  |
|
|
|
The protein folding problem is very significant in a variety
of fields of endeavor, including molecular biology, computational
chemistry, and even medical areas such as pathology. While
there are experimental methods available to determine protein
structure, it is clear that they will not be able to keep up
with the sheer volume forever. As more and more proteins are
discovered, it is becoming increasingly important for simulation
scientists to be able to lend a hand to their "wet" lab
counterparts in eliciting structure and functionality.
We've spent a couple of weeks reviewing some of the simulation
studies being conducted to examine protein folding, and this
week we'll conclude our discussion with a review of a fairly
comprehensive trade paper on the subject.
The introduction of the paper gives a general overview of the folding problem, with some mention to what has been attempted in the way of computational studies. In the second section, the authors present a more detailed discussion of protein folding, including an explanation of the structures involved. Armed with this introductory information, the authors go on to describe the basic computational approaches which might provide solutions to the problem. They conclude that any simulation which explicitly treats all solvent molecules would be too computationally overwhelming to complete with the resources currently available, while a simple lattice-type representation of a protein is not adequate to describe the important features. The remaining possibilities are all fairly intriguing. The first scenario involves treating the protein atomistically but representing the surrounding solvent as a continuous dielectric medium. The second suggestion is to treat the main chain of a peptide atomistically but then represent the side chains in some reduced form. (This is usually thought of as a unified model and can be imagined by thinking of a methane molecule. Methane consists of one carbon atom bonded to four hydrogens. A unified model will represent the entire methane as one point is space and give it a symbol such as "CH4". This is obviously not as rigorous as treating each atom as a distinct point, but the important properties are generally well-known for these small constituent molecular pieces.) The third possibility is the use of large protein data banks to calculate average properties which may be used in the simulation. The paper then goes on to choose among these methods before further developing the details of the simulations that Go To Page: 1 2
The copyright of the article Reviewing Protein Folding
in Scientific Computing is owned by . Permission to republish Reviewing Protein Folding
in print or online must be granted by the author in writing.
|
|
|
|
