Hi Daniele,
i'm experiencing a weird problem with phenix.refine. After a phenix.refine run with a pdb file that doesn't contain the atoms N, CA, C, and O all together (with sidechains atoms between C and O) the O atoms are displaced from their position. Is this a know problem? Is there anyone that have suggestions how to re-edit the pdb file accordingly instead of doing cut-past for too many atoms?
Could you please send me a file-example with the structure you work with. I'm not sure what's actually happening and how your file looks like? Most likely the restraints are not set properly for broken chains (I'm guessing). Also, you can look in .geo file that explicitly lists all restraints used in refinement. Does that restraint present in this file? If not then it's not surprising that the geometry is distorted. Thanks! Pavel.