Accelerated Drug Design Through Computational Biology

Sequencing the human genome has uncovered numerous putative novel targets for which 3D structures are not available and which have no known inhibitors. We proposed a system to use computational methodologies to generate 3D molecular models of target proteins, and to use these models to design novel chemical inhibitors. Currently, the system is capable of accurately folding a target's amino acid sequences into 3D structures and identifying its ligand binding sites, docking ligands into identified binding sites, and scoring docked ligands. The system was validated by demonstrating its ability to predict experimentally determined 3D structures of PfPK5 and its intermolecular interactions with bound inhibitors, as well as predicting experimentally determined structure-activity relationships of Pfmrk-inhibiting oxindole ligands. Enabling the system to design novel high affinity ligands, predict target selectivity, and integrate the targets' conformational flexibility into ligand design would complete its development. Additionally, acquiring a cluster of networked computers would enable execution and high throughput of large-scale compute intensive jobs. This system would be used to develop models and design inhibitors for several targets of interest to WRAIR including PfPK5 and Pfmrk, as well as FAB H, UMP Synthase, OMPDC, OPRTase, Dihydroorotase, Dephospho Co-A Kinase, Pantothenate Kinase, and Shikimate Kinase.