ABSTRACT
A wide variety of docking programs are currently employed in the pharmaceutical and biotechnology industries to advance lead discovery and lead optimization projects, the most widely used of which have been GOLD [1],
FlexX [2], and DOCK [3]. Over the past several years, considerable success has been reported for these programs in virtual screening (VS) applications [4-6]. However, neither these programs nor others that are available can be viewed as offering a robust and accurate solution to the docking problem, even in the context of a rigid protein receptor. Moreover, although many methods readily recognize favorable protein-ligand interactions, current-generation scoring functions do not adequately (if at all) penalize nonphysical interactions that oppose binding, and hence cannot efficiently rule out false positives. We and other developers of Glide are working to enable Glide, in reasonable computational time, to find the correctly docked structure in the vast majority of cases and to distinguish reliably between active ligands and ligands that could not plausibly bind to a given receptor. In this chapter, we will summarize the progress we have made to date in developing Glide 2.7 and will describe what users need to do or to know to employ Glide to best advantage in VS applications.
