ABSTRACT
Due to the increase in target protein structures available as well as the increased quality of homology-based protein models, molecular docking and structure-based virtual screening (VS) are two key technologies in current drug discovery (see [1-4] for recent reviews). Their clear advantage is the direct elucidation of binding properties without any bias of the chemical structure of the small molecules. Solving the docking problem, namely the prediction of the geometry and binding affinity of a protein-ligand complex given a three dimensional (3D) structure of the protein and the ligand molecule, is at the heart of structure-based design. Several variants of the base problem exist that often allow, sometimes request, differentiated method and tool development [4,5]. Some targets show significant protein flexibility in the active site. Frequently, models have a high degree of uncertainty concerning particular atom positions. Water molecules and metal ions might play an important role. When it comes to the docking of multiple molecules, the efficient handling of combinatorial libraries as well as the incorporation of prior knowledge such as enforcing a certain interaction pattern gain importance.
