2.03 *******************
* FSFOUR WRITE-UP *
*******************
PURPOSE- To calculate three dimensional Fourier transforms (maps)
when given a set of Fourier coefficients and control cards.
This program will calculate maps by using a multivariate variable
radix fast Fourier transform algorithm. the only restrictions are that
the number of grid points along each axis is even, and is a product of
the factors 2, 3, 4, or 5. Each factor can be used more than once.
The program is fully general so that all space groups can be handled.
The input structure factors must fall in the following range:
-NX/2 < h < NX/2
-NY/2 < k < NY/2
-NZ/2 < l < NZ/2
where NX, NY, NZ are the number of grid points along the a, b, and c
axes, respectively. Input structure factors outside the range will be
omitted from the calculation.
INPUT DATA (UNIT 5)
CARD 1 PAMFIL (free format)
PAMFIL = Name of input file containing cell and symmetry
information.
CARD 2 TITLE (free format)
TITLE = anything
CARD 3 NCENT,NX,NY,NZ,MAPTYP,IPRINT,NPIC,NORN,INPF,GSP,DCUT
(free format)
NCENT = 0 for noncentrosymmetric
space groups
= 1 for centrosymmetric
space groups
NX = number of grid points along the
a,b and c axes, respectively. If
NY = an input value is inconsistant
withthe factoring scheme, the
NZ = next largest acceptable value
will be used. If zero, see GSP
below.
MAPTYP = Fourier coefficient selection integer
= 1 for FO*exp(i*PHIC)
= 2 for FC*exp(i*PHIC)
= 3 for (FO-FC)*exp(i*PHIC)
= 4 for (2*FO-FC)*exp(i*PHIC)
= 5 for (FO-FC)**2
(difference Pattersons)
= 6 for FO**2
= 7 for FC**2
= 8 for -i*(FH+ - FH-)*exp(i*PHIH+)
(Bijvoet difference Fourier)
= 9 for (3*FO-2*FC)*exp(i*PHIC)
IPRINT = 0 for no printing of map
= 1 for printing of map
NPIC = number of non-hydrogen atoms
in the asymmetric unit. (Not
used within the program but
is passed on to program PSRCH
via the map file. Should not
exceed 140).
NORN = 0 for XZ sections
= 1 for YZ sections
= 2 for XY sections
***** CAUTION *****
If the map file is to be
input to programs PSRCH,
MAPINV, MAPVIEW, GMAP or CTOUR,
NORN must be 0.
INPF = 0 for binary reflection file input.
= 1 for formatted reflection file input.
GSP = Desired grid spacing in angstroms.
Defaults to 1.0, applied only if
NX=NY=NZ=0 to determine number of
grid points along each axis.
DCUT = minimum d spacing cutoff, in
angstroms, for acceptance of
input reflections.
CARD 4 INPREF (free format)
INPREF = Name of input reflection file.
CARD 5 MAPFIL (free format)
MAPFIL = Name of output map file.
CARD 6 LEVEL, (XLIM(I), I=1,3) (free format)
***** this card should be included ONLY if IPRINT is nonzero *****
LEVEL= scan level, if peaks are greater
than scan level, the peak will be
underlined with **, if zero,
defaults to 100
XLIM(1) =
XLIM(2) = printing limits. map will be
printed from 0 to XLIM (fractional)
XLIM(3) = along each axis
********* NOTES ON THE PROGRAM **********
The input reflection file is terminated by an end of file, and
should contain records with H, K, L, FOBS, FCAL, PHI where the first
three variables are INTEGERS and the remainder REALS. PHI should be in
degrees. The file may either be formatted or binary as indicated by
the parameter INPF. If it is formatted the format is assumed to be
( 3I4, 2F10.2, F7.2). If the input file contains records with
H,K,L,FPH,FP,PHI then MAPTYP=5 can be used to compute isomorphous
difference Pattersons (PHI is not used). If the input records contain
H,K,L,F(H,K,L),F(-H,-K,-L),PHI(H,K,L), then MAPTYP=5 will compute
anomalous difference Pattersons, and MAPTYP=8 can be used to
compute Bijvoet difference Fouriers. Note that if a binary file is
input each record must contain six words even if all of them are not
used in the calculation, i.e. PHI is not needed if MAPTYP=5,6 or 7,
but some value still must be supplied.
The output map file is binary and contains NSYM + 2 header records
followed by the map. If NORN = 0, the map is written such that each
record contains NX density values (one row along x), with NZ
consecutive records constituting each section of constant y, i.e. y is
slowest varying. If NORN = 1, the positions of x and y are
interchanged. If NORN = 2, the positions of y and z are interchanged.
All map values are integers scaled as described below. When NORN=0
the map file is suitable for input to programs PSRCH for locating
peaks, to MAPINV for modification followed by inversion, to MAPVIEW
for interactive contouring and display, to GMAP for conversion to
TOM/O or CHAIN formats and for creation of skeletons, or to CTOUR to
create hard copies of contoured plots. If NORN is nonzero the only
recourse is to print the map within this program.
***** SCALING THE DATA *****
Two scales are used, one for the binary map file and one for the
printed map output,if requested. If the input coefficients were on an
absolute scale, then the absolute electron density is obtained as
follows:
rho (absolute) = 10.*(printed map value)/(V*scale) + F000/V
rho (absolute) = (value on binary map output file)/(V*scale) + F000/V
where V is the unit cell volume and scale is given on the lineprinter
output. F000 is the total number of electrons in the unit cell. Note
that even if F000 is supplied on the input file, it will not be used
in the program, and must be added as indicated above. Also note that
the PRINTED output is limited to two digits per density value, but the
density is NOT rescaled to a maximum of 99. This means that values of
99 merely imply a density of AT LEAST 99.
***** BIJVOET DIFFERENCE FOURIERS *****
When maptyp=8 is selected, and the input reflection file contains
records with H, K, L, F(H,K,L), F(-H,-K,-L), PHI(H,K,L) then a
"Bijvoet difference Fourier" will be computed. In this case the map
consists of only the "imaginary" part of the electron density, and
should show strong positive peaks only at the sites of anomalous
scatterers (if the hand is correct). The multiplication factor -i is
applied only after expansion to a hemisphere to effectively
interchange real and imaginary parts of the density, as the program
would normally only compute the "real" part.
***** FILES *****
INPREF - Input Fourier coefficient file, can be either formatted or
binary as determined by input parameter INPF. Records
should contain h, k, l, Fobs, Fcal, Phi with h,k,l
INTEGERS and Fobs, Fcal, Phi REALS. Phi is in degrees. If
INPF=1, then file should be formatted with
FORMAT(3I4,2F10.2,F7.2)
MAPFIL - Binary map file output. Contains NSYM+2 header records
followed by the map, as described earlier.
2.04 *******************
* MAPINV WRITE-UP *
*******************
PURPOSE- To calculate three dimensional Fourier coefficients
(structure factors) when given an electron density
map and control cards.
This program will calculate structure factors by using a
multivariate variable radix fast Fourier transform algorithm to invert
an electron density map. The program is fully general so that all
space groups can be handled. It is assumed that the input map was
prepared by program FSFOUR. Structure factors may be calculated for
reflections in the following range:
h .ge. -NX/2 and h .lt. NX/2
k .ge. -NY/2 and k .lt. NY/2
l .ge. 0 and l .lt. NZ/2
where NX, NY, NZ are the number of grid points along the a, b, and c
axes, respectively, in the input map.
INPUT DATA (UNIT 5)
CARD 1 PAMFIL (free format)
PAMFIL = Name of input file containing cell and symmetry
parameters.
CARD 2 TITLE (free format)
TITLE = anything
CARD 3 MAPFIL (free format)
MAPFIL = Name of input map file.
CARD 4 SFOUT (free format)
SFOUT = Name of output structure factor file.
CARD 5 IPRNT, IPAIR, HMIN, HMAX, KMIN, KMAX, LMAX (free format)
IPRNT = 0 for no printing of structure
factors.
= 1 for printout
IPAIR = 0 for no pairing of calculated
structure factors with observed
data.
= 1 to combine calculated structure
factors with observed data
(supplied on auxilliary file) and
output R factor to the line printer
= 2 same as 1, but a separate file with
the combined data is also written.
HMIN =
limiting values defining range of
HMAX =
indices for which structure factors
KMIN =
will be calculated
KMAX =
(LMIN is always 0)
LMAX =
CARD 5A AUXINP (free format)
***** This card should be included ONLY if IPAIR > 0 *****
AUXINP = Name of input file containing auxilliary structure
factors for scaling.
CARD 5B AUXOUT (free format)
***** This card should be included ONLY if IPAIR = 2 *****
AUXOUT = Name of output file to contain calculated structure
factors scaled to (and paired with) the auxilliary
structure factors.
CARD 6 SC, F000, IMOD, IRHOMN (free format)
SC = scale factor applied to calculated
structure factors (see below).
If 0. defaults to 1.
F000 = total number of electrons in the unit
cell (see below).
IMOD = 0 for no modification of map prior to
transformation
= 1 to modify map prior to transformation
according to input criterion
=-1 same as 1 but the resulting density
is also squared prior to transformation.
IRHOMN = modification criterion (applied if
IMOD .ne.0). If (rho input + IRHOMN) < 0,
rho will be reset to 0. If F000 is
supplied, IRHOMN is automatically set to
correspond to non-negativity of electron
density.
********** NOTES ON THE PROGRAM **********
The input map file MAPFIL is assumed to have been generated with
program FSFOUR. It is binary, terminated with an end of file, and
after a few header records, contains the electron density map
represented as records (of integers) along x. y is the slowest varying
coordinate.
All calculated structure factors within the index range specified will
be output to file SFOUT. Note that this may include redundant
(symmetry related) as well as systematically absent reflections, if
they they fall within the specified index range. The output file is
binary, with records of H,K,L,FCALC,FCALC,PHI and is terminated by an
end of file. H,K and L are INTEGERS whereas FCALC and PHI are REALS.
PHI is in degrees. Note that FCALC is duplicated within each record so
that the file structure is consistant with the input required by
program FSFOUR.
If IPAIR > 0, then in addition to file SFOUT, an input file AUXINP of
observed structure factor amplitudes will be paired with the
corresponding calculated amplitudes and phases, and the combined data
used in one cycle of least squares refinement of a scale factor. The
resulting R factor between observed and calculated amplitudes is then
output to the lineprinter. Note that the input reflection data on file
AUXINP is restricted to the same range of indices as the calculated
data. If input values fall out of bounds they will be ignored.
Therefore, if data were collected with L negative, it will have to be
transformed by symmetry before it can be used successfully on file
AUXINP.
If IPAIR = 2, the results are identical to those obtained with IPAIR =
1, except that the combined (and rescaled) data is also output to a
separate file AUXOUT. The new file is of the same form as SFOUT, but
with records consisting of H,K,L,FOBS,FCALC,PHICALC for only those
reflections which were input on file AUXINP.
***** SCALING THE DATA *****
It is often desirable to control the scale of the calculated structure
factors. If the input electron density map was generated from
structure factors which are related to an absolute scale by:
F(input to FSFOUR) = k * F(absolute)
then k should be input for SC to obtain calculated structure factors
on an absolute scale. If sc= 0. (or 1.), then the calculated structure
factors will be on the same scale as those used to generate the map
(unless IMOD = -1, in which case they will be much larger). Note that
this scaling applies only to the output on file SFOUT (and the
lineprinter, if IPRNT .ne.0). If IPAIR .eq. 2, then the calculated
structure factors on file AUXOUT will always be scaled for best
agreement with those supplied on file AUXINP.
***** MODIFYING THE MAP *****
The following applies only if IMOD .ne. 0. Inclusion of F000 will
result in imposing non-negativity of electron density everywhere in
the map prior to inversion, provided SC is reasonably well known. If
SC is unknown, then F000 and SC on card 3 should be zero and IRHOMN
should be input to control the type and degree of modification.
Intelligent use of this parameter would then require knowledge of the
input map values prior to running the job. IRHOMN should be equal to
F000/V on the same scale as the input map. If IMOD = -1, IRHOMN is
first added to each density value, resulting values below zero are set
to zero, and each value is then squared prior to Fourier
transformation. This is equivalent to imposing non-negativity,
followed by one cycle through the tangent formula. Phases can
therefore be tangent formula refined or extended.
***** FILE REQUIREMENTS *****
MAPFIL - input map file, binary, as output by program FSFOUR
SFOUT - output file with all calculated structure factors,
binary, six word records as described earlier
AUXINP - auxilliary input structure factor file (required only
if IPAIR .ne. 0), binary, six word records in same form
as SFOUT (only H,K,L and FOBS are used)
AUXOUT - output auxilliary structure factor file, binary, six
word records as described earlier.
2.05 ********************
* PAMFILE WRITE-UP *
********************
This is not a program, but rather a description of a "standard
parameter file" which is read by all programs in the PHASES package,
and several auxilliary programs as well. The main purpose of this file
is to insure consistency in cell constants, symmetry, lattice type etc
throughout all programs, and to eliminate redundant input of these
parameters by the user. In addition one can optionally specify the
name of a "running log file." If this is done then in addition to
normal output to either the screen or individual log files for each
program, all printed output is also appended to a single file,
preceeded by a time stamp indicating what program was run and when.
Thus one can maintain a complete history of all computations and
results in a single log file. The standard paramater file is often
referred to generically in program write-ups as "PAMFIL." One should
select a name for it which is indicative of the particular structure
being worked on, and rapidly communicates to the user that it is a
parameter file. For example, PDC.PAM might be a good choice for
phasing pyruvate decarboxylase.
Each standard parameter file should contain the following
information in the indicated sequence.
LOGFILE=FILNAME Where FILENAME is the name of the
desired "running" log file. If no
cumulative log is desired, enter
LOGFILE=NULL
There must be no spaces immediately
preceeding or following the "=". Upper
or lower case is permitted.
LATTICE=X Where "X" is either P,A,B,C,I,F or R
There must be no spaces immediately
preceeding or following the "=". Upper
or lower case is permitted for the word
LATTICE, but only UPPER case for the
single character symbol.
A, B, C, ALPHA, BETA, GAMMA Unit cell constants, in angstroms and
degrees. Readable in free format, i.e.
at least one blank or comma separating
entries.
NSYM Number of equivalent positions in the
space group. Do NOT include
additional translations associated
with centering conditions for
non-primitive lattices, i.e. for
space group C2 NSYM=2. (this entry
read in free format).
The NSYM symmetry operators follow, one operator per line EXACTLY
as indicated in the International Tables for X-Ray Crystallography.
The first operator should ALWAYS be X,Y,Z. Note that for rhombohedral
lattices the HEXAGONAL AXES AND SYMMETRY OPERATORS SHOULD BE USED,
along with the lattice type R.
The following sample serves as a complete template for a parameter
file, for space group P2(1)2(1)2(1)
LOGFILE=seb.rlog
LATTICE=P
45.331 68.33 79.62 90. 90. 90.
4
X,Y,Z
1/2-X,-Y,1/2+Z
1/2+X,1/2-Y,-Z
-X,1/2+Y,1/2-Z