[gmx-users] Fine tune the RDF of water around a dummy metal

[Nash, Anthony]

Hi all,

I¹m trying to fine tune the rdf of tip3p water molecules around a central
metal dummy molecule ("Force Field Independent Metal Parameters Using a
Nonbonded Dummy Model²), essentially a central metal (with vdw parameter
and -1 charge) covalently bonded to six Œdummy¹ atoms (no vdw parameters,
each with a +0.5 charge). Despite all my efforts at recreating the RDF
from the published work, my first peak deviates by almost 0.8 Angstroms.
The original publication showed that a deviation from their rdf resulted
in a large deviation from the correct solvation free energy. Also, a metal
demonstrating a deviation of around 0.8 angstrom in a water box will more
than likely prevent the metal from settling in plane to the nitrogen atoms
of a heme group when the parameters are put into practice.

I have looked into adjusting the following in my attempt at bringing the
first hydration shell closer to the dummy atom:

-adjusting the sigma & epsilon (started from the published values I¹ve
iteratively gone through tens of paired values, with sigma never getting
below 0.48129. Any attempt at doing so causes steepest descent to crash
during an energy minimisation)
-shortened the equilibrium bond lengths of the central metal atom to its
bonded dummy atoms
-increased the positive charge on the dummy atoms whilst reducing the
negative charge on the central metal atom

0.32 Angstrom is as close as I can get it without the energy minimisation
from crashing out (excessive force).

A bit of a long shot but any suggestions on how I could tune these
parameters further?


Dr Anthony Nash
Department of Chemistry
University College London

Skeletal Tissue Dynamics Group
Committee member of London Matrix Group @LondonMatrixGrp