Re: [phenixbb] Planarity restraint ignored when dist_esd equal to 0

Here is a 3pg PDB file and a CCP4 restraints file. I should point out that Phenix applies the planar esd individually to each atom in the plane allowing some atoms to be more out of plane than others. I'm working on using this feature to better advantage at the moment. Cheers Nigel --- Nigel W. Moriarty Building 33R0349, Molecular Biophysics and Integrated Bioimaging Lawrence Berkeley National Laboratory Berkeley, CA 94720-8235 Phone : 510-486-5709 Email : [email protected] Fax : 510-486-5909 Web : CCI.LBL.gov On Tue, Sep 27, 2016 at 8:35 AM, Keitaro Yamashita < [email protected]> wrote:

Dear Pavel,

Yes, it would be problematic if 0 is not treated specially. What we normally expect might be the planarity constraint if sigma is 0. I don't know how ccp4 treats it (for example refmac5 keeps atoms planar in this case), but maybe just the value of the first atom is used?

Thank you very much for your help but I am sorry that I cannot give a PDB file because I do not have it.. Could you please generate a test data yourself?

Best regards, Keitaro

...

Hi Keitaro,

...

My colleague noticed that planarity restraint of a ligand was ignored when a cif file in CCP4 monomer library was given to phenix.refine. The cif file was $CLIBD_MON/3/3PG.cif.

This was because phenix (mmtbx) ignores planarity restraint when dist_esd is 0. In the line 2055 of mmtbx/monomer_library/pdb_interpretation.py (phenix-1.10.1), plane_atom is not added to the list when (plane_atom.dist_esd in [None, 0]) is True.

yep, that's right. This is because restraints are T = SUM w*(p_model - p_ideal)**2, with w = 1/esd**2. So instead of bombing with division by zero error, the software ignores instances of restraints that would cause numeric troubles.

I agree a warning may be helpful! Could you please send me a PDB file

2016-09-28 0:23 GMT+09:00 Pavel Afonine : that

...

contains this ligand?

Pavel

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb Unsubscribe: [email protected]