RxnSim: a tool to compare biochemical reactions

Area of Science:

• Computational chemistry
• Cheminformatics
• Bioinformatics

Background:

• Quantitative assessment of chemical reaction similarity is crucial for database searching, reaction classification, and identifying candidate enzymes.
• Existing methods primarily focus on chemical transformation patterns for similarity evaluation.
• A novel approach is needed to incorporate molecular signatures for a more comprehensive similarity assessment.

Purpose of the Study:

• To introduce RxnSim, a novel tool for computing chemical reaction similarity.
• To enable reaction comparison based on molecular signatures of substrates and products, alongside transformation patterns.
• To provide a flexible method allowing weighted similarity computations through moiety masking.

Main Methods:

• RxnSim computes reaction similarity by analyzing the molecular signatures of participating molecules (substrates and products).
• The tool integrates similarity assessment of substrates and products with the reaction's transformation.
• It incorporates a feature for masking user-defined chemical moieties to enable weighted similarity calculations.

Main Results:

• RxnSim provides a quantitative measure of reaction similarity based on molecular signatures.
• The tool facilitates reaction comparison by considering both molecular components and the chemical transformation.
• User-defined moiety masking allows for focused and weighted similarity assessments.

Conclusions:

• RxnSim offers a robust and versatile approach to quantifying chemical reaction similarity.
• The tool enhances capabilities in chemical database searching, reaction classification, and enzyme candidate identification.
• Its implementation in R and availability via CRAN promote accessibility for researchers.