[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [ccp4bb]: scale failure

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

Dear Yuan Ping,

I am not sure if it still happens but back in the dark ages it was quite easy
to give the wrong centre position, so that the pattern was indexed with the
spindle axis (here l) wrong by + or -1. Then the pattern could still be
predicted but with incorrect crystal angles and cassette angles giving worse
agreement than the true matrix (but you may not have that for comparison).
David Stuart went as far as moving the centre systematically and looking for
the best fit but it is worth noting that the centre needed to be moved and
fixed until the angles have adjusted. In your case it is quite easy to check as
the image shows a mirror plane and XDISP gives the index if you click on the
spot. So is the mirror at l=0?

Best wishes 

John Jenkins

Institute of Food Research     Tel:   +44 (0)1603 255319
Norwich Research Park          Fax:   +44 (0)1603 507723
Colney Lane                    E-mail john.jenkins@bbsrc.ac.uk
Norwich NR4 7UA

Edward A. Berry wrote
>Yuan Ping wrote:
>> Dear all,
>>      I got hexagonal crystals of a protein. My previous data show the space
>> is P6122. Recenly I put one of my best crystals for data collection. The
>> frame looks good (I have attached it). And I use DENZO program to index the
>> data. All previous step go on smoothly, until I scale the data. The scale
>> result is very bad ( Part is attached). The linear R factor is about 0.516
>> and square R factor is 0.922 and the chi**2 is obviously wrong. But I
>> notice that the chi**2 of each frame during process is near 1 which is OK.
>> Could anyone help me to figure out the possible cause?
>> Thanks a lot!
>Beautiful data! (the diffraction pattern, I mean).
>Chi^2 of each frame are positional (x, y) chi^2, low value indicates your 
>lattice fits well. Chi^2 in scalepack (between "initial scale factors" and
>"final scale factors") is scaling error, and indicates "equivalent"
>after scaling have very different value. Have you tried using a lower symmetry
>(P6) in scalepack? (No need to re-run denzo integration with the new space
>Maybe the true space group is P61 and the previous data were twinned with the
>polar axis running both directions giving false 2-fold axes?
>Another possibility is that something you did to get around the close spot
>interferes with accurate integration. You didn't use "overlap none", did you?
>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