Any study of the properties of proteins by means of theoretical methods requires a well prepared starting structure.
The coordinates of proteins deposited in a database suffer from the lack of hydrogens. Thus some essential properties have to be reconstructed before use: protonation state of ionizable functions (e.g. termini, acidic and basic side chains), orientation of ambivalent side chains (e.g. HIS, ASN etc.), ionization and coordination of metal centers, structure and charge distribution in hetero-groups.
The resultant structure has to be validated by MD simulations to judge the quality of the various assignments during the structure preparation. The deviations between the experimental structure and the representative structure from the simulation have to be determined and the reasons assessed.
With the flexible protein structure now available the essential structure forming contacts between the residues and to the hetero-groups have to be determined. With these patterns the relevant interactions to form the active sites can be summarized and a reliable active site model derived.
This model can subsequently be used to determine the specific functionalities available for interactions with complementary functions in other (possibly inhibiting) molecules and their flexibility. Thus the binding of possible drugs (that were even not yet synthesized) can be determined and novel molecules designed.