Illustration of current challenges in molecular docking: An application of docking and virtual screening to thymidine kinase

Abstract

Structure based drug design is now an established approach in drug discovery. Computational methodologies are used to facilitate structure based drug design at various stages of the process. One of the most important and routinely adopted methods is molecular docking. Molecular docking refers to the prediction of the binding mode of a specified compound within the active site of the protein target of interest (Taylor et al., 2002). A related application is virtual screening (VS) where a database of compounds is docked against a protein active site and a scoring function is used to rank the compounds in the database (Kitchen et al., 2004). This ranking is used in conjunction with other properties to identify compounds for synthesis or acquisition as part of an ongoing drug discovery program. This chapter will focus on the application of molecular docking and VS to structure based drug design and will focus on the current challenges that limit the accuracy and utility of these methods. The key challenges will be outlined in the next section and the remainder of the chapter will be concerned with a practical example of docking against thymidine kinase (TK), using the docking program GOLD (Jones et al., 1995; Jones et al., 1997). This example has been chosen because it illustrates many of the difficulties associated with accurate molecular docking and VS. We describe a number of theoretical and pragmatic strategies that we have found useful in addressing the challenges associated with docking against TK. © 2007 Springer.