Crazy rigid water molecules

From: Brian Radak (radak_at_uchicago.edu)
Date: Thu Nov 19 2015 - 15:40:55 CST

This is an admittedly weird hack that I want to do, but I figured I'd
post it here in case anyone has any interest or suggestions.

I want to do an alchemical transformation of a sodium "ion" into a TIP3P
water molecule. Presumably this will be a little more stable than
exnihilating a sodium ion; ultimately we want to use this for
nonequilibrium switching of ions on/off. My hybrid residue is thus a
sodium ion with rigid bonds to two hydrogens in the TIP3P geometry (so
the DOF and bond energy are conserved) and then a regular TIP3P
molecule. I suppose these need not be a single residue, but I used a
patch and so it is one residue.

The problem comes in getting SHAKE to handle the rigid TIP3P geometry.
Using "rigidBonds all" will keep the SOD-H distances fixed, but not the
H-H distances, while "rigidBonds water" keeps nothing rigid. Digging
into the code, it seems that H-H bonds are /only/ constrained by SETTLE
and /only/ when the molecule is flagged as a water molecule.

The criteria for a water molecule appears to be: 1.) does the molecule
contain 3 atoms (4 or 5 for TIP4P and SWM4)? 2.) Is the first letter of
the name of the first atom "O"? 3.) is the mass of the first atom
between 14 and 16?

Thus if I make my crazy sodium ion have the name "OSOD" and give it a
mass of 16, suddenly "rigidBonds water" now constrains everything just
fine (I of course defined new bonds to dummy types HTD with no LJ
parameters and having the same bonded terms as HT, etc.)

Is this the most optimal solution NAMD can offer? Are all of the
restrictions on three membered rings with an H-H bond really necessary?
I guess my "ideal" solution would be to expose these criteria to the
user in someway and also to permit multiple definitions. Then I wouldn't
have to have restrictive names and be limited in my choice of mass.

I should note that this won't break the obviously very useful assertion
that all rigid waters have the same shape.

Brian

//

-- 
Brian Radak
Postdoctoral Scholar
Gordon Center for Integrative Science, W323A
Department of Biochemistry & Molecular Biology
University of Chicago
929 E. 57th St.
Chicago, IL 60637-1454
Tel: 773/834-2812
email: radak_at_uchicago.edu

This archive was generated by hypermail 2.1.6 : Thu Dec 31 2015 - 23:22:15 CST