Predicting the structure and energetics of protein-ligand interaction

Abstract

Molecular docking has been a crucial component and remains a highly active area in computer-aided drug design (CADD). In simple terms, molecular docking uses computer algorithms to identify the "best" match between two molecules, a process analogous to solving three-dimensional jigsaw puzzles. In more rigorous terms, the molecular docking problem can be defined as predicting the "correct" bound association state for the given atomic coordinates of two molecules. Docking is an important tool for structure and affinity predictions of molecular association, which would lead to the mechanistic understanding of the physicochemical interactions at the atomic level. Protein-small molecule (referred to as "ligand") docking, in particular, has broad application to structure-based drug design, as drug compounds are usually small molecules. In this dissertation, I present my studies on protein-ligand docking. In the background introduction, I reviewed the docking methodology and the key recent developments in the field. Next, I applied an ensemble docking algorithm onto 14 protein kinases to study ligand selectivity, a major issue for the development of kinase inhibitors as anticancer drugs. In Chapter 3, I developed a web server for automated, in silico screening of multiple targets for a given ligand query. Finally, I integrated the new methods for protein-ligand binding mode prediction and applied the integrated method to a large-scale, blind prediction competition named Continuous Evaluation of Ligand Pose Prediction (CELPP).