Machine learning approaches for drug combination therapies.

Area of Science:

• Computational biology
• Pharmacology
• Bioinformatics

Background:

• Drug combination therapy shows promise for complex diseases like cancer and infectious diseases.
• Current knowledge is limited by adverse effects, toxicity, and cell line heterogeneity, making screening infeasible and expensive.
• Computational approaches, especially machine learning, are needed to overcome drug resistance and improve therapeutic efficacy.

Purpose of the Study:

• To review and categorize state-of-the-art machine learning approaches for analyzing personalized drug combination therapies.
• To discuss relevant databases and data analysis portals for drug combination studies.
• To highlight the importance of data integration and multi-source data analysis in identifying effective drug combinations.

Main Methods:

• Categorization of machine learning methods into three groups for personalized drug combination analysis.
• Description of relevant benchmark databases and publicly available data analysis portals.
• Experimental comparison demonstrating the advantages of combining multiple data sources.

Main Results:

• Machine learning approaches provide an effective strategy to overcome drug resistance and improve therapeutic efficacy.
• Data integration is crucial for the identification of effective drug combinations.
• Combining multiple data sources enhances drug combination analysis.

Conclusions:

• Machine learning offers a viable computational strategy to advance personalized drug combination therapies.
• Leveraging diverse datasets and integrated analysis is key to unlocking the full potential of drug combinations.
• This review provides a framework for understanding and applying machine learning in this critical area of research.