[phenixbb] tightening geometry

[Pavel Afonine]

Hi Patrick,

I think I have an answer for you (at least partial)...

In your refinement run you do refinement of only group B-factors: 
strategy=group_adp
so the parameter wxc_scale=0.??? (which is used in restrained individual 
coordinate refinement) is not used.

Did you try to run it with the default strategy and using 
"optimize_wxc=true optimize_wxu=true". It may take sometime (depending 
on the size of your structure) but it is likely to give you the best 
result. Alternatively, you can specify your own numbers for wxc_scale 
and/or wxu_scale but make sure that you actually asked phenix.refine to 
do refinement of individual coordinates and/or individual ADPs (by 
making sure that individual_sites and/or individual_adp are in the 
"strategy").

Please let me know if you have any other questions or problems!
Pavel.

PS> By default, phenix.refine refines individual coordinates and 
B-factors, and occupancies of atoms in alternative conformations or 
those having partial occupancies in input PDB file.




On 7/10/09 1:43 PM, Patrick Loll wrote:
> Hi,
>
> I'm a little baffled. We have a 3 Å structure, and while the R values 
> seem OK, I'd like to tighten up the rmsd values for the geometry. 
> However, altering wxc_scale seems to have no effect.
>
> Specifics:  In a clean directory, I run the following command
>
> phenix.refine X.mtz Y.pdb Z1.cif Z2.cif Z3.cif wxc_scale=0.??? 
> strategy=group_adp       (the cif files are libraries for different 
> ligands)
>
> When I vary wxc_scale from 0.5 to 0.005, the geometry doesn't change 
> at all, which makes no sense to me:
>
> For  wxc_scale=0.5       r_work = 0.2672 r_free = 0.2900 bonds = 0.028 
> angles = 3.151
> For  wxc_scale=0.05      r_work = 0.2672 r_free = 0.2900 bonds = 0.028 
> angles = 3.151
> For  wxc_scale=0.005     r_work = 0.2672 r_free = 0.2900 bonds = 0.028 
> angles = 3.151
>
> ...you get the idea. 
>
> These were run with phenix-1.4-3. With phenix-1.4-56, we get subtly 
> different r values, but the bonds and angles numbers are the same as 
> shown above (to 3 decimal places!), and the numbers don't change upon 
> changing wxc_scale.
>
> Am I missing something here? Is there a better way to tighten up the 
> geometry?
>
> Thanks,
>
> Pat
>
> ---------------------------------------------------------------------------------------
>
> Patrick J. Loll, Ph. D.  
>
> Professor of Biochemistry & Molecular Biology
>
> Director, Biochemistry Graduate Program
>
> Drexel University College of Medicine
>
> Room 10-102 New College Building
>
> 245 N. 15th St., Mailstop 497
>
> Philadelphia, PA  19102-1192  USA
>
>
> (215) 762-7706
>
> pat.loll at drexelmed.edu <mailto:pat.loll at drexelmed.edu>
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://www.phenix-online.org/mailman/listinfo/phenixbb
>   
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://phenix-online.org/pipermail/phenixbb/attachments/20090710/8be4744d/attachment-0003.htm>