PNMRNP: Predicted Carbon-13 NMR Data for 211,280 Natural Products

211,280 natural product compounds with predicted carbon-13 NMR chemical shifts, structures, and classifications. The dataset was created by Jean‐Marc Nuzillard of the Centre National de la Recherche Scientifique, starting from the UNPD database and processed using RDKit and nmrshiftdb2. A three-level classification system based on substructure presence is included.

Use Cases

• Train machine learning models for NMR chemical shift prediction based on molecular structure data.
• Classify natural products into hierarchical categories based on the described substructure classification system.
• Validate computational chemistry tools using the provided 2D structures with configuration data.
• Link compound identifiers to external databases like PubChem using the included synonyms and keys.

Strengths

• Contains data for 211,280 natural product compounds.
• Provides a three-level hierarchical classification (e.g., terpene, sesquiterpene, eudesmane) for compounds.
• Integrates data from multiple sources, including UNPD, PubChem, and nmrshiftdb2.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.
• Row count is confirmed, but specific file formats and data size are unknown.

Provenance

Source

Derived from the UNPD database (Gu J et al., PLOS ONE 2013) and processed with RDKit and nmrshiftdb2.

Collection Method

Structures generated from InChI strings using RDKit; NMR shifts predicted via nmrshiftdb2; PubChem identifiers used for naming.