Folding Forum

How does F@H work, How does it help?
I want more detail.
I want to know more about it.

Start here http://folding.stanford.edu/ and follow the links for more and more details. You'll find details at whatever level of scientific complexity that you're comfortable with.

I cant find exactly how it works though. does any one have a link or can they explain?

Protein folding is an N-body problem. Wikipedia has more information on this type of mathematical problems.

Basically you have bodies (atoms) with mass that interact through forces (e.g. electrostatic forces). Each atom interacts with all the others. So what they do is they simulate very small time steps, accumulate all the forces on each atom and move the atoms according to Newton's laws. That is done several million (billion?) times to cover a timespan of a several microseconds. If you're lucky, the protein has changed shape from one conformation to another.

Those shape changes are then analyzed with statistical tools. For this they need lots of samples (which is what we compute on the CPUs and GPUs). From analyzing this large number of samples researchers gather information how proteins change shapes and in what time frame. And if they are lucky, they find the reason why (and under which circumstances) certain protein take unwanted shapes which lead to disease. And that again might lead to development of new medicine.

Take this with a grain of salt, as I am an engineer of electrical engineering. So I can tell you how a CPU or GPU works, but I know diddly squat about protein folding

cbuchner1 is fundamentally correct, but he left out a couple of very significant details.

First, the force fields are a lot trickier at the atomic level than they are at the distances that we used when we did experiments in physics lab. (Please don't press me for details -- my knowledge of atomic level chemistry/physics is very limited, too.)

Second, there's the the issue of Brownian motion. The atoms only follow Newton's Laws at a very low temperature and protein folding takes place at body temperature or there-abouts. The statistical tools get applied before, after, and during the simulation calculation.

Codysluder is fundamentall correct, too. Please read http://folding.stanford.edu/English/Science and if that's too easy for you, you can start on the technical papers http://folding.stanford.edu/English/Papers. When I read one of the technical papers, there's an awful lot that I don't understand, but I often understand a small fraction of what they're trying to say and that helps in making more sense of what is going on than before I read the paper.

cbuchner1 wrote:Protein folding is an N-body problem. Wikipedia has more information on this type of mathematical problems.

Basically you have bodies (atoms) with mass that interact through forces (e.g. electrostatic forces). Each atom interacts with all the others. So what they do is they simulate very small time steps, accumulate all the forces on each atom and move the atoms according to Newton's laws. That is done several million (billion?) times to cover a timespan of a several microseconds. If you're lucky, the protein has changed shape from one conformation to another.

Those shape changes are then analyzed with statistical tools. For this they need lots of samples (which is what we compute on the CPUs and GPUs). From analyzing this large number of samples researchers gather information how proteins change shapes and in what time frame. And if they are lucky, they find the reason why (and under which circumstances) certain protein take unwanted shapes which lead to disease. And that again might lead to development of new medicine.

Take this with a grain of salt, as I am an engineer of electrical engineering. So I can tell you how a CPU or GPU works, but I know diddly squat about protein folding

A-.

Very nice work.



Dan

F@H is not only about protein folding study but also on simulation models study. Here is my take on what they are trying to study : viewtopic.php?p=49324#p49324

At least, I got a solid A from Dan on my try :p