Obol: integrating language and meaning in bio-ontologies

Area of Science:

• Bioinformatics
• Computational Biology
• Ontology Engineering

Background:

• Ontologies formalize biological knowledge but often lack computable definitions.
• This limits the utility of automated reasoning tools for managing complex ontologies like the Gene Ontology (GO).
• Existing OBO ontologies implicitly encode definitions within natural language term names.

Purpose of the Study:

• To develop a method for deriving computable definitions from OBO term names.
• To enable automated reasoning over biological ontologies.
• To improve ontology maintenance and identify inconsistencies.

Main Methods:

• Developed the Obol language to parse natural language term names in biological ontologies.
• Applied Obol to extract computable definitions from the Gene Ontology (GO).
• Utilized automated reasoners to analyze the derived definitions and identify relationships.

Main Results:

• Successfully derived unique, non-trivial computable definitions for approximately 50% of GO terms.
• Identified and facilitated the addition of 223 missing relationships within the GO.
• Demonstrated Obol's utility as an ontology maintenance and definition generation tool.

Conclusions:

• Obol provides a formal language to extract computable definitions from biological ontologies.
• Automated reasoning over these definitions enhances ontology quality and consistency.
• The Obol approach offers a scalable solution for ontology maintenance and knowledge discovery.