A Parallel Computing Platform For Multiscale Modeling of Platelets

Antiplatelet pharmacotherapy currently focuses on biochemical mechanisms involved in platelet activation and aggregation. However, it ignores the impact of blood flow and shear stress which is mediated via mechanical transduction mechanisms. We develop a multiscale platform to mimic the phenomena of platelet activation adhesion and aggregation under the influence of shear stress in Poiseuille flow. Models based on continuum mechanics and mesoscopic methods fail to capture the molecular level phenomena of platelet activation and aggregation. To alleviate these problems, we use a multiscale model based on coarse grained molecular dynamics (CGMD) for the platelet activation embedded in the dissipative particle dynamics (DPD) for the blood flow model. This platform is a computationally feasible approach to get a good resolution and accuracy into such cellular processes in fluid flow while also being able to characterize the mechanical transduction mechanisms involved. Furthermore, it has the ability to measure aggregation and adhesion properties like contact frequency, contact area and characteristic contact time that are verified by experimental results. We further investigate properties for a deformable platelet model and correlate results with those obtained from rigid model. The platform will be used for scalability test on some supercomputers and will also be optimized for them.