Abstract: Structure-based docking can be used to screen compound libraries for novel ligands. Recently, docking libraries have expanded from three million “in-stock” to billions of make-on-demand (“tangible”) molecules. Docking these new libraries versus receptor structures has revealed novel scaffolds with nM and sub-nM potencies directly from the docking. I will discuss recent applications to several targets where in vivo active leads have been developed for analgesia and metabolic control.
Methods questions will also be considered: the effect of bias toward bio-like molecules in the virtual libraries, how and if docking score improves as the libraries grow, how number tested affects the quality of the experimental actives, and whether we have reached a plateau in the results we can expect from large library docking, or if bigger remains better.