Mechanism-based Applications of Toxicology Ontology

Gordana Apic, Cambridge Cell Networks, UK

Any ontology or vocabulary resource is only useful if it is in a strong correlation with a specific intended application. Mechanism-based application for an ontology is a real challenge, as it requires complex and structured information. It is also a chellenge to deal with different levels of details available for certain effects, in some cases inferred protein-chemical interaction networks are sufficient while in other cases detailed models of how a particular pathway brings about a particular outcome is required.

CCNet’s Toxontology has been developed in order to organize and manage information about toxic effects and to enable to provide as much insight as possible about the mechanisms underlying a toxic effect and to help elucidate underlying biological pathways. This ontology is a set of controlled vocabularies used to describe and interconnect three major categories: effects (toxic and biological effects - including a full set of around 1000 histopathology terms covering all organs); target proteins/genes and chemical compounds to each other. The controlled vocabulary have been derived from about 150 man years of literature curation on mechanistic aspects of toxic effects reported in the scientific literature and regulatory reports and their hierarchical organisation.  

Using such an ontology, chemicals can be linked to a histopathological observation, and at different levels of granularity of the observation (hierarchy). In order to obtain a hypothesis about underlying biological pathways and mechanisms, histopathologies can be linked back to associated chemicals and at the same time linking histopathologies to associated proteins/genes, such as nuclear metabolizing enzyme, nuclear recetor or disregulated gene (proteins/genes in generic terms), enables studying biological pathways and mechanisms.