Dear phenix.bb

How do I specifically exclude particular amino acids from torsion-angle ncs?

From the GUI, it is not obvious.

In a couple of my chains, one or two amino acids are in definite crystal contacts, and are therefore in a different rotamer/conformation compared to the other ncs-related copies.  

Unfortunately, I can't seem to find the right option / keyword to stop the refinement job moving these side-chains to agree with the rotamers found in the other molecules of the asymmetric unit. 

Many thanks,

Tony.