BONITA

BONITA models and scores pathway-level signal integration in large-scale omics datasets using Boolean (discrete-state) network modeling to infer logic rules and quantify pathway modulation.

Key Features:

• Boolean Rule Inference: Infers Boolean rules that govern gene interactions within pathways to represent node-level signal integration.
• Discrete-State Network Modeling: Employs discrete-state (binary) network dynamics to model cellular heterogeneity across cross-sectional transcriptomic data, emphasizing intercellular variability over intracellular stochasticity.
• Signal Propagation and Integration: Models signal propagation by repeatedly applying binary signals across network topology to capture pathway activity patterns in cross-sectional data.
• Genetic Algorithm for Logic Rule Refinement: Uses a genetic algorithm to infer initial logic rules and refines them via local search to improve rule accuracy.
• Pathway Activity Scoring: Assesses the impact of individual nodes to score the probability of pathway modulation and rank pathways based on modulation of key regulatory genes.

Scientific Applications:

• Transcriptomics Data Analysis: Applied to transcriptomic data from translational studies for pathway-level interpretation.
• Pathway Prioritization: Prioritizes relevant pathways with increased sensitivity at lower levels of source node modulation compared with existing methods.
• Disease-Specific Pathway Detection: Detects disease-specific pathway modulations by comparing RNA-sequencing data from diseased and healthy controls.
• Drug Target Identification: Identifies high-impact nodes within pathways, including known drug targets, to inform therapeutic research.

Methodology:

Uses discrete-state (Boolean) network modeling and propagation of binary signals across networks; infers Boolean logic rules with a genetic algorithm and refines them via local search; scores and ranks pathways based on node impact and modulation of key regulatory genes.