[phenixbb] Duration of a phenix.elbow run

Dirk dirk.reiter at uni-tuebingen.de
Fri Jul 3 17:27:20 PDT 2009


thanks for the suggestions I tried to run it overnight, but without any 
result. I also tried to use GAMESS, but even so the GAMESS run is 
started I still get the message "GAMESS doesn't seem to be installed" 
and "mv: Aufruf von stat für „elbow.drgfin_pdb.012.gamess.gam“ nicht 
möglich: No such file or directory" even so the file exists. Since I'm 
not experienced in python I wasn't able to find the reason for that by 
inspecting the GAMESSMixins.py script.
I would like to try the forcefield approach, but don't know if there is 
any freeware available for that purpose - I would be glad about 
suggestions .
For the moment I "solved" the problem by manually editing the cif file 
using values from the refmac monomer and link libraries for the 
restraints. I still would be really interested in getting elbow to work 
to get better restraints for this ligand and to avoid the time consuming 
manual creation of cif files in the future.

Thank you again for the replies


Schubert, Carsten [PRDUS] schrieb:
> What I do in these cases is a MM minimization of the ligand using a high
> quality forcefield and using that output as input into elbow and not
> perform an optimization, but have it generate the restraints directly
> from the input. The differences between the forcefield and the AM1
> calculation in elbow are negligible for protein crystallography
> purposes. I would not waste time on o.n. runs. 
> 	Carsten
>> -----Original Message-----
>> From: phenixbb-bounces at phenix-online.org [mailto:phenixbb-
>> bounces at phenix-online.org] On Behalf Of Nigel W Moriarty
>> Sent: Wednesday, July 01, 2009 7:26 PM
>> To: PHENIX user mailing list
>> Subject: Re: [phenixbb] Duration of a phenix.elbow run
>> Dirk
>> That is quite a large molecule but it shouldn't take a few hours.  I'm
>> not sure what is happening but I'd guess that a convergence is failing
>> and the upper limit on the number of steps is too great.  If I was
> you,
>> I'd run it overnight and if that fails run it in another quantum
>> package
>> to get the geometry you desire.  A different run MAY generate a
>> different starting geometry for the --opt option and that may work
>> better.  Each step of the optimisation should take less than 2
> minutes.
>> Nigel
>> On 7/1/09 2:43 PM, Dirk wrote:
>>> Hi,
>>> I'm trying to create restraints for a carbohydrate ligand using
>>> phenix.elbow. If I run it without the --opt command everything works
>>> fine, but I'm not completely happy with the geometry of my ligand
>> after
>>> refinement. I tried to get an optimized cif by using --opt command,
>> but
>>> the program just runs up to the following output:
>>> MoleculeClass :  H:38  C:23  N: 1  O:19 (PDB format)
>>>         81 atoms
>>>         83 bonds
>>>         151 angles
>>>         238 dihedrals
>>>         --------------------
>>> Time      Energy        delta E    RMS         Max      Step size
>>> and then nothing happens for several hours. Does the optimization
>> needs
>>> such a long time and if I let it run overnight (or perhaps longer) I
>>> will get an output or is something going wrong?
>>> The input is a pdb (I tested several) and I tried both the stable
>> phenix
>>> release (1.3) and the current version (1.4).
>>> Any help would be very appreciated
>>> Thanks
>>> Dirk
>>> _______________________________________________
>>> phenixbb mailing list
>>> phenixbb at phenix-online.org
>>> http://www.phenix-online.org/mailman/listinfo/phenixbb
>> --
>> Nigel W. Moriarty
>> Building 64R0246B, Physical Biosciences Division
>> Lawrence Berkeley National Laboratory
>> Berkeley, CA 94720-8235
>> Phone : 510-486-5709
>> Fax   : 510-486-5909
>> Email : NWMoriarty at LBL.gov
>> Web   : CCI.LBL.gov
>> _______________________________________________
>> phenixbb mailing list
>> phenixbb at phenix-online.org
>> http://www.phenix-online.org/mailman/listinfo/phenixbb
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://www.phenix-online.org/mailman/listinfo/phenixbb

More information about the phenixbb mailing list