Application of GPU-Accelerated Molecular Dynamics Simulations of G-Quadruplexes with the Drude Polarizable Force Field
TimeTuesday, July 306:30pm - 8:30pm
LocationCrystal Foyer and Crystal B
DescriptionIntegrating computational tools in life science research is beneficial due to their low cost and ability to provide unique insights into biological and chemical phenomena. Molecular dynamics (MD) simulations is one such tool that can further characterize atomic-scale behaviors of biomolecules. MD simulations rely on force fields (FFs) to compute the potential energy of a given system of atoms, as such the quality of the chosen force field can greatly affect simulation results. The recent development of polarizable FFs is a significant advancement in MD that provides new insight on complex biomolecular systems. Our research group performs MD simulations of DNA G-Quadruplexes (GQs), which are fundamental to maintaining genomic stability and are implicated in cancer and neurodegenerative disease. Previous studies suggest that polarization is important for modeling GQs, so we use both the CHARMM36 (C36) nonpolarizable and Drude-2017 polarizable FFs to contextualize the role of polarization in the simulations. The Drude FF assigns negatively charged particles, called “Drude oscillators,” to all non-hydrogen atoms in the system to allow for the quantification of induced electronic polarization. Adding these auxiliary charged particles to the system increases computational burden and the time it takes to produce trajectories, which is a primary shortcoming of polarizable FFs. Therefore, accelerating computational performance with GPU-equipped clusters is important to produce results in a practical time frame. Our work has utilized GPU- and CPU-based clusters to produce simulations with the C36 and Drude FFs, allowing us to compare cluster hardware and expand our understanding of nucleic acid dynamics.