AMBER Archive (2005)

Subject: RE: AMBER: group-based cutoff

From: Ross Walker (ross_at_rosswalker.co.uk)
Date: Wed Aug 10 2005 - 20:57:35 CDT

Dear John,

> I'm curious if a group-based cutoff can be implemented with the AMBER
> force fields (I'm looking at all_amino02.lib in particular).
> It seems
> that when I try to group atoms together in alanine, for
> example, I can't
> make charge neutral groups. Are non-neutral groups used? Is a
> residue-based cutoff used at all or is it done based on
> atom-atom distances?

Versions of Amber prior to version 6 used a residue based cutoff. The
reasoning behind this was that if you use an atom based cutoff you get
instabilities unless the cutoff is very big because you split dipoles and so
as individual atoms (which have non-zero charges) move in and out of the cut
off so you get big jumps in energy. Residues on the other hand have integer
charges, typically zero, -1 or +1. Thus you get much better stability using
a cut off based on residues. Neutral groups would give you even better
stability but I don't think this was done. However, this requires
considerably more book keeping than a simple atom based cutoff. Although if
coded correctly can be faster for doing list updates.

When Amber moved to version 6 a particle mesh ewald approach to long range
electrostatics was implemented which is stable regardless of whether you use
a residue based or atom based cutoff, since the error function is setup in
such a way to make the real space potential go to zero at the cutoff (8
angstroms) and everything else gets included in the reciprocal part. Since
gas phase calculations with a cutoff are not recommended so the decision was
made to drop support for residue based cutoffs and just have an atom based.
If you are doing gas phase calculations then you should not really use a
cutoff. Note Amber8 uses a residue cutoff for QM/MM interactions when doing
hybrid simulations but this is via a completely different set of subroutines
to what are used for classical simulations. Amber 9 will have Ewald support
for QM/MM interactions and so will use an atom based cutoff here as well.

Thus if you want a residue based cutoff then you need to go back to versions
of amber before version 6 (or use sander_classic that shipped with v6???).
If you are running periodic boundaries with PME though there is nothing to
be gained from using a residue based cutoff.

I hope this helps.
All the best
Ross

/\
\/
|\oss Walker

| Department of Molecular Biology TPC15 |
| The Scripps Research Institute |
| Tel: +1 858 784 8889 | EMail:- ross_at_rosswalker.co.uk |
| http://www.rosswalker.co.uk | PGP Key available on request |

Note: Electronic Mail is not secure, has no guarantee of delivery, may not
be ready every day, and should not be used for urgent or sensitive issues.

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu