metadata
license: cc-by-sa-3.0
language:
- en
tags:
- chemistry
- drug-discovery
- bioactivity
- chembl
- smiles
- molecular-properties
size_categories:
- 1M<n<10M
ChEMBL Full Dataset
A flat, fully-joined export of the ChEMBL database (1,980 rows, 139 columns).
What's included
Each row is one bioactivity measurement (IC50, Ki, EC50, Kd, GI50 etc.) enriched with:
| Source | Key columns |
|---|---|
| Activity | pchembl_value, standard_type/value/units, activity_comment, ligand_efficiency |
| Molecule | canonical_smiles, standard_inchi, all physicochemical properties, max_phase, indication_class, synonyms |
| Assay | description (free text), confidence_score, assay_type, cell/tissue/organism context |
| Target | pref_name, target_type, uniprot_accession, component_description |
| Mechanism | mechanism_of_action, action_type, mechanism_comment, selectivity_comment |
| Drug indications | mesh_headings, efo_terms, max_phase_for_ind |
| Document | doc__title, doc__abstract, doc__doi, doc__pubmed_id |
Usage
from datasets import load_dataset
ds = load_dataset("juppy44/chembl-full")
df = ds['train'].to_pandas()
License
ChEMBL data is provided under CC BY-SA 3.0. Cite: Zdrazil et al., Nucleic Acids Research 2023. DOI: 10.1093/nar/gkad1004