Re: [ccp4bb]: CCP4 : Space group transformations

[Edward Berry <eaberry@lbl.gov>]

***  For details on how to be removed from this list visit the  ***
***    CCP4 home page http://www.dl.ac.uk/CCP/CCP4/main.html    ***

David A R Sanders wrote:
> 
> ***  For details on how to be removed from this list visit the  ***
> ***    CCP4 home page http://www.dl.ac.uk/CCP/CCP4/main.html    ***
> 
> I have a sorta double-barrelled/loaded question. I'm working on a protein
> that we have managed to collected data for in 2 different space-groups.

(:>) Could it be the subscribers are more interested in PC's for crystallography
than crystal space group transformations? Or more likely the answers are so
obvious to the experts that they all answered in private, to avoid boring
the rest of us? I hope you will summarize your responses.

We have one "crystal from hell" that crystallizes in multiple, obviously
related, space groups. The part we understand is transformation from
orthorhombic (with a=b within experimental error) to tetragonal (with
identical cell dimensions). A sharp Post-doc recognized this as an 
indication of twinning, which he confirmed by the twinning test
at http://www.doe-mbi.ucla.edu/Services/Twinning/ and by helpful communications
with its author, Todd Yeates. The a=b condition makes the symmetry of the
lattice tetragonal but the symmetry of the cell contents is orthorhombic.
Auto-indexing couldn't distinguish a from b, so in the case of the un- or
partially twinned data we had to reindex half the datasets before merging.
In the tetragonal space group, the crystal growth process also failed to 
distinguish, or maybe used both orientations in pairs, and we had
R-merge below 5% when reducing as tetragonal.

The part that still puzzles us, and makes me interested in
your crystal: on one dataset the c axis increased by a factor of three, as
if the new cell consists of three of the original cells, i.e. breakdown of 
translational symmetry between cells but with periodicity-3 repeat. I'm having
a hard time imagining what could cause that, and I wonder if there are 
any cases in the literature or people's notebooks where such a crystal has 
been solved and the cause of the space-group or unit cell transformation is 
understood. (The c-axis is now ~450 A, and we haven't yet reduced the data
well enough to say whether it is tetragonal or orthorhombic).


> 
> The second question is that having figured out the transformation, how do
> you fool AMORE into doing a transformation search only along the one axis
> necessary ??
> 
For the second question about amore, how to enforce translation along one axis,
I think you can do the following:
 

1. do tabling from the model in your solved crystal with keyword NOROTATE (and 
noshift if the centre of mass is near the origin), or if you haven't got
a model yet use structure factors from cut-out but un-rotated electron density
of one monomer.

2. Skip the ROTing step, but create an RF solution with no rotation:
SOLUTIONRC 1 0. 0. 0.  0. 0. 0.

3. Do translation search as usual. This involves calculating a map of the
translation function and picking peaks, and I don't there there would be 
significant savings in CPU time from restricting the map to one axis. But you 
can screen the solutions by choosing only those with zero translation along 
two axes after the REORIENT step (or without reorient, if you tabled with 
NOROTATE NOSHIFT)

4. FITing as normal
-- 
Edward A. Berry, MailStop 3-250
Lawrence Berkeley National Laboratory
1 Cyclotron Road, Berkeley, CA 94720
Phone: +1-510-486-4335
Fax: +1-510-486-6059
Jfax +1-530-323-9836 (you send fax, I receive email)
e-mail: EABerry@lbl.gov
> =======================================================================
> It is characteristic of thoughtful people that they don't understand some
> things that to others are as plain as a pikestaff.  -  A. G. Cairns-Smith
> =======================================================================
>                         David J. Schuller
Hey, that was my excuse! - EAB