Perturbation-based Markovian Transmission Model for Macromolecular Machinery in Cell

The study of the dynamics of a complex system is an important problem that includes large macromolecular complexes, molecular interaction networks, and cell functional modules. Large macromolecular complexes in cellular machinery can be modeled as a connected network, as in the elastic or Gaussian network models as demonstrated by Bahar and colleagues. Here we propose the Perturbation-based Markovian Transmission Model for studying the dynamics of signal transmission in macromolecular machinery. The initial perturbation is transmitted by a Markovian processes, and the dynamics of the probability flow is analytically solved using the master equation. Due to the large size of macromolecular complexes, it is very difficult to obtain analytical time-dependent Markovian dynamics of all atoms from the first perturbation until stationary state. To overcome it, we decrease the level of complexity of the transition matrix using a Krylov subspace method. This method is equivalent to integrating all eigen modes, and we show it can provide a globally accurate solution to the dynamics problem of signal transmission for very large macromolecular complexes with reasonable computational time. We give results of the dynamics of the GroEL-GroES chaperone system by applying uniform perturbation to all residues. We are able to identify experimentally found important residues and provide a set of predicted pivot, messenger, and effector residues, each with distinct dynamic behavior. Further results of selective perturbation on the surface of ATP binding pocket identifies the path of maximal probability flow of signal. Our method can also be applied to other large systems, for example, virus capsid, ribosome, and large allosteric proteins.