RNADynBench

A standardized all-atom RNA molecular dynamics benchmark.

Browse data ↗Search & DownloadQuick start
2,585MD trajectories
1,469Unique RNAs
100 nsPer trajectory
258.5 μsTotal simulation time

Each trajectory spans 100 ns with 1,001 frames at 100-ps intervals; standard evaluation uses 101 frames at 1-ns intervals. File formats & units →

Simulated in explicit water and ions; released coordinates contain RNA only. Equilibration is excluded.

A traceable data foundation.

01 / Curate

Experimental RNA structures from RNA-Solo, with molecular context checked against original PDB/mmCIF records.

02 / Simulate

All-atom MD with Amber RNA OL3 and OPC water, at 300 K, 1 bar and 150 mM NaCl.

03 / Validate

Production completion, coordinate continuity, topology and multi-chain behavior checks, with retained diagnostic records.

RNADynBench dataset characterization from manuscript Figure 1
RNA size coverage and finite-window dynamics. Water RDFs characterize the simulations.

Two benchmark tasks.

Task 01

Trajectory generation

Start from an initial conformer and generate 101 frames over 100 ns.

Protocol & results →

Task 02

Single-conformer prediction

Predict MD-derived residue covariance, RMSF and normalized motion coupling.

Protocol & results →

Paper

First page of the RNADyn arXiv paper

RNADyn: A Benchmark for Generating and Understanding RNA Dynamics

Yiming Huang, Lennart Bastian, Hanqun Cao, Luis Vollmers, Tolga Birdal

arXiv preprint · 2026 · arXiv:2610.03712

BibTeX

@misc{huang2026rnadyn,
  title         = {{RNADyn}: A Benchmark for Generating and Understanding {RNA} Dynamics},
  author        = {Yiming Huang and Lennart Bastian and Hanqun Cao and Luis Vollmers and Tolga Birdal},
  year          = {2026},
  eprint        = {2610.03712},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  url           = {https://arxiv.org/abs/2610.03712}
}