PubChem Transformations: Chemical Reaction and Metabolite Data

Transformations in PubChem is an archive of chemical transformation data from the PubChem database, intended for integration into workflows like patRoon. The dataset contains over 11,000 unique transformation entries, including reactions and compounds from sources like NORMAN-SLE and ChEMBL. It provides calculated properties such as XlogP and mass differences to support cheminformatics and environmental fate studies.

Use Cases

• Predicting metabolite structures based on known chemical transformation rules.
• Training machine learning models for reaction outcome prediction using SMILES and InChI keys.
• Studying environmental fate of chemicals using transformations from the NORMAN-SLE list.
• Integrating transformation data into non-target screening workflows for patRoon.
• Analyzing property changes (e.g., XlogP, mass) between chemical predecessors and successors.

Strengths

• Contains 11,260 unique transformation entries, indicating a large, curated collection.
• Integrates data from authoritative sources including PubChem, NORMAN-SLE (8,426 unique compounds), and ChEMBL (1,418 unique compounds).
• Includes calculated chemical descriptors like XlogP differences (range: -12.5 to 10) and mass differences.

Limitations

• Column names and exact data schema are not provided in the source descriptions, limiting understanding of the structure.
• Multiple platform entries have incomplete metadata (e.g., missing row counts, size, specific update dates).
• The dataset's temporal coverage and the recency of the underlying PubChem data are not specified.

Provenance

Source

PubChem 'Transformations' section, with contributions from the NORMAN-SLE list, ChEMBL, and the ECI (Environmental Cheminformatics) team.

Collection Method

Archived and processed from PubChem, with RInChI calculations and integration code written in R.

Freshness

The most specific date mentioned is a stats update from 21 March 2026, and a version (v0.2.0) was released in March 2025.