liuganghuggingface/moldata-bace

BACE-1 inhibitor potency and binary activity

BACE-1 inhibitor potency and binary activity. DeepChem/MoleculeNet release.

Part of the Molecular Datasets collection.

Get started

from datasets import load_dataset
dataset = load_dataset("liuganghuggingface/moldata-bace", "default")

Where the data comes from

• Original data: published with the original study
• Original study: Subramanian et al. Computational modeling of beta-secretase 1 (BACE-1) inhibitors using ligand based approaches (2016)
• Redistributed by: MoleculeNet (DeepChem) (dataset page)

Tables

Config	Splits	Rows	Targets
default	train, validation, test	1,513	pIC50 BACE-1 inhibition (pIC50) (regression; -log10(IC50 in mol/L)), Class BACE-1 inhibitor (active/inactive) (classification; binary)

data/ contains rows with readable molecular structures; identical observations are collapsed. observations/ contains all source rows.

Source label columns are text, including values such as <10. SMILES_original retains the input representation; SMILES_canonical provides a canonical representation. structure_key groups molecular components, and structure_status records parsing outcomes.

Data quality

• experiment: Labels are measured in a lab.
• many sources: Compiled from several labs or papers, whose protocols differ. Independent IC50 values for the same compound and target typically differ by a factor of about 5 (Kalliokoski et al. 2013).
• all structures readable: All 1,513 SMILES were read with RDKit.
• 71.2% undefined stereo: 71.2% of molecules have a stereocenter whose configuration is not given, so mirror-image forms cannot be told apart.
• 35 of 56 stereoisomer sets differ: Found 56 groups of molecules that differ only in stereochemistry. In 35, the values are 10-fold apart (log-scale labels), more than one standard deviation apart, or in different classes. That can be a real effect or a mistake in the source; a model that ignores stereochemistry treats these molecules as the same.
• 0.1% PAINS alerts: 0.1% of molecules contain a PAINS substructure, which often looks active in screening assays without real activity.
• 45.6% in rarer class: In a typical task, 45.6% of labeled molecules are in the rarer class.
• 100% measured per task: In a typical task, 100% of molecules have a value. A missing value means not measured, not inactive.
• 1,513 molecules per task: A typical task has 1,513 labeled molecules.
• source split: train / validation / test: The source divides the data into train (203 rows), validation (45 rows), test (1,265 rows). Use this division to compare with published results.

Dataset notes

• pIC50 and Class are two forms of the same endpoint. The Model column holds the source's train, validation and test assignment.

License

Dataset-wide license: unspecified. Original source.

Citation

Cite the original study and data source:

• Subramanian et al. Computational modeling of beta-secretase 1 (BACE-1) inhibitors using ligand based approaches (2016)
• Wu et al. MoleculeNet: a benchmark for molecular machine learning (2018)
• Official DeepChem loader and task definitions