[phenixbb] Refining an ion on a special position

[Pavel Afonine]

Just for the archive record: I think this only works well if the entity 
that sits on special position is single-atom. For anything else (like a 
ligand crossing a symmetry element) the trick with reducing occupancy 
and excluding atoms from non-bonded repulsions with their symmetry mates 
is what works.

Pavel

On 1/11/19 11:44, Paul Adams wrote:
> Maybe something has changed, but originally phenix.refine would automatically determine if an atom was on a special position, adjust the occupancy accordingly, and keep the ion on the axis during refinement. There was no requirement to set the occupancy to something other than 1, nor turn off symmetry interactions. What happens if you give the ion an occupancy of 1 and leave the symmetry interactions as normal?
>
>
>> On Jan 11, 2019, at 11:33 AM, Pavel Afonine <pafonine at lbl.gov> wrote:
>>
>> Hi Shane,
>>
>>> I've got a question related to refinement of an ion on a special position. I have a structure with an atom placed squarely on the C2 axis. I've been able to get a mostly-reasonable model by dropping in the ion at 0.5 occupancy and turning off repulsion for the ion in question, as Pavel describes here. It almost completely works.
>> OK, all sounds good so far!
>>
>>> The thing I find now is that during refinement there can still be some random drift of the atom away from the symmetry axis, splitting from its symmetry ghost, while I know that it should stay perfectly coincident.
>> Hm.. this is odd and I'm pretty sure the code that handles this makes sure atoms on special position should stay there perfectly (no drifting).
>>
>>> My resolution is 3.1 and there's no indication of a multiple conformation, so a split isn't justified. The most parsimonious model should place the ion exactly on the C2.
>> All true!
>>
>> Looks like I need to investigate this in greater detail.. Would it be possible to send me model and data files, as well as any ligand cif (if used) and that will be a good start for me. I'm traveling next week, so if I get files today I may be able to respond before my travels.
>>
>> Pavel
>>
>> _______________________________________________
>> phenixbb mailing list
>> phenixbb at phenix-online.org
>> http://phenix-online.org/mailman/listinfo/phenixbb
>> Unsubscribe: phenixbb-leave at phenix-online.org