Hi Jeremy,
Just looking at the stats:
1- you could try to remove the low intensity reflections (are they
reflections, or are they just indices - it's difficult to say), i.e. the
first 4 shells in your tables dealing with intensities;
2 - you could try to truncate the high resolution to 1.6 A.
The latter will not solve your problem of a high R-factor though.
Further:
what happens (R-sym) when you integrate in P2 with a = 39.633, b =
18.232, c = 77.526, beta = 104.823 ?
Are there any extinctions present on the (new) b* axis, indicative of a
2(1) screw axis then ? (you could also check, in your current list of
reflections, if there are apparent extinctions along the a* axis - this
would strongly suggest a P2(1) space group).
Fred.
?????? wrote:
Dear Kelly and all who replied,
Thanks very much for your valuable suggestions. I'm so sorry that the
0.6 Rsym is refering to the highest resolution shell. Please find the
attached log file (only the bottom sections) for more details. I've
also attached a picture of the refined density with Rfree=0.4.
I've tried all possible spacegroups and P1 gives the best results
(Rfree=0.4) so far.
Dear Ken, I'm sorry that I could not release the pdb at this moment.
However, we may try your suggestion at later stage.
Regards,
Jeremy
>Given your unit cell parameters + high Rsym I'd say you have an indexing
>problem. If you try P2, what happens? I suspect that you might have
>something as simple as incorrect beam center position and while
>integration works, scaling fails (the only way you are getting away with
>it is by choosing P1 ans thus low redundancy).
>
>Show us your scaling statistics. Some programs are more sensitive than
>others to incorrect beam center, your best bet is to try something like
>Labelit and verify that you are using correct values. On a diffraction
>pattern, sometimes you can identify rings of reflections, one of them
>may pass through the beam center. What software you are using for data
>processing?
>
>On Mon, 2010-07-12 at 11:00 +0800, ?????? wrote:
>> Dear ccp4 experts,
>>
>> I'm a beginner and have a problem with the following structure
>> solution. I need great help from you experts. Thanks in advance!
>>
>> The dataset is 1.6A resolution and the spots looks ok, except some
>> smear in the higher resolution. Space group is P1(could be P2 or P21).
>> Completness is above 95% and mosaicity is less than 1. Rsym is 0.6.
>>
>> Cell parameters are: a=18.2320 b= 39.6330 c=77.5260 alpha=104.8230
>> beta=90.0780 gamma=90.0400
>>
>> The model is about 60 amino acids snake venom protein with 60%
>> identity. However, no solution with an initiall R less than 0.7 was
>> obtained by Molrep, Phaser, or Phenix .
>>
>> Using a molrep search result (R=0.75, mtz in P1, 2 moelcules in
>> AU), with Refmac5 twin refinement, we could refine the structure to
>> 40% Rfree and 38%Rwork. The density looks good (can see hole in rings
>> structures) for all residues and no continous unexplained density is
>> observed. However, the R and Rfree does not go down no matter what we
>> try...
>>
>> We tested the twinning by CNS or the crystal twinning server
>> (http://nihserver.mbi.ucla.edu/Twinning/) and found no twinning
>> present.
>>
>> I'm sure the sequence is correct.
>>
>> I've tried hard on this for quite some time. Really appreciate if you
>> could help.
>>
>> Thanks and regards,
>>
>> Jeremy
>>
>>
>>
>
>--
>"I'd jump in myself, if I weren't so good at whistling."
> Julian, King of Lemurs
>
------------------------------------------------------------------------
???????????????????????????????????????? <http://qiye.163.com/?163>
------------------------------------------------------------------------
