[phenixbb] coordinates and map shifted after phenix.refine without simulated annealing

[hari jayaram]

Hello Tom ,
Thanks for your speedy reply

My space group is P1 , and the shift is by around an 1.2 A away and very
much in the vicinity of the input model and experimental map
I suspected the rigid body step (
rigid_body+individual_sites+individual_adp+tls) might be causing this and so
I am trying a refinement without the rigid body.
Also I am quite close to convergence and that step should not be necessary
as a first step.

But I am still confused as to why it shifts the molecule very close by , but
slightly off , like a off kilter MR solution.


Hari


On Sat, Jun 6, 2009 at 10:40 AM, Thomas C. Terwilliger <terwilliger at lanl.gov
> wrote:

> Hi Hari,
> What is your space group?  I am guessing that the shift is along an axis
> where the origin is not fixed, for example the b axis in P21. This could
> happen with rigid-body refinement where a molecule can move as a whole so
> that the center of mass could move along that axis.  Perhaps someone else
> can comment as to whether there is or could be an optional term in
> phenix.refine to maintain the center of mass at the same coordinate in
> such cases.
> All the best,
> Tom T
>
> >> Hi i am using phenix version 1.4 release tag 70
> >>
> >> hari at charlie:~$ phenix.refine --version
> >>
> -------------------------------------------------------------------------------
> >>   PHENIX: Python-based Hierarchical ENvironment for Integrated
> >> Xtallography
> >>   Version: 1.4
> >>   Release tag: 70
> >>   Platform: intel-linux-2.6-x86_64 linux
> >>   User: hari
> >>
> -------------------------------------------------------------------------------
> >>
> >>
> >> I am seeing a strange shifting of my coordinates "out of my experimental
> >> phases map" before and after refinement without simulated annelaing
> >>
> >> I am at the final stages of my refinement. When I refine without
> simulated
> >> annealing the resulting model is shifted relative to my experimental
> >> phases
> >> map into which I build my model
> >> The phenix command line strategy i am using are
> >> rigid_body+individual_sites+individual_adp+tls .or the same strategy
> with
> >> simulated_annealing=true
> >>
> >> The output map coefficients fortunately match my model in both cases .
> But
> >> I
> >> am wondering why there is a difference.
> >> I am using the mlhl target for refinement to 3.1 A , My R and rfree is
> >> 28/31
> >> .
> >>
> >> Any ideas why my model shifts without simulated annealing . All pdbs
> have
> >> the same CRYST  record .
> >> I am using coot  version 0.61 for my building and map display using fft
> >> inside of coot .
> >>
> >>
> >> hari
> >> _______________________________________________
> >> phenixbb mailing list
> >> phenixbb at phenix-online.org
> >> http://www.phenix-online.org/mailman/listinfo/phenixbb
> >>
>
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://www.phenix-online.org/mailman/listinfo/phenixbb
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://phenix-online.org/pipermail/phenixbb/attachments/20090606/44db0b5d/attachment-0003.htm>