Re: [ccp4bb]: Twin problem

[Phil Jeffrey <phil@xray2.mskcc.org>]

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

On Wed, 23 Oct 2002, Katja wrote:


> Measuring my second, less twinned crystal (B) yielded a dataset that I
> couldn't successfully integrate using the space group C222. Instead, this
> crystal appeared to belong to space group P2.
>
> I concluded, that our crystals belong to space group P2 and that the
> additional axis is an artifact caused by the twinning.

Without knowing cell dimensions this sounds like a lot like twinning that
occurs when the P2 or P21 lattice approximates C222 or C2221 via a
combination of a, c, beta lattice values.  Typically there would be 2
(or more) molecules per asymmetric unit, with the non-crystallographic
symmetry axis parallel to the twin-operation axis.

If this is correct, the pseudo-orthorhombic axis would be coincident with
the monoclinic a, the pseudo-ortho c/c* axis would be along the monoclinic
c* axis.  I don't have the math for the desired a/c/beta relationship to
hand, but if this is the case, the twin relationship is

(h,k,l) -> (h,k,-h-l)

If this is the case I'd expect that your "P2" data would show pseudo-mmm
symmetry that would be relatively obvious in an (e.g.) self-rotation
function using POLARRFN (which would also confirm the direction of the
twin-axes and therefore the twin-operator).

> Since molecular replacement using the (admittedly not very good) search
> model again gave no result, and modifying and changing my search model
> didn't help either. I didn't get an AMORE solution that I could use to
> determine the twin-operator.

It may be possible to solve the structure using the data with the lower
twin fraction using AMORE and the space group P21 or P2 even without
twinning.  Detwinning data with a relatively high twin fraction is prone
to introducing error as the difference between the twin-related
intensities approaches the noise level in the data, but there's no reason
not to try it.

The article to read would be
T.O. Yeates, Detecting and Overcoming Crystal Twinning (1997)
Meth in Enzymology vol. 276, page. 344

There are a bunch of scripts in CNS to do what you want, and
http://www.doe-mbi.ucla.edu/Services/Twinning/intro.html
http://www.doe-mbi.ucla.edu/Services/Twinning/
may come in useful.

Regards
Phil Jeffrey
-------------------------------------------------------------------------------
| Phil Jeffrey                                  |                             |
| Crystallography Facility Manager              | If you lie to the compiler, |
| Memorial Sloan-Kettering Cancer Center, NYC   | it will get its revenge     |
| phil@xray2.mskcc.org                          |     - Henry Spencer         |
| Voice: (212) 639 8547   Fax: (212) 717 3135   |                             |
-------------------------------------------------------------------------------