Re: AMBER: explicit implicit solvent

From: Carlos Simmerling <>
Date: Wed, 10 Jan 2007 09:53:53 -0500

Amber is not set up yet to generally do hybrid solvent models during MD,
since restraints are needed to keep the explicit water near the solute.
Hybrid explicit-implicit water models are still very much under development
and none has yet been shown to work well without artifacts. We recently
reviewed these methods :

**Okur, A. and Simmerling, C., /“Hybrid Explicit/Implicit Solvation
Methods”/, Annual Reports in Computational Chemistry, 2:97-109 (2006).

We also used a hybrid GB/TIP3P model in REMD, the work was published
in JCTC.

**Okur, A., Wickstrom, L., Layten, M., Geney, R., Song, K., Hornak, V.
and Simmerling, C., /“Improved Efficiency of Replica Exchange
Simulations through Use of a Hybrid Explicit/Implicit Solvation Model”/,
J. Chem. Theory Comput., 2:420, 2006.

We noted very large differences between GB models, so you
need to be very careful in validating your results.

Ji-Lai Li wrote:

> Dear all:
> I have questions how to consider the solvente effects when runing
> targeted md (nonperiodic).
> Some JMB papers "A shell of explicit solvent molecules was included
> in the simulation and a modified TIP3P water model." (J. Mol. Biol.
> (2006) 356, 237–247 ant its references).
> In AMBER package, the system can be prepareed by using "solvateshell"
> to solvate the protein in the WAT molecules. And then minimization,
> dynamics...
> However, in the AMBER archive,
>> I prefer to use solvateshell instead of solvatebox to
>> decrease the time, but I am not sure if it is really
>> better and I don't know what I should do for periodic
>> boundary conditions.
> I don't know what time it is you want to decrease. Solvateshell
> creates just
> a sphere of water molecules about some point. It is not very accurate,
> since
> the polarization of waters outside the "shell" is not accounted for,
> and there
> can be unphysical effects at the water-vacuum boundary at the edge of the
> water shell.
> Solvateshell cannot be used for periodic simulations; use solvateBox or
> solvateOct for that purpose.
> ..good luck...dac
> Now, I am puzzled: how to do with the solvent effects for a targeted
> md simulations?
> I plan to do follow belows, but don't know if the protocol is
> acceptable (the results are accurate?). 1. Solvate the solute with a
> water shell (solvateshell);
> 2. The minimization, dynamics, etc. employed the Generalized Born
> model (igb=5).
> It is similar with the hybrid explicit-implicit solvent used in
> Poisson-Boltzmann calculations.
> I am looking forward to getting your advice.

The AMBER Mail Reflector
To post, send mail to
To unsubscribe, send "unsubscribe amber" to
Received on Sun Jan 14 2007 - 06:07:09 PST
Custom Search