AMBER Archive (2008)

Subject: Re: AMBER: "Assuming uniform neutralizing plasma"

From: Thomas Cheatham (tec3_at_utah.edu)
Date: Tue Feb 26 2008 - 21:31:39 CST

Search through the archives for there are a number of discussions on
this topic; in particular, see

  http://amber.ch.ic.ac.uk/archive/200712/0223.html

> and I am hoping to learn more about this. My guess is that this is done to
> avoid infinite energies of a periodic system with a charged unit cell, but I

Although the energies in principle should diverge, the forces (what we
need for the MD) are just fine... No extra (energy or force) term is
added; this is a by-product of the standard Ewald construction omitting
the zeroth order term in the reciprocal sum.

> My specific interests are in performing MD simulations on small molecules that
> may have 0, +1, or -1 molecular charges. These simulations should be run in
> TIP3P water (ie. not in GB) to compare with other simulations I have done of
> large systems. I would like to keep the simulation boxes small (say 1000
> water molecules) to keep computational cost low, and I don't know if adding
> explicit ions makes physical sense for such a small system.

In addition to finite size effects (effectively high concentration and/or
possible, but in practice rather small, periodicity artifacts), there are
extra sampling issues with explicit ions, not to mention that many of the
ion parameters do not well-reproduce dGsolvation, water-ion energies, etc.

-- tec3
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu