Hello, I am trying to use phenix.real_space_refine to improve the fit between my protein's ligand and the electron density. I include a CIF file that has worked well for reciprocal space refinement, made through JLigand. It seems to me that the CIF file is read properly, it does not produce errors and the restrains are stated correctly in the output .cif file that includes all protein + ligand atoms. However, the output pdb shows a distorted ligand, where specific bond angles and bond lengths have been modified past what should be allowed by the restrains. Any insights on how to improve/the causes of this effect would be very welcome! My command is : phenix.real_space_refine model.pdb map.ccp4 ligand.cif resolution=1.63 Thank you, Alisia