Dear All, If my protein is rather large, and I have the mtz file of the whole protein, when I do phenix.refine of a single domain PDB of the whole protein against the whole protein mtz, after several cycles of phenix.refine the domain will not fit in the map at all (or even the domain moves far away from its initial position in map which it fits best), perhaps because the phenix.refine has made the domain to seek suitable fit in the map. Is any possibility at the whole process of phenix.refine we keep the domain relatively surround the map position where it fit best, so that phenix.refine will really do the refine in this situation? Best regards. Smith