Things you need:

1) native x-ray data;  2) derivative x-ray data.

A) First run merge.inp to merge the native and derivative data
into one hkl file:   cns_solve < merge.inp > merge.log &

Input files are CNS format reflection files. In this case these files
are  e100c_nat3.cns (native data) and e100c3i.cns (iodine derivative data).

Output is a merged reflection file called merge.hkl.
 

B) Next, run scale.inp to scale the native and derivative data:
        cns_solve < scale.inp > scale.log &

Input is the merged reflection file merge.hkl.
Output is a new reflection file called scale.hkl, where the native and
iodine derivative data are scaled

C) Use patterson_map.inp to calculate the difference Patterson map.
         cns_solve < patterson_map.inp > patterson_map.log &

Input is a scaled reflection file such as scale.hkl.
Note that Patterson maps may appear very different depending on what
resolution range is used in patterson_map.inp.

Use plot_patterson  to make postscript plots of the Harker sections
(type "plot_patterson" and answer the questions).  This program requires
that the files PSmapx, PSmapy and PSmapz be available.

D) heavy_search.inp may be used to  to look for solutions to the difference Patterson.

Input is a scaled reflection file such as scale.hkl.
Within heavy_search.inp, set the expected number of heavy atoms and heavy atom types
to be searched for.

Output is a set of ".sdb" files which contain possible heavy atom coordinates.
 

Finally, working examples of "merge.inp", "scale.inp", "patterson_map.inp" and
"heavy_search.inp" can be found on hermione in   /home/dave/cns_examples/L9_diff_patt/

You can copy them from here into your directory if you want to run them.

These files are set up to calcalate a difference patterson map for L9 native
and iodine derivative data.  There are two iodine sites.