Roger kornberg the importance of basic science
- 1. The importance of basic science Major advances in medicine from apparently unrelated discoveries:X-rays, antibiotics, nonivasive imaging, geneticengineering
- 4. The importance of basic science Major advances in medicine from apparently unrelated discoveries:X-rays, antibiotics, nonivasive imaging, geneticengineering The pursuit of knowledge for its own sake
- 5. Solve problems indirectlySupport for discoveryDo not: Define specific areas or priorities – let the best ideas win
- 6. Judge proposals on detailsDo: Fund individual investigators, directly, based on merits of proposals
- 7. Review and re-review
- 8. Play the odds - the law of large numbersConfidential and ProprietaryDrug design: Designing a ligand to fit into a protein and interfere with functionHenry Moore, Two FormsPynkado wood, 1934Metropolitan Museum of Art, New York© MMA, N.Y.
- 9. Industrial Drug DevelopmentHigh throughput screeningLong and elaborate Hit-to-Lead processConfidential and Proprietary
- 10. Confidential and ProprietaryIf we did a space launch the way we design drugs we would:Shoot 100 rockets to the Moon in hope of landing one.
- 11. Shoot another 500 rockets to get one to Mars.
- 12. Each new satellite would have to be very similar to a previous one. This is one of the reasons for Pharmaceutical industry’s troubles.Confidential and ProprietaryMoney spent on R&DApproved DrugsT.T. Ashburn and K.B. Thor, Nature Rev. Drug Discov. 3, 673-683 (2004)
- 13. Confidential and ProprietaryCrystallography -> Design -> Synthesis -> Evaluation -> Crystallography -> …
- 14. Why can’t theory help?The underlying physical laws necessary for the mathematical theory of a large part of physics and the whole of chemistry are thus completely known, and the difficulty is only that the exact application of these laws leads to equations much too complicated to be soluble.P. A. M. Dirac, Proc. R. Soc. London 123, 714 (1929)This is no longer true, computers can solve the equations. Given 10^12 years. Confidential and Proprietary
- 15. Current force-fields are too simpleConfidential and Proprietary
- 16. Current force-fields are parametrized depending on application: not extensibleDo not respond to environment
- 17. Parameters change with different applications: a sign of an incomplete model
- 18. Polarizability fundamentally impossible Confidential and Proprietary
- 19. Polarization is crucial in drug designConfidential and Proprietary
- 20. QUANTUM MECHANICAL POLARIZABLEFORCE FIELD (QMPFF)All major current force fields including AMBER and Merck Molecular Force Field (MMFF) are based on similar principles that date back decades agoQMPFF is designed to be physically more realistic and therefore more accurate by fully integrating quantum physics – made possible by vast increases in computer speed to do necessary calculations
- 21. +1+1QMPFF CONCEPTConventional Molecular Representation QMPFF Molecular Representation H2OH2O+ 0.4- 0.45- 7.1+6- 0.8- 0.45+ 0.4Point charges known to be physically unrealistic; no natural way to include important polarization energy termCharge clouds approximate quantum reality; polarization naturally introduced via shifting of electron clouds
- 22. QMPFF technology captures the missing detailsQuantum Mechanical Polarizable Force Field (QMPFF)Polarizable force field models polar interactions (see below)Can model water, gas phase, ligand binding better than any existing computational tools using the same parameters. Produced DipoleElectric Field+Confidential and Proprietary
- 23. Interacting entities are cloud-like, polarizableOParameterized from QM calculations. Polarizability is essential to modeling biological interactions. Electron clouds adjust during the simulation-> (real VdW).+---++HHConfidential and Proprietary
- 24. WATER SIMULATIONS WITH QMPFF2
- 25. QMPFF vs MMFF94 RING-STACKING OF BENZENE DIMER
- 26. QMPFF AND DRUG DESIGNQMPFF via MD simulation can be applied to calculate relative binding free energy (and therefore relative binding affinity) of two related ligands for a proteinX-ray structure of protein is required, and structure of complex with representative ligand(s) is very helpful
- 27. We mutate one ligand into anotherThermodynamic integrationEach alchemical mutation requires running a series of simulations. Takes approximately 8 days per calculation (10 cores). Most suitable for lead and drug optimization. MutationConfidential and Proprietary
- 28. 2BZAbenzyl-ammoniumBINDING FREE ENERGY CALCULATIONS FOR 1TNH AND 2BZA IN TRYPSIN1TNH4-fluoro-benzyl-ammoniumΔΔG (kcal/mol)QMPFF3 0.85 0.17MMFF94 -0.4 Exper. 0.78
- 29. Does it work?Ligand substitution pilot study (trypsin, thrombin and uPA inhibitor families). Nothing else can predict free energies accurately (see comparison to Merck results below)QMPFFMMFFConfidential and ProprietaryActual
- 30. Does it work?
- 31. Does it work?Drug design – blind prediction---------PredictedMeasuredReference compoundL
- 32. Speed-accuracy tradeoffthe exceptionConfidential and ProprietarySpeed10^12 years1 week0.01 secQMQMPFFNecessary accuracyPrecisionConventional Force FieldsDocking(Snapshot)
- 33. Impact on drug designConfidential and ProprietarySpeeding up binding optimization
- 34. Difficult synthesis can be virtual
- 35. Scaling with available computers
- 36. Making best of class drugs SOME APPLICATIONS OF QMPFFHydrogen storage and fuel cellsBatteriesOptical materialsCatalysts