metadata
license: cc-by-4.0
A Dataset of Small Protein Conformational Ensembles from All-Atom Molecular Dynamics Simulations
Authors
Zilin Ren, Ying Hu, Xin Yang
Dataset Overview
This dataset, hereafter referred to as DynoDB, consists of 8,385 molecular dynamics (MD) trajectories of small proteins (5–100 amino acids). Each entry contains 1,000 ensemble frames extracted from 100 ns production runs, providing a systematic characterization of small protein conformational landscapes at the atomic scale.
License
This dataset is licensed under the Creative Commons Attribution 4.0 International License (CC BY 4.0). By using this dataset, you agree to cite the original work as specified below.
Citing The Dataset
If you use this dataset in your research, please cite the following paper:
[... (Scientific Data, 2026)]
DOI: https://doi.org/10.57967/HF/7057
Dataset Structure
The data is organized as follows:
static_structures/: Initial PDB conformations.processed_structures/: Preprocessed and repaired models.trajectory_frames/: Extracted frames (1,000 per system).metadata/: Multi-sheet Excel/CSV files providing comprehensive structural, dynamic, and functional annotations.
Dataset Size
| Dataset Metric | Value |
|---|---|
| Number of trajectories | 8,385 |
| Length per trajectory | 100 ns |
| Cumulative simulation time | 838.5 µs (~0.84 s) |
| Total number of atoms | 3,990,516 |
| Average system size (atoms) | ~476 |
| Total number of residues | 491,395 |
| Average system size (residues) | ~59 |
| Storage size | ~611 GB |