Coincidentally, I just noticed this problem while perusing a validation report last night and am tooling up to see if I can get this problem fixed.
The origin of the flaw is the asymmetry of the NH1 and NH2 atoms in Arginine. If one considers only bonding it would appear that these two atoms are interchangeable, but one clashes with CD and the other does not. The clash causes one bond angle to be larger than 120 deg and the other to be smaller. To allow this situation to be described the IUPAC convention says that the atom near CD should be named NH1 and the other one name NH2. This naming convention is one of the things that Coot begs to be allowed to fix every time you start it. In the Engh & Huber library (2001) this convention was not imposed and their analysis was performed with the assumption that these two angles are equivalent. The result is a symmetrized value midway between the two correct values, which as you note are about 123 deg and 117 deg. The bundling together of these two angles shouldn't cause a validation problem expect for the fact the Engh & Huber assigned the impossibly small sigma of 0.5 deg to this angle. A 50/50 mixture of angles each centered on 123 and 117 should have a sigma upwards of six degrees. Such a sigma would give a pass to correct models, but a 0.5 deg sigma will flag correct models with 6 sigma deviations. Until we can get the validation report generator fixed you should tell this story to anyone who complains about your model. Your model is fine and the validation software needs better validation itself. Dale Tronrud On 2/9/2018 10:18 AM, Oganesyan, Vaheh wrote: > Dear crystallographers, > > > > Lately when refining a structure (at 2.8A) with Refmac5 I’ve found that > nearly all Arg residues get distorted at one angle: NE-CZ-NH1(2). > Starting model has 120°, final model 123°(117°), which validation server > considers a major issue. May any of you recognize why is this happening? > Don’t remember seeing anything like this before. > > Current CCP4 version 7.0.050; Refmac5 version 5.8.0189. > > > > Last structure deposited in December’17 did not have those issues. CCP4 > version then was 7.0.047; Refmac5 version was the same. > > > > Thank you for your time. > > > > Regards, > > > > /Vaheh Oganesyan/ > > /MedImmune, ADPE/ > > /www.medimmune.com/ > > > > > > To the extent this electronic communication or any of its attachments > contain information that is not in the public domain, such information > is considered by MedImmune to be confidential and proprietary. This > communication is expected to be read and/or used only by the > individual(s) for whom it is intended. If you have received this > electronic communication in error, please reply to the sender advising > of the error in transmission and delete the original message and any > accompanying documents from your system immediately, without copying, > reviewing or otherwise using them for any purpose. Thank you for your > cooperation.