A key requirement for protein-family-based drug design  is the ability to compare and contrast active sites within and across families.   For example,  in the design of family-targeted combinatorial libraries it is desirable to ensure that energetically important interactions exist between the scaffold and portions of the active site that are common to all members of a family.  In the design of highly specific drugs it is desirable to ensure that energetically important interactions are formed to parts of the active site that are unique to the protein of interest.  We have developed structural bioinformatics tools to assist in these endeavors that providing  facile  surface-shape, surface-property, and interaction-based comparisons of active sites, specifically designed to guide drug discovery. 

The starting point is the careful superposition of  active sites for proteins to be compared.    This superposition is done in a sequence-independent manner, using an hierarchical description of each active site that includes its surface shape, surface properties, and underlying residues.   Relative weights can be assigned, depending on the problem, to surface-shape , surface-property , or underlying amino acid superpositions. This allows for an approach to families  where there is significant flexibility, where amino acids differ, or where surfaces are similar but where underlying amino acids are contributed from different parts of the protein structure (e.g. from insertions or deletions).

Once a collection of active sites is superimposed, visual maps are generated to highlight regions of similarity and/or of differences. Differences can be highlighted in surface shape, surface properties ( such as hydrophobicity), or interaction potentials.  Interaction potentials that can be examined include  hydrogen-bond donor potential, hydrogen-bond acceptor potential,  ring-stacking potential, and salt-bridge potential.   Again - regions that are similar for multiple members of a family can drive the development of family-targeted screening libraries or of family-targeted combinatorial chemistry libraries.  Regions that are different can drive the development of target-specific compounds, a strong requirement in lead optimization.