A variety of different protein targets have been used to test a virtual screening approach for docking small molecule ligands. Over 18 isoforms of the cyclophilin family are present in the worm C.elegans. A store of over 5 million available molecules stored in a relational database is used to identify potential ligands. Over five chemically distinct families of inhibitor have been identified. Subsequent chemical modification and testing show theses families are biologically active and cause phenotypic changes in the worm similar to those caused by cyclosporine. A similar approach has been applied to a series of cell cycle regulators including CDK,cyclin and PCNA resulting in novel and specific inhibitors. Biophysical binding studies using surface plasmon resonance and ITC have been used to characterise the protein-ligand interactions and provide experimental measures of binding energies. X-ray structures of selected complexes have been determined to provide experimental poses. The program LIDAEUS has been parallelised to run on the supercomputer BlueGene which allows significantly higher throughput. In massively parallel docking runs it is possible to store top scoring poses for millions of compounds docked into all available classes of protein binding pockets. Such data can be used to generate novel chemical descriptors and provides new ways of classifying chemical families.