Molecular Modeling


The molecular modeling facility offers several computational biology services such as the generation, dynamics and analysis of molecular structures of proteins (including membrane proteins) and nucleic acids, investigating the interaction of protein:ligand, protein:protein, protein:nucleic acid interactions, nucleic acid:ligand complexes, drug discovery including virtual screening against any target, and limited bioinformatics.

The Molecular Modeling Facility was created in 1999 when the Director, Dr. John Trent, was recruited to the Brown Cancer Center. It comprises three servers with a total of 450 processors, six graphics workstations, six state of the art GPUs, and, in partnership with Dataseam, a ~20,000 processor distributed computing grid for virtual screening. The Facility is an integral part of experimental therapeutic efforts and works closely with the NMR Facility, biophysicists, molecular biologists, chemists, computer scientists, and biochemists. The core has evolved into a highly collaborative center that provides expertise and resources not available elsewhere in the region. The philosophy of the Facility involves the rationalization of experimental data, the prediction of new empirical experiments, and the structure-based drug design of new agents. These aims are being met by maintaining state-of-the-art software and computational resources that has become the major source of drug discovery at the Brown Cancer Center.

The in silico drug discovery effort uses Autodock, DOCK, and Surflex with several databases of over 20,000,000 compounds for virtual screening. We use both Core, UofL CRC, and distributed grid processors.

We have successfully virtual screened and tested biologically in over 50 systems where we have experimentally obtained inhibitors with activity at less than 10 µM.

Current Molecular modeling software in modeling lab include Amber 5.0-14.0, complete Schrodinger Suite, DOCK 4.0, 5.1.1,6.2 Autodock 3.0.5, 4.0.1, Surflex, InsightII 2000.1 modules Discover, affinity, ludi, biopolymer, builder. Sybyl 8.1, Sybylx 2.1, Computational informatics software for molecular modelers including Advanced computation CoMFA Advanced, Biopolymer, CombiChem, Concord, Confort, CScore, Dynamics, FlexS, FlexX, FlexXPharm, GALAHAD, GASP, HINT, HQSAR, Legend, MOLCAD, MOLCONNZ, PMA, QSAR, RACHEL, SurFlexDock,Tuplets, Unity, Unity3D, VolSurf. Other programs are MODELLER, GRAMM. Ab initio, semiempirical, and Poisson-Boltzman electrostatic calculations, GAMESS, MOPAC, Delphi. Molecular Graphics and visualization packages, Grasp, Midas-plus, chimera. DNA analysis program Curves. Other programs include, MC-SYM, Namot, moil-view, mfold, VMD, molekel and rasmol.

Routine techniques involve ab initio, semi-empirical, empirical (in vacuo, implicit and explicit solvation), QM/MM calculations. Conformational searching: grid searches, simulated annealing, monte carlo, JBW, MC/MD, transition states. Electrostatic surface potentials, partial charges, visualization. Docking. DNA, modified DNA, RNA modeling and analysis and drug design including sequence selectivity, Protein Modeling “sequence to drug design” Structure, homology modeling, ala-scanning, stability, dynamics, protein-ligand, protein-protein, and protein-DNA interactions, membrane proteins, electrostatics, hydration, ion association, docking, virtual screening, pharmacophore generation and searching, NMR structures using state of the art advances in Molecular Modeling, Database searching. The Core will work with investigators in all aspects of computational biology.

Contacts:

John O. Trent, Director – john.trent@louisville.edu or 502-852-2194

Jonathan M. Maguire, Manager – jon.maguire@louisville.edu or 502 852-3653