Hi Peter,
I have a partial model at low resolution. Is there a way let Phenix.refine know the model is not complete and that when a mask is calculated for bulk solvent correction it should take this fact into consideration ?
No. Bulk solvent mask is set up everywhere where there is no model. May be it is a reasonable approximation at early stage of refinement to have something in that area than vacuum?
Alternatively, is there a program that will build dummy atoms (zero occupancy) automatically in a map in all peaks above a certain sigma (i.e. atoms just for mask calculation).
I'm not aware of such program although it's a couple of hours to write one. I would probably even write it if I ever see a clear need for it or evidence for its usefulness. So far I don't understand why you really need it? Pavel.