enviRule

enviRule extracts and generalizes reaction rules from biotransformation pathways to predict environmental transformation products (TPs) of man-made chemicals.

Key Features:

• Automatic rule extraction: Automatically extracts and generalizes reaction rules from biotransformation pathways.
• Reaction clustering: Clusters biotransformation reactions based on similarities in their reaction fingerprints.
• SMARTS rule representation: Outputs generalized transformation rules in SMARTS (SMiles ARbitrary Target Specification) format.
• Rule optimization: Optimizes the genericity of extracted rules specifically for TP prediction tasks.
• Performance improvement: Models trained with rules generated by enviRule showed a 30% increase in area under the curve (AUC) compared to manually curated rules.
• Automated updating: Provides an automated mechanism to update the rule set with new reactions.

Scientific Applications:

• TP prediction in environmental chemistry: Predicts transformation products of man-made chemicals in environmental contexts.
• Microbially mediated transformation analysis: Supports analysis of microbially mediated biotransformations and their products.
• Biotransformation pathway generalization: Generalizes reaction rules for incorporation of new reactions into predictive frameworks.

Methodology:

enviRule clusters biotransformation reactions using reaction fingerprints, automatically extracts and generalizes transformation rules for each cluster, formats rules in SMARTS, optimizes rule genericity for TP prediction, evaluates performance using AUC comparisons, and supports automated updating of the rule set with new reactions.