RMechDB: A Public Database of Elementary Radical Reaction Steps

Area of Science:

• Computational Chemistry
• Chemical Kinetics
• Data Science

Background:

• Accurate modeling of radical reactions is crucial for chemical research and prediction.
• Existing data on elementary radical reaction steps are fragmented and lack standardization.
• Computational tools require reliable, curated datasets for effective radical reaction modeling.

Purpose of the Study:

• To introduce RMechDB, an open-access platform for aggregating, curating, and distributing elementary radical reaction step data.
• To provide a centralized, structured resource for computational radical reaction modeling and prediction.
• To facilitate community contributions for expanding the dataset of radical reaction mechanisms.

Main Methods:

• Developed an open-access platform (RMechDB) with a database schema and interactive search interface.
• Manually curated over 5,300 elementary radical reaction steps from textbooks, publications, and the Master Chemical Mechanism (MCM) website.
• Encoded reactions in SMIRKS format with atom mapping and arrow-pushing mechanism annotations.

Main Results:

• RMechDB contains over 5,300 manually curated elementary radical reaction steps.
• Data includes reactions with single transition states at room temperature, sourced from diverse origins.
• Reactions are standardized using SMIRKS format with detailed mechanistic annotations.

Conclusions:

• RMechDB provides a unified, structured representation of radical elementary steps, crucial for computational modeling.
• The platform's open-access nature and community contribution model foster data-driven advancements in chemical prediction.
• This resource is expected to enhance the development of predictive models in chemistry and chemoinformatics.