[gmx-users] CH1 vs. CS1, advice on silanol angles and dihedrals

[Sherwin J. Singer]

I would appreciate help with force field constructions:

1) beta-D-glucose:

When fed a pdb file for , the Dundee PRODRG server assigned 3 of the
ring carbons as CH1, and the two carbons adjacent to the ring oxygen
as CS1.  Shouldn't they all be CS1 ("sugar CH groups" according to the
manual)?  The CH1/CS1 distinction does seem to make a difference in the
carbon-hydroxyl oxygen bond parameters.

2) Silanols at a silicate surface (diagram below might only look right
with fixed width font):

...O1\
...O2-SI-O-H
...O3/

I need to come up with OS-SI-OA and SI-OA-HO angle functions,
OS-SI-OA-HO dihedral functions, and charges.  To save the trouble of
performing ab initio on fragments, I would appreciate input anyone can
offer, including your thoughts on the following approximations I am
considering:

* Approximate OS-SI-OA by OS-SI-OS, which can be found in
ffgmxbon.itp?  This seems pretty reasonable.

* Approximate SI-OA-OH by SI-OS-Si?

* Approximate *-SI-OA-* dihedral by *-SI-OS-* dihedral?


I also have a question about how your convention for dihedrals:

Is the *-SI-O-* dihedral force constant in ffgmxbon.itp designed for
including 3 dihedral interactions across the OS-SI-OA-OH bond, namely
(see above diagram) O1-SI-O-H, O2-SI-O-H, O3-SI-O-H?

Or is the force constant given in ffgmxbon.itp more appropriate for
adding just one dihedral interaction controlling azimuthal motion of
the terminal H?

At this point my "anchor" oxygen atoms (O1,O2,O3 in the diagram) are
frozen, so adding the interactions either should not make a
significant difference.

Thanks,
  Sherwin Singer

-- 
Sherwin J. Singer
Department of Chemistry
Ohio State University
Columbus OH, 43214
singer at chemistry.ohio-state.edu
614-292-8909
614-292-1685 (fax)