Date: Fri, 6 Dec 2002 23:18:53 -0500
From: Rick Venable
Subject: CCL:ion concentration in protein simulation


The ion concentration question has a couple nuances.

It's typically too costly (CPU time) to simulate a system with enough
water to match the experimental protein concentration, i.e.

[protein]sim > [protein]expt

Instead, it might be better to match the salt/protein mole ratio, rather
than the concentration based on the simulation cell volume. Depending
on the density of charged residues on the surface of the protein, this
may be an underestimate. There may be considerable ion condensation
around the protein, so that the ion distribution in the solution would
be non-uniform and higher near the protein.

For Ewald electrostatic methods, a system with a net zero charge is
recommended. In addition to adding an equal number of K+ and Cl- ions
to at least match expt salt/protein, one should also add enough of one
type of ion to neutralize the net charge of the protein.


On Fri, 6 Dec 2002, Ioana Cozmuta wrote:
> My question relates to adding ions when simulating a large protein
> system. The experiments were performed using an ionic solution with
> a 1M concentration and with an applied electric field. I was
> wondering what is the usual approach in this kind of simulations: to
> bring the ion concentration close to the physiological number, for
> example? Also the solution is one of KCl so when adding ions would
> it be required to add in equal amounts both K+ and Cl-?
>