The following procedure creates the database used by the Search Fragment Database utility.
1. Select from the Brookhaven database a set of protein coordinates files, that have good resolution and include different structure types.
2. Construct a file, dmlist, containing a list of these protein coordinate files.
The following format is used in the file construction:
Number of proteins to be used (NPROT)
Name of coordinate file 2
.
.
.
3. Run the program $HYD_MSF/dmprep.
The program prompts for the name of the file, dmlist, which contains the list of proteins, and asks for a name for the distance matrix file to be created, dmfile.new. The program then reads each protein coordinate file and constructs the distance matrix file. The program also creates a QUANTA input command file. This file is later used from within QUANTA to generate an MSF for each of the protein coordinate files. You are prompted for a name for this file.
The dmprop executable distributed with QUANTA is dimensioned for 2000 proteins with limits of 2000 residues and 100,000 alpha carbon distances per protein. The FORTRAN sources for dmprep, dmprep.f and dmsubs.f, are also distributed. This allows you the flexibility to increase the dimensions as needed. To create a modified dmprep executable, type:
4. Move the distance matrix file to the $QNT_ROOT/dmatrix directory, renaming it to dmfile.
Because the variable $HYD_DMF is already defined in the QUANTA environment as $QNT_ROOT/dmatrix/dmfile, this is easily accomplished by typing:
dmfile.new = the filename of the distance matrix file
created in step 3.
5. To create the MSFs, which are also required, start QUANTA and type @command_file, where command_file is the name given to the QUANTA command file.
Respond appropriately to the dialog boxes. Treat the sixth character in the atom field as a disorder using the no-hydrogen dictionary file, and exclude symmetry in the molecular structure file.
6. Move the newly created MSFs to the directory $MSF_LIB.