ClassyFire

ClassyFire classifies chemical entities using an ontology-based, rule-driven system to provide systematic structural classification and a computable chemical taxonomy.

Key Features:

• Ontology-Based Framework: Employs an open ontology-driven methodology to categorize and name chemical compounds systematically.
• Rule-Based Classification: Uses predefined logical rules to ensure consistent classification across diverse chemical structures.
• Integration with Cheminformatics: Combines expert knowledge with cheminformatics techniques to enhance accuracy and consistency of classifications.
• PFAS Subdivision and Validation: Implements the splitPFAS approach to subdivide PFASs by a CnF2n+1-X-R pattern and was applied to evaluate classification and naming on 770 PFASs from the OECD list, revealing limitations of open cheminformatics approaches.

Scientific Applications:

• PFAS categorization and nomenclature: Supports subdivision and systematic categorization of per- and polyfluoroalkyl substances (PFASs) to address complex nomenclature and naming issues.
• Large-scale chemical dataset standardization: Provides a standardized terminology framework useful for assimilating and comparing reported information across large molecular datasets.

Methodology:

The methodology uses the splitPFAS approach to subdivide PFASs according to a CnF2n+1-X-R pattern focusing on X = CO, SO2, CH2, and CH2CH2; ClassyFire was tested for categorizing and naming these subdivided structures and the workflow was applied to a dataset of 770 PFASs from the OECD list.