MappedCRD / README.md
lukaskim's picture
Update README.md
c45ac06 verified
|
Raw
History Blame Contribute Delete
3.36 kB
metadata
license: cc-by-4.0
task_categories:
  - other
language:
  - en
tags:
  - chemistry
  - reactions
  - smiles
  - atom-mapping
  - graphies
  - reaction smiles
  - rxnmapper
  - ITS
pretty_name: MappedCRD
size_categories:
  - 1M<n<10M
configs:
  - config_name: rxnmapper
    data_files: rxnmapper/train-*.parquet
  - config_name: its
    data_files: its/train-*.parquet
  - config_name: its-nomod
    data_files: its-nomod/train-*.parquet

MappedCRD — Mapped Chemical Reactions Database

This dataset is a processed version of the Chemical Reactions Database (CRD) (Zenodo), providing atom-mapped reaction SMILES and imaginary transition state (ITS) graph encodings for large-scale cheminformatics and machine-learning use.

Dataset Description

The CRD is a curated collection of chemical reactions. This dataset distributes two subsets derived from it:

Subset Examples Description
rxnmapper 1,446,478 Full reaction SMILES mapped with RXNMapper
its 1,405,109 Reactions encoded as imaginary transition state (ITS) graphs using graphies and SynKit
its-nomod 1,405,109 ITS graphies without charge and hydrogen modifiers

Subsets

rxnmapper

Atom-mapped reactions produced by RXNMapper. Each row contains:

Column Type Description
reaction string Original reaction SMILES (unmapped)
mapped_reaction string Atom-mapped reaction SMILES
confidence float64 RXNMapper confidence score (0–1)

its

Reactions that were successfully processed into ITS graphs and encoded with graphies. Each row contains:

Column Type Description
original_rsmi string Original reaction SMILES
mapped_rxn string Atom-mapped reaction SMILES
its_graphies string graphies-encoded ITS graph string

The ITS (imaginary transition state) representation encodes both reactant and product bond changes in a single molecular graph, enabling direct graph-based learning on reaction data.

Usage

from datasets import load_dataset

# Atom-mapped reactions
ds_mapped = load_dataset("lukasmki/MappedCRD", name="rxnmapper", split="train")

# ITS graph encodings
ds_its = load_dataset("lukasmki/MappedCRD", name="its", split="train")

# ITS graph encodings without charge/hydrogens
ds_its = load_dataset("lukasmki/MappedCRD", name="its-nomod", split="train")

Source Data

Tools

Citation

If you use this dataset, please cite the original CRD:

@misc{crd,
  doi = {10.5281/ZENODO.18109268},
  url = {https://zenodo.org/doi/10.5281/zenodo.18109268},
  author = {van der Lingen,  Rik},
  language = {en},
  title = {Organic reaction SMILES 1.44M dataset},
  publisher = {Zenodo},
  year = {2025},
  copyright = {Creative Commons Attribution 4.0 International}
}