[gmx-users] Re: simulation crash with GROMOS96 force field

[Justin A. Lemkul]

Vitaly V. Chaban wrote:

There are several potential problems with your .mdp file, mostly due to 
parameters that seem inappropriate for use with Gromos96.

> ; NEIGHBORSEARCHING PARAMETERS
> ; nblist update frequency
> nstlist                  = 50

This updating is very infrequent.  Usually one sees a value of 5-10 here 
(usually 5 for Gromos96).  If your neighborlist is updated infrequently, you can 
have problems (poor energy conservation and perhaps incorrect calculation of 
interactions).  Perhaps this is why your short timestep (0.025 fs) leads to 
stability - the neighborlist gets updated every 12.5 fs instead of every 50 fs. 
  Usually the neighborlist should be updated every 10-20 fs.

> ; ns algorithm (simple or grid)
> ns_type                  = grid
> ; Periodic boundary conditions: xyz (default), no (vacuum)
> ; or full (infinite systems only)
> pbc                      = xyz
> ; nblist cut-off
> rlist                    = 1.45
> 

What motivated this choice?  The original derivation of G96 parameters used 0.8 
or 0.9, IIRC.

> ; OPTIONS FOR ELECTROSTATICS AND VDW
> ; Method for doing electrostatics
> coulombtype              = Reaction-Field-zero
> epsilon_rf               = 0
> rcoulomb-switch          = 0
> rcoulomb                 = 1.3
> vdw-type                 = Shift
> ; cut-off lengths
> rvdw-switch              = 1.1
> rvdw                     = 1.3

The derivation of G96 didn't use shifted vdW; rather a plain cutoff at 1.4 nm 
was used.  I don't know about the implementation of Reaction-Field-zero, but the 
force field was also originally derived with plain Reaction Field, and has also 
been used with PME, but I realize some papers have proposed some problems.  I 
would definitely recommend reading the original derivation of your desired 
parameter set and using values that are suggested for the force field.  This all 
seems kind of random, unless you have some reason to be testing different 
parameters.

-Justin

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

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

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