[phenixbb] Phaser and anisotropy question
det102 at uoxray.uoregon.edu
Thu Sep 10 09:46:00 PDT 2009
Sharpening the map to aid the human eye is one thing, modifying
the data you put into refinement is another matter completely.
People have been sharpening low resolution maps for model-building
for years, anisotropy has nothing to do with it. It seems to be
easier for the eye to pick out features in the contours of a
sharpened map when the resolution is low. There is no reason to
expect that the sharpness of the data in the best diffracting
direction is the optimal sharpness for model-building. You may
wish to sharpen more depending on your data set and your eyes.
Refinement programs don't care. If you up weight the high
resolution terms your average B will drop, the Fcalc's also up
weight and you end up with the same difference coefficients, the
same shifts, and the same results. All that happens of significance
is that you have inappropriately lowered your atomic B factors
and your Table 1 gives the reader an overoptimistic view of the
quality of your model.
If you have a low resolution map, by all means go ahead and
look at a sharpened version (as I said, anisotropy has nothing
to do with it). You must, however, refine your model against
the data you actually collected.
Peter Grey wrote:
> Dear Prof. Read,
> Thank you very much. Will this improved anisotropic scaling be
> implemented in other modules of phenix.refine ? If not do you advise
> using F_ISO, SIGF_ISO from the newer version of phaser as the input
> labels for refinement ?
> On Thu, Sep 10, 2009 at 2:31 PM, Randy Read <rjr27 at cam.ac.uk
> <mailto:rjr27 at cam.ac.uk>> wrote:
> Yes, the data used in the Phaser map coefficients have been rescaled
> to remove the anisotropy, so the map should look more isotropic than
> the one you get from running SIGMAA. But note that neither program
> does a proper bulk solvent correction.
> Until recently (e.g. in version 2.1.4 of Phaser), the rescaling was
> done so that the overall average falloff of diffraction was preserved,
> i.e. the weakest direction was scaled up and the strongest direction
> was scaled down. However, we were inspired by a paper from Mike
> Sawaya to look at this again. He showed some convincing results that
> the maps are more interpretable if the weak data are all scaled up to
> the falloff of the strongest direction, and the tests we did agreed
> with this. So that is the behaviour you'll get in recent nightly
> builds. This agrees with your worry that the scaling could diminish
> the strongest reflections too much, as happened in the older versions
> of Phaser.
> I hope that helps!
> Randy Read
> On 10 Sep 2009, at 12:58, Peter Grey wrote:
> > Dear Phenix users,
> > I have a very anisotropic data as phaser reports anisotropic deltaB
> > = 60.2. I would be grateful for advice of several issues.
> > 1.Could you please tell me if phaser map coefficients FWT,PHWT take
> > into account the anisotropic scaling ?
> > 2.This means that these coefficients will be different from those
> > calculated from a partial model in sigmaa because sigmaa has no
> > anisotropic scaling (and no bulk solvent correction) ?
> > 3.In the case of such severe anisotropy can the scaling diminish too
> > strongly the well measured high resolution reflections ? If so
> > should I calculate the coefficients my self by sigmaa and not use
> > pahser mtz output or is there a better solution ?
> > Many thanks in advance for sharing your thoughts and experience,
> > Peter.
> > _______________________________________________
> > phenixbb mailing list
> > phenixbb at phenix-online.org <mailto:phenixbb at phenix-online.org>
> > http://www.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 <mailto:rjr27 at cam.ac.uk>
> Cambridge CB2 0XY, U.K. www-
> structmed.cimr.cam.ac.uk <http://structmed.cimr.cam.ac.uk>
> phenixbb mailing list
> phenixbb at phenix-online.org <mailto:phenixbb at phenix-online.org>
> phenixbb mailing list
> phenixbb at phenix-online.org
More information about the phenixbb