Date: Fri, 20 Sep 2002 12:01:39 -0700 (PDT)
From: Thomas Cheatham
Subject: Re: Dynamic Movement of Molecule


> I am simulating DNA binding protein on Amber6. This protein moves
> its pair of arms very dynamically, and seems to hold DNA.

This seems like an excellent application for molecular dynamics (MD)
simulation assuming that the motion of the arms is "fast" compared to the
MD time scale which currently (and realistically) limits accurate
simulation to the 1-100 ns time scale unless you have massive
computational resources.

> Question is about the equation<a href=http://www.amber.ucsf.edu/amber/eqn.txt> here</a>.
> Is the first part of V uses some approximation? If so, dynamic movement
> of my protein would be nonsense?

It seems to me that you need to learn a bit more about biomolecular
simulation and the underlying approximations inherent. I would suggest
looking at:

A Leach "Molecular Modeling: Principles and Applications" 2nd edition
(2001) Prentice-Hall

...broad overview of computational chemistry


F Jensen "An introduction to Quantum Chemistry" (1998) Wiley

...some molecular mechanics but mostly QM


Allen & Tildesley "Computer Simulation of Liquids" 1989 (Oxford)

...methods for simulation using MD and Monte Carlo

I have also written a series of "how to" guides for nucleic acid
simulation that are published in the Current Protocols in Nucleic Acid
series by Wiley (1999-2001), units 7.5, 7.8, 7.9, and 7.10 that may be
useful.

There is also a host of primary literature available that discusses
various aspects of the application of molecular mechanics and molecular
dynamics to protein and nucleic acid structure and dynamics...

Answering your question: The first three terms of the equation represent
the covalent connectivity, i.e. for each physical bond, angle and rotation
about a bond that is included in your model, the terms approximate the
energetic penalties as we move from the ideal geometry(s). Motion of the
arms of your protein is still inherently possible and this is due to,
primarily, collective rotation about bonds that will allow the arms to
move. The last term of the equation represents pair interactions among
all the atoms, and respectively, repulsion, dispersion attraction, and
charge interactions.

Overall, these methods can work surprisingly well-- despite the
approximations-- for representing biomolecular structure and dynamics
(although this simple "force field" or molecular mechanics representation
does not allow for watching real chemistry such as electron transfer or
bond breaking/forming).