[phenixbb] ligand refinement

[Pavel Afonine]

Hi Christine,

I think you need to do this (Ralf please correct me if I'm wrong):

1) Run phenix.elbow to create a cif file for your ligand (if you don't 
have one):

% phenix.elbow model.pdb --do-all

elbow will go though your pdb file and create a dictionary file for 
whatever is not present in the standard library. One of the output cif 
files with "all" (xxx_all.cif) in its name is the file that you need to use.

2) Now you need to tell phenix.refine that your ligand is covalently 
bonded to the macromolecule:

% phenix.refine model.pdb data.mtz xxx_all.cif may_params.par

where may_params.par defines the bond and contains these lines, for example:

refinement.geometry_restraints.edits {
   bond {
    action = *add
    atom_selection_1 = chain X and resname HIS and resid 117 and name NE2
    atom_selection_2 = resname LIG and resid 98 and name C
    distance_ideal = 1.45
    sigma = 0.02
  }
}

You can also define an angle. Look "Definition of custom bonds and 
angles" in phenix.refine manual for more information:
http://www.phenix-online.org/documentation/refinement.htm

Please let us know if you have any questions.

Pavel.




 
On 2/21/2008 5:52 PM, Christine Gee wrote:
> Dear Phenix developers
> I have a covalently bound ligand in my structure and I would like to  
> use phenix to refine the structure.  The ligand is bound to the  
> sulfur of a cysteine.  What is the best method for defining the  
> ligand.  Should I generate a cif file for the whole amino acid and  
> the ligand, or can I make one for just the ligand and add in some  
> restraint somewhere which tells phenix to make a bond between the  
> ligand atom and a nearby sulfur in a cysteine.  If I generate a cif  
> file with the whole amino acid and ligand, I assume I will I have to  
> follow some convention in atom naming so that the main chain bonds  
> and angles will form and refine OK?
> Regards
> Christine
>
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://www.phenix-online.org/mailman/listinfo/phenixbb
>