Uncertainty reduction in biochemical kinetic models: Enforcing desired model properties

Area of Science:

• Systems Biology
• Computational Biology
• Metabolic Engineering

Background:

• Kinetic models of metabolism are crucial but hindered by enzyme kinetic property uncertainty.
• Existing methods indirectly address uncertainty by integrating diverse biological data.
• Accurate kinetic parameters are vital for predicting and controlling metabolic functions.

Purpose of the Study:

• To present and validate the iSCHRUNK computational approach for characterizing and reducing uncertainty in metabolic kinetic models.
• To identify specific kinetic parameters and their values that drive desired metabolic behaviors.
• To demonstrate the application of iSCHRUNK in metabolic engineering for optimizing cellular functions.

Main Methods:

• Utilized iSCHRUNK, integrating Monte Carlo parameter sampling with machine learning within a Bayesian inference framework.
• Employed Monte Carlo methods to generate diverse kinetic models consistent with data and physical laws.
• Applied machine learning for data mining kinetic parameters and deriving posterior distributions aligned with physiological observations.

Main Results:

• Successfully identified a small set of critical kinetic parameters influencing metabolic behavior.
• Demonstrated that only three kinetic parameters are essential for describing the studied physiology.
• Quantified flux control coefficients for xylose uptake (XTR) in S. cerevisiae.

Conclusions:

• The iSCHRUNK framework effectively reduces uncertainty in metabolic kinetic models.
• Accurate characterization of a few key kinetic parameters enables precise control over metabolic responses.
• This approach facilitates informed decision-making in metabolic engineering and synthetic biology by leveraging omics data.