SurGBSA: Molecular Dynamics Simulation Data for Representation Learning

SurGBSA is a dataset developed for learning representations from molecular dynamics simulations, as described in a 2025 arXiv paper. The data was curated by researchers at Lawrence Livermore National Laboratory, indicated by the identifier LLNL-DATA-2016815. The dataset was last updated on the Hugging Face platform in March 2026.

Use Cases

• Training machine learning models for molecular property prediction based on simulation trajectories.
• Developing novel representation learning methods for biomolecular systems.
• Benchmarking computational methods for analyzing molecular dynamics data.
• Studying protein-ligand interactions through learned simulation features.

Strengths

• Associated with a peer-reviewed 2025 arXiv paper detailing its development and purpose.
• Originates from a reputable national laboratory (LLNL), suggesting scientific rigor.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and dataset scale are unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Lawrence Livermore National Laboratory (LLNL).

Collection Method

Data used in the development of the SurGBSA method, likely derived from molecular dynamics simulations.

Freshness

Last updated 2026-03-14 01:34:23; freshness should be verified.