Re: AMBER: Sander NOE violations

From: David A. Case <>
Date: Thu, 6 Jul 2006 18:01:19 -0700

On Thu, Jul 06, 2006, Andrew Box wrote:
> I am running nmr restraints during my dynamics runs in sander. When I look
> for the NOE violations in the out put file generated, I notice I have
> violations between atoms that have no constraints placed on them.

You don't say how the restraint file was created...Are you sure that there
are no distance restrints between (say) C24-BEN-13 and C25-BEN-13? Basically,
the only thing that should be printed here are pairs of atoms that are in the
distance restraint file. This is why I am asking if you are sure that no
such restraints were present.

> a small part of my out put file:
> First atom Last atom curr. value target deviation penalty
> ------------------------------------------------------------------------------
> C24 BEN 13 -- C25 BEN 13: 1.500 1.800 0.300 1.799 d
> 0: 0
> C24 BEN 13 -- *O3 BEN 13: 5.992 4.600 1.392 22.845 d
> 0: 0
> In the first line, it denotes a problem between two carbon atoms that are
> connected with in aromatic ring. The second line shows a violation
> between a carbon and oxygen atom that are separated by more than 8 covalent
> bonds and have no NOE constraint between them.

> Also, I have no idea what
> the * in front of the atoms names mean, as I have used no * in any atom
> names.

A "*" in front of an atom name means that that atom was part of a "group"
of atoms (see the igr1 and igr2 variables). The atom name printed is the last
atom of the group, and the "*" is a signal that it is member of a group. The
hope is that knowing the last atom of the group will be enough to identify
which group it is.

Overall, it sounds like sander is not interpreting your RST file in the way
you intended. If you set pencut to a negative value, and run (say) a
single-point caclulation, you can see all of the restraints that are being
applied, in the order that they appear in the RST file. You can examine
each one, and try to figure out what is going on.

..good luck...dac

The AMBER Mail Reflector
To post, send mail to
To unsubscribe, send "unsubscribe amber" to
Received on Sun Jul 09 2006 - 06:07:17 PDT
Custom Search