3D Activity Cliffs and 2D Analogs from PDB and ChEMBL

236,292,595 three-dimensional activity cliffs were systematically identified from X-ray structures in the Protein Data Bank. Norbert Furtmann at the University of Bonn compiled these cliffs based on Ki, IC50, and combined Ki/IC50 measurements. The dataset also includes 2D structural analogs of the cliff compounds sourced from the ChEMBL database release 19.

Use Cases

• Identifying activity cliffs based on 3D structural similarity and potency data from PDB.
• Comparing 2D and 3D molecular representations for cliff compounds sourced from ChEMBL.
• Training models to predict drastic activity changes from small structural modifications using matched molecular pairs.

Strengths

• Provides 236,292,595 identified 3D activity cliffs, a large-scale systematic collection.
• Integrates data from two authoritative sources: the Protein Data Bank and ChEMBL release 19.
• Offers both 3D structural cliffs and their 2D structural analogs for comparative analysis.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the 2D analog subset is unknown, which may limit suitability assessment.
• Last update date is unknown; freshness unverified.

Provenance

Source

Protein Data Bank (PDB) and ChEMBL database (release 19).

Collection Method

Systematic determination based on X-ray structures and matched molecular pair analysis.