Modeling Path Importance for Effective Alzheimer's Disease Drug Repurposing

Area of Science:

• Computational biology
• Neuroscience
• Pharmacology

Background:

• Drug repurposing is a cost-effective strategy for Alzheimer's disease (AD) drug discovery.
• Network-based methods leverage complex interactions for identifying drug candidates.
• Existing methods often overlook the varying importance of network paths of the same length.

Purpose of the Study:

• To propose a novel network-based method, MPI (Modeling Path Importance), for AD drug repurposing.
• To address the limitation of equal path importance assumption in current network-based approaches.
• To improve the accuracy and efficiency of identifying potential anti-AD drugs.

Main Methods:

• Developed MPI, a novel network-based method utilizing learned node embeddings to prioritize network paths.
• Employed learned node embeddings to capture rich structural information and differentiate path importance.
• Evaluated MPI against a baseline method relying on shortest paths between drugs and AD.

Main Results:

• MPI demonstrated superior performance in prioritizing anti-AD drug candidates compared to the baseline.
• Among the top-50 ranked drugs, MPI identified 20.0% more drugs with existing anti-AD evidence.
• Cox proportional-hazard models suggested etodolac, nicotine, and BBB-crossing ACE-INHs as potential AD-reducing candidates.

Conclusions:

• MPI offers a more effective approach to AD drug repurposing by modeling path importance.
• The findings highlight the potential of etodolac, nicotine, and BBB-crossing ACE-INHs for further investigation in AD.
• This work advances network-based drug discovery methodologies for neurodegenerative diseases.