[phenixbb] bad geometry problem

Nathaniel Echols nechols at lbl.gov
Tue Jun 19 06:02:46 PDT 2012


On Tue, Jun 19, 2012 at 1:17 AM, Qian Sarah <oceanlineqwm at gmail.com> wrote:
> Hi, this may not be a new problem, but I tried every stretagy that I know
> from bb, but the geometry is still very bad. So I wish I can get some help
> from here.
> Crystal information:
>  SAD data. Resolution 3.0.  There is no anisotropy in the data.
>  R/Rfree : 22.7/26.8 (Rfree-flag is always from the Refine_data.mtz file
> generated after the first refinement.)
>    RMS(bonds):0.02
>    RMS(angles): 1.453
>    Ramachandran outliers: 3.0%
>    Ramachandran favored: 91.2%
>    Rotamer outliers: 18.7% !!
>    Clashscore: 31.84
>    overall score: 3.47
>    I tried different strategies, like autobuild and then refine; set
> wxu/wxc=0.3/0.3; turn on "Ramachandran restrains", "fix bad sidechain
> romaters"etc, however the paramters are still very bad, especially the
> rotamer outliers. Then I rebuild them manually in coot,after refinement the
> parameters are even worse.

A few suggestions:

1. Don't use Ramachandran restraints in Phenix - please!
2. Do use Ramachandran restraints when rebuilding in Coot - fix the
problems manually, perform local real-space refinement, and check that
the result makes sense.  Ensuring that the geometry is as good as
possible before you run another round of refinement will make Phenix
work better.
3. Don't adjust the weights manually; use the automatic weight
optimization.  It takes a long time to run (unless you have lots of
CPU cores), but it's going to save you time in the end.  My guess is
this is going to help with some of the
4. At this resolution the rotamer fitting in phenix.refine is unlikely
to work very well.  You need to fix these problems manually too, I'm
afraid.

If you're willing to send us the model and data (off-list), we may
have additional suggestions.

> p.s. phenix version: phenix-1.7.3-dev-1046.

You should update to the latest official release.

-Nat


More information about the phenixbb mailing list