[gmx-users] Still force field issues: dihedral angle parameters

[Berk Hess]

>OK, but my point is that I don't understand why the manual on page 59 says
>"with the Ryckaert-Bellemans potential the 1-4 interactions must be 
>excluded
>from the non-bonded list." (btw, this remark is repeated 3 times, so I 
>guess
>there was supposed to be some emphasis on it ?)
>To my knowledge in other forcefields (Amber, MM3, ..) RB type dihedral
>potentials (in the form of a series expansion of the 1-cos(n p - p0) term)
>and 1,4-LJ interactions are used simultaneously.

I think the remark was made for the RB dihedrals as they are used
in alkanes, where no 1-4 interactions are added.
If one uses the RB dihehdral to 'simulate' a 1-cos(np-p0) expansion
I guess it is no longer a RB dihedral in the alkane forcefield sense.
In a forcefield with explicit 1-4's one should of course add 1-4's.

>Also, this opens another question: if you know what your dihedral angle
>potential is supposed to look like, how do you derive corresponding
>dihedral and LJ-1,4 parameters (eg how do you separate the function
>V(phi) for i-j-k-l into V_dihedral(j-k) and V_LJ,1-4(i-l) ?

That is a very good question. I would say this is very difficult.
The dihedral should refect the dependence on the inner two atoms
and the 1-4 interactions the dependence on the outer two atoms.
The only way is to systematically parametrize a large number of
related compounds.

>A last question (I hope) on dihedrals:
>If I use  [ dihedrals ] to define eg 2 or more entries for the same
>dihedral angle, will I get a sum of several cosine terms for that bond ?
>(which would be a sort of faking a multiple term cosines series)

Yes, but one could also use an RB dihedral, as is explained in the
manual.

Berk.




_________________________________________________________________
Join the world’s largest e-mail service with MSN Hotmail. 
http://www.hotmail.com