[gmx-users] Acetonitrile with CHARMM ff

[Justin Lemkul]

On 7/7/17 10:11 AM, Sonia Milena Aguilera Segura wrote:
> Dear Justin,
> 
> Thank you very much for your answer. I tried both approaches: erasing the dihedral from the topology (first approach) and modifying the ffbonded.itp as described in the previous mail, without adding any improper (second approach). I ran a minimization, 200 ps NVT, 200 ps NPT, and 1 ns MD. In general the results seem pretty much similar with respect to the energy terms:
>     
>                             1st             2nd
> E.Pot			-22425.4	-22452.9
> E. Kin			6940.56	         6928.76
> E. Tot			-15484.8	-15524.2
> Temp (K)	        301.302	           300.79
> Pressure (bar)	        2.35245	        -1.23941
> Density			788.145 	788.793
> diffusivity x10-5 cm2/s	2.0584 (+/- 0.0720)	2.1406 (+/- 0.1427)
> dielec. Constant	20.6455	        19.8552
> 
> However, I compared the density and it's higher with respecter to the experiment. Also, I noticed that the values T are slightly higher and P seems to show really different behavior. I checked the results for the NPT equilibration and the values were 8.8 and 14.5, respectively. The temperatures during the NVT were  297.565	and 297.799. Moreover, dielectric constants are underestimated. Can it be that I am using the wrong parameters?  

I don't know the history of the acetonitrile parameters but very rarely to force 
field parameters reproduce every quantity of interest (heck, no one can even get 
water right).  There is a limit to how accurate a fixed-charge simulation can 
be, and perhaps you're finding that out.

Find the literature reference for the parameters and see what quantities are 
reproduced well and what the limitations might be.

Here is my mdp file for the MD (same for NVT and NPT, adjusting the coupling terms):
> 
> ; Run control
> integrator               = sd       ; Langevin dynamics
> tinit                    = 0
> dt                       = 0.0005
> nsteps                   = 40000000   ; 20 ns
> nstcomm                  = 100
> ; Neighborsearching and short-range nonbonded interactions
> cutoff-scheme            = verlet
> nstlist                  = 20
> ns_type                  = grid
> pbc                      = xyz
> rlist                    = 1.2
> ; Electrostatics
> coulombtype              = PME
> rcoulomb                 = 1.2
> ; van der Waals
> vdwtype                  = cutoff
> vdw-modifier             = potential-switch
> rvdw-switch              = 1.0
> rvdw                     = 1.2
> ; Apply long range dispersion corrections for Energy and Pressure
> DispCorr                  = EnerPres

CHARMM uses a force switch, and only apply dispersion correction in the case of 
lipid monolayers.

http://www.gromacs.org/Documentation/Terminology/Force_Fields/CHARMM

-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

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