Machine and deep learning approaches for cancer drug repurposing

Area of Science:

• Oncology
• Computational Biology
• Pharmacology

Background:

• Recent exponential growth in understanding cancer initiation, progression, and metastasis.
• Limitations in current anti-cancer therapeutics necessitate novel strategies.
• High cost and time associated with developing new anti-cancer drugs drive interest in drug repurposing.

Purpose of the Study:

• To review computational and machine learning methods for identifying drug repurposing opportunities in oncology.
• To discuss the application of these methods in understanding cancer biology and immunomodulation.
• To highlight strategies for overcoming challenges in oncologic drug repurposing.

Main Methods:

• Utilizing advanced 'omics' data (genomics, proteomics, etc.).
• Applying machine learning and artificial intelligence (deep learning) for biological process modeling.
• In-silico computational strategies for target and drug-phenotype association discovery.

Main Results:

• Machine and deep learning methods have significantly advanced the identification of novel drug targets and associations.
• Computational approaches aid in modeling complex cancer biology and identifying pharmacologically relevant pathways.
• These methods facilitate the discovery of new uses for existing and investigational drugs.

Conclusions:

• Computational strategies, particularly AI and deep learning, are crucial for accelerating drug repurposing in oncology.
• Integrating 'omics' data with advanced computational models can uncover new therapeutic avenues.
• Drug repurposing holds significant promise for cost-effective and efficient expansion of the anti-cancer therapeutic armamentarium.