hercules



NAME

     hercules - searches for heavy atom sites.  This is  a  brute
     force  application  of Patterson correlation methods.  It is
     slow, but extremely robust.



USAGE

     hercules control-file print-file



DESCRIPTION

     control-file

     The control-file describes what you want  done.   The  input
     data  descriptions  are  listed below.  The first five lines
     are string values for file names and the crystal  name,  and
     the rest of the lines in the file contain numeric data.


          1.   fin file with (merged native and derivative data)

          2.   solution file (NONE  for  the  first  round,  then
               solution  file  output from xheavy or xpatpred for
               the second round)

          3.   correlations file (output)

          4.   correlation map file (output in xcontur/xfit  for-
               mat)

          5.   crystal (crystal name)

          6.   grid  (should  be  1/4  to   1/6   resolution   in
               Angstroms)

          7.   resmax, resmin (resolution limits  on  diffraction
               data)

          8.   xmin, xmax (in fractions of  unit  cell,  x  asym-
               metric unit)

          9.   ymin, ymax (in fractions of  unit  cell,  y  asym-
               metric unit)

          10.  zmin, zmax (in fractions of  unit  cell,  z  asym-
               metric unit)

          11.  occupancy (see description)




     print-file

     The print-file will contain some of the output from the pro-
     gram.   Other output goes into the correlations and correla-
     tion map files.



DISCUSSION

     If you put "NONE" in the heavy solution file name spot (line
     2),  then  atom  data is read from the control-file.  Other-
     wise, it is read from  the  file  specified  in  that  field
     (which  may  be results from a previous run).  The atom data
     in the solution file is in the same format as that  used  by
     xheavy.

     For example, to use hercules to search for an initial  site,
     enter  "NONE"  for the solution file (line 2), run hercules,
     and select the highest hit.  You can put this  initial  site
     into  a solution file using xpatpred.  Then, to search for a
     second site, enter the new solution file in line  2  instead
     of "NONE" and run hercules again.

     To guess an occupancy for the second site  relative  to  the
     first,  examine  the Patterson map.  It's not very sensitive
     to the occupancy, so numbers like 1, .75,  .5,  or  .25  are
     sufficient.

     Use xpatpred to display the predicted  Patterson  positions.
     To  do  this, enter the site and then write out a prediction
     file.  Read this file into xcontur as a  labels  file.   The
     vectors  are  put  into the volume as 0-1, 0-1, 0-1, so they
     won't show up on negative sections.  For multiple site solu-
     tions, pay particular attention to the cross-peaks.

     The search volume should be the  asymmetric  volume  of  the
     Patterson map.  The input format only allows for rectangular
     asymmetric units.  This may cause some redundant computation
     in  high symmetry space groups.  For single sites there will
     be extra symmetry due to the fact that both  hands  and  all
     origin  shifts  are  valid  interpretations of the Patterson
     map.  For a second site there is a two-fold ambiguity due to
     the  hand  choice.  You can reduce the search volume accord-
     ingly.  If not, note that there will be  several  equivalent
     hits   (They   will  all  give  identical  predictions  with
     xpatpred).

     The solution can then  be  read  directly  into  xheavy  for
     refinement.   If  you  want  to  use the PHASES package from
     Furey, then xheavy can rewrite  the  solution  into  a  file
     that's  more  or  less  ready to use depending on the PHASES
     version you use.

     Hercules writes an fsfour style map  that  may  be  examined
     with  xcontur.   To get the coordinates in xcontur, click on
     the peaks.  Sites are added to the solution file as they are
     found  and  the  program is rerun.  The peaks can be entered
     into xpatpred and displayed on the Patterson map  to  verify
     their correctness.

     Hercules only takes a .fin file, not a .df  file.   You  can
     split off part of the data in a .df file to from a .fin file
     with awk.  If you do this, you can extract the  first  entry
     in the .df file with the command:

            awk '{print $1,$2,$3,$4,$5,$6,$7}' < file.df > file.fin

     You can extract the second entry with the command:

            awk '{print $1,$2,$3,$8,$9,$10,$11}' < file.df > file.fin

     (For information  about  .fin  and  .df  file  formats,  see
     Chapter 3 of the XtalView user guide.)

     You can view the results of a hercules run  with  a  command
     such as:

            sort -nr +3 test.list | head -20 > top20.hits

     This command produces a sorted list of the top 20 hits.



EXAMPLES

     Example control-file

     The following is an example control-file.

          NONE

          ex.list

          ex.map

          cvccp

          2.5

          1000.0, 6.0

          0.0 0.5

          0.0 0.3

          0.0 0.05

          1.0

     On an SGI workstation, you may need to enclose each  of  the
     first  five  lines  in  double quotes to make your file look
     like this:

          "ccpniau1.fin"

          "NONE"

          "ex.list"

          "ex.map"

          "cvccp"

          2.5

          1000.0, 6.0

          0.0 0.5

          0.0 0.3

          0.0 0.5

          1.0


NOTES

     Runs can take several hours  and  are  proportional  to  the
     number of sites that are input.



VERSION

     Release 3.2 of XtalView



SEE ALSO

     xtalview(1), xhercules(1)