Addressing the challenges posed by combinatorial chemistry: 3D databases, pharmacophore recognition and beyond

The flow of information in combinatorial chemistry is much like that which medicinal chemists have traditionally faced; however, the speed and quantity of that flow is much greater. In traditional chemistry, computational tools have been applied essentially as a replacement for Dreiding models; with combi-chem, the use of computational tools now becomes compelling. In our environment, the primary focus is less on the use of combi-chem in the development of universal libraries for screening, and more on its use in the development of targeted libraries. Hence, the four primary challenges we face are: (1) gaining an understanding of the structure-activity relationship (SAR), (2) using that understanding in the iterative design of new targeted libraries, (3) incorporating any structural data into that iterative design process, and (4) registering the compounds produced by combi-chem into a database. For computational tools to be effective, they must address these challenges more quickly and more accurately than the combi-chemist is able to achieve independently, i.e, high-quality answers must be developed in less than a few days. We describe how the first challenge may be addressed by pharmacophore recognition tools (Catalyst, DANTE), how the second and third may be addressed by 3D database searching, and how the fourth may be addressed by 'virtual chemistry' using databases of compounds and tables of reactions which may be applied to those compounds.