Hi everyone, New to this forum, I usually lurk in the European counterpart, which makes me a novice on the many tools of phenix. I’m trying to refine a low-res structure that has a strong signal in the active site, for which I modeled a putative reaction intermediate. However, this molecule is both linear and long (14C) and thus very flexible, and only the part nested shows clear electron density, leaving out 4-5 atoms. When I run refinement, seems like the molecule gets crammed into the available density and this strains geometry. A clipped version of the ligand fits the density nicely and doesn’t appear strained. Hence, my attempt to use Amber on phenix.refine. AmberPrep gives back an error saying that the file is empty and the leap.log complains both about the ligand and a distance in the protein. The ligand was generated with eLBOW using SMILES and AM QM 1, while the enzyme is split in an alpha and beta subunits each monomer due to autoproteolytic activation and the in-between is highly flexible and thus absent from the model. Is there a way to fix this? Do I have to split each subunit into different chains in the PDB file (e.g. diner with ABCD chains instead of just AB)? Thanks! Best, Nikolas