[phenixbb] bug? iotbx.reflection_statistics apparent space group error reading scalepack .sca file

Ralf W. Grosse-Kunstleve rwgk at cci.lbl.gov
Mon Mar 6 13:53:46 PST 2006


> iotbx.reflection_statistics returns space group 195
> for a scalepack 'no merge original index' .sca file for
> space group 123 (i.e space group 197)
>
> the first line of the .sca file is
>    12 i23

You are right, this is a bug. Sorry. However, it went unnoticed because
we usually supply the correct space group (along with the missing unit
cell parameters) via

  --symmetry=some_other_file

You could also use

  --unit_cell=10,20,30,90,90,90 --space_group=197

Mostly I prefer --symmetry because there is less of a chance to
mistype. Also, quite often the space group in a reflection file
is not correct anyway (e.g. P222 vs. P212121).

> the CCP4 program scala can output a 'no merge original index' format
> file, which has all 24 i23 symmetry operators in the header - but
> iotbx.reflection_statistics cannot read that either

I think you should be able to use the .sca file with the --symmetry or
--space_group option. I've never tried to work with scala 'no merge
original index' files. If you could send me your files I will look into
fixing both problems.

Peter Zwart wrote:
> Note that iotbx.reflection statistics itself performs a merging step,
> which is less sophisticated then those available in SCALA or scalepack.
> Therefor, it might be worthwile considering supplying merged data to
> iotbx.reflection_statistics.

That's right. The simple merging in the cctbx is meant to be a stop-gap
measure. It is best to merge externally with a program specifically
designed for this purpose.

Thanks for reporting the bug!

Cheers,
        Ralf



More information about the phenixbb mailing list