Combinatorial chemistry and parallel synthesis techniques can generate orders of magnitude more compounds than can be practically synthesized or screened. Therefore, there is a real need for efficient and reliable methods to rationally select the optimal library members for synthesis. In recent years, combinatorial library design has shifted toward the creation of smaller, more focused libraries that exhibit optimal drug-like physiochemical properties and are biased toward a specific target or class of targets.

High-throughput docking methodology that can accurately estimate ligand binding affinities can significantly reduce time and cost by identifying potent members of a virtual library and eliminating unpromising compounds early on.