Quoting Paul Emsley from ccp4-dev ml: It seems to me that the dictionary for several standard residues need to be updated [1]. My reading of the IUB-IUPAC convention is that the primary torsion is relative to the xx1 atom (not the xx2 atom). This caused me confusion for a while when I was using the dictionary to set the torsions of roamers by using the dictionary definitions of the torsions. Relatedly, it seems to me that to reduce the occurrence of nomenclature errors in VALs and LEUs, the chiral centre sign should be changed from the (too liberal) "both" to "negativ". The combination of the incorrect atom for setting torsions in LEUs, combined with the nomenclature errors in the model have lead to many "decoy leucines" [2]. Let's clean things up. The patch is attached. Paul. [1] following conversations with Jeff Headd and Gert Vriend, I should add [2] See for example residue 105 from 1A0J.
Created attachment 208760 [details] standard-residue-torsions.patch standard-residue-torsions.patch
Decided with dberkholz to stick to Garibs (sci-chemistry/refmac) nomenclature.