Hi Genevieve,
The first thing I would do in this case is to examine the diffraction images to see that there are no strange artefacts. For fine-sliced Eiger or Pilatus data, you may have to merge say 10 images to get images that can be examined by humans. Maybe the synchrotron people could help you here.
Is there any possibility, as Michael suggested, that your molecule has an internal 6-fold screw axis?
IDXREF (in XDS) produces a list of possible space groups and I guess the other programs will do the same. Is there any possible space group listed with 6-fold symmetry? In that case you could force the program to take this space group for processing.
How many molecules do you expect in the asymmetric unit? Are there many, which could cause all kinds of funny NCS, or just one or a few, which should behave somewhat better?
Best,
Herman
-----Ursprüngliche Nachricht-----
Von: CCP4 bulletin board [mailto:[log in to unmask]] Im Auftrag von Genevieve Evans
Gesendet: Mittwoch, 31. August 2016 11:48
An: [log in to unmask]
Betreff: [ccp4bb] 2-fold screw axis or not?
Hi All,
I've been working with datasets for a protein crystal that I originally thought would be solved using MR, but so far I haven't been able to find a solution.
There appears to be something not quite right about the space group.
The highest symmetry Bravais Lattice with best quality of fit is oP (= orthorhombic primitive) with a unit cell of 29 92 94 90 90 90 in both XDS & iMosflm. POINTLESS auto-processes the datasets as P212121. However, in all the datasets I've worked on, the axial absences (from POINTLESS analysis) along one axis look more like 6-fold screw axis than a 2-fold screw axis .
An example of this odd axial absences pattern from my best resolution dataset (AIMLESS, with conservative cut-off at 1.66 Angstrom, gives Rpim= 0.567 in outer shell & Rpim= 0.023 overall):
Index I sigI I/sigI I'/sigI
1 -2 2 -0.94 0.47
2 4 2 1.61 1.96
3 4 3 1.36 1.67
4 5 4 1.11 1.50
5 3 5 0.68 0.00
6 5292 54 97.68 97.67
7 27 7 3.88 0.00
8 14 8 1.87 1.78
9 5 8 0.58 1.35
10 -7 8 -0.86 0.92
11 5 8 0.68 0.00
12 4126 43 96.33 96.32
13 18 18 1.04 0.00
Additional when I look at the mtz file produced by AIMLESS, in reciprocal space using Phenix 2D map viewer, the data appears to have "stripes" in one direction, which correspond to this axis. I believe the stripes are due to the index numbers along one axis being 4-5 fold less than the index numbers along the other axes.
Any ideas or suggestions? Perhaps someone else has come across a similar situation?
Thanks,
Genevieve
Dr Genevieve Evans
------------------------
School of Chemistry and Biochemistry
The University of Western Australia (M310)
35 Stirling Highway
CRAWLEY WA 6009
Australia
|