General description

Starting from the adf2005.01 version the utility pdb2adf is available in the official release. Previously this utility could be found on the contributed software page.

The pdb2adf utility was written to read a PDB file, which contains the atomic coordinates of a protein structure, and transform it into an ADF inputfile, particularly for use with QM/MM calculations. Starting from the current release it can also be used for setting up a solvent shell around a solute molecule.

The PDB files are generally used for protein structures, and are formatted according to certain rules, see: http://www.rcsb.org/pdb/docs/format/pdbguide2.2/guide2.2_frame.html, and the part about the official PDB format below.

For every residue/molecule present in the PDB file, there should be a fragment file available, either in the general ADF library ($ADFRESOURCES/pdb2adf directory), or in the local directory where the pdb2adf program is being called. Fragment files in the local directory take higher priority than those in the general ADF library. The fragment files are formatted, based loosely on AMBER parameter files, and contain information about the residues; e.g., the atoms present, with their general and forcefield atomnames, atomic charges, connections to other atoms for creating their positions when not found on the PDB file, etc.; see part about fragment files below. Available in the ADF library are fragment files for amino acid residues, including those at the N- or C-terminal residue, three solvents (water, methanol, chloroform), some ions that are present frequently in protein structures (copper, fluoride), etc.

Also present in the ADF library are solvent box files that can be used to place a layer of solvents surrounding the protein, or a solute. Available are the three solvents mentioned above.

After reading the PDB and corresponding fragment files, the program tries to figure out which atoms are missing, and will add those; it uses the information provided on the fragment files to do so. For certain amino acid residues, there are several protonation states possible, e.g. histidine can be protonated at the N-delta position, at the N-epsilon position, or on both. The default option is to choose the fully charged option for aspartate (Asp), glutamate (Glu), lysine (Lys) residues, and decide for each histidine (His) and cysteine (Cys) residue individually what the protonation state should be. In those individual cases, the distances of neighboring molecules/residues are given that may help determine the protonation state. See the protein example below.

After all that is setup properly, a list is given with residue names/numbers, from which you can choose those that should be placed in the QM system; afterwards, for each of the selected QM residues, a choice should be made where to cut-off the QM part. The most appropriate point to cut-off seems to be at the C-alpha position, except when dealing with a proline (Pro). The latter residue is cyclic, e.g. the sidechain is connected to the C-alpha carbon ! For that residue, it may be better to include the C-alpha, H-alpha, and backbone carbonyl group of the preceding residue in the QM part.

The program will try to use to replace the ".pdb" extension of the PDB file by ".pdb2adf" for the ADF inputfile to be made; for convenience, the program also writes out an ".p2a.pdb" file with the complete system as it being made by the program. This file can then be visualized by conventional viewer programs (such as iMol, VMD, Molekel, ADFview) for visual inspection if everything has been carried out correctly.

Given below are two examples, one for the application of a protein, the other how to set up a solvent shell run.