Re: how to set up force parameters of a novel residue?

From: Axel Kohlmeyer (akohlmey_at_gmail.com)
Date: Thu Aug 27 2009 - 16:14:17 CDT

On Thu, 2009-08-27 at 15:37 +0800, He Yuhui wrote:
> Dear all,

dear yuhui,

> I am now trying to simulate the motion of a guanine (G) base adhered
> to Au electrodes via sulfur atom, i.e, some bond structure as follows:
> Au-S-(G)
>
> Therefore, I have to find bond, angle and dihedral parameters related
> to this novel residue, e.g., Kb for Au-S bond, Ktheta for Au-S-N angle,
> and so on.

have you searched the literature for parameters?
people have done classical MD of sulfur containing compounds
stuck to gold surfaces for quite a while and i would be surprised
if there is nothing that would work on some level for your system.

however, there is one complication: the gold surface is quite
dynamic and has all kinds of dynamic reconstructions, e.g. atoms
jump out of the surface to form ad-atoms with two sulfur bonds
and ad-atoms can go back into vacancies and many more of these
things that cannot be modeled with simple classical potentials.

> With Dr. Branko's suggestion, I have read the force-field tutorial. It
> seems that I should do some quantum chemistry calculation of this
> "molecule", and then extract those force-field parameters.
>
> Now it comes my question: tutorial of MOPAC (or GAMESS) tells that it
> can calculate force parameters of a molecule system. I guess what it
> presents are the normal modes of molecule vibration. How can I
> "translate" these normal modes to force parameters?
>

you have to keep in mind that you don't bind to a single gold
atom but to a gold surface which has quite different properties
than a single gold atom.

but if you consider the problems with a classical model altogether,
the accuracy of the parameters is of a lesser concern, since you
simply cannot get a very accurate description in the first place.

so rather than worrying about the exact parameters (again, always
search the literature first) you should worry about what you can
actually extract as useful information from such purely classical
MD simulations.

cheers,
   axel.

>
> I am not clear of it, since normal modes are presented in generalized
> coordinates while Kb, Ktheta and Kphi are related with localized atoms.
>
> Besides, I found the following description in intro of GAMESS:
> "The numerical force constant computation and normal mode analysis was
> adapted from Andy Komornicki's GRADSCF program, with decomposition of
> normal modes in internal coordinates written at NDSU by Jerry Boatz"
>
> Does it means force constant Kb, Ktheta and Kphi can be obtained from
> internal coordinate transformation of normal modes?
>
> Can anyone give me some hints?
>
> Thanks and best wishes!
>
> Yuhui
>

-- 
Dr. Axel Kohlmeyer  akohlmey_at_gmail.com 
Research Associate Professor
Institute for Computational Molecular Science
College of Science and Technology
Temple University, Philadelphia PA, USA.

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:51:25 CST