[gmx-users] Using CHARMM36 in GROMACS

[Justin Lemkul]

On 3/30/16 5:17 PM, Gregory Poon wrote:
> Hello everyone,
>
> I am learning how to perform dynamics on double-stranded DNA sequences using the
> CHARMM36 FF I downloaded from the MacKerell lab.  I have consulted the GROMACS
> documentation page regarding specific settings in the .mdp file for CHARMM36
> (http://www.gromacs.org/Documentation/Terminology/Force_Fields/CHARMM), and
> adopted those for the energy minimization step.  In these settings, constraints
> = h-bonds.  My question is: for the subsequent equilibration steps, I have read
> that one should use constraints = all-bonds, but I have also read that might

The constraint settings are not dependent upon the phase of the simulation; 
these should be kept constant.  For CHARMM, constraints = h-bonds.

> adversely affect 1-4 interactions.  And what about DispCorr: should I keep =
> none or one of the other options -- or does it matter?
>

I don't know that anyone has done a systematic comparison, but I doubt that it 
matters in any significant way for normal MD.  Dispersion correction is 
necessary for some systems with CHARMM36 (e.g. lipid monolayers) but not with 
bilayers.  AFAIK no one has looked at this with DNA.

> Thank you in advance.  If there is some other resource that gives all the
> FF-specific settings, I would love to know about it.
>

The force field derivation papers are the first port of call.  The methods there 
should be considered canon unless otherwise revised later.

-Justin

-- 
==================================================

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==================================================