Estimating parameters and hidden variables in non-linear state-space models based on ODEs for biological networks...

Area of Science:

• Systems Biology
• Computational Biology
• Bioinformatics

Background:

• Biological networks, including gene regulatory, signaling, and metabolic networks, are crucial for understanding cellular interactions.
• Estimating parameters and hidden variables in dynamical, non-linear, and partially observed biological systems presents significant challenges.

Purpose of the Study:

• To develop and apply a robust method for parameter and hidden variable estimation in non-linear biological network models.
• To adapt Unscented Kalman Filtering (UKF) for inferring states and parameters within complex biological systems.

Main Methods:

• Non-linear state-space models were derived from Ordinary Differential Equations (ODEs) describing biological networks.
• Unscented Kalman Filtering (UKF) was employed for the estimation of parameters and hidden variables.
• The method was tested on transcriptional regulatory (Hill kinetics) and signaling pathway (mass action kinetics) models.

Main Results:

• Successful estimation of parameters and hidden variables was achieved using both synthetic and experimental data.
• The approach demonstrated efficacy on diverse biological network models, including transcriptional and signaling pathways.
• The method provides direct Bayesian uncertainty estimates for parameters and hidden states.

Conclusions:

• The proposed UKF-based approach offers a versatile and efficient tool for analyzing a wide range of biological network models.
• The method yields simple and fast estimation algorithms with inherent uncertainty quantification.
• This framework can be integrated with structure inference methods from Graphical Probabilistic Models for enhanced network analysis.