drug-db-chembl

Retrieve ChEMBL bioactivity data for specific targets by querying ChEMBL web services with filters for potency type, assay type, standardized nM units, and equality relations. You can generate an EGFR IC50 binding dataset with pChEMBL filtering to ensure potency labels are suitable for QSAR and ML training.

pChEMBL is a standardized potency metric computed from nM equality measurements during drug discovery dataset curation. You need it when assembling uniform potency-labeled datasets for QSAR or ML, as it normalizes diverse activity measurements into a single comparable scale for model training.

Search ChEMBL molecules by SMILES using similarity or substructure matching to retrieve molecule records. The Skill queries ChEMBL web services to support chemical lookup by SMILES, ChEMBL ID, or InChIKey for drug discovery dataset building.

Find candidate ChEMBL targets by querying ChEMBL web services with a gene or protein name, or resolve targets directly by UniProt accession. This retrieves curated target records needed to anchor bioactivity data retrieval for drug discovery datasets.

The dataset curation approach exports standardized bioactivity tables in both JSON and CSV formats. This enables quick inspection and integration into downstream QSAR or ML pipelines for drug discovery applications.