> As far as I can tell, there isn't a gap in the input pdb file, and > phenix.refine is creating the gap. Could you send us the fragment from the PDB file with the two residues that aren't linked correctly? Maybe there is just something unusual about the chainids or segids? Ralf