Content-type: text/html Man page of X-GEN

X-GEN

Section: X-GEN Commands (1)
Updated: April 2005
Index Return to Main Contents
 

NAME

X-GEN - scalek  

SYNOPSIS

scalek [-abdefghjlknrtvwx] [-c<val> . . . ]  

DESCRIPTION

"scalek" is the X-GEN scaling functionality. It reads in the output of "reduce" or "mrmerge", i.e. a MULTIREF file containing multiple observations of the reflections in a set of diffraction data, and determines scaling functions that minimize the squared differences among the intensities of the symmetry-related observations.
-a
Use all observations, even those coming from scaling groups for which non input scale factor was provided. This is default ONLY if no scale-input file is specified.
-b
Scale using a nonlinear, Simplex algorithm with the scale factors and (if requested) spherical harmonic parameters acting as the adjustable parameters in the Simplex algorithm. If this is not specified the ordinary diagonal linear least-squares approach is taken. The Simplex scaling is MUCH slower than the diagonal linear least-squares, and should not be used on large datasets unless you are willing to wait for ten minutes or more.
-d
Scale using a spherical harmonic expansion in the diffraction vectors of the input observations. A total of ten parameters per data run will be calculated, along with the conventional scaling parameters that are calculated on a finer grid.
-e
Scale using a spherical harmonic expansion in the exit-beam vectors of the input observations. A total of nine parameters per data run will be calculated, along with the conventional scaling parameters that are calculated on a finer grid.
-f
The weight factor used in scaling is the inverse of the variance of each observation. Normally the variance is computed as the sum of the observation's own variance and a term proportional to the square of the Rsym of the scaling group: this term represents the variance in the scale factor itself. If this Boolean is on, the scale-factor variance term is set to zero; this has the effect of up-weighting contributions from poorly-determined scaling groups but it leaves the scaling problem somewhat stabler.
-g
Scale using a spherical harmonic expansion in the diffraction vectors of the input observations, but with the term that would multiply (sx^2 + sy^2 + sz^2) set to zero.
-h
Start the scaling process by performing diagonal linear least-squares determinations of the scaling parameters. If this approach reaches a stage where the overall R value is going up rather than down, it will switch to a Simplex nonlinear optimization. If this switch and -b are both specified, this one takes precedence.
-j
Stop cycling once the scaling functions converge to within the convergence criterion. This is the default, and is the converse of the "-k" flag.
-k
Continue cycling for as many cycles as the user has requested, even if the scaling parameters and symmetry R values have converged.
-l
Do _not_ perform spherical-harmonic scaling in any of the variants available in this package. This is the default.
-n
Treat Bijvoet-related reflections (e.g., (6,4,3) and (-6,-4,-3)) as independent in calculating scaling factors. By default, Bijvoet-related reflections are treated as the same reflection.
-r
Do not perform ordinary spherical-harmonic scaling. This is the default.
-t
Merge the top and bottom of the detector for scale-factor determinations. If we do not specify this option, the top and bottom of the detector are treated as separate scaling groups.
-u
Do not merge the top and bottom of the detector for scale-factor determinations; instead, treat the top and bottom as separate scaling groups. This is the default.
-v
If this Boolean is on, a moderate amount of additional diagnostics beyond the default will be written to the XLOG log file and stdout.
-w
Determine the initial scale factors from a constructed Wilson-plot rather than setting them to 1.
-x
Do not switch from linear scaling to nonlinear Simplex scaling, even if the R-factor starts to creep up. This option is the default, and is the converse of the /-h/ option.
-c<val>
This specifies the maximum number of cycles to be
performed. Cycling will quit early if the scaling converges unless the "keep cycling" flag is on. Default value: 10.
-i<val>
This specifies the minimum value of the mean intensity to
be used in scaling. Default value: 0.
-o<val>
This specifies the minimum number of observations per
reflection required for inclusion in the scaling. Default value: 2.
-p<val>
This specifies the number of parameters calculated per
ordinary scaling group. Legal values are 1, 2, and 3. The first parameter is always a linear scale factor; the second and third are terms multiplied by sin(theta) squared and sin(theta), respectively. Default value: 1.
-q<val>
This specifies 10000 times the convergence criterion. Thus
if you specify this value as 10, then scaling will be defined to have converged if the largest scale-factor change since the previous cycle is (10/10000) = 0.0001. Default value: 100.
-s<val>
This specifies the fraction of the computed changes in the
scaling parameters that will actually be applied at the end of each scaling cycle. Thus if this is specified as 1, the full computed change will be applied; if it is specified as 0.05, only 0.05 * the computed change will actually be applied to the scale factors. Default value: 1.
 

EXAMPLES

scalek
Perform ten cycles of scaling using all reflections with two or more observations, for which the mean intensity is non-negative, using one parameter per scaling group and exiting if the scale factors converge to within 0.0001:
scalek -ktghnp1
Scale using spherical harmonic expansions in the
diffraction vectors and switching to a Simplex optimization if the linear least-squares approach becomes unstable, merging the top and bottom halves of the detector together but separating Friedel mates during scaling:
 

REPORTING BUGS

Report bugs to Andy Howard at howard@iit.edu or 312-567-5881.  

COPYRIGHT

Copyright © 2002, Illinois Institute of Technology. See the file 'LICENSE' for information on usage and redistribution of this file, and for a DISCLAIMER OF ALL WARRANTIES


 

Index

NAME
SYNOPSIS
DESCRIPTION
EXAMPLES
REPORTING BUGS
COPYRIGHT

This document was created by man2html, using the manual pages.
Time: 02:08:09 GMT, October 03, 2005