All input parameters needed for scaling data sets (obtained from processing with xds) are collected in the file named XSCALE.INP which must reside in the current directory where XSCALE will be invoked. To simplify the task of preparing the input file, a file template (example) for XSCALE.INP is included in the xds package that can easily be edited according to the actual case.
This chapter explains the meaning of all parameters used by the XSCALE program. Each parameter name consists of a string of characters without intervening blanks or exclamation marks and includes an equal sign as its last character. The value must follow the parameter name on the same line. The parameter names cannot be abbreviated; they are case sensitive, too. Characters in a line to the right of an exclamation mark are comment.
The parameters may be given in arbitrary order - except for the parameters defining input and output refelction files (INPUT_FILE=, OUTPUT_FILE=). Here, an output file is defined first by the parameter OUTPUT_FILE= that will include the scaled and merged reflections from all following input files specified by the parameters INPUT_FILE= until the next occurence of OUTPUT_FILE= in XSCALE.INP. Several output files can be specified (together with their set of input files) in a single run of XSCALE. All output files are then on the same scale - a program feature recommended for MAD data sets.
Input parameters of XSCALE
This parameter defines the maximum number of cpu's that can be employed by the parallel version of XSCALE. Up to 32 cpu's can be used. The default of 1 cpu will be assumed in the single processor version of XSCALE.
Example: MAXIMUM_NUMBER_OF_PROCESSORS= 16
Resolution shell limits (Å). Only the high resolution limit of each shell is given. Up to 20 resolution shells will be accepted. The shell limits must be specified in decreasing order. The resolution shells are used to report statistical properties of the data sets as a function of resolution. Note also that the last resolution shell defines the high resolution limit of the data that are included in the scaled output data set(s).
Example:
RESOLUTION_SHELLS=20.0 10.0 6.0 3.0
Completeness will be reported for the resolution shells infinity-20.0, 20-10, 10-6, and 6-3 Å. The scaled output file(s) will include no reflections beyond 3 Å, even if the input files contain reflections to a higher resolution.
Space-group number of the crystal. The numbers corresponding to each possible space group are defined in the "INTERNATIONAL TABLES I". All 230 space groups are implemented. From the space group number and the unit cell parameters XSCALE provides a standard set of symmetry operators. In case a reindexing transformation is specified, the space group symmetry refers to the new cell.
Example: SPACE_GROUP_NUMBER=77
This specifies the tetragonal space group P42
Unit cell parameters a, b, c (Å) and alpha, beta, gamma (degrees). The cell constants must meet the requirements implicated by the space group. First and second setting of monoclinic crystals must be distinguishable by the cell constants. In case a reindexing transformation is specified, the unit cell parameters refer to the new cell.
Example:
UNIT_CELL_CONSTANTS=125.9 125.9 144.7 90.0 90.0 90.0
SPACE_GROUP_NUMBER=77
This specifies the cell constants of a tetragonal crystal obeying P42 space group symmetry. Note that the a and b axes must have identical length and all angles must be exactly 90 degrees as required by the space group.
This optional transformation provides a possibility for reindexing the reflections of all input data files (except for an optional reference data set, see below). The meaning of the 12 integer numbers that make up the parameter value is defined as:
h' = REIDX(1)*h + REIDX( 2)*k + REIDX( 3)*l + REIDX( 4)
k' = REIDX(5)*h + REIDX( 6)*k + REIDX( 7)*l + REIDX( 8)
l' = REIDX(9)*h + REIDX(10)*k + REIDX(11)*l + REIDX(12)
where h', k', l' are the new indices.
In case this transformation is omitted it is assumed that reindexing is not required, that is h'=h, k'=k, l'=l. XSCALE issues a warning message if the reindexing transformation implies a change of hand (negative determinant). In this case the sign of anomalous intensity differences will flip.
Example REIDX= 0 -1 0 0 -1 0 0 0 0 0 -1 0 SPACE_GROUP_NUMBER=77 UNIT_CELL_CONSTANTS=125.9 125.9 144.7 90.0 90.0 90.0
The new reflection indices are related to the old h, k, l by the transformation h'=-k, k'=-h, l'=-l. Space group and cell constants refer to the new cell. Note, that the cell constants have to be "clean"; that means they have to satisfy the constraints of the symmetry group.
File name of an optional reference data set and its type. Type can be XDS_ASCII or the old formats DIRECT ANOMAL NORMAL OLDHKL UNIQUE. If a reference data set is not specified explicitly, XSCALE will generate one from all input data sets. This reference data set is put to an approximate absolute scale (assuming a protein with 50% solvent fraction). The reference data set is used for:
Example:
REFERENCE_DATA_SET= ../XDS_ASCII_native.HKL XDS_ASCII
This specifies a native data set of type XDS_ASCII serving as a reference. Note, that the reflection indices of the reference data set are not affected by a reindexing of the input reflection files.
The value of the parameter OUTPUT_FILE= is the name of an output file to be generated by XSCALE. The file name is restricted to at most 50 characters with no intervening blanks or exclamation marks. The reflections included in the scaled output data set are taken from files whose names are specified by subsequent parameters INPUT_FILE= up to the next occurence of OUTPUT_FILE=.
The file name must be followed by the parameter FRIEDEL'S_LAW= whose value (either TRUE or FALSE) specifies whether Friedel's law holds true or not.
Symmetry related reflections will be merged by default - unless the parameter MERGE=FALSE has been specified. In case Friedel's law holds true, reflections h, k, l and -h,-k,-l are considered to be symmetry related.
The parameter STRICT_ABSORPTION_CORRECTION= controls the calculation of the absorption correction factors. The parameter value can be TRUE (default) or FALSE. If STRICT_ABSORPTION_CORRECTION=FALSE, Friedel-pairs are treated as symmetry-equivalent reflections in the calculation of the absorption correction factors. In the presence of anomalous scattering effects this could lead to an underestimate of the anomalous differences.
If STRICT_ABSORPTION_CORRECTION= TRUE and FRIEDEL'S_LAW=FALSE, Friedel-pairs are treated as different reflections in the calculation of the absorption correction factors, otherwise not.
Format of the output file is always XDS_ASCII. However, the number of data items in each reflection record and their meaning depend on whether Friedel's law holds true or not and whether symmetry related reflections are merged or not.
Example:
OUTPUT_FILE=myo.ahkl FRIEDEL'S_LAW=FALSE MERGE=TRUE
STRICT_ABSORPTION_CORRECTION=TRUE
The parameters state that the scaled output data set will be given the name myo.ahkl, anomalous intensity differences are expected to be present in the data, and symmetry related reflections should be merged. In the calculation of the absorption correction factors Friedel-pairs are treated as different reflections.
File name of the xds input file, its format, and a resolution window for accepting reflections. Up to a total of 50 input files are allowed.
Example
INPUT_FILE=*../XDS_ASCII.HKL XDS_ASCII 50 1.2 WEIGHT=1.0 DELFRM=500 NBATCH=36
This specifies an input file, named ../XDS_ASCII.HKL. As the name is preceeded by a '*', the file also serves as a reference to other input files which will inherit the intensity fall-off as a function of resolution. The input file is of type XDS_ASCII and the subset of reflections within the resolution range 50 ... 1.2 Å should be included in the scaled output data set. A weighting factor of one is applied to the standard deviation of each reflection intensity, which is the default and could have been omitted. The value of 500 for limiting inclusion of Friedel pairs is rather permissive. NBATCH=36 allocates the maximum number of correction factors allowed for a single input data set. As the input file contains many reflections, there is no risk of overfitting the data.
This parameter (default value 0.1) defines the maximum probability the user is willing to risk that a 0-dose extrapolation is carried out when this correction should not have been done. A smaller value would reduce this risk on the expense that an increasing number of reflections are not corrected then - even when this should have been done.
Wolfgang Kabsch