From: Anton Arkhipov (anton_at_ks.uiuc.edu)
Date: Thu Jan 15 2009 - 10:00:04 CST
The choice of a dielectric constant for CG simulations is really a
tricky thing to do, and after all it depends on the problem you are
trying to address. In the original RBCG model by Marrink et al. they
used dielectric constant of 20, and ionic charges 0.7, but later on
they changed that to dielectric constant 15 and ionic charges 1.0.
Either way, these settings do not describe electrostatics interactions
too well, and if those are important for the problem you are studying,
you may have to do something about the dielectric constant.
For some of our work we used dielectric constant 1. Please see this
paper for the explanation why this choice has been made:
Four-scale description of membrane sculpting by BAR domains. Anton
Arkhipov, Ying Yin, and Klaus Schulten. Biophysical Journal,
And yes, you are right, essentially you have to tune your CG
parameters, including the dielectric constant, so that CG simulations
reproduce alla-tom simulations or some experimentally known behavior.
In short, if electrostatics is not important for the problem you are
studying, using the dielectric constant of 10-20 should be fine. If
electrostatics is important, you have to check carefully what value of
the dielectric constant leads to a reasonable behavior in CG
simulations, as compared to ll-atom simulations and/or experiments. In
the latter case you may need to use dielectric constant 1, or maybe
some other value would work better for your case.
On 15 Jan 2009, at 01:10, BIN ZHANG wrote:
> Hi all:
> Can someone explain to me why usually dielectric is not left as
> default (set to 20) in CGMD (rbcg) ? To get better agreement with
> All-Atom Simulation?
> Also, how is the parameter chosen? Is it the same for only water
> system and water lipid system?
> Thanks a lot.
> The tree of liberty must be refreshed from time to time with the
> blood of patriots and tyrants.
This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:50:22 CST