Why misfolds?
Posted: Thu Apr 21, 2011 12:28 pm
Hello, I am studying proteomics right now, and I have a question for you FAH experts:
Wikipedia declares:
(see wikipedia article "Protein_tertiary_structure")
<<No currently existing algorithm is yet able to consistently predict a proteins' tertiary or quaternary structure given only its primary structure; learning how to accurately predict the tertiary and quaternary structure of any protein given only its amino acid sequence and the pertinent cellular conditions would be a monumental achievement.>>
I understand that sometimes folding takes a lot of time because every time the protein flaps randomly against itself there is only a certain probability of it attaching to itself in the correct final shape.
However if this was the only problem, the only problem would be false negatives, i.e., no folding obtained.
I got the impression, by reading from elsewhere, that the real primary problem in ab-initio folding is the false-positive thing: a protein that folds in-silico in an incorrect folding. So you think it's correct but it's not. And that's why you need actual confirmation by X-ray cristallography.
Ok then, this is the question: why the in-silico folds are not validated with quantum mechanics simulations? Refer to
(See wikipedia article "Molecular_dynamics")
I understand that the whole folding simulation is way too long to be performed fully in quantum mechanics, however only a few tens of picoseconds of quantum mechanics simulation would be needed in order to validate the fold you have obtained with classical MD. You need to re-simulate only the, say, 50ps in which the protein flaps against itself and attaches to itself. If it really attaches, ok, if not, it was a false positive and the folding simulation has to be discarded/repeated...
Why isn't this done?
Thanks for your insight
Wikipedia declares:
(see wikipedia article "Protein_tertiary_structure")
<<No currently existing algorithm is yet able to consistently predict a proteins' tertiary or quaternary structure given only its primary structure; learning how to accurately predict the tertiary and quaternary structure of any protein given only its amino acid sequence and the pertinent cellular conditions would be a monumental achievement.>>
I understand that sometimes folding takes a lot of time because every time the protein flaps randomly against itself there is only a certain probability of it attaching to itself in the correct final shape.
However if this was the only problem, the only problem would be false negatives, i.e., no folding obtained.
I got the impression, by reading from elsewhere, that the real primary problem in ab-initio folding is the false-positive thing: a protein that folds in-silico in an incorrect folding. So you think it's correct but it's not. And that's why you need actual confirmation by X-ray cristallography.
Ok then, this is the question: why the in-silico folds are not validated with quantum mechanics simulations? Refer to
(See wikipedia article "Molecular_dynamics")
I understand that the whole folding simulation is way too long to be performed fully in quantum mechanics, however only a few tens of picoseconds of quantum mechanics simulation would be needed in order to validate the fold you have obtained with classical MD. You need to re-simulate only the, say, 50ps in which the protein flaps against itself and attaches to itself. If it really attaches, ok, if not, it was a false positive and the folding simulation has to be discarded/repeated...
Why isn't this done?
Thanks for your insight