Quick
Start III-1. Quick Heavy-Atom
Sites Searching with Parallel SHELXD
{0}. i)
Log on to the cluster using: ‘ssh –X username@idc11’.
ii) Go
to the directory with your *.sca data file(s).
iii) sgxpro
·
Try
login as ‘staff’ and copy/scp your ‘.sca’ file there, if your account on the cluster has
problems.
Searching
heavy atom sites is the first step in novel-structure-solving process. The newly developed Parallel SHELXD module in
SGXPRO would quickly (in a couple of minutes on the cluster) tell if there is
any anomalous signal or the anomalous signal is good enough to find heavy
atoms. If not, you might have to check
your data collection (wavelength, beam intensity, exposure time, redundancy …),
data processing etc. to find and solve those experimental problem(s) that you
can’t do anything about after your data collection is done.
B E D C A
![]()

![]()

A. Click on “HA Search”, then “Parallel ShelxD” buttons to bring up SHELXD GUI.
B. Browse your .sca file (output from HKL2000 with anomalous on) or .hkl file of
the SHELX HKLF 3 format. Check to make sure the space group is right (‘spgr4d’
can help).
C. Select heavy-atom type, and give the number of sites (in one AU) to search. If totally
unknown as in cases of soaked derivatives, give a reasonably larger number.
SGXPRO has a built-in algorithm to output the only sites that are reasonable.
D. Click on “Save” button.
E. Click on “Run” button.
Results: The ‘*.shelxd.log’ contains
the best “CC all/weak” (an above
30/above 10 is a good
sign). The HA site(s) if found are saved
in
the
‘*.xyz’ file in PDB format.
Ref: Fu, Chrzas,
Sheldrick, Rose, Wang (2007): J. Appl.
Cryst. 40: 387-390.
Quick Start III-2. Evaluate
Anomalous Scattering Data Quality with Ras
{0}. i)
Log on to a computer (using ‘ssh –X’ when remotely).
ii) Go
to the directory with your *.x files.
Run command:
spgr4d Npix startXfile [Nframes] ras SP
Example:
spgr4d
4000
test_1_0001.x ras p41212
This command will
use all frames of data set 'test_1_####.x' to evaluate the anomalous
signal/noise ratio with space group P41212. The number ‘4000’ is the detector dimension
in pixels of MARCCD300, on which the data were collected.
For more details on how to use the program, just type
the command spgr4d from a terminal.
Notes to SER-CAT Users: SGXPRO is available on all the SER-CAT
computers for data processing. Functions
like Ras evaluation described in this section and other utility
tools can be done on any of these computers.
To reduce the workload, please use the cluster only if it is necessary.
About the Results:
Based on tests, a to-shell (cumulated) Ras of 1.5 may be significant
and have a good potential for successful phasing with most of the current
heavy-atom searching and phasing computer programs. The larger, the better. However, if the in-shell values are very
bumpy, it may not be good enough.
Statistically, the anomalous signal-to-noise ratio decreases with
resolution. So, a smoothly decreasing Ras
with resolution is expected for a normal data set. If the Ras values in the
resolution shells are significantly up and down, it may suggest a low
redundancy etc..
Tips:
1). The ‘.merx’ data file output from spgr4d can be read into SGXPro -> Data Reduction -> 3DScale, where more
control options (such as resolution cutoff) are available.
2). Data
integrated by programs such as d*Trek, Proteum etc. can be directly input into
3DScale as in Tips 1).
Warning: The ‘spgr4d’ calls ‘merx’ to harvest
reflections from .x files, then ‘3DSCALE’ for scaling. As ‘merx’ is a very simple program lack of
post-refinement to appropriately merge partial reflections, the merged data in ‘.merx'
file are not accurate, leading to the scaled data that are not the best you
could get from 3DSCALE. However, Ras
calculated using the data harvested by ‘merx’ is found reliable based on the
tests done so far.
Ref: Fu, Z.-Q., Rose,
J. & Wang, B.-C. (2004): Acta Cryst. D60:499-506.
Quick Start V.
Other Utility Tools
SGXPRO has many other modules
and built-in utility tools. For example,
the ‘FpFpp’ utility tool provides an f’,
f’’ vs energy/wavelength plot of any selected atom/ion type, providing a
general idea what the anomalous scattering looks like, what anomalous signal
level you would expect for a certain heavy-atom type etc., which may help
selecting a wavelength and planning your data collection ahead of a trip. Shown bellow is the plot for Ho.
Click on the curve will point to the corresponding data
values in the table on the left.
![]()

Ref: Fu, Z.-Q.,
Rose, J. & Wang, B.-C. (2005): Acta Cryst. D61:951-959.