In silico machine learning methods in drug development.

Area of Science:

• Computational chemistry
• Pharmacology
• Bioinformatics

Background:

• Machine learning (ML) is increasingly used in drug discovery for predicting pharmacological activity and ADMET properties.
• Techniques like ANNs, SVMs, and genetic programming identify protein inhibitors, agonists, and substrates for therapeutic targets.

Purpose of the Study:

• To provide an overview of ML strategies and progress in drug design.
• To discuss the potential of ML model development tools.
• To examine ML algorithm applications across common disease classes.

Main Methods:

• Review of ML techniques including artificial neural networks, support vector machines, and genetic programming.
• Application of ML for predicting compound activity, pharmacodynamics, and ADMET properties.
• Utilizing ML to screen compound libraries and complement QSAR and structure-based methods.

Main Results:

• ML methods demonstrate significant potential in predicting drug candidates.
• ML is valuable for screening diverse chemical structures and handling high-dimensional data.
• ML complements traditional methods when 3D receptor structures are unavailable.

Conclusions:

• ML offers powerful computational tools for modern drug design and evaluation.
• The review highlights the strategic application and progress of ML in identifying potential therapeutics.
• ML algorithms show promise for drug discovery across various disease areas.