---
dataset_info:
  features:
    - name: compound_id
      dtype: string
    - name: smiles_input
      dtype: string
    - name: smiles_standardized
      dtype: string
    - name: source
      dtype: string
    - name: provenance
      dtype: string
    - name: passed_mpo_filter
      dtype: bool
    - name: mw
      dtype: float64
    - name: logp
      dtype: float64
    - name: tpsa
      dtype: float64
    - name: hbd
      dtype: float64
    - name: hba
      dtype: float64
    - name: heavy_atoms
      dtype: float64
    - name: fsp3
      dtype: float64
    - name: pka_representative
      dtype: float64
    - name: cns_mpo_score
      dtype: float64
    - name: efflux_MDR1_HUMAN
      dtype: float64
    - name: efflux_A4D1D2_HUMAN
      dtype: float64
    - name: efflux_ABCG2_HUMAN
      dtype: float64
    - name: efflux_MRP1_HUMAN
      dtype: float64
    - name: efflux_MRP2_HUMAN
      dtype: float64
    - name: efflux_MRP4_HUMAN
      dtype: float64
    - name: efflux_S47A1_HUMAN
      dtype: float64
    - name: efflux_S22A8_HUMAN
      dtype: float64
    - name: p_bbb
      dtype: float64
    - name: heuristic_reason
      dtype: string
    - name: heuristic_veto
      dtype: string
    - name: score_total
      dtype: float64
  splits:
    - name: train
      num_examples: 1023158502
  dataset_size: 140000000000
license: cc-by-nc-4.0
task_categories:
  - tabular-classification
tags:
  - chemistry
  - drug-discovery
  - blood-brain-barrier
  - bbb
  - cns
  - molecular-properties
  - screening
pretty_name: "BBB-Nuke: 1B Compound BBB Permeability Screen"
size_categories:
  - 1B<n<10B
---

# BBB-Nuke: 1 Billion Compound Blood-Brain Barrier Permeability Screen

## Overview

1.02 billion small molecules screened for blood-brain barrier (BBB) permeability using the [BBB-Nuke](https://github.com/ATTN-Lab) pipeline (v0.12.0).

## Screening Pipeline

Each compound was scored through:
1. **Standardization** - SMILES canonicalization via RDKit
2. **Physicochemical properties** - MW, LogP, TPSA, HBD, HBA, Fsp3, heavy atom count
3. **pKa prediction** - Representative pKa via MolGpKa (batched GCN inference)
4. **CNS-MPO scoring** - Multi-parameter optimization (6 properties)
5. **Efflux transporter prediction** - 8 efflux proteins (S1 fingerprint RF classifiers)
6. **Classifier + heuristic** - Final P_BBB probability score (5-fold CV AUROC 0.933)

## Data Sources

| Source | Parquet Dir | Compounds |
|--------|-------------|-----------|
| Enamine REAL | data/enamine_real/ | ~199M |
| Existing (ZINC/ChEMBL) | data/existing/ | ~666M |
| PubChem | data/pubchem/ | ~59M |
| CPU Screen (ZINC) | data/existing_cpu/ | ~99M |
| **Total** | | **1,023,158,502** |

## Key Columns

- `smiles_standardized` - Canonical SMILES
- `p_bbb` - BBB permeability probability (0-1)
- `cns_mpo_score` - CNS multi-parameter optimization score
- `mw`, `logp`, `tpsa` - Key physicochemical descriptors
- `pka_representative` - Predicted pKa
- `efflux_*` - Efflux transporter affinity scores (8 proteins)
- `provenance` - Data source (enamine_real, existing, pubchem)
- `score_total` - Composite heuristic score

## Citation

If you use this dataset, please cite the ATTN-Lab BBB-Nuke project.

## License

CC BY-NC 4.0