Working Groups 4 . 1 Simulating macromolecular crowding with particle and lattice - based methods ( Team 3 )

Many simulation algorithms have been developed to help model spatial structure in cellular systems, each of which is intended to represent reaction dynamics with high spatial resolution. In this study, we simulated the e ects of macromolecular crowding on biochemical reaction rates to investigate which method actually performs best in practice. All 5 simulators investigated showed that di usion-limited reaction rates decreased monotonically with the fractional crowder occupancy and activation-limited reactions exhibited an initial reaction rate increase with crowder occupancy (due to excluded volume e ects). The eGFRD simulations were presumably highly accurate, but were too computationally intensive to be ideal for this problem. The Smoluchowski method as implemented in Smoldyn had simulation parameters that could be connected directly to physical parameters, and did not appear to exhibit simulation artifacts. The Smoluchowski method as implemented in NL-space produced qualitatively similar di usion-limited results, but did not show a change of dynamics when changing to activation-limited conditions. Spatiocyte used a microscopic lattice, which enabled it to run very fast but introduced lattice artifacts in the results. Finally, we did not collect quantitative results with Kappa, but instead observed that Kappa can be used for this type of problem. Overall, this study showed that the detailed simulation methods substantially a ect the results and that each of these simulators can still be improved.