Symbolic approaches for finding control strategies in Boolean Networks

Area of Science:

• Computational Biology
• Systems Biology
• Computer Science

Background:

• Boolean Networks (BNs) are widely used to model complex biological systems, such as gene regulatory networks.
• Finding effective control policies for BNs is computationally challenging, with recent studies highlighting its inherent hardness.
• Automated modeling and control are crucial for understanding and manipulating biological processes.

Purpose of the Study:

• To develop an exact algorithm for finding control policies in Boolean Networks with control nodes.
• To guarantee the deterministic driving of a BN from initial states to goal states within a specified time.
• To extend the method for automatic identification of Boolean transfer functions that model gene regulatory networks.

Main Methods:

• Utilized Model Checking techniques to design an exact algorithm for BN control policy synthesis.
• The algorithm takes a BN, initial states (I), goal states (F), and a target time (t) as input.
• Applied the algorithm to learn BN models from data, specifically for Drosophila melanogaster embryogenesis.

Main Results:

• The algorithm successfully finds control policies or proves their non-existence for BNs.
• Demonstrated practical efficiency, running in seconds to minutes on over 13,400 BNs, including a large model of Drosophila embryogenesis.
• Successfully learned a BN model of Drosophila melanogaster embryogenesis from a vast search space in 5.3 seconds.

Conclusions:

• The developed Model Checking-based algorithm provides an efficient and exact solution for BN control policy synthesis.
• The method is scalable and effective for complex biological network modeling and control.
• This approach significantly advances automated modeling and analysis of gene regulatory networks.