Hi,

In your specific case, you can predict that reflections for which the l index 
is a multiple of 9 will be particularly strong, while others will be weaker.  
So you could use sftools to select reflections that are a multiple of 9 and 
write them into one MTZ file.

   sftools
   read my.mtz
   select index l zone 9n
   write my_stronger.mtz
   stop
   yes

To get the rest of the reflections, you can add the "select invert" command 
after the initial "select index…" command to get everything else.

A more general way is to run a recent version of Phaser, choosing the NCS mode. 
 Presumably there are 9 copies related by the translational NCS.  The default 
in Phaser is to assume there are two copies related by tNCS, if there's a large 
peak in the native Patterson, so you would have to tell it there are 9 copies.  
Running from a script, you would use the command "TNCS NMOL 9".  The NCS mode 
produces an MTZ file containing a column labelled NcsEps, which is the factor 
by which the tNCS increases the expected intensity for each reflection.  The 
log file has a histogram of NcsEps values, so you could decide on a cutoff 
between weak and strong reflections, then use sftools to select them.  To get 
the reflections with greater than average intensity, you could use something 
like this:

   sftools
   read NCSanalysis.mtz
   select column NcsEps > 1
   write bigeps.mtz
   stop
   yes

I hope that helps!

Randy Read

-----
Randy J. Read
Department of Haematology, University of Cambridge
Cambridge Institute for Medical Research    Tel: +44 1223 336500
Wellcome Trust/MRC Building                         Fax: +44 1223 336827
Hills Road                                                            E-mail: 
[email protected]
Cambridge CB2 0XY, U.K.                               
www-structmed.cimr.cam.ac.uk

On 26 Feb 2013, at 03:06, Yuan SHANG <[email protected]> wrote:

> Dear all,
>    currently, I have a data set scaled in P22121 which containing a PST of 
> (0.5,0.5,0.111). The structure were successfully solved by molecular 
> replacement. However, the R free factors remained as high as ~33% in the 
> refinement. I search the literature and found that it was common to have such 
> high R free factors in case of PST (Felix F.Vajdos,etc.,protein 
> science,1997;Arthur H.Robbins,etc.,Acta D,2010;Florence 
> Poy,etc,NSMB,2001;Cory L.Brooks etc,Acta D,2008;). In the 2001 NSMB 
> paper(doi:10.1038/nsb720), the authors split the dataset into 'weak,medium 
> and strong' reflections, and showed good refinement statistcs in the 'medium 
> reflection dataset'. Although I had good electron density maps to show my 
> solution is correct. To further convince the reviewers, I also want to split 
> my data set into such sub-datasets according to the symmetry. Did anyone know 
> how to split the data set in this case?
> 
> Best regards,
> Yuan 

Reply via email to