Date: Thu, 17 Oct 2002 10:58:38 -0700
From: "David A. Case"
Subject: Re: dummy torsional parameters again.

On Thu, Oct 17, 2002, X. Tan wrote:
>
> 1. what is the exact meaning of "dummy torsional parameter"?

I'll make another stab at explaining this...

Suppose your real system has atoms A-B-C-D that have two torsions (a 2-fold
and a 3-fold term).

Then suppose these atoms "mutate" into A'-B'-C'-D' in the perturbed state,
and you want the A'-B'-C'-D' torsion to have just a 2-fold term.

*Then*, you have to add a "dummy" (zero-force constant) 3-fold term to
A'-B'-C'-D'. For bookkeeping reasons, LEaP and sander/gibbs require the
same sorts of terms for both perturbed and non-perturbed systems. Of course,
in this example, you only "really" have the 2-fold term you want, since
the three-fold term is always zero.

LEaP flags the above situation, and tells what to do, i.e. it suggests which
torsion to add, with zero force constant.

...dac