[gmx-users] G96 alkane params

Marcos Villarreal arloa at dqb.fcq.unc.edu.ar
Wed Jun 4 13:16:01 CEST 2003

On Tuesday 03 June 2003 12:46, you wrote:
> On Tue, 2003-06-03 at 18:24, Marcos Villarreal wrote:
> > Hello,
> > I am trying to reproduce the simulations used to develop the G96 45A3
> > (new alkanes params).
> > When I perform the NVT simulations with a 0.8/1.4 cut-off, the resulting
> > pressure is to low. So I decided to use a 0.8/1.3 cut-off and then, the
> > calculated pressure, dHvap and Vol (in a different NPT simul) are exactly
> > as the published results for dodecane and for hexadecane.
> >
> > Is it possible that the different neighbor searching algorithms
> > implemented in GMX and in Gromos are the responsable for this?

> To be sure you'd want to check *every* possible parameter, like nstlist
> (should be 5 according to gromos), but first check all the energy terms.

the MD setup is the same as they use, including charge-group definition.

> Is the difference really in the LJ/Coulomb only?

I can not check that since they report only dH vap. 
But I would say that for these n-alkanes molecules (no partial charges), the 
LJ interaction would be the dominant factor to achieve the right presssure in 
a NVT simulation, provided the other bonding terms are reasonable.

Marcos Villarreal
Grupo de Biofisica
Departamento de Quimica Biologica - CIQUIBIC.
Universidad Nacional de Cordoba.
Cordoba. Argentina.

More information about the gromacs.org_gmx-users mailing list