Abstract:
|
Cellular intrinsic noise plays an essential role in the regulatory interactions between genes. Although a variety of quantitative methods are used to study gene regulation system, the role of intrinsic noises has largely been overlooked. Using the Kolmogorov backward equation (master equation), we formulate a causal and mechanistic Markov model. This framework recognizes the discrete, nonlinear and stochastic natures of gene regulation and presents a more realistic description of the physical systems than many existing methods. Within this framework, we develop an associated moment-based statistical method, aiming for inferring the unknown regulatory relations. By analyzing the observed distributions of gene expression measurements from both unperturbed and perturbed steady-states of gene regulation systems, this method is able to learn valuable information concerning regulatory mechanisms. This design allows us to estimate the model parameters with a simple convex optimization algorithm. We apply this approach to a synthetic system that resembles a genetic toggle switch and demonstrate that this algorithm can recover the regulatory parameters efficiently and accurately.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.