The files in this directory are derived from public databases and are shared under the licence the most restrictive of them requires:

  Creative Commons Attribution-ShareAlike 3.0 (CC BY-SA 3.0), https://creativecommons.org/licenses/by-sa/3.0/

Sources, as released: chembl 37, bindingdb 202609, pdb 2026-09-26-r2.
  ChEMBL (CC BY-SA 3.0), https://www.ebi.ac.uk/chembl/
    Credit: Zdrazil B, Felix E, Hunter F, et al. The ChEMBL Database in 2023: a drug discovery platform spanning multiple bioactivity data types and time periods. Nucleic Acids Res. 2024;52(D1):D1180-D1192. doi:10.1093/nar/gkad1004
  BindingDB (CC BY 3.0; the rows it imported from ChEMBL keep CC BY-SA 3.0), https://www.bindingdb.org/
    Credit: Liu T, Hwang L, Burley SK, et al. BindingDB in 2024: a FAIR knowledgebase of protein-small molecule binding data. Nucleic Acids Res. 2025;53(D1):D1633-D1644. doi:10.1093/nar/gkae1075
  wwPDB (CC0), https://www.wwpdb.org/
    Credit: Burley SK, et al. RCSB Protein Data Bank (RCSB.org): delivery of experimentally-determined PDB structures alongside one million computed structure models of proteins from artificial intelligence/machine learning. Nucleic Acids Res. 2023;51(D1):D488-D508. doi:10.1093/nar/gkac1077
  PubChem BioAssay, release 20260929 (public; NCBI places no restriction on it, and each depositor was checked one by one for a stated restriction), https://pubchem.ncbi.nlm.nih.gov/
    the calls of 36 depositors are kept: Broad Institute; Buhrlage Lab, Dana-Farber Cancer Institute and Novartis Institutes for BioMedical Research (Cambridge, Mass); C. David Weaver Laboratory, Vanderbilt University; CC_PMLSC; Columbia University Molecular Screening Center; Emory University Molecular Libraries Screening Center; Finley and King Labs, Harvard Medical School; Fisher Drug Discovery Resource Center; Gregory J. Crowther; ICCB-Longwood Screening Facility, Harvard Medical School; Innovative Therapeutics Lab; Johns Hopkins Ion Channel Center; MTDP; Meiler Lab, Vanderbilt University; Molecular Libraries Program, Specialized Chemistry Center, University of Kansas; NMMLSC; National Center for Advancing Translational Sciences (NCATS); PCMD; PDSP; RESOLUTE - Research Empowerment on Solute Carriers; SGCOxCompounds; SGCStoCompounds; SRMLSC; Sanford-Burnham Center for Chemical Genomics; Santagata Laboratory - Brigham and Women's Hospital and Harvard Medical School; SciLifeLab; Southern Research Institute; Southern Research Specialized Biocontainment Screening Center; The Scripps Research Institute Molecular Screening Center; Tox21; University of Kansas High Throughput Screening Laboratory; University of Pittsburgh Molecular Library Screening Center; Vanderbilt High Throughput Screening Facility; Vanderbilt Specialized Chemistry Center; Victor Tsetlin Lab, Institute of Bioorganic Chemistry RAS; Wong's Lab, Department of Immunology, School of Medical Sciences, Health Campus, Universiti Sains Malaysia.
    Credit: Kim et al., PubChem 2025 update, Nucleic Acids Res., doi:10.1093/nar/gkae1059.
  COCONUT (CC BY 4.0), https://coconut.naturalproducts.net/
    Credit: Chandrasekhar V, Rajan K, Kanakam SRS, et al. COCONUT 2.0: a comprehensive overhaul and curation of the collection of open natural products database. Nucleic Acids Res. 2025;53(D1):D634-D643. doi:10.1093/nar/gkae1063
  LOTUS (CC0 1.0 (Wikidata); CC BY 4.0 on the Zenodo export), https://lotus.naturalproducts.net/
    Credit: Rutz A, Sorokina M, Galgonek J, et al. The LOTUS initiative for open knowledge management in natural products research. eLife. 2022;11:e70780. doi:10.7554/eLife.70780
  Natural-product-likeness model (BSD, RDKit Contrib (Novartis)), https://github.com/rdkit/rdkit/tree/master/Contrib/NP_Score
    Credit: Ertl P, Roggo S, Schuffenhauer A. Natural product-likeness score and its application for prioritization of compound libraries. J Chem Inf Model. 2008;48(1):68-74. doi:10.1021/ci700286x

What was done to them: each molecule is reduced to its Morgan fingerprint (radius 2, 2048 bits) and a hash of its flat InChIKey, and linked to the targets it was measured on or seen bound to,
with the best pActivity of each link (an exact Ki, Kd, IC50 or EC50), the number of measurements of each database and the PDB entries. Targets keep their UniProt names and accessions. Once the ligands are shown
(step 2 of the web version) each molecule also keeps its flat SMILES (no stereochemistry) and its ChEMBL identifier.
PubChem: only the pairs a depositor called active on those targets (381,932 pairs of 188,780 molecules): each molecule keeps its fingerprint and flat SMILES, each pair the kind of assay, the first year, the number of assays, the potency when one was reported and the first two depositors. They are shown as evidence and move no rank.
Natural products: the structures of COCONUT and LOTUS keep their flat SMILES and fingerprint, and what the two say about each: names, classes, organisms and references (no activity). They are shown as what a molecule is and which known natural products are nearest to it, and rank no target.
Built by target-id 0.19.0; release efe2ff8bee79.