[ccp4bb]: Twinning problems

[Dr Nicholas Keep <n.keep@mail.cryst.bbk.ac.uk>]

***  For details on how to be removed from this list visit the  ***
***          CCP4 home page http://www.ccp4.ac.uk         ***

Eleanor is right.  We solved a P42212 MAD structure using data from a
crystal that only went to 3.5 A.  (We did have better native).  There
was no evidence for an anomalous signal except possibly in the  low
resolution bins (~15-8).  We did have huge redundancy (approx 12 from
150 degrees) not really a problem unless your crystal is very sensitive.
There was a discussion on twinning in P4 /P422 a few weeks ago.  The
conclusion was that P422 data looks twinned in P4 (including the data
for this structure) but this data gives correct moments in P422 see
http://student.cryst.bbk.ac.uk/~ebrig02/twin/
and the BB archive
Best wishes
Nick 

"Eleanor J. Dodson" wrote:
> 
> ***  For details on how to be removed from this list visit the  ***
> ***          CCP4 home page http://www.ccp4.ac.uk         ***
> 
> Mark Roe wrote:
> >
> > ***  For details on how to be removed from this list visit the  ***
> > ***          CCP4 home page http://www.ccp4.ac.uk         ***
> >
> > Dear All,
> >
> > Not CCP4 I know, but ccp4bb seems to be the best forum to get an intelligent
> > response.
> >
> > The problem is that we appear to be getting twinned crystals, but that
> > neither truncate nor the twinning server shows this up.  We have tetragonal
> > crystals, apparent space group P41212 or P43212. The crystals show 100%
> > incorporation of Se by mass-spec and the fluorescence scan shows a Se edge.
> > We collected Se-SAD data sets at the peak wavelength for five crystals, all
> > diffracting to 2.8 - 3.0A.  The data was processed with mosflm. Parts of the
> > scala and truncate logfiles for one are reproduced below. As you can see, the
> > anomalous R merge is lower than the normal R merge, indicating (as I
> > understand it) that there is little or no anomalous signal. This (as I also
> > understand it) indicates twinning and the twinning cancels out any anomalous
> > signal. The truncate output, though, clearly indicates an
> > untwinned crystal.
> 
>  That isnt necessarily true - Rmerge will increase with the number of
> observations whilst Ranom will probably get smaller as your data
> improves.
> 
> And in fact Ranom and Rmerge are defined quite differently - I think the
> definitions are listed in the log file, but alas, NOT in the
> documentation! Sorry abt that!!
> 
> The cumulative intensity more or less proves it is not a twinned crystal
> - you dont include the 2nd moment plots which are usually the clearest
> indicator.
> 
> >
> > The questions are these:
> >
> > 1) Are the crystals twinned, or is there another explanation?
> 
>  Almost certainly not..
> 
> > 2) If so, why doesn't truncate or the twinning server show this?
> 
> See above..
> 
> > 3) Can any useful info about the twinning be gained from the above two
> > questions?
> 
> see above..
> 
> > 4) Is there any other way of showing the twinning, without the need to
> > collect anomalous data (because otherwise it is going to be a hard slog
> > screening to find untwinned crystals)?
> 
>  What evidence do you have of twinning?
> 
> > 5) Back to finding another crystal form?
> 
>  This one is prob OK.. persistence is a virtue..

-- 
Dr Nicholas H. Keep
School of Crystallography,
Birkbeck,  University of London,
Malet Street,
Bloomsbury
LONDON
WC1E 7HX

email     n.keep@mail.cryst.bbk.ac.uk
Telephone 020-7631-6852  (Room 1083 Office)
          020-7631-6868  (Rosalind Franklin Laboratory)
          020-7631-6800  (Department Office)
Fax       020-7631-6803