Network inference using steady-state data and Goldbeter-Koshland kinetics. [corrected].

Area of Science:

• Systems Biology
• Molecular Systems Biology
• Biochemical Network Analysis

Background:

• Network inference is crucial for understanding molecular interactions but often overlooks nonlinear biochemical kinetics.
• Existing methods for steady-state data primarily use statistical approaches, neglecting valuable mechanistic information.
• Nonlinear biochemical processes govern many biological networks, such as gene regulation and protein signaling.

Purpose of the Study:

• To develop a network inference approach for steady-state data that incorporates nonlinear biochemical mechanisms.
• To enable more accurate estimation of molecular interaction networks by utilizing chemical kinetics.
• To provide a method applicable to gene expression and proteomic data without prior knowledge of network topology or kinetic parameters.

Main Methods:

• Developed an approach rooted in the equilibrium analysis of chemical kinetics to derive functional forms for network inference.
• Applied the method to steady-state gene expression and proteomic data.
• Utilized simulated data from a mechanistic model and real proteomic data from cancer cell lines for validation.

Main Results:

• The proposed method effectively infers network topology from steady-state data by leveraging nonlinear biochemical descriptions.
• Demonstrated superior performance in estimating network topology compared to traditional methods based on linear models.
• Successfully applied the approach to protein phosphorylation networks using both simulated and experimental proteomic data.

Conclusions:

• Network inference for steady-state data can be significantly improved by incorporating nonlinear chemical kinetics.
• The developed methodology offers a powerful tool for deciphering complex molecular regulatory networks.
• This approach enhances the understanding of biological systems by integrating mechanistic insights with data-driven inference.