[phenixbb] Question about restraints

[Nigel Moriarty]

Carsten

I'm sure you are aware of all the things I'm going to mention but I
mention them for completeness. Also, there will be a lot of differing
opinions so I state mine with the coda that its by no means an
absolute and others experience may suggest differing assertions. I'm
sure there will be some discussion.

Your mention of over-tightening is important. The ligand is a small
part of the restraints model and therefore the restraints must be
viewed in light of the entire set.  Making the Estimated Standard
Deviations (ESD) smaller does have the result of tightening the
restraints but making them more than hundred times smaller than the
corresponding ESD in the protein portion is not numerically sensible.

Regarding the tightness of restraints, the ESD are in the units of the
restraint. The majority of the restraints will be within the ESD of
the ideal value. The variation of a restraint is dependent on the type
of restraint (bonds vary less than dihedrals) and on the different
coordinates being restrained (single bonds and double bonds).

Naturally, the quality of the data has an influence.  You don't want
to restrict the geometry if the density says something different.

Having said all that, I would suggest that for bonds the 'loose',
'medium' and 'tight' (LMT) is something like 0.05, 0.01, 0.005.  For
angles you could use 10., 3., 1. The remaining restraints are less
likely to need adjusting but similar values could work for dihedrals
and planes.

Cheers

Nigel

On Fri, Nov 12, 2010 at 6:19 AM, Schubert, Carsten [PRDUS]
<CSCHUBER at its.jnj.com> wrote:
> Hi,
>
> I was wondering what the default values for restraints in reel/elbow
> represent in terms of ‘tightness’ for lack of a better word.
>
> I am trying to tighten up the geometry of a badly behaving ligand and want
> to avoid over tightening. What is the consensus out there what kind of ESDs
> represent a ‘loose’ ‘medium’ and ‘tight’ restraint for bond, angle,
> dihedral, planar and chiral restraints.
>
> Thanks
>
>         Carsten
>
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://phenix-online.org/mailman/listinfo/phenixbb
>
>



-- 
Nigel W. Moriarty
Building 64R0246B, Physical Biosciences Division
Lawrence Berkeley National Laboratory
Berkeley, CA 94720-8235
Phone : 510-486-5709     Email : NWMoriarty at LBL.gov
Fax   : 510-486-5909       Web  : CCI.LBL.gov