# 3BPA Flexible Drug-Like Molecule: Mixed-Temperature MD Training Set for ACE Force Fields (Kovacs et al. 2021), ColabFit Conversion

## Facts

- Provider: David Peter Kovacs, Cas van der Oord, Jiri Kucera, Alice Allen, Daniel Cole, Christoph Ortner, Gabor Csanyi; ColabFit conversion
- License: MIT (https://github.com/davkovacs/BOTNet-datasets/blob/main/LICENSE)
- Size: 4 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/colabfit/3BPA_train_mixed
- Access: public, readable without an account
- Category: Chemistry & Materials
- Source: https://doi.org/10.1021/acs.jctc.1c00647
- Page: https://scigantic.com/archive/e54b0a00-edda-5b5d-87f6-bd285fb7408d
- Keywords: Chemistry & Materials, molecular dynamics, DFT, interatomic potentials, Parquet, Hugging Face

## About

Training configurations of 3BPA, 3-(benzyloxy)pyridin-2-amine, a flexible drug-like organic molecule, taken from molecular dynamics at 300 K, 600 K and 1200 K and labelled with DFT energies and forces. The authors used 3BPA to show how linear atomic cluster expansion (ACE) force fields predict the potential energy surface of an organic molecule, with test sets at higher temperatures and along dihedral scans to probe extrapolation. This repository is the 500-configuration mixed-temperature training set; ColabFit's sibling repositories hold the 300 K training set, the 300 K, 600 K and 1200 K test sets, the beta-dihedral scans and the isolated atoms. ColabFit's card counts 500 configurations and 13,500 atoms, labelled with energy and atomic forces; about 2 MB. Files are the ColabFit Exchange conversion: Parquet files under co/ hold one row per configuration (atomic numbers, positions, cell, computed properties and metadata), ds.parquet holds dataset-level information, and cs/ groups configurations into named sets where defined; read them with pyarrow, pandas, polars or Hugging Face datasets. The authors' BOTNet-datasets repository (dataset_3BPA) is released under the MIT License; cite Kovacs et al., Journal of Chemical Theory and Computation 17, 7696 (2021).

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("colabfit/3BPA_train_mixed")
print(ds)
```

## Data structure

File types in sampled files: Parquet

Example file: `co/co_0.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/e54b0a00-edda-5b5d-87f6-bd285fb7408d
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/e54b0a00-edda-5b5d-87f6-bd285fb7408d
- Over MCP: `get_schema_card` with id `e54b0a00-edda-5b5d-87f6-bd285fb7408d` at https://api.scigantic.com/mcp

## Related topics

- [Parquet datasets](https://scigantic.com/topics/parquet)
- [Chemistry & Materials datasets](https://scigantic.com/topics/chemistry-and-materials)
- [DFT datasets](https://scigantic.com/topics/dft)
- [Interatomic potentials datasets](https://scigantic.com/topics/interatomic-potentials)
- [Molecular dynamics datasets](https://scigantic.com/topics/molecular-dynamics)
