One problem that has arisen is the treatment of dimers in Shelx. The problem that arose was that of numbering.
Shelx doesn't recognize SEGIDs or CHAINIDs from CNS, so it will have trouble with the fact that there are two of each residue. There are a few things to try:
1. Renumber the second molecule so that if the first molecule is numbered 1-300, you'd start the second molecule at 301. This adds another problem. Shelx will try to connect residues 300 and 301 unless you tell it:
FREE C_300 N_301
2. Renumber the second molecule with an "offset" of 1000, so that if the first molecule is numbered 1-300, the second one is numbered 1001-1300.
3. Put molecule 1 into PART 1 and molecule 2 into PART 2, keeping the numbering for each the same (ie. 1-300 in the example given).
Before going to Shelx, do the following rather time-consuming but useful procedure in CNS:
i. Take your CNS output .pdb file and run the shift_solvent.inp script. This will move all your water molecules as close as possible to the two protein molecules.Using Shelxpro,
ii. Then, use contact.inp to find out which of the water molecules are associated with each protein of the dimer.
iii. Create two new .pdb files, one for each protein of the dimer and its accociated water molecules, ligands, etc.
iv. Take the information from the contact.inp results and, in your two new protein .pdb files, add the water molecule coordinates that are associated with each protein.
v. Create separate .ins files from the two .pdb files
vi. Combine the two .ins files, placing a PART 1 before the first molecule and a PART 2 before the second. You only need one set of header information in your .ins file, so the easiest way to do this is to copy the coordinates from the second .ins file onto the end of the coordinates for the first molecule (make sure the END instruction only occurs at the very end of the combined file).