On Wed, Mar 26, 2014 at 11:04 PM, Dongshan Wei <dswei0523.gmail.com> wrote:
> Hi Everyone,
>
> I'm converting the Amber force field to DL_POLY force field for Trp-cage
> protein. I dealed with the Amber ff99SB FF, I found in the .prmtop file,
> the "SCEE_SCALE_FACTOR" and "SCNB_SCALE_FACTOR" for some dihedral angles
> are 0.0. If the SCEE factor is 0.0, it means the actually scaling is
> 1.0/0.0 infinitely. So how to understand the 0.0 factor? Thakns!
>
Not all dihedrals have their 1-4 interactions 'counted' toward the 1-4
nonbonded and 1-4 electrostatic energy terms. In particular, you don't
want to compute the 1-4 nonbonded interactions for EVERY term in a
multiterm dihedral expansion (since you'll be including the same nonbonded
energy multiple times). Likewise for improper terms, the dihedrals do not
represent atoms connected by 3 bonds (the central atom is connected to all
others, so each atom is fully excluded from each other).
There are other cases like 6-member (and fewer) rings in which atoms can be
1-4 partners as well as 1-3 partners and as a result connected by both a
torsion and an angle. Because they're 1-3, they're fully excluded from
each other, so the 1-4 interaction should _not_ be counted. These are
treated the same as all but one term in a multi-term dihedral.
Look at the topology file format in
http://ambermd.org/formats.html,
specifically at the NOTE: * In the case of the dihedrals, if the fourth
atom is negative, this implies that the dihedral is an improper. If the
third atom is negative, this implies that the end group interations are to
be ignored.*
The "zeros" in the scaling factor sections correspond to the third atom
being negative in the dihedral list. As a result, any "correct"
implementation of variable 1-4 scaling will simply ignore the zero values.
This is actually an easy way of figuring out whether or not an
implementation is correct. If the energy blows up (because you are
dividing by 0), then you are including terms you shouldn't.
HTH,
Jason
P.S. Please exercise more patience in the future. Hannes' suggestion of
(at least) 24h is a good one.
--
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Mar 27 2014 - 03:30:03 PDT
