Crystallographic model building

The strategy used to build structures with O [8] will not be repeated here. Copies of the publication are available. The user requires a minimum of 3 files:

1. An electron density map. At present this must be a special file created with the mappage program. You can use the Frodo map file (defined with the DSN6 command) created on a Vax if you follow the example in chapter 5. The version of Frodo distributed by Dr. Phil Evans uses a different structure for this map and is not at present supported.

   Our file is a binary, direct access file that is created by a program with the generic name mappage. A version of this program is included with the distribution tape in export_source. This program can read in protein, CCP4, X-PLOR, FFT-Y maps. On the Unix workstations, we do not have Protein running at present so we have been unable to check if the program works.

   The mappage distributed with X-PLOR produces standard files but works only on the VAX. It is a much abused version of the original program and should be burned.

   If the map is made on a VAX and needs to be transferred to a UNIX machine, use Ftp with the binary option. Then run swapbytes on the UNIX machine.

2. A skeletonized electron density. This is a 3 step procedure run in the one program bones; namely make the actual skeletonization, assign the main chain/side chain status codes, output an O style formatted data structures. The program needs a level and step value. The step value can be the standard deviation of the map. The base level may need some experimentation. Too low a value produces too many bones atoms and becomes confusing. Too high a value gives too few. Try 1.25 sigma to start. The program handles the same maps as mappage.

3. A sequence. The user must prepare the sequence in a form suitable for input to O, i.e. a formatted datablock. This describes the residue property _residue_type of the molecule to be built. The option Sam_init_db will then generate the rest of the datablocks. If you want to use your own names for each residue then also load in a property _residue_name. Remember that both datablocks will be read in under format control and must be left justified. See the tutorial for detailed use. This will be simplified in a future release, allowing one to read in the sequence from external databases.