SimNMR-PubChem: 106 Million Molecules with Simulated NMR Spectra

SimNMR-PubChem Database provides a collection of 106 million molecules from PubChem, each paired with corresponding simulated NMR spectra. The dataset was created by yqj01 and was last updated on 2025-09-03. It is preprocessed for spectral analysis, molecular structure elucidation, and machine learning applications in chemistry.

Use Cases

• Train machine learning models for NMR spectral prediction based on molecular structure data.
• Perform molecular structure elucidation by comparing simulated spectra with experimental results.
• Benchmark computational chemistry algorithms using a large-scale, preprocessed collection of molecules and spectra.

Strengths

• Contains 106 million molecules, indicating a large scale.
• Each molecule is paired with a simulated NMR spectrum, providing a structured data relationship.

Limitations

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

Provenance

Source

PubChem

Collection Method

Simulated NMR spectra paired with molecules from the PubChem database.

Freshness

Last updated 2025-09-03 05:36:25; freshness should be verified.