Parameter estimation for biochemical reaction networks using Wasserstein distances

Abstract

We present a method for estimating parameters in stochastic models of biochemical reaction networks by fitting steady-state distributions using Wasserstein distances. We simulate a reaction network at different parameter settings and train a Gaussian process to learn the Wasserstein distance between observations and the simulator output for all parameters. We then use Bayesian optimization to find parameters minimizing this distance based on the trained Gaussian process. The effectiveness of our method is demonstrated on the three-stage model of gene expression and a genetic feedback loop for which moment-based methods are known to perform poorly. Our method is applicable to any simulator model of stochastic reaction networks, including Brownian Dynamics.

arXiv comment22 pages, 8 figures. Slight modifications/additions to the text; added new section (Section 4.4) and Appendix

References in corpus (5)

• Analytical distributions for stochastic gene expression
• Approximation and inference methods for stochastic biochemical kinetics - a tutorial review
• Steady-state fluctuations of a genetic feedback loop: an exact solution
• Comparison of different moment-closure approximations for stochastic chemical kinetics
• Validity conditions for moment closure approximations in stochastic chemical kinetics

Cited by in corpus (6)

• Stochastic modeling of auto-regulatory genetic feedback loops: a review and comparative study
• Tensor-train approximation of the chemical master equation and its application for parameter inference
• Model Reduction for the Chemical Master Equation: an Information-Theoretic Approach
• Systematic model reduction captures the dynamics of extrinsic noise in biochemical subnetworks
• Identifiability of Chemical Reaction Networks with Intrinsic and Extrinsic Noise from Stationary Distributions
• Non-equilibrium time-dependent solution to discrete choice with social interactions