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 - reduce  

SYNOPSIS

reduce [-behijkoqrsuvxz] [-a<val> ...] [runno] [minsi] [resmi] [resmax]  

DESCRIPTION

"reduce" converts the output of "integrate" into a more compact form, and can be used to filter out possibly faulty reflections. The initial output format of integrate produces 80 bytes of information per observation, and the output format from reduce merges that down to 16 bytes of information per unique reflection plus 16 additional bytes for each individual observation. Data can be excluded from output based on agreement between profile-fitted and summed intensities, the actual I/sigma values, resolution range, consistency with the space-group symmetries, and other factors.
-b
Use the "best" I/sigma value; that is, look at the profile-fitted and summation intensities for each reflection and write out whichever one has the larger I/sigma. The resulting MULTIREF will be partially profile-fitted, partially summed.
-e
Kabsch output, i.e. choose the integration results based on geometrical rebinning into diffraction vector space (alpha,beta,gamma) rather than using the direct experimental space (X,Y,omega).
-h
Replace the current polarization correction with that associated with a fully-polarized input X-ray beam, e.g. from a synchrotron, in which the goniostat's rotation axis is in the plane of polarization.
-i
Re-index the reflections by a 3x4 matrix transformation on the indices before output. The transformation matrix is specified in a file with environment variable REINDEX. If the matrix is specified as
1
0 0 -1
0
-1 0 0
0
0 -1 0
then the output reflections will obey h(out) = h(in) - 1, k(out) = -k(in), l(out) = -l(in).
-j
Causes the output to exclude reflections in specified resolution ranges. Typically this would be used to eliminate reflections in ice-rings. The file containing the resolution ranges to be eliminated is specified with the environment variable ICERING. Each line contains a lower and an upper resolution value in Angstroms, e.g.
3.67 3.63 2.24 2.20 1.93 1.89 1.48 1.44
-k
If this is specified, the program will the recompute indices of all the reflections in the UREFLS file according to the crystal and detector parameters contained in the UPARAMS file. If any of a reflection's indices differs from an integer by greater than the index-error limit in the UPARAMS file (typically 0.1-0.25) the reflection is not output. This becomes a way of filtering out reflections that have been moved too far from their starting position; it is probably less effective than the XYOmega filter (below).
-o
If this is specified, the program will read a file with environment variable name REFLECTIONS as a output file from the XENSYS / BUDDHA program SCAN rather than reading UREFLS as an output from the integration program herein.
-q
Ordinarily the "reduce" operation prints out the first 10 reflections rejected by the algorithm within each category of rejection; after that it suppresses printout. If this Boolean is specified, ALL rejected reflections will be printed to the log file.
-r
If this is specified, the program will recompute the indices of all reflections in the file according to the crystal and detector parameters contained in the UPARAMS file. Unlike the "recompute and reject" operation described under /-k/, this one does not then reject any reflections.
-s
Use intensity and sigma values obtained by simple summation rather than by profile-fitting. In combination with the "transformed coordinate" option; this option allows for 4 different intensity and sigma measurements: profile profile fitted and summed, direct-space or transformed-coordinate integration.
-u
This flag tells the program to /remove/ the monochromatized polarization correction from the intensity measurements and replace it with an unmonochromatized polarization correction. Thus if the data were erroneously integrated with the monochromatized polarization formula "on", this flag allows the user to undo that erroneous correction.
-v
Replace the current polarization correction with that associated with a fully-polarized input X-ray beam, e.g. from a synchrotron, in which the goniostat's rotation axis is perpendicular to the plane of polarization.
-x
If this is specified, the program will write out only reflections for which |Xob - Xp | < dX, |Yob - Yp | < dY, and |Zob - Zp | < dZ, where (Xob ,Yob ,Zob ) is the observed centroid, (Xp ,Yp ,Zp) is the predicted centroid, and (dX,dY,dZ) are the error limits in (X,Y,Z) set in the UPARAMS file. Thus this provides a filter preventing inclusion of spots whose centroids moved too far during integration.
-z
If this is specified, we recompute the indices of all reflections in the file according to the crystal and detector parameters contained in the UPARAMS file. If any of the reflection's indices differs from an integer by greater than the index-error limit specified in the UPARAMS file (typically 0.1-0.25 degrees) or if the new indices are different from the old indices, do not output the reflection.
-a<minfm,maxfm>
This specifies the range of frames over which
integrated data will be output. The values are given in the form minfm,maxfm. Thus if you wish to output the data from frame 312 to frame 381, enter "-a312,381". Default value: no restriction imposed.
-c<val>
A non-blank, nonzero value <val> here specifies that we
should write out a CENTROIDS file rather than a MULTIREF. The resulting file can be used as input to the refinement functionality, so that we can use the superior centroid-finding and background-subtraction algorithms of "integrate" as input to the refinement effort. The value <val> given should be the maximum number of centroids written out in each of the roughly 400 regions of the detector face within which output will occur; thus specifying "7" here will cause at most 7 spots to be written out per region, or about 2800 altogether. Default value: if "-c" is not specified, no CENTROIDS file is created. If "-c" is specified without a corresponding numeric argument, then 10 is the corresponding value.
-d<val>
If you specify a value <val> here, the reduction routine
will delete from output all reflections with
|Ifit - Isum | > <val>* sqrt(sigfit)^2 + sigsum)^2).
Default value: 4.
-f<val>
A nonzero value here specifies that we should only output
reflections for which I/sigma is greater than that value. Thus specifying "-2" here would cause the program to output only reflections with I/sigma > -2. Default value: -3.
-g<val>
If you specify a value <val> here, the reduction routine
will delete from output all reflections with
|Ifit - <val> * Isum | > <val> * sqrt(sigfit^2 + (G*sigsum )^2),
Thus if <Ifit > = <Isum > (G == 1) this option is identical to to the -d option given above; if not, it will result in the deletion of a slightly different list of spots. Default: no restriction imposed.
-l<val>
Inflate the Lorentz correction values by this specified
value so that
I = (Iraw / p) * (val + 1 / L) instead of I = (Iraw / p) * (1 / L).
The purpose of this option is to allow for an artificial expansion of the Lorentz correction (1/L). There is some evidence with certain kinds of data that this correction may be useful. Default value: 0.
-m<val>
Examine the data to find the dependence of <I/sigma>
on sin(theta)/lambda and impose a resolution cutoff at the sin(theta)/lambda value where <I/sigma> falls to <val>. Thus if you specify -m1.5 and the average intensity/sigma value falls to 1.5 at a D spacing of 2.35 Angstroms, then the data will be output only out to 2.5 Angstroms. Default: 1.4.
-n<val>
Output a combination of profile-fitted and summed data.
Any reflection whose I/sigma value is above <val> will have its intensity output according to summation; any other reflection will have its intensity output according to profile-fitting. Default: 999999, i.e. this option is not put in effect.
-p<val>
Ordinarily the "reduce" operation prints out the first 10
reflections rejected by the algorithm within each category of rejection; after that it suppresses printout. If a nonzero value <val> is specified here, then for each category of rejected reflections the program will output <val> reflections to the log file rather than 10. Default value: 10.
-q<val>
Turn on a variety of options to "reduce" depending on
the specified <val>. There are more than 26 run-time options for "reduce", which is the reason this approach must be used. The options available and their meanings are: val Meaning Equivalent
1
Output larger I/sigma -b
2
Kabsch intensities -e
3
Input from SCAN/BUDDHA -o
4
Simple-summation output -s
5
Change polarization to synchrotron-horizontal -h
6
Omit reflections within ice-rings -j
7
Omit reflections with non-integer indices -k
8
Print all screw-axis-deleted observations -q
9
Recompute reflection indices using UPARAMS -r
10
Undo the polarization correction -u
11
Change polarization to synchrotron-vertical -v
12
Omit reflections with large (X,Y,omega) errors -x
13
Keep duplicate observations -y
14
Recompute hkl and omit if non-integer -z
Thus reduce -q5 -q9 would alter the polarization to that appropriate to synchrotron data with a horizontal axis (assuming the data had been initially integrated with a monochromator polarization formula), and would recompute the reflection indices based on the contents of the UPARAMS file.
-t<val>
Adjust all intensity and sigma values by a
multiplicative value of 1 - <val>*cos(2 * pi * omega). This corrects for certain obscure goniostat-related errors. Default value: 0.
-w<val>
Set the nominal width of the scaling groups generated by
the reduction step to be <val> degrees rather than the default 5 deg. The actual scaling group width will be slightly larger than the value specified here so that we don't end up with a ragged scaling group at the end. Thus if the data span 60 degrees and you specify 8 degree scaling groups, the program will create 60 / 8 = 7 pairs of scaling groups, and each will be 60 / 7 = 8.57 degrees wide. Default value: 4, 1.5, or 1, depending on how many reflections are present in the data set (4 with fewer than 5000 reflections, 1.5 with more than 5000 reflections but a run-width larger than 100 degrees; 1 with a run-width smaller than 100 degrees). With very long runs the nominal width is set to ensure that there are fewer than 200 scaling groups associated with a run.
-z<val>
Turn OFF a variety of options to "reduce" depending on
the specified <val>. There are more than 26 run-time options for "reduce", which is the reason this approach must be used. In the list of equivalents below, a notation like !-x would indicate the complement of what option -x does. The options available and their meanings are: val Meaning Equivalent
1
Disable output of larger Isigma !-b
2
(X,Y,omega) intensities !-e
3
No SCANBUDDHA input !-o
4
Profile-fitted output !-s
5
Don't change polarization to synchrotron-horizontal !-h
6
Use reflections within ice-rings !-j
7
Keep reflections even if they have non-integer indices !-k
8
Don't print all screw-axis-deleted observations !-q
9
Don't recompute reflection indices !-r
10
Don't undo the polarization correction !-u
11
Don't correct polarization to synchrotron-vertical !-v
12
Keep reflections with large (X,Y,omega) errors !-x
13
Don't keep duplicate observations !-y
14
Don't recompute hkl and omit if non-integer !-z
Thus reduce -z3 -z9 would ignore data from a SCAN or BUDDHA run, and would decline to recompute reflection indices.
<runnumber> This specifies the run number associated with the run.
Default value: 0.
<minsigoi>
This specifies a machine error correction to the data.
If the input variance is v and the value specified here is m, then the output variance is v + (m*I) * (m*I). Default value: 0.
<resmin>
This specifies the low-resolution cutoff in
Angstroms, below which reflections will not be output. Default: no restriction imposed.
<resmax>
This specifies the high-resolution cutoff in Angstroms,
above which reflections will not be output. Default value: 0.
 

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
REPORTING BUGS
COPYRIGHT

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