Thanks for all the great suggestions. They included beamline problems,
detector-specific problems, translational NCS, narrow dynamic range problem
for strong low resolution reflections,radiation damage, and more. The
agreement seems to be that no matter what the reason for the high Rsym is,
it is probably save to publish the structure. I am curious about one
specific suggestion about interference observed in mostly helical proteins:

 Helices, due to their regular structure, cause an interference effect that
reduces the intensities of reflections in that resolution range. It should
be visible in the diffraction images. It appears as if there is a broad
region with weak or even missing reflections. The reflections are stronger
again at higher resolution.

The effect on diffraction is exactly what I am seeing and my proteins are
probably about 80% helical. Does anybody have an opinion about this?

Ursula




On Tue, Oct 15, 2013 at 2:13 AM, Kay Diederichs <
kay.diederi...@uni-konstanz.de> wrote:

> Hi Ursula,
>
> if you look at the Wilson plot, you see a minimum near 6A resolution. This
> means that in the range 7-5A, intensities are low, and Rsyms are therefore
> elevated.
>
> best,
>
> Kay
>
> On Mon, 14 Oct 2013 11:52:03 -0700, Ursula Schulze-Gahmen <
> uschulze-gah...@lbl.gov> wrote:
>
> >Here is some more info on the data:
> >
> >lowest res. shell (50-8.0A): Rsym=4.4%, I/sig=100.
> >detector ADSC315, 1 degree oscillations, mosaicity around 0.3
> >
> >The Rsym stays high whether I merge just the first 30 frames or 200
> frames.
> >Wouldn't that speak against radiation damage?
> >
> >The crystals were grown from 2.5 M sodium formate and cryoprotected in 2.8
> >M sodium formate, 30% glycerol.
> >
> >Ursula
> >
> >
> >
> >On Mon, Oct 14, 2013 at 11:39 AM, Dominika Borek <domin...@work.swmed.edu
> >wrote:
> >
> >> There are several possibilities:
> >>
> >> (1) Radiation damage causes such bump in R-merge at medium resolution
> >> (2) Crystallization in very high salt that absorbs causes such
> behaviour,
> >> e.g. if your crystallized from 2M sulfate.
> >> (3) Very short exposure time -- very short means shorter than 1 second
> per
> >> image -- that results in large errors due to goniostat instability.
> >> (4) Very short exposure may also result in errors due to shutter
> problems.
> >> (5) Very weak diffraction. All errors will be "better" visible with weak
> >> diffraction.
> >>
> >> Any combination of the above factors is also possible.
> >>
> >> What are R-merges in the lowest resolution shell i.e. 50-8 A? If they
> are
> >> high too -- this is probably goniostat, shutter or beam instability. If
> >> they are low in the lowest resolution shell -- this probably radiation
> >> damage.
> >>
> >> What is the mosaicity? Low means lower than 1 degree for some people,
> for
> >> others less than 0.1 degree?
> >>
> >> What detector was used to collect data? Oscillation step? What is I/s(I)
> >> in the lowest resolution shell?
> >>
> >> Dominika
> >>
> >>
> >> Ursula Schulze-Gahmen wrote:
> >> > I have a data set with high Rsym in the lower resolution ranges, and I
> >> > don't understand what is going on.
> >> >
> >> > The crystal diffracts to about 3.0 A and has large cell dimension (
> space
> >> > group P6522 a= 185., c= 360.) Mosaicity is low.  I processed the data
> in
> >> > P6522, solved the structure and refined it. The maps look good and the
> >> > structure refines very well to R/Rfree of 20.5, 23.5%.
> >> >
> >> > But the dataset has a total Rsym of 22%, a redundancy of 20, an Rpim
> of
> >> > 7.6%, and CC1/2 of 0.55 in the highest resolution shell.
> >> > The problem seems to be in the lower resolution region around 8-5.5A.
> The
> >> > Rsym there is around 20%, it actually has a bump in this resolution
> >> region
> >> > and is slightly lower at 5.0 A before it steadily increases to higher
> >> > resolution. The high Rsym region between 8-5.5 A correlates with very
> low
> >> > I/sig which also increases again around 5.0 A and then steadily
> >> decreases.
> >> > The diffraction image shows very weak spots too.
> >> >
> >> > My question is: Is it likely that the data are processed correctly and
> >> the
> >> > crystal packing causes this strange diffraction pattern, or is there
> >> > something wrong with the data processing or the crystal?
> >> >
> >> > Any suggestions would be greatly appreciated.
> >> >
> >> > Ursula
> >> >
> >>
> >>
> >> Dominika Borek, Ph.D. *** UT Southwestern Medical Center
> >> 5323 Harry Hines Blvd. *** Dallas, TX 75390-8816
> >> 214-645-6378 (phone) *** 214-645-6353 (fax)
> >>
> >>
> >
> >
> >--
> >Ursula Schulze-Gahmen, Ph.D.
> >Assistant Researcher
> >UC Berkeley, QB3
> >356 Stanley Hall #3220
> >Berkeley, CA 94720-3220
> >
>



-- 
Ursula Schulze-Gahmen, Ph.D.
Assistant Researcher
UC Berkeley, QB3
356 Stanley Hall #3220
Berkeley, CA 94720-3220

Reply via email to