Multiscale ab initio QM/MM and machine learning methods for accelerated free energy simulations

Q Chem is a state of the art commercial computational quantum chemistry program that has aided aboutusers in their modeling of molecular processes in a wide range of disciplinesincluding biologychemistryand materials scienceIn this proposalwe seek to significantly reduce the computational timenow aroundCPU hoursrequired to obtain accurate free energy profiles of enzymatic reactionsSpecificallywe propose to use a multiple time stepMTSsimulation methodwhere a low leveland less accuratequantum chemistry method is used to propagate the systemi emove all atomsat each time stepusuallyorfsand then a high leveli emore accurate and expensivequantum chemistry method is used to correct the force on the atoms at longer time intervalsIn this waythe simulation can be performed at the high level energy surface in a fraction of timecompared with simulations performed only using the high level quantum chemical methodIn the Phase I proposalour goal is to allow the high level force update only once everyfs by identifying appropriate lower level theoriesAimand incorporating machine learning techniquesAimThis will accelerate accurate free energy simulations byfoldreducing the overall computer time to aroundCPU hoursThusour new MTS simulation method will make it feasible to routinely perform computational studies on enzymatic reaction mechanismThe addition of these new tools will also further strengthen Q Chemandapos s position as a global leader in the molecular modeling software marketmaking our program the most efficient and reliable computational quantum chemistry package for simulating largecomplex chemical biological systems In this projectwe seek to significantly reduce the computational timecaCPU hoursrequired to obtain accurate free energy profiles of enzymatic reactions to caCPU HoursBuilding upon sophisticated quantum mechanicsthis can lead to reliable and quick predictions of enzyme activities