Hybrid regulatory models: a statistically tractable approach to model regulatory network dynamics

Area of Science:

• Systems Biology
• Computational Biology
• Molecular Biology

Background:

• Computational modeling of gene regulatory networks is crucial in systems biology.
• Ordinary differential equations (ODEs) are commonly used but difficult to calibrate.
• A need exists for robust statistical inference frameworks for complex biological networks.

Purpose of the Study:

• To develop a general statistical inference framework for stochastic transcription-translation networks.
• To enable scalable inference and learning of model parameters for gene regulatory networks.
• To provide a flexible method for making testable biological predictions.

Main Methods:

• Developed a coarse-grained approach representing systems as networks of stochastic promoter and continuous protein variables.
• Derived an exact inference algorithm for precise analysis.
• Implemented an efficient variational approximation for scalable inference and parameter learning.

Main Results:

• Demonstrated a general statistical inference framework for stochastic transcription-translation networks.
• Successfully applied the method to two biological case studies.
• Showcased the framework's flexibility and capability for generating novel biological predictions.

Conclusions:

• The developed framework offers a powerful and flexible approach to modeling gene regulatory networks.
• Scalable inference and parameter learning are achievable for stochastic transcription-translation systems.
• This method facilitates data-driven biological discovery and hypothesis generation.