[phenixbb] altloc non-bonded interactions

[Paul Smith]

Is it possible to tell phenix.refine to ignore certain non-bonded interactions?

I have a structure at 1.1A with some interesting sulfate molecules that refuse to refine well.  I suspect two positions are alternately occupied by sulfate in some cases and two water molecules in others.  I would like to refine with both (a sulfate and two waters) with each component at 1/2 occupancy.  With alternate conformations of a given sidechain, the two alternate locations automatically ignore one another w.r.t. non-bonded interactions.  However, different residues, it would seem, do not.  Has anyone encountered a similar situation or do the phenix gurus have any input?

As an alternative, can I define an alternate sulfate residue with several missing atoms?  

Thanks in advance,

--Paul