[buster-discuss] special position in refine v1.10.0

[Oliver Smart]

On Fri, 3 Feb 2012, Dale Tronrud wrote:

>
>   I agree with Peitro.  I don't know of a reason why an atom must be exactly
> on the symmetry axis.  Often the density on the axis is not spherical so either
> you have anisotropic B's (for a small deviation) or the atom has to shift a
> bit off center.
>
>   If a bad contact restraint is pushing the atom away, then the EXCLUDE
> statement should do the trick.
>
> Dale Tronrud


It can be a reasonable choice to model an atom or ion on a special 
position with a single atom that lies exactly on axis or position.
I think that SHELX allows this to be done with constraints. I would
like to allow BUSTER users to choose this with the distance restraints
(next release).

If the atom is allowed to drift from axis then it is a bit like allowing 
two alternates on a "normal" water. It is a perfectly reasonable modelling 
choice. But additional refinement parameters are introduced by allowing 
the atom to move off axis. In my limited experience density on symmetry 
axes is difficult to interpret and can have artefacts.

The special position handling in BUSTER is not yet completely sorted. The 
current state-of-play is described:

http://www.globalphasing.com/buster/manual/gelly/manual/gelly7.html

The distance cut-off used does mean that drift can suddenly mean fully 
non-bonded terms appear and the atom flys away from the axis.

Regards

Oliver

| Dr Oliver Smart                             |
| Global Phasing Ltd., Cambridge UK           |
| http://www.globalphasing.com/people/osmart/ |