MappedCRD: Atom-Mapped Chemical Reactions Database

MappedCRD is a processed version of the Chemical Reactions Database (CRD) from Zenodo. It provides atom-mapped reaction SMILES and imaginary transition state (ITS) graph encodings for large-scale cheminformatics and machine-learning use. The dataset includes at least one subset with over 1.4 million examples and was last updated on April 25, 2026.

Use Cases

• Training machine learning models for reaction outcome prediction based on atom-mapped SMILES.
• Developing graph neural networks for chemical property prediction using ITS graph encodings.
• Benchmarking cheminformatics algorithms on a large-scale, curated collection of chemical reactions.

Strengths

• Contains over 1.4 million examples in the 'rxnmapper' subset.
• Provides processed, atom-mapped reaction data derived from a curated source (CRD).

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the full dataset is unknown, which may limit suitability assessment.

Provenance

Source

Chemical Reactions Database (CRD) from Zenodo.

Collection Method

Processed version providing atom-mapped SMILES and ITS graph encodings.

Freshness

Last updated 2026-04-25 00:16:00.