[phenixbb] inputing tNCS into phenix?
rjr27 at cam.ac.uk
Wed Apr 23 00:33:12 PDT 2014
There could be a number of things going on, depending on other factors. If you had a relatively poor starting model and relatively low resolution, then that could be responsible. However, if you’ve got a reasonably good (high sequence identity) starting model and reasonable resolution, then I’d expect to do better even in the presence of tNCS — especially if the Patterson peak was weak enough that xtriage didn’t clearly flag it as indicating tNCS.
tNCS often arises when a local symmetry operator (e.g. a dimer axis) is parallel to a crystallographic symmetry axis. In a subset of these cases, the NCS symmetry is close to crystallographic and sometimes it’s not obvious which is the NCS and which is the crystallographic symmetry. The old version of Phaser (before the tNCS analysis and correction) often made the wrong choice, but the new version is usually pretty good at making the right choice, at least if it is asked to check all possible space groups and the tNCS is being accounted for in the search. Did you let Phaser try both P2 and P21 when you ran the molecular replacement search?
On 22 Apr 2014, at 19:52, Yarrow Madrona <amadrona at uci.edu> wrote:
> Thanks Randy,
> I guess my next question is whether, a stalled Rfree of 40% for data with otherwise good processing stats and a P21 space group could possibly be due to tNCS or if another reason is also likely?
> On Tue, Apr 22, 2014 at 7:34 AM, Randy Read <rjr27 at cam.ac.uk> wrote:
> Dear Yarrow,
> Phenix.refine can impose NCS restraints for general NCS, and that is handled pretty automatically without having to specify a rotation and translation, but that’s probably not what you’re asking. In the particular case of tNCS, there are problems arising from the fact that there is an overall modulation to give systematically strong and systematically weak intensities, whereas the refinement likelihood targets implicitly assume that all the intensities (and, more importantly, the errors in predicting the intensities from the model) have a smooth distribution.
> For molecular replacement, the performance of Phaser has been improved significantly by accounting for these tNCS effects and, in principle, it would be great to do the same in refinement. That’s pretty high up on the to-do list — and the same theory used in Phaser would apply, but at the moment there isn’t any way yet to account for this in refinement.
> Best wishes,
> Randy Read
> On 22 Apr 2014, at 14:51, Yarrow Madrona <amadrona at uci.edu> wrote:
> > Hello,
> > I recently posted a question about stalled R-factors and and obtained excellent help resulting in the realization that I have tNCS even though it is not picked up by X-triage. I know the transformation matrix of one molecule onto another but I am not sure how to input this into phenix (both the translation and the rotation) to improve refinement. Is there a way to do this? I'm sorry if this has already been covered.
> > -Yarrow
> > _______________________________________________
> > phenixbb mailing list
> > phenixbb at phenix-online.org
> > http://phenix-online.org/mailman/listinfo/phenixbb
> 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: rjr27 at cam.ac.uk
> Cambridge CB2 0XY, U.K. www-structmed.cimr.cam.ac.uk
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: rjr27 at cam.ac.uk
Cambridge CB2 0XY, U.K. www-structmed.cimr.cam.ac.uk
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the phenixbb