Data Reduction using SORTAV

1
Data Reduction using SORTAV
Jyväskylä Summer School on Charge Density August
2007
Louis J Farrugia
2
Jyväskylä Summer School on Charge Density August
2007
Data Reduction using SORTAV
SORTAV is a multi-purpose data reduction program,
capable of handling data from area-detectors and
scintillation point-detectors. It performs the
following tasks 1. Time dependent sample
decomposition 2. Inter sub-set scaling 3.
Empirical absorption correction based on multiple
observations 4. Empirical TDS corrections
(point-detectors only) 5. Data averaging and
merging 6. Bayesian treatment for weak
reflections Input data are raw intensities as
taken from integration program. For a charge
density analysis, we need to have multiple
observations (especially with area-detector data)
to improve the precision of experimental
structure factors.
R. H. Blessing (1987) Crystallogr. Rev. 1, 3
3
Jyväskylä Summer School on Charge Density August
2007
Data Reduction using SORTAV
Absorption corrections
Numerical analytical or by Gaussian
quadrature Requires measurement of face indices
can be difficult to measure thin plates
accurately. If the sample is medium-strongly
absorbing, i.e. ? gt 3 mm-1, then this method is
virtually mandatory for highly accurate
work. J. de Meulenaar H. Tompa (1966) Acta
Cryst. A19, 1014 P. Coppens, L. Leiserowitz D.
Rabinovich, D. (1965) Acta Cryst. A18 1035. G.
T. DeTitta (1985) J. Appl. Cryst. 18, 75 (crystal
in a capilliary)
4
Jyväskylä Summer School on Charge Density August
2007
Data Reduction using SORTAV
Absorption corrections
Empirical - requires multiple measurements to
determine absorption surface. Aaniso Ah,i for
each ith measurement, Ih,i among nh measurements
which are equivalent by symmetry (or through
azimuthal rotation). The ylm are real spherical
harmonics, and their arguments (-uo and u1) are
the reversed-incident and diffracted beam unit
direction vectors referred to the crystal-based
orthonormal axial system. The alm are refinable
coefficients obtained by a least squares fit to
minimise the residual ?2(summed over the
equivalent reflections R. H. Blessing (1995)
Acta. Cryst. A51, 33
5
Jyväskylä Summer School on Charge Density August
2007
Data Reduction using SORTAV
Data averaging
An advanced feature of SORTAV is the outlier
downweighting used, which biases the sample mean
towards the sample median. The most appropriate
for average data sets is the robust resistant
Tukey downweighting (which is chosen by default
by the SORTAVGUI). The outliers can also be
identified and rejected on input the program
will then repeat the process of downweighting.
This can be carried on in a cycle until no new
outliers are identified. For a high quality
data-set, a few cycles suffice to reach this
stage R. H. Blessing (1997) J. Appl. Cryst. 30,
421
6
Jyväskylä Summer School on Charge Density August
2007
Data Reduction using SORTAV
Analysis of variance improves the estimate of
?(F2)
7
Jyväskylä Summer School on Charge Density August
2007
Data Reduction using SORTAV
Rmerge indices
Several merging agreement indices are printed,
which including adjustments for small sample
statistics. normalised mean absolute
deviation R1 the normalised RMS
deviation R2
and the RMS standardised deviation Z
K. Diederichs P. A. Karplus (1997) Nature
Struct Biol. 4, 269
8
Jyväskylä Summer School on Charge Density August
2007
The SORTAVGUI
9
Jyväskylä Summer School on Charge Density August
2007
The SORTAVGUI
10
Jyväskylä Summer School on Charge Density August
2007
The SORTAVGUI
11
Jyväskylä Summer School on Charge Density August
2007
The SORTAVGUI