[phenixbb] LABIN labels for Autosol/resolve NCS scores

Florian Schmitzberger schmitzberger at crystal.harvard.edu
Thu Sep 16 10:59:10 PDT 2010


Dear Tom,

Thank you very much for the prompt reply!

I think I understand. One more question, still being somewhat  
unfamiliar with the Phenix GUI/command line concept; can I actually  
specify the correct_aniso=True keyword within the GUI or do I have to  
do that via an .eff file or via the command line. If the first  
applies, where do I specify this in the GUI ("Refine eff file list" ?)

Would you also recommend to explicitly specify a range and ratio for  
the anisotropy correction, such as "delta_b_for_auto_correct_aniso=20"  
or let Phenix take care of that?

I am putting in heavy atom sites from ShelxD, so use Autosol primarily  
for phasing and density modification right now; so I think including  
slightly more noisy data should not make as big a difference.

Best regards,

Florian



On Sep 16, 2010, at 11:57 AM, Tom Terwilliger wrote:

> Dear Florian,
>
> Normally I would recommend letting AutoSol deal with the anisotropy  
> of your data, as it already has the capability of anisotropy  
> correction. It is usually best to keep your "data" uncorrected, so  
> that the "data" are not changing as you develop new interpretations  
> of them. You can control the application of anisotropy corrections  
> in AutoSol as you noticed with correct_aniso=True.  AutoSol will  
> apply an anisotropy correction during structure solution and map  
> generation, and will keep your uncorrected data for refinement (an  
> anisotropy term is automatically applied as a parameter in  
> refinement as well).
>
> There is a possible advantage to using the UCLA anisotropy server in  
> that it throws away data that is weak, and it is possible that in  
> some cases that weak data could interfere with structure solution.
>
> If you do  use anisotropy-corrected data for AutoSol I would  
> recommend:
>
> 1. Specify a "refinement file" explicitly, and supply a file that is  
> not anisotropy-corrected (i.e., your original data).
> 2. Specify "correct_aniso=False" for AutoSol (it would probably do  
> that anyway as your corrected-data would have no anisotropy left).
>
> AutoSol would not normally use FP or SIGFP from the data file, only F 
> +SIGF+ F-SIGF-; however if you are supplying a "refinement file"  
> then this file would need F and SIGFP (and you would would want to  
> use uncorrected data there)
>
> The NCS overlap is not an absolute scale, so greater than one is  
> ok.  The NCS correlation goes from -1 to 1, and 0.5 is ok, but not  
> that great.
>
> You can supply AutoBuild with a sharpened map with  
> map_file=Mymap_coeffs.mtz and it will use this map in the first  
> cycle of building (you might only run one cycle in this case).  At  
> 4.1 A, I would only use the helices_strands_only option in AutoBuild  
> however, as model building at this resolution, particularly with a  
> relatively difficult map, will not work well.
>
> I hope that helps!
>
> All the best,
> Tom T
>
> On Sep 16, 2010, at 8:23 AM, Florian Schmitzberger wrote:
>
>> Dear All,
>>
>> I would have a number of Phenix related questions:
>>
>> Firstly, I used the anisotropy server ( to anisotropically truncate  
>> my mtz file with anomalous signal (SAD data). My question is if the  
>> labels (for F+/F- and SIGF+/SIGF- are correctly recognized by  
>> Phenix (Autosol).
>>
>> The columns labels are: F_ref  for FP; QSIGF_ref for SIGFP; GF(+)pk  
>> for F(+); LSIGF(+)pk for SIGF(+), GF(-)pk for F(-); LSIGF(-)pk for  
>> SIGF(-).
>>
>> It seems QSIGF and GF(+)pk/LSIGF(+)pk etc. are , but I am uncertain  
>> if it recognizes F_ref, because for the LABIN INPUT LABELS it says  
>> "NONE" for"F".
>>
>> Phenix does not seem to detect the anistropy ("Not using aniso- 
>> corrected data files as the range of aniso b  is only  19.52  and  
>> 'correct_aniso' is not set); so I suppose I should use the  
>> correct_aniso=True keyword.
>>
>> Secondly, what are good values for the evaluation of the NCS mask  
>> and NCS averaging Resolve is using in Phenix;  NCS overlap values  
>> (e.g. in my case 1.1) and NCS_CORR (in my case 0.52) in the  
>> "resolve.scores" file. Does an NCS overlap value of greater than 1  
>> make sense?
>>
>> I am working with a (anisotropic) dataset with 4.1/4.5 A  
>> resolution, 78 % solvent content, 2-fold NCS;  Perhaps a naive  
>> question: It is not possible to use B-factor sharpened maps for  
>> autobuilding with Phenix Autobuild is it?
>>
>> Thank you for any comments!
>>
>> Regards,
>>
>> Florian
>> -----------------------------------------------------------
>> Florian Schmitzberger
>> Biological Chemistry and Molecular Pharmacology
>> Harvard Medical School
>> 250 Longwood Avenue, SGM 130
>> Boston, MA 02115, US
>> Tel: 001 617 432 5602
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> _______________________________________________
>> phenixbb mailing list
>> phenixbb at phenix-online.org
>> http://phenix-online.org/mailman/listinfo/phenixbb
>
>
> Thomas C. Terwilliger
> Mail Stop M888
> Los Alamos National Laboratory
> Los Alamos, NM 87545
>
> Tel:  505-667-0072                 email: terwilliger at LANL.gov
> Fax: 505-665-3024                 SOLVE web site: http:// 
> solve.lanl.gov
> PHENIX web site: http:www.phenix-online.org
> ISFI Integrated Center for Structure and Function Innovation web  
> site: http://techcenter.mbi.ucla.edu
> TB Structural Genomics Consortium web site: http://www.doe-mbi.ucla.edu/TB
> CBSS Center for Bio-Security Science web site: http://www.lanl.gov/ 
> cbss
>
>
>
>
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://phenix-online.org/mailman/listinfo/phenixbb

-----------------------------------------------------------
Florian Schmitzberger
Biological Chemistry and Molecular Pharmacology
Harvard Medical School
250 Longwood Avenue, SGM 130
Boston, MA 02115, US
Tel: 001 617 432 5602










-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://phenix-online.org/pipermail/phenixbb/attachments/20100916/604b96d9/attachment-0003.htm>


More information about the phenixbb mailing list