peace?
ET: Use *this*, but now you're bored outta your minds.

M@

Comment removed

[account_deleted]

Comment removed based on user account deletion

Good results

[ShooterNeo]

Well, on a related topic has anyone thought of this explanation for why SETI has found no results :

Most current radios, based on decades old tech, broadcast a very orderly signal. It is confined to a narrow band range, only one transmitter is allowed per channel, the data being transmitted is uncompressed and so has many repeating orderly patterns.

To increase capacity future radios will do the opposite.. They will broadcast compressed data that seems completely random, they will use a large swath of spectrum, they will repeat parts of the same signal across a large portion of the spectrum using a "chipping" algorithm. Even farther in the future, so many radios at once may be talking on the same spectrum that to identify a particular sender in order to communicate you'll have to use multiple antennaes and know his location (you'll share spectrum based on location).

What is the end result of advanced communication gear that intelligent minds develop? What is the optimal result? To an outside observer the signal will seem like pure, almost totally random noise. Only to the electronics of a particular receiver that has the correct encryption and chipping key will it seem like anything else.

THAT's why we can't hear anything. Trillions of sentient beings could communicate using this method and we wouldn't hear a thing.

Re:They could work on the screensaver version

[Vijay Pande]

in the latest clients, you can turn off the logos if you like. We've been pretty responsive to people's feature requests like that.

Re:The results were off by 20 percent!

[skeedlelee]

That's pretty good. Lab to lab variation when repeating measurements like this is usually this bad or worse. Factor of two would have been reasonable. When comparing in silico work to experimental its usually considered good when you're within a factor of ten. More commonly a series of related proteins is ranked in order (by some property) and then compared to experimental measurements of said rankings.

Also, keep in mind that this is a microsecond folding rate. Rates in the ms regime of folding are routinely measured with high accuracy, the microsecond regime is really hard as it usually takes longer than that to do what ever you're going to do to trigger folding in the first place. The number in the paper is 7.5us +/- 3.5us, so they got as close as could be expected.

Re:I fold at home

[Webmoth]

Go check out Yamaha Papercraft for some neat origami, including a few motorcycles.