[phenixbb] refinement error with ligand

[Pavel Afonine]

Hi Christian,

a possible solution using the command line:

1) take your PDB file with all the ligands,

2) create a CIF file:

phenix.ready_set model.pdb

3) edit *.eff file to specify correct input file names for model, data, 
CIF file (and other parameters if necessary) and run phenix.refine as 
following:

phenix.refine my_file.eff --overwrite

If it still give you that error message then you may need to send us 
your PDB file so we can fix the problem. Meanwhile you can try running 
your refinement without H atoms:

- remove H:

phenix.reduce model.pdb > model_noH.pdb

- run refinement as above.

Pavel.



On 2/17/10 8:58 AM, Christian Roth wrote:
> Dear all,
>
> I started a refinement of one of my structures using the GUI of phenix (Version 
> 1.6). I used the options of adding hydrogens and dis a suimulated annealing. I 
> used the output file and corrected it manually using Coot. I incorporated a few 
> Ethylenglycol Monomers which might have bound.  I merged everything wrote the 
> file out  and loaded this pdb in the GUI. I got an error which might be du to 
> the fact that I didnt specify a CIF file for the ligand. I started ready set 
> from the GUI to check the pdb and used REEL to create a CIF File. I added the 
> CIF-File to the  File list and tried to start again phenix.refine using the 
> hydrogen option but no sim. annealing. However I got always the following 
> error log:
>
> # Date 2010-02-17 Time 12:00:35 CET +0100 (1266404435.31 s)
> #phil __OFF__
>
> Command line arguments: 
> "/homes/roth/clcf/049b32mlsyncref/phenix/phenix/project_data/refine_15.eff" "--
> overwrite"
>
> HOST = bbzws339
> HOSTTYPE = i486-linux
> VENDOR = intel
> USER = roth
> PID = 29983
>
> -------------------------------------------------------------------------------
>   PHENIX: Python-based Hierarchical ENvironment for Integrated Xtallography
>   Version: 1.6
>   Release tag: 289
>   Platform: intel-linux-2.6 linux
>   User: roth
> -------------------------------------------------------------------------------
>
>               phenix.refine: Macromolecular Structure Refinement
>
>  Monomer Library directory:
>     "/software/linuxsoft/phenix-1.6-289/chem_data/mon_lib"
>   Total number of atoms: 16881
>   Number of models: 1
>   Model: ""
>     Number of chains: 19
>     Chain: "A"
>       Number of atoms: 1575
>       Number of conformers: 1
>       Conformer: ""
>         Number of residues, atoms: 94, 1575
>           Classifications: {'peptide': 94}
>           Modifications used: {'COO': 1, 'NH3': 1}
>           Link IDs: {'PTRANS': 6, 'TRANS': 87} 
> and so on for other chains
>  Chain: "R"
>       Number of atoms: 7
>       Number of conformers: 1
>       Conformer: ""
>         Number of residues, atoms: 1, 7
>           Unusual residues: {'EDO': 1}
>           Unexpected atoms: {'EDO,HO1': 1, 'EDO,HO2': 1}
>           Classifications: {'undetermined': 1}
>   Number of atoms with unknown nonbonded energy type symbols: 16
>     "HETATM16821  HO2 EDO K   1 .*.     H  "
>     "HETATM16826  HO1 EDO K   1 .*.     H  "
>     "HETATM16828  HO2 EDO L   1 .*.     H  "
>     "HETATM16834  HO1 EDO L   1 .*.     H  "
>     "HETATM16836  HO2 EDO M   1 .*.     H  "
>     "HETATM16840  HO1 EDO M   1 .*.     H  "
>     "HETATM16842  HO2 EDO N   1 .*.     H  "
>     "HETATM16848  HO1 EDO N   1 .*.     H  "
>     "HETATM16850  HO2 EDO O   1 .*.     H  "
>     "HETATM16857  HO1 EDO O   1 .*.     H  "
>     ... (remaining 6 not shown)
>
> I looked in the pdb for this atoms, but I cannot find them.
> When I use a cif file created directly by phenix elbow using command line and 
> when I did not use the update hydrogen button the refinement works. But the new 
> residues as well as the EDO molecules have no hydrogens in riding positions. 
> Do I have to remove all hydrogens and attach them again for every cycle?
> One Refinement runs for more then 5 hours regardless I use sim annealing or 
> not. This seems to be very long for me. Maybe I want  too many optimization 
> steps?
>
> Could anyone explain me what goes wrong here. I am no so much experienced with 
> phenix and now I run out of ideas. 
>
> Thanks in advance 
>
> Christian
>
>
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://phenix-online.org/mailman/listinfo/phenixbb
>