# AIMD-Chig Chignolin Ab Initio MD, Random Split Training Set, ColabFit Conversion

## Facts

- Provider: Wang, He, Li, Shao and Liu, Microsoft Research AI for Science; converted by ColabFit Exchange
- License: CC-BY-4.0 (https://api.figshare.com/v2/articles/22786730)
- Size: 5 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/colabfit/Chig-AIMD_random_train
- Access: public, readable without an account
- DOI: https://doi.org/10.60732/fac841ac
- Category: Chemistry & Materials
- Source: https://doi.org/10.1038/s41597-023-02465-9
- Page: https://scigantic.com/archive/2c5860d6-a717-5f84-b81d-854a65ee2a05
- Keywords: Chemistry & Materials, molecular dynamics, DFT, quantum chemistry, interatomic potentials, Parquet, Hugging Face

## About

ColabFit Exchange conversion of the training configurations from the random split of AIMD-Chig, published by Tong Wang, Xinheng He, Mingyu Li, Bin Shao and Tie-Yan Liu of Microsoft Research (Scientific Data 2023). AIMD-Chig explores the conformational space of chignolin, a 10-residue, 166-atom mini-protein, at DFT precision: replica exchange molecular dynamics (REMD) and conventional MD seeded 10,000 starting conformations covering folded, unfolded and metastable states, and ab initio molecular dynamics (AIMD) then sampled 2 million conformations. Per the upstream record, AIMD was driven by M06-2X/6-31G* with a Berendsen thermostat at 340 K, and coordinates, energies and forces are reported. This repo holds only the training portion of the random split; ColabFit lists 1,592,677 configurations (264 million atoms) of C, H, N and O with energy and atomic forces. Measured 2026-11-13: ds.parquet plus a co/ folder of 2 parquet files (12.5 GB); repository storage 26.1 GB. Rows follow the ColabFit configuration schema (positions, atomic_numbers, cell, pbc, energy, atomic_forces, cauchy_stress, software, method and per-row metadata) and open with pandas.read_parquet, pyarrow or datasets.load_dataset; ds.parquet holds the dataset summary. Use it to train protein-scale machine learning force fields such as AI2BMD and ViSNet. Licensed CC BY 4.0; cite Wang et al. 2023 and the ColabFit Exchange.

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("colabfit/Chig-AIMD_random_train")
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/2c5860d6-a717-5f84-b81d-854a65ee2a05
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/2c5860d6-a717-5f84-b81d-854a65ee2a05
- Over MCP: `get_schema_card` with id `2c5860d6-a717-5f84-b81d-854a65ee2a05` 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)
- [Quantum chemistry datasets](https://scigantic.com/topics/quantum-chemistry)
