metadata
license: unknown
task_categories:
- tabular-classification
- graph-ml
- text-classification
tags:
- chemistry
- biology
- medical
pretty_name: MoleculeNet BBBP
size_categories:
- 1K<n<10K
MoleculeNet BBBP
BBBP (Blood-Brain Barrier Penetration) dataset [1], part of MoleculeNet [2] benchmark. It is intended to be used through scikit-fingerprints library.
The task is to predict blood-brain barrier penetration (barrier permeability) of small drug-like molecules.
| Characteristic | Description |
|---|---|
| Tasks | 1 |
| Task type | classification |
| Total samples | 2039 |
| Recommended split | scaffold |
| Recommended metric | AUROC |
References
[1] Ines Filipa Martins et al. "A Bayesian Approach to in Silico Blood-Brain Barrier Penetration Modeling" J. Chem. Inf. Model. 2012, 52, 6, 1686–1697 https://pubs.acs.org/doi/10.1021/ci300124c
[2] Wu, Zhenqin, et al. "MoleculeNet: a benchmark for molecular machine learning." Chemical Science 9.2 (2018): 513-530 https://pubs.rsc.org/en/content/articlelanding/2018/sc/c7sc02664a