UC Riverside

Gliding assays, reduced in vitro model systems where motor proteins adsorbed onto a planar substrate bind to and move filaments, exhibit large-scale dynamic patterns like coherent swarming motion and density waves. These systems are sensitive to microscopic behavior such as the motor protein binding and unbinding dynamics, which take place on a faster timescale than the direct and fluid-mediated filament interactions. A micro-macro multiscale modeling and simulation framework for gliding assays is developed, allowing for detailed microscopic motor modeling as well as both steric and hydrodynamic interactions between filaments. A scalable hybrid CPU and multi-GPU implementation alleviates the cost associated with tracking the high-dimensional microstructure, achieving performance over 500 times greater than a single-threaded implementation.