Dear Jason,
as a follow-up to Randy's excellent suggestions, you could try to use
the anomalous scattering from intrinsic sulphur atoms (if there are any
in the structure, and you have good anomalous completeness) to improve
your phases. You can use Phaser in MR_SAD mode and your current model to
find these sites. The signal is probably not very strong, but may be
sufficient in case you only need a little push to obtain an
interpretable map for the missing domain.
BW, Gabor
On 2016-05-26 13:50, Randy Read wrote:
> Dear Jason,
>
> With data to only about 3A resolution and a model missing more than a
> third of the structure, I'm not surprised that you can't get any
> further with the missing part. Depending on how conservative you were
> when deciding on the resolution limit, it's possible that you could
> improve things a bit by processing the data to the highest possible
> resolution.
>
> Your best hope would normally be to get some additional phase
> information. The top possibility would, of course, be experimental
> phasing, but that's clearly difficult if you've run out of crystals
> and can't reproduce them! Another theoretical possibility is
> cross-crystal averaging, if (for instance) you collected data from
> multiple crystals and some of them were poorly isomorphous.
>
> If you were using PDB BLAST to find homologues, you might have missed
> some distant relatives. Have you tried a more sensitive search, such
> as HHpred? If you had something with the right fold, then you might
> either be able to add it to your existing model by molecular
> replacement, or do real-space molecular replacement into the density
> you're seeing.
>
> Do you have anything left of your original crystal to use for seeding
> experiments?
>
> Good luck!
>
> Randy Read
>
>
>> On 26 May 2016, at 13:29, Jason <[log in to unmask]> wrote:
>>
>> Dear ccp4 users,
>>
>> I have been trying to solve the structure of a protein which has 760
>> residues. The protein has three domains and there is no homologue
>> structure for the first domain which is about 300 residues. The top
>> hit from PDB blast gave an identity of 36%. The resolution was 2.97 A.
>> I have tried different Molecular Replacement programs with different
>> models, the best solution, which was a partial solution, was from
>> MrBump and had only the last two domains in the structure. These two
>> domains fit in the density well and I have manually built that part a
>> few time, so I am afraid that there is nothing I could do with these
>> two domains at this stage. The Rfactor and Rfree are about 31% and 37%
>> now. I have also tried many tracing and autobuild programs from both
>> ccp4 and phenix, nothing really helped. Things like phase improvement
>> and solvent flattening have also been tried. I could definitely see
>> some density map for the first domain but it is difficult to trace the
>> chain from coot at this resolution. The statistics are all ok with a
>> P21 space group. I did check other possibilities of the space group.
>> This crystal was from a screening and I could not reproduce it at this
>> stage.
>> So does anyone have any experience in solving this kind of problem?
>> Actually, I also have another similar case at the moment.
>> Many thanks in advance.
>>
>> All the best,
>>
>> Jason
>>
>> University College London
>
> ------
> Randy J. Read
> Department of Haematology, University of Cambridge
> Cambridge Institute for Medical Research Tel: + 44 1223 336500
> Wellcome Trust/MRC Building Fax: + 44 1223 336827
> Hills Road E-mail: [log in to unmask]
> Cambridge CB2 0XY, U.K.
> www-structmed.cimr.cam.ac.uk
--
##################################################
Dr Gabor Bunkoczi
Cambridge Institute for Medical Research
Wellcome Trust/MRC Building
Addenbrooke's Hospital
Hills Road
Cambridge CB2 0XY
##################################################
|