Inferring Gene Ontologies from Molecular Networks

Ontologies have proven very useful for capturing knowledge as a hierarchy of terms and their interrelationships. In biology a major challenge has been to construct ontologies of gene function given incomplete biological knowledge and inconsistencies in how this knowledge is manually curated. Here I will discuss our result that large networks of gene and protein interactions in Saccharomyces cerevisiae can be used to infer an ontology whose coverage and power are equivalent to those of the manually curated Gene Ontology (GO). Our first attempt, the Network eXtracted Ontology (NeXO, www.nexontology.org) contains 4,123 biological terms and 5,766 term-term relations, capturing 58% of known cellular components. New terms in NeXO were supported with quantitative genetic interaction profiling and chemogenomics and several were added as updates to GO. I will also discuss our recent work which has advanced our ability to infer ontologies from molecular networks by factoring the challenge into two steps: 1) combining experimental molecular networks into an integrated, quantitative similarity score network and 2) inferring an ontology in the form of a directed acyclic graph (DAG) from this integrated network. I will discuss the applications of this recent advance, the Clique Extracted Ontologies (CliXO) algorithm, focusing on how researchers can evaluate where a high-throughput 'omics dataset provides new biological information not contained in a reference ontology such as GO. This work enables a shift from using ontologies to evaluate data to using data to construct and evaluate ontologies.