(Modeling proteins)

Choosing a format: MOPAC or PDB

The two main formats used by MOPAC are the native MOPAC format and the PDB format, so a reasonable question would be, "Which is better?"  The easiest way to answer this is to compare the features of the MOPAC and PDB formats as they apply to data sets that can be run using MOPAC.

PDB format


Exact compatibility with the PDB format.


The user cannot specify individual atoms for optimization, however sets of atoms can be selected using NOOPT OPT-H, etc. 
Coordinates are limited to three decimals, so calculated properties would be slightly different to those predicted if the MOPAC format were to be used.  These differences are quite small, a few tenths of a Kcal/mol for chymotrypsin, and if a geometry optimization is involved, are really quite unimportant.

MOPAC format


Very fine control of geometry optimization is possible, for example gradient minimization of the individual atoms involved in a reaction can be specified.


The geometry does not include all information that is present in the PDB geometry.  Extra information, such as the positions of missing residues and chain ends and chain letters must be given by keyword START_RES  at the start of the data set.  This extra information is only needed if an exact comparison with the original PDB file has to be made.

Suggested use of PDB and MOPAC formats

Set up data sets for modeling using the MOPAC format.  This is easy to use, and when atom numbers are replaced by PDB or Jmol labels, editing optimization flags and adding or deleting atoms is straightforward.  Include keyword PDBOUT or HTML, so that the job will produce a PDB file suitable for use with a GUI. This gives the best of both formats: MOPAC format is easy to edit, and PDB format is easy to use with GUIs.