We use a slimmed version of the Experimental Factor Ontology (EFO) to describe diseases available in the Open Targets Platform. This modified ontology consists of a subset of EFO terms in addition to terms from other ontologies (e.g. Human Phenotype Ontology - HPO) for which an EFO term is not available.
We map diseases to our Open Targets Platform ontology using OnToma to integrate evidence from different sources and describe relationships between diseases.
The reasons for choosing EFO as our core ontology are listed below:
It is an application ontology, therefore it can incorporate parts (or the whole) of other ontologies, such as the rare disease ontology ORDO
It is able to bring in additional (not disease only) ontologies
It can easily accommodate our requests for new ontology terms in addition to updates and changes, as Open Targets and the EFO team collaborate closely
It is the ontology of controlled vocabulary used by the GWAS catalog and the Expression Atlas. Therefore, no extra step was needed when incorporating those two data sources in our pipeline