DBT is mainly used
to create a highly magnified portion of a 3D MAP so ROBEM can
produce publication quality shaded-surface renderings. Such maps
earn the publication quality rating because voxels in the new 3D
MAP are much smaller than in the original 3D MAP.
Note:

RWA
currently has a beta version of this routine working in RobEM.
When you first enter RobEM, go to the main pulldown menu Option and
then click the button labeled Calculate structure factors for
incoming maps. EXIT and proceed as normal and in Display you can
change the pixel to any size (within reason of course!).
EMSF3DBT evolved from a crystallographic Fourier summation
program, FOURIER, which I believe was written in the 1960s by R.
J. Dellaca of the University of Canterbury, Christchurch, N.Z.
Except for some hard core subroutines, the program has been
revised and rewritten numerous times to work on various machines.
EMSF3DBT is intended for use on VAX and Unix computers and is
specifically tailored for electron crystallography applications.
The program FOURIER was designed to compute several useful
crystallographic functions, including difference Fouriers and
Patterson maps. EMSF3DBT
only computes a Fourier summation. Other
options from FOURIER are
not available in EMSF3DBT, although
interested parties may have the original code if they so desire
(
and
I can locate it!).
The latest version of EMSF3DBT has done away with the infamous
CONTROL data file (*.HED), which was used to store input
parameters. All input is now entered interactively or via a BATCH
command file, as you wish. Though the input to EMSF3DBT at first
glance appears daunting, the program is normally run with a
standard set of simple input parameters.
MOST of the program input,
especially line 2, and

lines 5 through 11, can be entered by typing
a <CR> and accepting the DEFAULT values. B. PROGRAM INPUT
1. Filename for 3D structure factor data input (A)
2. RES_HI (F)
3. Filename for 3D MAP data output (A)
4. 3D MAP output file header (80A1)
5. ITRANS, NEQUIV, SCALEF (2I,F)
6. ((TS(n),FS(n,m),m=1,3),n=1,3) (12F)
7. U MATRIX (9F) (Only ITRANS=1)
8. P1,P2,P3,D,CELL(3) (7F) (Only ITRANS=2)
9. XMIN,XMAX,YMIN,YMAX,ZMIN,ZMAX,DX,DY,DZ (9F) (Only ITRANS=2)
10. NREF_DUMP (I)
11. XMIN,XMAX,YMIN,YMAX,ZMIN,ZMAX (6F) (Only ITRANS.LE.1)
12. DX,DY,DZ (3F) (Only ITRANS.LE.1)
13. NORIEN (I) (Only ITRANS.LE.1)
Detailed descriptions of program input:
1. Filename for 3D structure factor data input (A FORMAT)
Enter the name for the file that contains the input 3D structure
factor data. SF data are currently stored in BINARY or ASCII
format. EMPROGS.DOC describes the SF data and how they are stored.
2. RES_HI (F)
RES_HI defines the highest resolution structure factor to use in
computing the new 3D MAP. The default is to read and use all the
SF data stored in the file. RES_HI
must be specified in the same
units (Å, nanometers, pixels, yards, etc.) as was used to create
the SF data, otherwise strange results can occur. If the
resolution of the input data is
lower than your request, then
RES_HI will DEFAULT to the lower value (
i.e.
you cant make
something out of nothing!).
3. Filename for 3D MAP data output (A FORMAT)
Enter a filename for storing the new 3D MAP.
4. 3D MAP output file header (80A1 FORMAT)
Enter any header information (up to 80 characters) you like
here. This information is stored in the header record of the PIF
format 3D MAP data file. 5. ITRANS, NEQUIV, SCALEF (2I,F)
DEFAULTS:
0, 1, 1.0
ITRANS specifies the orientation of the sections. The DEFAULT
(used 99.99% of the time) is to compute sections parallel to the
unit cell axes:
= 0 Sections calculated parallel to unit cell axes
= 1 Sections calculated parallel to a general plane, the U
matrix of which is supplied
= 2 Sections calculated parallel to a general plane, the
equation of which is supplied
NEQUIV specifies the number of symmetry cards (input line #6).
This is the number of equivalent positions in the space group
EXCLUDING those related by a center of symmetry or by a unit cell
centering condition. Since P1 space group is used 99.99% of the
time, NEQUIV is normally set = 1.
SCALEF is used to scale the height of density peaks if the input
data are not on an absolute scale. For example, for a C-centered
cell, SCALEF = 2.0. If the input data are doubled, for example
with data sets containing Friedel pairs, SCALE = 0.5. For P1 data,
SCALEF is normally set to the DEFAULT value of 1.0.
6. ((TS(n),FS(n,m),m=1,3),n=1,3) (12F) (one for each of the NEQUIV
positions)
DEFAULT = 0.,1.,0.,0., 0.,0.,1.,0., 0.,0.,0.,1.
Each of the NEQUIV input lines contains a [3x3] rotation matrix
[FS] and a [3x1] translation vector [TS] such that:
P' = [FS]*P + [TS],
where P' and P are the transformed and initial positions,
respectively.
Take for example, (-x, x-y, 1/2+z):
| 0 | |-1 0 0|
[TS] = | 0 | and [FS] = | 1 -1 0|
|0.5| | 0 0 1|
and the data card should appear: ((TS(n),FS(n,m),m=1,3),n=1,3)
The above example would be:
0., -1., 0., 0., 0., 1., -1., 0., 0.5, 0., 0., 1.
Hence, the DEFAULT corresponds to the simplest case: (x,y,z) 7. U MATRIX (9F) (Only for ITRANS = 1)
Supply transformation matrix, U (U11,U12,U13,...U33), where
| a'| | a |
| b'| = U * | b |
| c'| | c |
8. P1, P2, P3, D, CELL(1-3) (7F) (Only for ITRANS = 2)
Supply the general equation of the plane as:
P1*X + P2*Y + P3*Z = D,
where D is the distance (in Å, nms, etc.) of the plane from the
origin of the unit cell.
For this option, the new axis X' is set normal to the plane, Y'
is set parallel to the plane and to the projection of the original
unit cell vector which has the largest component on this plane
(
i.e.
the unit cell axis which has the minimum P value). Z'
completes a right-handed orthogonal system.
In this instance the program automatically sets NORIEN = 100.
The length of each new axis is set to the maximum of four
quantities:
|2*XMAX(I)|, |2*XMIN(I)|, 10.0, CELL(I)
The longer the axis, the greater the accuracy,
i.e.
unless the
elements of U are integers, errors will be introduced when the
indices are transformed by the U matrix. For a section through the
plane, set XMIN = XMAX = |D|.
The U matrix is printed out as:
NEWA1 = U11*A1 + U12*A2 + U13*A3 etc.
The determinant of the U matrix is equal to the ratio of the
volumes of the transformed and original unit cells.
9. XMIN,XMAX,YMIN,YMAX,ZMIN,ZMAX,DX,DY,DZ (9F) (Only for ITRANS =
2)
The limit and increment information (in Å, nms, etc.) refer to
the NEW TRANSFORMED unit cell.
10. NREF_DUMP (I) NREF_DUMP specifies the number of reflections for which the
equivalent forms and Fourier coefficients are to be listed. This
is normally used to check the calculation. The DEFAULT setting
(NREF_DUMP = 0) bypasses this option.
11. XMIN,XMAX,YMIN,YMAX,ZMIN,ZMAX (6F)
DEFAULT: 0.0,1.0, 0.0,1.0, 0.0,1.0
These specify the limits along the X-, Y-, and Z-axes in
fractional coordinates. If, for example you only wish to examine
the central portion of the top half of the original MAP, values
like the following might be entered:
0.25,0.75, 0.25,0.75, 0.5,1.0
12. DX,DY,DZ (3F) (Only for ITRANS.LE.1)
DEFAULTS:1.0,1.0,1.0
These specify the intervals (
i.e.
voxel dimensions in Å, nms,
etc.) between the sampling points parallel to the X, Y and Z axes
respectively.
WARNING: Use dimensions in the same units as those
used to specify the unit cell axis dimensions (as stored in one of
the header records of the SF data file). If for example, the
original 3D MAP was based on (10Å)
3
pixels, you might want to
produce a magnified portion of the MAP with (2.5Å)
3
voxels. Be
aware that if you attempt to create the entire 3D MAP with (2.5Å)
3
voxels, it will take up 64 times more memory and/or disk space! If
you make (1.0Å)
3
voxels, it would be 1000 times larger!!! Hence,
the smaller you make the voxels in the MAP, the smaller the region
you are realistically able to view with programs like ROBEM.
If DY or DZ are not specified, they DEFAULT to the value of
DX. Hence, if you only need to enter a value for DX if you want DY
and DZ to be the same (saves a little typing!).
13. NORIEN (I) (Only for ITRANS.LE.1)
NORIEN defines the orientation of the sections as follows:
= 100 Calculate sections along the X-axis
10 Calculate sections along the Y-axis
1 Calculate sections along the Z-axis
The following points should be noted:
(a) For each value of NORIEN, the orientation of the remaining
two axes is as follows: NORIEN
SECTION
ACROSS PAGE DOWN PAGE
100
X
Z
Y
10
Y
X
Z
1
Z
X
Y
(b) The program adjusts the increments (DX,DY,DZ) so that there
are an integral number of points parallel to each axis, as
follows:
(XMAX-XMIN)
DX = -----------------------------
AINT((XMAX-XMIN)*A_CELL / DX)
where A_CELL is the length of the X-axis (in Å, nms, etc.). The FORTRAN code for EMSF3DBT is in DEXTRO3:[TSB.FOR]EMSF3DBT.FOR.
Documentation at
http://bilbo.bio.purdue.edu/~baker/programs/programs.html
C. FLOW CHART FOR EMSF3DBT PROGRAM
*
* |- STRING_UPPER STRING_LEN !
*- SF_OPEN3 SF$OPEN |- REAL_TO_REC !
* |- FILE_CHECK !
*
*- SF3DBT_GETMEM1 - MALLOC !
*
*- SF_FILL2 !
*
*- FREE !
* |- PIRADDEG_D !
*- SF3DBT INFO -|- SF3DBT_STOAA !
* |- SF3DBT_M13 !
* |- SF3DBT_ROT !
* |- SF3DBT_M13 !
* |- SF3DBT_VMV !
*- SF3DBT_TRANS -|- SF3DBT_GP1 ------|- SF3DBT_VM !