AMBER Archive (2007)

Subject: Re: AMBER: BUG? Does TLEAP understand SPC/E water?

From: Thomas Steinbrecher (steinbrt_at_scripps.edu)
Date: Fri Jun 01 2007 - 13:30:54 CDT


Hi David,

while leap understands that you want a HOH residue to be SPC and sets the
right charges, this doesn't change that the water molecule is made up of
OW and HW atom types for which the standard parameters correspond to
TIP3P. You should also source the corresponding frcmod file:

$AMBERHOME/dat/leap/parm/frcmod.spce

to get the right bond and angle parameters.

Regards,

Thomas

Dr. Thomas Steinbrecher
The Scripps Research Institute
10550 N. Torrey Pines Rd.
San Diego CA 92037, USA

On Fri, 1 Jun 2007, David Cerutti wrote:

> Hello,
>
> I'm now being very careful about the topology files generated by TLEAP
> because the results I'm getting don't match some of those reported by others.
>
> When I change a leaprc file to reference the SPC water in solvents.lib
> rather that TIP3P, by adding the following lines:
>
> HOH = SPC
> WAT = SPC
>
> rather than "HOH = TP3" / "WAT = TP3" as the default seems to be, it seems
> that this should direct TLEAP to look in the same solvents.lib file where it
> finds the TIP3P parameters and find instead the SPC/E parameters and put
> those in the topology wherever it finds water in the simulation.
>
> The part I'm scratching my head about is that it DOES obtain the correct
> charges (I've had to figure out what AMBER's unit conventions are in that
> topology file, but I've verified that generating the topology with "WAT =
> SPC" gives the charges I'd expect for SPC/E, and they are different that
> those you get with "WAT = TP3"). But, TLEAP still inserts the TIP3P
> parameters for Lennard Jones A and B coefficients, as well as bond
> lengths--the TIP3P OH bond length is 0.9572, and this is still what I see for
> the SPC/E topology, even though the OH bond length should be 1.0000.
>
> I found this when I was trying to convert my own PDB, a box of 1728 water
> molecules, into a topology and coordinate files. However, when I used a
> "solvateBox x SPCBOX 14.0" command in another study to solvate a protein in
> SPC/E water, I find that I actually obtained similar results--the charges in
> the topology were correct, but the bond lengths remained those of TIP3P and
> the Lennard-Jones A and B coefficients were also those of TIP3P, not SPC/E.
>
> It seems that there is a major problem somewhere. Either I didn't know
> about some subtle, additional command that I needed to insert to include the
> rest of SPC/E water in the setup, in which case TLEAP needs to be made more
> user-friendly, or there is some bug in TLEAP that makes it think that SPC/E
> water has the same geometry and Lennard-Jones parameters as TIP3P (I checked
> the solvents.lib file, and it has the correct geometry and LJ parameters for
> both TIP3P and SPC/E).
>
> The result is the same whether I use TLEAP from Amber8 or Amber9.
>
> Please help!
>
> Dave
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber_at_scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu
>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu