[phenixbb] Omit map

xaravich ivan xaravich.ivan at gmail.com
Tue Nov 22 23:13:23 PST 2011


Hi,
 I am also a bit confused about your suspicion. If your Rwork/Rfree
22.7/27.9 for a 2.7 angstrom data, you have probably reached the so called
endpoint of your refinement, unless Molprobity shows serious deviation.
 What is the basis of your skepticism? Were you expecting this structure to
be different from the bacterial model to fit your hypothesis? I would be a
bit careful in such situations!

On Tue, Nov 22, 2011 at 10:01 PM, Pavel Afonine <pafonine at lbl.gov> wrote:

> **
> Hi Rajesh,
>
> model bias, in a nutshell, is when you see something in the map that is
> conveyed by the model (that may not be correct) and not by the data.
>
> I am not sure how one can get an idea about model bias looking at triplet
> of numbers: Rwork, Rfree and Molprobity percentile. By the way, your
> Rwork/Rfree look pretty good (given the 2.7A resolution).
>
> Anyway, if everything else: bunch of model and data statistics, local and
> global model-to-data quality fit all look well (*), and you are still
> suspecting something isn't right, then you can invest from a few days to a
> week or so into computing an "Iterative Build OMIT map" as described here:
>
> Iterative-build OMIT maps: map improvement by iterative model building and
> refinement without model bias. T. C. Terwilliger, R. W. Grosse-Kunstleve,
> P. V. Afonine, N. W. Moriarty, P. D. Adams, R. J. Read, P. H. Zwart and
> L.-W. Hung Acta Cryst. D64, 515-524 (2008)
>
> This map is supposed to be bias free.
>
> Pavel
>
> (*) For more details see here:
> http://www.phenix-online.org/presentations/latest/pavel_validation.pdf
>
>
>
>  I have a structure 2.74 A structure with R/freeR 22.7/27.9%. But the
> molprobity score is 50th percentile.
> If I fix the outliers they wouldn't go away and molprobity score remains
> same or get worst.
> This structure was solved by molecular replacement using a bacterial
> protein with 38% identity.
> I am wondering if there is an model bias? I have not identified particular
> region in model for bias.
>
>  My questions are: Am I right in thinking if there is model bias based on
> above info?
> I  added my .pdb and .mtz and do SA-composit omit map. This is running
> from morning and I have feeling that I will get results after thanksgiving.
> Is this the correct way of doing a omit map for the purpose I mentioned?
> I saw lot of discussions on omit map but couldn't grasp everything and I
> am sure same questions might have been asked before.
> I have couple of classical papers next to me but I need help.  I would
> appreciate if you could help me out either suggesting me how and what to do
> or pointing to previous thread over the this BB.
>
>
>
>
> _______________________________________________
> phenixbb mailing list
> phenixbb at phenix-online.org
> http://phenix-online.org/mailman/listinfo/phenixbb
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://phenix-online.org/pipermail/phenixbb/attachments/20111122/63014d71/attachment.htm>


More information about the phenixbb mailing list