Bayesian Model Calibration and Sensitivity Analysis for Oscillating Biological Experiments

Area of Science:

• Systems Biology
• Computational Biology
• Biophysics

Background:

• Understanding biological oscillations requires integrating experimental and computational approaches.
• Significant statistical challenges exist in harmonizing these experimental types, including identifiability issues and high-dimensional instability.
• Oscillating biochemical models are crucial for studying biological processes like circadian rhythms.

Purpose of the Study:

• To develop a novel Bayesian calibration framework for oscillating biochemical models.
• To address statistical challenges in parameter inference and sensitivity analysis for these models.
• To effectively link simulated and observed oscillatory biological data.

Main Methods:

• A Bayesian calibration framework is proposed for oscillating biochemical models.
• Advanced Markov chain Monte Carlo (MCMC) techniques are employed for parameter inference.
• An intervention posterior approach is used for sensitivity analysis.

Main Results:

• The framework efficiently infers parameter values matching simulated and observed oscillatory processes.
• Sensitivity analysis quantifies the influence of individual parameters on biological processes.
• The methodology is successfully illustrated using circadian oscillations in *Neurospora crassa*.

Conclusions:

• The proposed Bayesian framework offers a robust method for calibrating oscillating biochemical models.
• The MCMC technique and sensitivity analysis provide valuable tools for systems biology research.
• This approach enhances the integration of computational and experimental data in understanding biological oscillations.