Re: [AMBER] Force constant

From: Nicholas Musolino <musolino.MIT.EDU>
Date: Sun, 18 Apr 2010 14:38:33 -0400

On Apr 18, 2010, at 12:56 PM, William Flak wrote:
> Dear Dr. NicholasThank you for an instant reply.I did optimization (minimization) of two fragments (two separated molecules) using the previously posted input. When I changed the force constant for bond length between N and EP from 500 to 600 (just to test it) I found the minimized geometry changed, i.e. the bond distance between EP and H-second fragment. I couldn't see why, changing the force constant may affect on the total energy of one fragment, but why it affected the electrostatic interaction and geometry.I used GAFF force field, where I minimized only two ligands (heterocyclic compounds).
> I have more questions, if you don't mind please:* How they determined the force parameters for the lone pair in the original paper:RICHARD W. DIXON, PETER A. KOLLMAN "Advancing Beyond the Atom-Centered Model in Additive and Nonadditive Molecular Mechanics"* Should I set dihedral parameter for the EP to zero because it is missing?
> Thank you in advance

Dear William,

I think it's natural that the geometry of the minimized structure will change in some small way: after all, the potential energy is a function of the (x,y,z) coordinates of _all_ atoms, and at a PE minimum, you will have a balance of net force on each atom (in this case, EP).

If you think of it this way: changing the bond distance between EP and Y (whatever it's bonded to) will change the position of EP, which may change the distance to H-second fragment, or which might change many more things over the course of minimization.

To get technical, another thing to consider is that minimization will take you to a _local_ minimum in the PE surface. If, in the changed PE function, there is a saddle point near the initial structure (introduced by change from K=500 to 600), which the system will now slide over, there is a chance you might get to a totally different energy minimum.

Without any details, it's hard to say. Why don't you try posting (on a website) a screenshot or picture of the geometries? Or changes in the distances/energies you're concerned about? I mean, a change of 0.1 A would not be significant, I'm sure you would agree.

As for your second questions:
* I haven't read that paper and can't offer any comments. As Lincoln might have said, it is far beyond my poor power to add or detract from a 14-page journal article in an e-mail message.
* As for how the lone pair fits into GAFF, I don't see any EP (or ep) types in the original GAFF paper from 2004. Does antechamber give any messages about this? What does the output from parmchk look like?

Maybe if you provide some more details and some of the experts weigh in, you'll get more explanatory advice. Sorry I can't be more helpful.


Nicholas Musolino
Ph.D. candidate, Department of Chemical Engineering, MIT

AMBER mailing list
Received on Sun Apr 18 2010 - 13:00:03 PDT
Custom Search