Title:Hessian Analysis Enables Prediction of Cytoquake Occurrence

Abstract:The recent experimental discovery of large, sudden motions of the cytoskeleton, called "cytoquakes," opens new questions about whether cytoskeletal dynamics exhibit self-organized criticality and whether such motions are the result of some underlying mechanical instability. We explore these questions using a computational model of actomyosin network self-organization, which we observe produces earthquake-like releases of mechanical energy that we identify as cytoquakes. We characterize in detail the statistics of these mechanical energy fluctuations, illustrating that for the physiological conditions tested here the system is subcritical. The large releases of stored energy are observed to correspond to large displacements of the filaments, i.e. global structural rearrangements of the network. We next probe the cytoquake mechanism by analyzing the network's vibrational normal modes using Hessian analysis. We describe several interesting observed trends resulting from the novel application of this analysis to cytoskeletal networks, including a relationship between normal mode delocalization and network connectivity. Finally, a machine learning model is implemented which successfully forecasts cytoquake occurrence, showing that signatures of mechanical instability present in the Hessian eigenspectrum are well correlated with the propensity for a network to experience a large fluctuation in mechanical energy.