An Ontology of Functions and Processes Derived from Molecular Networks

Ontologies have been very useful for capturing knowledge as a hierarchy of concepts and their interrelationships. In biology, a prime challenge has been to develop ontologies of gene function given only partial biological knowledge and inconsistency in how this knowledge is curated by experts. I will discuss how large networks of gene and protein interaction, as are being mapped systematically for many species, can be transformed to assemble an ontology with equivalent coverage and power to the manually-curated Gene Ontology (GO). Our network-extracted ontology contains 4,123 biological concepts and 5,766 relations, capturing the majority of known cellular components as well as many additional concepts, triggering subsequent updates to GO. Using genetic interaction profiling we provide further support for novel concepts related to protein trafficking, including a link between Nnf2 and YEL043W. This work enables a shift from using ontologies to evaluate data to using data to construct and evaluate ontologies.