On Thu, Oct 15, 2015 at 3:26 AM, Hadeer ELHabashy <
hadeer.elhabashi.gmail.com> wrote:
> Dear Sir
>
> wish you are fine !
>
> I need to run MD simulation on a small constant volume periodic box
> containing only a crystal of a protein without any solvation. However, when
> minimizing the structure, I had a problem with the cut of . the cut off
> needed to be very small about 1 Angstrom and the minimization was not able
> to run without higher or lower values of cut off .
Then your box is too small to run simulations using the minimum image
convention. Many bonds (especially between heavy atoms) have lengths
larger than 1 Angstrom. This is a rather ridiculous cutoff, and I wouldn't
expect any simulations employing it to actually work (it's certainly not
tested). If your unit cell is too small for an 8 A cutoff, then you need
to make your unit cell larger.
There is no alternative within Amber (you would need a simulation package
that *doesn't* use the minimum image convention, and would instead populate
the pairlist with multiple images of the same atom if necessary, and I
don't know of a single simulation package that behaves that way --
certainly not any of the "mainstream" ones like Amber, CHARMM, GROMACS,
OpenMM, LAMMPS, ...).
HTH,
Jason
--
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Oct 15 2015 - 05:30:03 PDT
