[phenixbb] C-beta outliers

Hi all,   I am running Phenix v. 1.20.1-4487 on Windows and I keep on getting C-beta outliers of methionine residues after phenix.refine (happened in different data from two different proteins). Even if I correct the methionines in Coot (and they fit perfectly in the density), I get the outliers at the end of refinement again (while these outliers don't fit the density and I can see a green difference map where the C-beta is).   Why is this? Is there a way to tell phenix.refine not to move a specific amino acid during refinement?   Thanks, Andrea