I wrote a little jiffy for doing this some years ago:
http://bl831.als.lbl.gov/~jamesh/pickup/adsc2pdb.com
However, I should note that this program relies on the DPS program:
dps_peaksearch to pick spots on each image in your data set. These
spots are then transformed into reciprocal space coordinates in the form
of a PDB file, which you can view and rotate around with your favorite
graphics program. At the time I wrote it, the most popular program was
"O", and so the jiffy writes input files for that program. If you don't
have DPS, then a similar jiffy for converting the *.spt files that
MOSFLM generates during the autoindexing process is here:
http://bl831.als.lbl.gov/~jamesh/pickup/spt2xyz.com
This is not exactly the 2D-3D-2D re-binning program that is now being
discussed as the "much harder job", but I have found that transforming
peaks will do in most cases. After all, you are generally looking for
how the spots transform anyway. Even this will make a very large PDB
file in typical cases. The main caveat here is that my jiffy is only
set up by default to work with a particular image type (ADSC) from a
detector and spindle set up in the "usual" way (the way it is at my
beamline). This is because I very rarely have access to other kinds of
images.
I think this underlies the reason why a general program for converting
any "diffraction image" into reciprocal space does not exist: you have
to know the camera geometry, and camera geometries are poorly
documented. Specifically, given the "fast" and "slow" dimensions of the
image file, you need to know not just the pixel size, but the "beam
center". The latter has at least 32 different conventions to represent
it. Nearly all are used by one facility or another, but no image file
format tells you which one it is using. For example, the "beam center"
is usually given as two numbers, but "0 0" can be any of the four image
corners (depending on the beamline), and "X" and "Y" could be "fast" or
"slow". You also need to know if you are in "cameraman view" or "sample
view". Both are popular, but invert the hand of the world if you get it
wrong. Oh yes! and you also need to know the orientation and rotation
direction of the spindle. This is often given as "vertical" vs
"horizontal" and "clockwise" vs "anticlockwise" (which is not enough
information), but most spindles are not precisely 90 degrees to the
x-ray beam, and yes that does impact precisely where the spots fall.
The detector is also not usually perfectly square with the beam either,
and is even sometimes purposefully offset at a "2theta" angle.
Add to all this the heated debates over which direction is "X", "Y" or
"Z" (some think that "X" is the spindle axis, others (like me) think
that "X" is the X-ray beam), and I think you can see the scope of this
problem. Phil makes a good point about memory management problems, but
I think that pales in comparison to the headaches of getting a poor,
defenseless biologist with a stack of images to come up with the 14
numbers needed to uniquely describe the camera geometry. This is
probably why most people just write a program for their favorite image
format (the one down the hall).
Part of the problem, I think, is that there are more combinations of
conventions than there are degrees of freedom in the camera. You really
only need 14 numbers (4 vectors and two scales) to completely describe
an image file from a flat 2D detector and spindle (independent of XYZ
conventions). But that is a topic for another time....
-James Holton
MAD Scientist
George M. Sheldrick wrote:
> As Phil says, constructing an undistorted slice through reciprocal space
> is much harder than displaying integrated intensities. The Bruker APEX2
> software does this nicely and I understand that they can also convert
> MAR CCD and possibly some other frame formats to Bruker format, which
> presumably would be necessary to apply it to your data. Although intended
> for - and widely used by - small molecule crystallographers this should
> work equally well for macromolecules.
>
> George
>
> Prof. George M. Sheldrick FRS
> Dept. Structural Chemistry,
> University of Goettingen,
> Tammannstr. 4,
> D37077 Goettingen, Germany
> Tel. +49-551-39-3021 or -3068
> Fax. +49-551-39-22582
>
>
> On Wed, 5 May 2010, Phil Evans wrote:
>
>
>> I think he's looking for a program which will extract a plane from the raw 3D reciprocal space, as sampled by the raw images (ie before integration, but with the plane defined by the indexed lattice). That's a much harder job
>>
>> Phil
>>
>> On 5 May 2010, at 16:50, Tim Gruene wrote:
>>
>>
>>> Hi Tillmann,
>>> what do you mean by 'raw intensities' as opposed to integrated data?
>>>
>>> Would xprep be an option for you? It reads XDS_ASCII.HKL, but that's of course
>>> after integration.
>>>
>>> But it should be easy to convert any (non-binary) file containing raw
>>> intensities into an hkl-file that you can read with xprep!?
>>>
>>> rlatt might be another program you are looking for.
>>>
>>> Tim
>>>
>>>
>>> On Wed, May 05, 2010 at 03:33:40PM +0200, Tillmann Heinisch wrote:
>>>
>>>> to my knowledge hklview just works with integrated data whereas I need to plot raw intensities along h, k and l to investigate reflection streakings. I heard such software is routinely used in small molecule crystallography.
>>>>
>>>> Tillmann
>>>> On May 5, 2010, at 3:18 PM, David Briggs wrote:
>>>>
>>>>
>>>>> Hi Tillmann
>>>>>
>>>>> Will the CCP4 program HKLview do what you want?
>>>>>
>>>>> http://www.ccp4.ac.uk/html/hklview.html
>>>>>
>>>>> Cheers,
>>>>>
>>>>> Dave
>>>>>
>>>>> ============================
>>>>> David C. Briggs PhD
>>>>> Father, Structural Biologist and Sceptic
>>>>> ============================
>>>>> University of Manchester E-mail:
>>>>> [log in to unmask]
>>>>> ============================
>>>>> http://xtaldave.wordpress.com/ (sensible)
>>>>> http://xtaldave.posterous.com/ (less sensible)
>>>>> Twitter: @xtaldave
>>>>> Skype: DocDCB
>>>>> ============================
>>>>>
>>>>>
>>>>> On 5 May 2010 14:03, Tillmann Heinisch <[log in to unmask]> wrote:
>>>>> Hi,
>>>>> I have problems solving the structure of a protein crystal which seems to be disordered. In order to investigate the disorder it would be useful to have a precision photograph that shows reflections only in the [0kl] plane. Does anyone know software that can transform raw data to give intensity distribution in distinct zones of hkl?
>>>>>
>>>>>
>>>>> Many Thanks for your help,
>>>>> Tillmann
>>>>>
>>>>>
>>> --
>>> --
>>> Tim Gruene
>>> Institut fuer anorganische Chemie
>>> Tammannstr. 4
>>> D-37077 Goettingen
>>>
>>> GPG Key ID = A46BEE1A
>>>
>>>
|
