The International Conference for High Performance Computing, Networking, Storage and Analysis
A Molecular Model for Platelets at Multiple Scales and Simulations on Supercomputers.
Student: Li Zhang (Stony Brook University)
Supervisor: Yuefan Deng (Stony Brook University)
Abstract: A molecular model at multiscale for human platelets is designed and implemented on large-scale supercomputers. The model characterizes the principal physiological functions of key components of single platelet. The organelle zone centralizes as the cytoplasm, the peripheral zone includes the bilayer and exterior coats, and the cytoskeleton zone forms the flow-induced filopodia. It is the first time such a massive model in biomedical engineering is proposed and high-performance computing is the only potential solution for initial-understanding of the model. Algorithmically, a coarse-grained molecular dynamics force field describes the molecular-level interactions in the membrane and cytoskeleton while a Morse potential is applied to the cytoplasm. Together, a coarse-grained stochastic dynamics is applied to simulating the macroscopic viscous fluid flows. Numerical experiments with our model require considerations of large number of particles and the physiological phenomena at multiscale and thus demand development of the-state-of-the-art parallel algorithms, the main thrust of this research.