folding@home

swampy — Wed, 11 Apr 2007 15:39:00 +0100

This may be old news to some people but the folding@home project run by stanford university (see http://folding.stanford.edu/) has been making some serious progress recently and I've been so impressed by its potentia.

For those that don't know folding@home is a 'massively parrallel distributed computing network' that aims to use spare domestic computer power to process really tough sums in order to work out what "protein folding" is all about and why it happens. When this happens in biology it has genetic consequences and has been attributed to the onset of Alzeimer's, Parkinson's and other degenerative diseases.

In the last year the project has come up with the following conclusion. Given (taken from one of their published papers) :-

" ...that the confinement-induced decrease in solvent entropy is directly responsible for the generally hydrophilic nature of confining spaces encountered by proteins. The decreased helical propensity of [the] peptide with decreasing vessel size, and thus decreasing solvent entropy, is analogous to the hydrophilic destabilization reported for this same peptide under “hydrophilic titration” computational experiments "

We can draw the conclusion that:-

" On the basis of the physical arguments discussed above, we suggest that the role of solvent entropy in protein confinement is likely a general phenomenon, as hydrogen bonding and the hydrophobic effect are both driven by solvent entropy. Thus, by greatly affecting solvent entropy, confinement directly alters many of the commonly held rules of protein stability. "

Which seemingly means that water plays an important role in confining protein spaces or something or like that. I'm no expert but I think their pleased with the outcome!

ANYWAY... The upshot of all this is that by running a program on your PC that crunches numbers in the background when you are not using it (it runs at a very low priority meaning it won't affect other things you do) you can help cure diseases.

This is the kind of thing that universities, research labs and goverments buy multi-million pound "super-computers" for. To give you some perspective as to how powerful this project is consider that the fastest computer in the world is IBM's BlueGene/L owned by the US department of energy (see http://domino.research.ibm.com/comm/research_projects.nsf/pages/bluegene.index.html) - also used for protein research. BlueGene/L has a capacity of 280.6 TeraFlops. That means it can process 280.6 trillion calculations per second. Oh and it cost $100m. The last time the statistics were in, folding@home was running at 567 TeraFlops - twice as powerful!! As more people join in with the project, it just keeps going up.

In an extraordinarily well-spirited move by Sony the Playstation 3 also comes preloaded with software that allows people to participate. Turns out the Playstation 3 with it's incredibly powerful Cell processor is amazingly well suited to this kind of number crunching. Since the PS3's release folding@home's power has doubled (see http://fah-web.stanford.edu/cgi-bin/main.py?qtype=osstats). So if you have a PS3 seriously consider taking part - it's trivally easy and all you have to do is leave it running when you aren't playing games.

If you don't have a Playstation but a PC at home then you can still take part (don't run it at work it'll only get me into bother). You can also form "teams" that consolidate the effort. For example Virgin Atlantic have a team (see http://fah-web.stanford.edu/daily_team_summary.txt) but they aren't doing very well. What about a Virgin Radio team?

swampy blog on Absolute Radio

folding@home