Re: How to confirm water box is big enough?

From: Giacomo Fiorin (
Date: Tue Mar 30 2010 - 16:38:00 CDT

Hi Jin, ideally, you'll want that two periodic images of the protein
are always farther than twice the van der Waals cutoff: this way they
would not only be out of range for their mutual interaction, but also
that they don't interact with the same water molecules.

If you used 10 as cutoff, you should be almost fine (give or take 1
water molecule). FYI, it is becoming much more commonplace to use a
larger cutoff (e.g. 12 ).

One thing you can do to make sure, is to compute the average
projection of the water molecules' dipole moments on the distance
vector from the protein. If the number goes to zero before the 10
maximum distance, in my opinion it's one of the best checks that you
can provide.


---- ----
 Giacomo Fiorin
   ICMS - Institute for Computational Molecular Science
     Temple University
     1900 N 12 th Street, Philadelphia, PA 19122
 work phone: (+1)-215-204-4216
 mobile: (+1)-267-324-7676
---- ----

On Tue, Mar 30, 2010 at 3:29 PM, Jin Liu <> wrote:
> Hi,
> I did a few MD simulations with the the minimum distance from the protein to
> the edge of the water box as 10 angstrom. The simulations ran smoothly
> without any problem. I am wondering what I could do to confirm the water box
> is big enough? Should I measure the distance between the protein and its
> image in the other boxes during the simulation? Or what else can I do to
> convince people that the water box is not too small?
> Thank you very much.
> Best regards,
> Jin

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:53:57 CST