KGBN: Augmenting and optimizing logical gene regulatory networks using knowledge graphs.

Area of Science:

• Systems Biology
• Computational Biology
• Genomics

Background:

• Logical gene regulatory network (GRN) models are crucial for understanding cellular regulation but often remain incomplete and context-specific.
• Limitations hinder their application in areas like drug-response prediction and precision medicine.

Purpose of the Study:

• To present KGBN (Knowledge Graph-augmented Boolean Network modeling), a novel computational workflow for systematically augmenting logical GRN models.
• To enhance the interpretability, context-specificity, and comprehensiveness of GRNs for advanced applications.

Main Methods:

• KGBN integrates regulatory interactions from curated knowledge graphs as alternative logical rules.
• It preserves the validated structure of existing GRN models.
• Rule probabilities are optimized against experimental data for data-driven calibration and to represent regulatory uncertainty.

Main Results:

• Application of KGBN to acute myeloid leukemia (AML) demonstrated its utility.
• Extending an existing GRN with drug-target pathways and training against ex vivo drug-response data.
• Generated mutation-specific models that accurately recapitulate known therapeutic sensitivities and signaling dependencies.

Conclusions:

• KGBN provides a powerful framework for building interpretable, context-aware, and data-driven GRN models.
• This approach advances precision medicine by enabling more accurate drug-response predictions.
• The workflow facilitates the extension of existing models for broader applicability.