How Useful Is F@H To The Public?

[Jeronimus]

Hello, i am thinking about joining this, but first i have a few questions which puzzled me.


Assuming we had a device that can scan the structure of already folded proteins on a atomic/molecular level:

Why would we then need f@h still? Would f@h give us any additional information and if yes, what kind of additional information would we get other than knowing the structure of the folded protein once it was folded successfully?

If such a device does not exist, then my question would be:

How do you know the proteins were folded correctly, as nature would have folded them according to the laws of physics?


Assuming the only extra information we get extra, is if the algorithm used to fold the proteins is accurate enough to map what nature/the universe does according to the laws of physics, then my question would be:

Does the public have access to the algorithm used and all it's updates? Is it open source? Or with other words. Could anyone start his own Folding@home v2 project using available open source software?


Because if a device exists which can scan *real* proteins on their folded structure, then f@h succeeding into folding a protein, the resulting structure is of little value to the public.
The only value would be in knowing the algorithm used to fold the protein.

[ChristianVirtual]

To my understanding the current F@H implementation is basically closed source; but you can find here an open source basic framework https://simtk.org/home/openMM (as also mentioned here http://folding.stanford.edu/home/the-software/ )
The closed source for FAH makes senses to ensure integrity of the results for their needs.

For the rest of your questions I need to pass to some more knowledgable members

[7im]

Most of the project is open sourced. See the Open Source FAQ on the F@h web site. The research data is publicly available by request to institutional sized organizations. And all the research results are publicly available in peer reviewed Papers, published In various scientific journals, copies also linked on the web site.

As for the machine question, it's kind of in between. They can get pictures of some biological folding, others they cannot. It's kind of like the problem of doing an autopsy on a live person. When they can do Crystallography, fah has matched those pictures very accurately, and has even won contests to prove it. See the news section of the web site.

[Zagen30]

Assuming we had a device that can scan the structure of already folded proteins on a atomic/molecular level:
Why would we then need f@h still? Would f@h give us any additional information and if yes, what kind of additional information would we get other than knowing the structure of the folded protein once it was folded successfully?

A large part of F@h's work is figuring out why proteins misfold, as that's the cause of many of the diseases that are being studied. I don't think you can get that information by scanning the folded structure of the protein in question.

[Jeronimus]

As for the machine question, it's kind of in between. They can get pictures of some biological folding, others they cannot. It's kind of like the problem of doing an autopsy on a live person. When they can do Crystallography, fah has matched those pictures very accurately, and has even won contests to prove it. See the news section of the web site.

That would make sense then, if not all proteins can be scanned adequately via high tech devices for their structure.

[Jeronimus]

A large part of F@h's work is figuring out why proteins misfold, as that's the cause of many of the diseases that are being studied. I don't think you can get that information by scanning the folded structure of the protein in question.

That makes sense, even though the "mis"fold might be just be one of the many possible outcomes as nothing on such a small scale is fully deterministic, so the algorithm used to fold probably also contains some kind of randomness in it, resulting in itself ending up "mis"folding a protein.

Whatever it is however, being able to watch the whole misfolding process step by step in slow motion, certainly will allow to make some guesses on what could possibly enable us to end up with more proteins folded the way we would like them.

[Jesse_V]

Just to clarify, a protein can fold via many, many different pathways.

This is a simplified example. See this link for the high-res version. The full results that came back from F@h contained 2,000 states.

Assuming we had a device that can scan the structure of already folded proteins on a atomic/molecular level:
Why would we then need f@h still? Would f@h give us any additional information and if yes, what kind of additional information would we get other than knowing the structure of the folded protein once it was folded successfully?

There's a big difference between knowing a protein's structure and knowing how it folds. In fact, protein structure prediction is a different scientific problem than the study of protein folding. Rosetta@home studies the first problem, Folding@home studies the second. Even if you knew the final structure, the information F@h can find is very important too. In certain environments and under certain conditions, proteins can misfold. If you know how they misfold, what causes them to do so, and the consequences for misfolding (such as the subsequent molecular processes that lead up to the disease), you can design drugs to help them fold properly.

If such a device does not exist, then my question would be:
How do you know the proteins were folded correctly, as nature would have folded them according to the laws of physics?

One way you can know that a protein folded correctly is that all the important pieces of the protein are located where they are supposed to be. The bits that are extremely reactive and important are stuck out, the pieces that are hydrophobic are tucked away inside the protein, etc. If the protein is misshapen, starts causing havoc with other proteins, or starts clumping up and forming dangerous filaments and structures, then you know that something is going on.

Assuming the only extra information we get extra, is if the algorithm used to fold the proteins is accurate enough to map what nature/the universe does according to the laws of physics, then my question would be:
Does the public have access to the algorithm used and all it's updates? Is it open source? Or with other words. Could anyone start his own Folding@home v2 project using available open source software?

F@h is very much based on open-source molecular dynamics projects. If you have the know-how, enough interest, and enough computational power, you could create and launch your own GROMACS projects if you wanted to. They wouldn't be on the F@h network however, since the servers would have to distribute them, compile the results, and there are non-trivial server-side work that is necessary too. Other research labs have joined F@h and they launch their own projects under whatever molecular dynamics engine they are interested in.

[PantherX]

...How do you know the proteins were folded correctly, as nature would have folded them according to the laws of physics?...

In addition to the explanations above, please note that sometimes, computational results are verified against lab experiments. If the results of the computer simulation match that which is obtained from the experiment, you know that everything is working fine.

PS -> Welcome to the F@H Forum.

[Jesse_V]

Yes indeed. Lab experiments don't reveal all the information, in fact they can be quite limited. Computational results are critical for getting more details, but they have to be demonstratably accurate and true to real life. Therefore is necessary to use the two together, which is what F@h does.

[codysluder]

Assuming we had a device that can scan the structure of already folded proteins on a atomic/molecular level:

Why would we then need f@h still? Would f@h give us any additional information and if yes, what kind of additional information would we get other than knowing the structure of the folded protein once it was folded successfully?

Several people have attempted to answer your "why" question, and that's the most important part anyway.

With respect to the "device" part of your question, the classical method of determining protein shape of any protein that can be purified and crystallized is by x-ray diffraction. Once science determines that there is more than one final shape, especially if one of causes a disease and one of which does not, you're led right back to the "why" question.

See also "prion"

[Jeronimus]

Several people have attempted to answer your "why" question, and that's the most important part anyway.

With respect to the "device" part of your question, the classical method of determining protein shape of any protein that can be purified and crystallized is by x-ray diffraction. Once science determines that there is more than one final shape, especially if one of causes a disease and one of which does not, you're led right back to the "why" question.

See also "prion"

I don't think there is a 'why' question. The reason proteins fold the way the fold is simply the laws of physics within our universe which allow the same initial protein to fold in various ways. The universe does not consider it a "misfold". It's just one of the many possible outcomes, each having a certain probability.


The question we are asking is "How can we change the protein or the environment it is within, to allow for a greater yield of what we consider "good" folds and less mis-folds?"

So by looking at it step by step in slow-motion, we might be able to figure out which state/states in the folding process are those where the proteins branch out to one of the many possible outcomes, and try to encourage or prevent such states depending on what we would like the outcome to be.
This would have to be done either directly by changing the initial protein in a way which grants more "good" folds, or by changing the environment the protein is in (biological nano-machines, chemicals, electromagnetic waves, pressure, temperature etc)


The algorithm itself remains the most valueable part of this project still in my view. At some point we might be able to just have some emulated DNA in a computer with the necessary bio-machines that work on it, and have a whole organism develop from it's very basic DNA.

If the algorithm was complex enough, given enough computational power in the future, even a human could develop out of it, and with it, emulated human level intelligence.
Granted, this is certainly not the most efficient way to get there, but possible in theory.