[phenixbb] reflection files, anomalous data, and test set

Mark Collins mcollins at convex.hhmi.columbia.edu
Wed Jul 23 15:39:26 PDT 2008


Hello again,
I'm still seem to have a problem with merging these files.  
I tried this in my .params file but phenix can't use my test set, any 
suggestions appreciated?

  input {
    xray_data {
      file_name = ../testmtz.mtz
      labels = "I(+),SIGI(+),I(-),SIGI(-)"
      r_free_flags {
        file_name = ../SePeak.hkl
        label = TEST
        test_flag_value = 1  }}

Sorry: R-free flags not compatible with F-obs array: missing flag for 2 
F-obs selected for refinement.

This appears to arise from the SePeak.hkl having 
Number of Miller indices: 9198
but the testmtz.mtz having
Number of Miller indices: 9199  (where is the extra reflection coming 
from?)
by converting the SePeak.hkl to an mtz and dumping, I get SePeak.hkl.mtz 
and testmtz.mtz
 
   label #valid  %valid  min    max type
    H       9198 100.00% 0.00  14.00 H: index h,k,l
    K       9198 100.00% 0.00  10.00 H: index h,k,l
    L       9198 100.00% 0.00 135.00 H: index h,k,l
    hkl     9198 100.00% 0.00   1.00 I: integer

    H               9199 100.00%   0.00    14.00 H: index h,k,l
    K               9199 100.00%   0.00    10.00 H: index h,k,l
    L               9199 100.00%   0.00   135.00 H: index h,k,l
    I(+)            9199 100.00% -39.40 32280.10 K: I(+) or I(-)
    SIGI(+)         9199 100.00%  -1.00   771.20 M: standard deviation
    I(-)            6529  70.98% -32.00 16383.20 K: I(+) or I(-)
    SIGI(-)         6529  70.98%   5.20   293.30 M: standard deviation

Thanks Mark


> Dear Mark,
> 
> I believe your could simply define explicitly the columns in your
> input .def file for phenix.refine, I have done so previously with
> great succes.
> 
> I guess the keywords you are looking for is:
> xray_data.file_name
> xray_data.labels
> xray_data.r_free_flags.file_name
> xray_data.r_free_flags.label
> 
> And phenix.refine should put it in a nice .mtz file for you :-)
> 
> Hope this helps.
> 
> Regards,
> Folmer
> 
> 
> 
> 
> > phenixbb mailing list
> > phenixbb at phenix-online.org
> > http://www.phenix-online.org/mailman/listinfo/phenixbb
> >
> 



More information about the phenixbb mailing list