BioTransformer 3.0-a web server for accurately predicting metabolic transformation products

Area of Science:

• Computational chemistry and cheminformatics
• Drug metabolism and pharmacokinetics
• Environmental chemistry

Background:

• Accurate prediction of small-molecule metabolism is crucial for drug development and environmental risk assessment.
• Previous versions of BioTransformer had limitations in scope and accuracy for predicting metabolic transformation products.
• There is a need for comprehensive in silico tools that can handle both exogenous and endogenous compounds.

Purpose of the Study:

• To introduce BioTransformer 3.0, an enhanced web server for in silico metabolism prediction.
• To improve the accuracy, speed, and comprehensiveness of small-molecule metabolism prediction.
• To extend prediction capabilities to endogenous metabolites and environmental microbiota.

Main Methods:

• Integration of machine learning approaches with a refined rule-based system.
• Inclusion of new databases, constraints, and prediction modules.
• Support for customized sequential transformations and multi-step simulations.

Main Results:

• BioTransformer 3.0 achieves 40-50% higher accuracy compared to previous versions.
• The system demonstrates reduced susceptibility to combinatorial 'explosions'.
• Expanded metabolite coverage and prediction capabilities for both xenobiotics and endogenous compounds.

Conclusions:

• BioTransformer 3.0 provides a powerful and accurate platform for in silico metabolism prediction.
• The enhanced capabilities facilitate a deeper understanding of xenobiotic and endogenous compound biotransformation.
• This tool supports research in drug discovery, environmental science, and toxicology.