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



Hi Markus, Juergen and Peter,

I agree and just adding my experience to what Peter has already written.

In addition to scream we can do some simple tests which can tell the 
possibility of refinement. In
my case, Britton plots (and relative height of the off-origin peak) did help 
for five of twinned
(pseudotranslated) data sets we had (Monoclinic P21 Vs Orthorhombic C2221). It 
could categorize
depending on the twinning fractions which was eventually visible in the 
refinement statistics. Two
data sets with very low twinning fraction (and lower off-origin peak) did well 
with Refmac5 only.
Rests of the three were refined with ShelxL and extent of the refinement 
matched with their
twinning fraction. Data would be worth working as much the twinning fraction is 
far from 0.5.
Please note that, the twining fractions suggested by Britton plots were much 
lower than the one
suggested by ShelxL in presence of pseudo-translation.

I am curious to know the unit cell dimensions and height of off-origin peak. I 
am sure Eleanor
would be equally inquisitive.

regards
Manish




Peter Zwart <[EMAIL PROTECTED]> wrote:

Hi Markus, Juergen,

I doubt slippage would result in a pseudo translational vector.
In some circumstances, two crystals sitting on top of each other can 
however produce this effect, it is called non-merohedral twinning.

trying not to embarrass myself, see here an (2D) unit cell, with *'s 
denoting the corners:

*-----------*
| |
| |
*-----------*


Now be creative and stick two together:

*-----------*
| |
| |
*-----+-----*
| |
| |
*-----------*

Note that this set of two unit cells has a higher symmetry than the 
single cell: a(n approximate) four fold axis is present at the +. If you 
rotate the two unit cells (or sublattices) you get


*-----------* *-----*-----*
| | | | |
| | | | |
*-----------* * | *
| | | | |
| | | | |
*-----------* *-----*-----*


overlaying these two you get something like:

*-----*-----*
| | |
| | |
*-----|-----*
| | |
| | |
*-----*-----*

If you have a situation like this, and if it is indexed like the small 
unit cell, you will have a large number of unaccounted spots, so you 
actually might end up with an indexing solution that has the 'super 
cell' shown above, albeit with a number of indices that have no 
significant intensity (given rise to large peaks in the Patterson: your 
pseudo translational symmetry)

Not sure what to do with non merohedrally twinned data, besides panick, 
scream and/or hide (especially on Halloween). I believe shelxl is (of 
course) able to deal with it during refinement.

There are two papers that inspired my explanation (or blatant copy) as 
seen above:

Parsons, Acta Cryst. (2003). D59, 1995-2003
(above example, but nicer and more thorough explanation)
Dauter, Acta Cryst. (2003). D59, 2004-2016
(general twinning stuff and nice images of non-merohedral
diffraction patters)

Note that the presence of pseudo translational symmetry is not that 
uncommon, and the excursion into non merohedral twinning as seen above 
might be simple paranoia....

Cheers,

Peter


Juergen Bosch wrote:
> *** For details on how to be removed from this list visit the ***
> *** CCP4 home page http://www.ccp4.ac.uk ***
> 
> 
> Hi Markus,
> 
> since you are in desperate mode, I would recommend XDS for integration 
> to rescue your dataset and hopefully come up with something decent to 
> work with.
> Regarding the pseudotranslation I would tend to say you can't get this 
> from a split crystal or from two crystals in the loop - but I'm sure 
> some expert will prove me wrong with my statement or perhaps not :-)
> 
> Juergen
> 
> 
> Markus Rudolph wrote:
> 
>> *** For details on how to be removed from this list visit the ***
>> *** CCP4 home page http://www.ccp4.ac.uk ***
>>
>>
>>
>> This may be a PCI question:
>>
>> Is it possible to introduce pseudo-translation by integrating a 
>> diffraction pattern from a split crystal or when two crystals in a 
>> loop with a defined orientation have been collected on? It might be 
>> that during integration the mask slips from one pattern to another. 
>> Some sharp reflections are integrated single when the box size is 
>> small enough while others are split so little that they are integrated 
>> as a single reflection. Is there software that could help in 
>> integrating and separating such a split pattern (Bruker software does 
>> not take marccd images)?
>>
>> Ah, and no, there is no other dataset, no other crystal, no more protein.
>>
>> Thanks for any input,
>>
>> Markus
>>
>> ___________________________________________________________________________________
>>  
>>
>>
>> Markus Rudolph phone: +49 551 39-14088
>> Dept. Mol. Struct. Biol. fax: +49 551 39-14082
>> Georg-August University e-mail: [EMAIL PROTECTED]
>> Justus-von-Liebig-Weg 11



Center for Advanced Research in Biotechnology 
University of Maryland Biotechnology Institute 
9600 Gudelsky Drive, Rockville 
MD 20850  USA 
Tel: +1-240-314-6130;  fax: 240-314-6255


 
____________________________________________________________________________________
Low, Low, Low Rates! Check out Yahoo! Messenger's cheap PC-to-Phone call rates 
(http://voice.yahoo.com)

Reply via email to