ChemSchematicResolver
ChemSchematicResolver decodes two-dimensional chemical schematic diagrams in scientific documents into machine-readable representations to extract carbon-based compound structures and resolve R-group substituents for integration into relational databases.
Key Features:
- Diagram Identification: Employs a proprietary algorithm to detect chemical schematic diagrams within figures of scientific documents for high-throughput processing.
- R-Group Resolution: Resolves R-group substituents and supports a variety of general R-group structures present in diagrams.
- Machine-Readable Conversion: Converts identified and resolved diagrams into a machine-readable format for computational analysis.
- Contextual Integration: Integrates diagram data with contextual document information to support the creation of relational databases.
- High Precision and Reliability: Has been evaluated with reported precision rates between 83–100% across assessed areas.
Scientific Applications:
- Chemical literature mining: Automates extraction of chemical structure information from scientific publications.
- Organic chemistry diagram interpretation: Extracts structural and labeling details from 2D diagrams of carbon-based compounds in organic chemistry.
- Relational database population: Populates relational chemical databases by combining diagram-derived structures with document context.
Methodology:
Detection of chemical schematic diagrams using a proprietary algorithm. Resolution of R-group substituents within detected diagrams. Conversion of resolved diagrams into machine-readable formats and integration with contextual document data.
Topics
Details
- License:
- MIT
- Tool Type:
- library, web application
- Programming Languages:
- Python
- Added:
- 1/18/2021
- Last Updated:
- 2/11/2021
Operations
Publications
Beard EJ, Cole JM. ChemSchematicResolver: A Toolkit to Decode 2D Chemical Diagrams with Labels and R-Groups into Annotated Chemical Named Entities. Journal of Chemical Information and Modeling. 2020;60(4):2059-2072. doi:10.1021/acs.jcim.0c00042. PMID:32212690.