Recent progresses in the exploration of machine learning methods as in-silico ADME prediction tools

Area of Science:

• Computational Chemistry
• Pharmacology
• Drug Discovery

Background:

• In-silico methods are crucial for early-stage drug discovery, particularly for assessing ADME properties.
• Machine learning (ML) excels at predicting complex ADME mechanisms and classifying diverse chemical structures.

Purpose of the Study:

• To review recent progress in applying ML to predict ADME and ADME regulatory properties.
• To discuss the performance, prospects, and challenges of ML-based ADME prediction tools.

Main Methods:

• Review of current literature on ML applications in ADME prediction.
• Analysis of advancements in ML model scope, training data, molecular features, and consensus modeling.
• Examination of emerging online ML ADME prediction servers.

Main Results:

• ML methods show significant promise for predicting ADME and regulatory properties.
• Improvements in ML models focus on broader ADME coverage and enhanced predictive accuracy.
• Several user-friendly online ML ADME prediction tools are now available.

Conclusions:

• ML is a valuable tool for predicting ADME properties in drug discovery.
• Continued research is needed to address challenges and expand the application of ML in ADME assessment.
• The development of diverse training data and robust models is key for future success.