Binding Pathways to Break Addiction
Date:
Invited Talk at National Institute on Drug Abuse: National Institute of Health, Washington D.C., United States
This talk showcased cutting-edge advances in combining multiplex network biology with deep learning–driven structural modeling to uncover mechanistic pathways and potential therapeutic targets in addiction.
Topics covered: - Reverse engineering predictive expression networks (PENs) with iRF-LOOPy to model brain-specific gene regulation from single-cell and bulk RNA-seq data. - Construction of a 25-layer prefrontal cortex multiplex network integrating HumanNet, transcription factor networks, and scPENs for smoking and opioid use disorder. - Application of the MENTOR framework (Multiplex Embedding of Networks for Team-Based Omics Research) to cluster addiction-associated genes into mechanistic clades using Random Walk with Restart. - Evidence that glucagon-like peptide-1 receptor agonists (GLP-1RAs) show promise in curbing addictive behaviors, with mechanistic insights from Boltz, a deep learning model for protein–ligand structural interaction prediction. - Identification of high-confidence interactions between GLP-1RAs and addiction-linked proteins such as PPP6C. - Future directions leveraging off-target network analysis, RWR connectivity measures, and broader gene set expansion to identify therapeutic mechanisms across addictions.
This presentation demonstrated how AI-enhanced systems biology—spanning exascale network modeling, predictive multiomic integration, and structural deep learning—can guide discovery of shared mechanisms and drug repurposing strategies for complex, comorbid addictions.