Stochastic Rate Parameter Inference Using the Cross-Entropy Method

Abstract

We present a new, efficient algorithm for inferring, from time-series data or high-throughput data (e.g., flow cytometry), stochastic rate parameters for chemical reaction network models. Our algorithm combines the Gillespie stochastic simulation algorithm (including approximate variants such as tau-leaping) with the cross-entropy method. Also, it can work with incomplete datasets missing some model species, and with multiple datasets originating from experiment repetitions. We evaluate our algorithm on a number of challenging case studies, including bistable systems (Schlögl’s and toggle switch) and experimental data.