Dataset Information

Model-based control of the temporal patterns of intracellular signaling in silico.

ABSTRACT: The functions of intracellular signal transduction systems are determined by the temporal behavior of intracellular molecules and their interactions. Of the many dynamical properties of the system, the relationship between the dynamics of upstream molecules and downstream molecules is particularly important. A useful tool in understanding this relationship is a methodology to control the dynamics of intracellular molecules with an extracellular stimulus. However, this is a difficult task because the relationship between the levels of upstream molecules and those of downstream molecules is often not only stochastic, but also time-inhomogeneous, nonlinear, and not one-to-one. In this paper, we present an easy-to-implement model-based control method that makes the target downstream molecule to trace a desired time course by changing the concentration of a controllable upstream molecule. Our method uses predictions from Monte Carlo simulations of the model to decide the strength of the stimulus, while using a particle-based approach to make inferences regarding unobservable states. We applied our method to in silico control problems of insulin-dependent AKT pathway model and EGF-dependent Akt pathway model with system noise. We show that our method can robustly control the dynamics of the intracellular molecules against unknown system noise of various strengths, even in the absence of complete knowledge of the true model of the target system.

SUBMITTER: Murakami Y

PROVIDER: S-EPMC5325056 | biostudies-literature | 2017

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Model-based control of the temporal patterns of intracellular signaling <i>in silico</i>.

Murakami Yohei Y Koyama Masanori M Oba Shigeyuki S Kuroda Shinya S Ishii Shin S

Biophysics and physicobiology 20170222

The functions of intracellular signal transduction systems are determined by the temporal behavior of intracellular molecules and their interactions. Of the many dynamical properties of the system, the relationship between the dynamics of upstream molecules and downstream molecules is particularly important. A useful tool in understanding this relationship is a methodology to control the dynamics of intracellular molecules with an extracellular stimulus. However, this is a difficult task because ...[more]

PMID: 28275530

Similar Datasets

Project description:Despite their small and simple structure compared with their hosts, virus particles can cause severe harm and even mortality in highly evolved species such as humans. A comprehensive quantitative biophysical understanding of intracellular virus replication mechanisms could aid in preparing for future virus pandemics. By elucidating the relationship between the form and function of intracellular structures from the host cell and viral components, it is possible to identify possible targets for direct antiviral agents and potent vaccines. Biophysical investigations into the spatio-temporal dynamics of intracellular virus replication have thus far been limited. This study introduces a framework to enable simulations of these dynamics using partial differential equation (PDE) models, which are evaluated using advanced numerical mathematical methods on leading supercomputers. In particular, this study presents a model of the replication cycle of a specific RNA virus, the hepatitis C virus. The diffusion-reaction model mimics the interplay of the major components of the viral replication cycle, including non structural viral proteins, viral genomic RNA, and a generic host factor. Technically, surface partial differential equations (sufPDEs) are coupled on the 3D embedded 2D endoplasmic reticulum manifold with partial differential equations (PDEs) in the 3D membranous web and cytosol volume. The membranous web serves as a viral replication factory and is formed on the endoplasmic reticulum after infection and in the presence of nonstructural proteins. The coupled sufPDE/PDE model was evaluated using realistic cell geometries based on experimental data. The simulations incorporate the effects of non structural viral proteins, which are restricted to the endoplasmic reticulum surface, with effects appearing in the volume, such as host factor supply from the cytosol and membranous web dynamics. Because the spatial diffusion properties of genomic viral RNA are not yet fully understood, the model allows for viral RNA movement on the endoplasmic reticulum as well as within the cytosol. Visualizing the simulated intracellular viral replication dynamics provides insights similar to those obtained by microscopy, complementing data from in vitro/in vivo viral replication experiments. The output data demonstrate quantitative consistence with the experimental findings, prompting further advanced experimental studies to validate the model and refine our quantitative biophysical understanding.

Project description:Computational approaches to tune the activation of intracellular signal transduction pathways both predictably and selectively will enable researchers to explore and interrogate cell biology with unprecedented precision. Techniques to control complex nonlinear systems typically involve the application of control theory to a descriptive mathematical model. For cellular processes, however, measurement assays tend to be too time consuming for real-time feedback control and models offer rough approximations of the biological reality, thus limiting their utility when considered in isolation. We overcome these problems by combining nonlinear model predictive control with a novel adaptive weighting algorithm that blends predictions from multiple models to derive a compromise open-loop control sequence. The proposed strategy uses weight maps to inform the controller of the tendency for models to differ in their ability to accurately reproduce the system dynamics under different experimental perturbations (i.e. control inputs). These maps, which characterize the changing model likelihoods over the admissible control input space, are constructed using preexisting experimental data and used to produce a model-based open-loop control framework. In effect, the proposed method designs a sequence of control inputs that force the signaling dynamics along a predefined temporal response without measurement feedback while mitigating the effects of model uncertainty. We demonstrate this technique on the well-known Erk/MAPK signaling pathway in T cells. In silico assessment demonstrates that this approach successfully reduces target tracking error by 52% or better when compared with single model-based controllers and non-adaptive multiple model-based controllers. In vitro implementation of the proposed approach in Jurkat cells confirms a 63% reduction in tracking error when compared with the best of the single-model controllers. This study provides an experimentally-corroborated control methodology that utilizes the knowledge encoded within multiple mathematical models of intracellular signaling to design control inputs that effectively direct cell behavior in open-loop.

Dataset Information

Model-based control of the temporal patterns of intracellular signaling in silico.

Publications

Model-based control of the temporal patterns of intracellular signaling <i>in silico</i>.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets