Model reduction, coherence, and information transfer in stochastic biochemical systems.

Area of Science:

• Biophysics
• Systems Biology
• Biochemical Kinetics

Background:

• Stochastic models are crucial for analyzing noise propagation and information transmission in biochemical reaction networks.
• Coherence and mutual information (MI) rate are key metrics for quantifying these processes.
• Reduced models are commonly used approximations of complex biochemical systems.

Purpose of the Study:

• To challenge the assumption that simplified models accurately approximate coherence and MI rates in biochemical networks.
• To investigate the discrepancies in frequency-resolved noise propagation between full and reduced models.
• To identify the factors contributing to these inaccuracies in simplified models.

Main Methods:

• Analysis of frequency-resolved noise propagation in biochemical reaction networks.
• Comparison of coherence spectra and mutual information (MI) rates between full and reduced models.
• Investigation of model reduction methods and their impact on statistical properties.

Main Results:

• Reduced models accurately reproduce low-order statistics of molecular counts.
• Significant discrepancies are observed in coherence spectra at intermediate and high frequencies.
• These discrepancies lead to substantial inaccuracies in estimated mutual information (MI) rates.

Conclusions:

• Simplified stochastic models may not reliably estimate coherence and mutual information (MI) rates, despite accurately capturing molecular counts.
• The accuracy of reduced models depends on network structure, reduction method, and asymptotic limits.
• Careful consideration is needed when using simplified models for quantifying information flow in biological systems.