AMBER Archive (2000)

Subject: Re: Additional Water Potentials

From: Bill Ross (ross_at_cgl.ucsf.edu)
Date: Wed Nov 29 2000 - 18:47:37 CST

        the TIP4P and TIP5P are slightly more accurate models than TIP3P. I
        would appreciate any information regarding parameters developed for
        these models

The models are defined by their parameters; however as far
as I know, no one has adapted amber to work with these models.

TIP3 uses an imaginary H-H bond plus SHAKE to keep the water
rigid, so I imagine you could add 1 or 2 extra bonds from H
to the charge center and hope that it would be kept sufficiently
planar. However if much cpu time will go into these calcs, I
wonder if the best thing might be to add a rigid molecule handler
to sander that would evaluate net force on the molecule before
translating/rotating it as a rigid piece.

        and how to add them to the xleap parameter set,

The atom types & default bond lengths would be added via a frcmod
file. The charge center would presumably have a new atom type.

        how well
        these models interact with the amber forcefield and amino acid
        parameterization, and any other comments on their use with Amber.
        
I suspect that the reason that TIP3/SPC are the only water models
used in amber (except for flexible waters omitting the H-H bond) is
that everyone who has evaluated this issue so far has concluded that
the improvement of TIP>3 doesn't matter given the other approximations
inherent in molecular mechanics. Possibly the incorporation of (periodic)
long-range cutoffs via Ewald sum would make a difference, now that this
method has become affordable and common.

Bill Ross