Machine Learning Applied in Small Molecule Drug Discovery: Models, Strategies, and Future Prospects.

Area of Science:

• Pharmacology
• Computational Chemistry
• Artificial Intelligence in Drug Discovery

Background:

• Machine learning (ML) is increasingly vital in small-molecule drug discovery.
• Pharmacologists and drug researchers require a strong grasp of ML principles and applications.
• Despite progress, challenges like poor data quality, feature selection issues, and limited model generalizability hinder ML adoption.

Purpose of the Study:

• To systematically review machine learning algorithms used in drug discovery.
• To categorize ML algorithms by model type.
• To highlight ML tools and techniques addressing specific challenges across the drug discovery pipeline.

Main Methods:

• Systematic literature review of machine learning applications in drug discovery.
• Categorization of ML algorithms based on their underlying models.
• Analysis of ML techniques tailored to overcome common drug discovery challenges.

Main Results:

• Comprehensive survey of diverse ML algorithms relevant to drug discovery.
• Classification of ML methods, aiding researchers in tool selection.
• Identification of ML strategies for improving data quality, feature selection, and model generalizability.

Conclusions:

• Machine learning offers significant potential to advance small-molecule drug discovery.
• Addressing challenges in data, features, and generalizability is key to unlocking ML's full capabilities.
• This review provides a structured overview to guide the effective application of ML in pharmaceutical research.