Hi Tony,
in my experience it is not uncommon for crystals that diffract to
resolution worse than 3 A to be anisomorphous - after all, there is a
reason for diffracting badly, and that must be the lattice defects.
Lattice defects e.g. often lead to differences in cell parameters, and
in that case the merging R-factors are bad, because the molecular
transform is sampled at different positions in reciprocal space.
So this may be a problem inherent to the current state of your project,
and not strictly curable in software.
Concerning XDS, my advice for such projects can be found in the XDS wiki
in the "Small molecules" article. In essence,
a) give XDS _less_ parameters to fit and adjust (people tend to let the
programs fit everything, but the results are better if as few free
parameters as possible are used)
b) iterate a few times (recycling GXPARM.XDS and the _E.S.D.
parameters), each time monitoring ISa [(I/sigma)^asymptotic]
c) for scaling in XSCALE, use the result of the iteration with highest ISa
d) maybe discard the most disagreeing wedges.
HTH,
Kay
Tony Wu schrieb:
> I am having problems processing data from multiple wedges of 5-20
> degrees each. I am working with data from microcrystals collected at a
> minibeam. I want to use the data for molecular replacement. I can
> usually only collect about 10-15 frames before the crystal is obviously
> radiation damaged. I need to scale and merge the data from multiple
> crystals. I typically get poor diffraction, maybe to 4A. I have been
> able to process the data with xds/xscale, but the final statistics are
> worse than I would expect. For example, I can see fairly strong spots
> out to 4A, but xds might say the data only goes to 5A. This might be
> accurate, but I would like to also try the data with mosflm and scala. I
> can usually get mosflm to autoindex if I give it 3+ frames, but the
> results are not as robust as with xds. Usually the residual is poor, >
> 0.3 mm. When I try scala (tried multiple options) with just one wedge,
> it fails with "too few reflections", "no observations", or "negative
> scales".
>
> xds says the crystals are fairly isomorphous, < 1%. Symmetry is
> low, space group C2.
>
> Does anyone have some advice for processing this kind of data
> successfully?
--
Kay Diederichs http://strucbio.biologie.uni-konstanz.de
email: [log in to unmask] Tel +49 7531 88 4049 Fax 3183
Fachbereich Biologie, Universität Konstanz, Box M647, D-78457 Konstanz
This e-mail is digitally signed. If your e-mail client does not have the
necessary capabilities, just ignore the attached signature "smime.p7s".
|