Moment-Based Parameter Estimation for Stochastic Reaction Networks in Equilibrium.

Area of Science:

• Systems Biology
• Computational Chemistry
• Biophysics

Background:

• Parameter calibration is essential for quantitative reaction network models.
• Current stochastic model methods are limited to small systems or require extensive statistical sampling.
• High-throughput data, such as from flow cytometry, presents new opportunities and challenges for model calibration.

Purpose of the Study:

• To present a novel inference procedure for calibrating parameters in stochastic models at equilibrium.
• To enable the use of high-throughput data for parameter inference in complex reaction networks.
• To provide a method that avoids approximating the equilibrium probability distribution.

Main Methods:

• The study introduces a moment matching scheme with optimal weighting for parameter inference.
• The procedure is designed to be compatible with high-throughput data acquisition methods like flow cytometry.
• Parameter estimation involves solving a linear system of equations when reaction rate constants are unknown.

Main Results:

• The proposed method effectively calibrates parameters for stochastic models in equilibrium.
• It demonstrates applicability to systems generating high-throughput data.
• The approach was validated through three distinct case studies, showing its practical utility.

Conclusions:

• The presented inference procedure offers an efficient and scalable solution for parameter calibration in stochastic reaction network models.
• It overcomes limitations of existing methods, particularly for large and complex systems analyzed with high-throughput data.
• The method provides a robust framework for advancing quantitative modeling in various scientific domains.