The clinical measurement, measurement method and experimental condition ontologies: expansion, improvements and new applications

Abstract

Background: The Clinical Measurement Ontology (CMO), Measurement Method Ontology (MMO), and Experimental Condition Ontology (XCO) were originally developed at the Rat Genome Database (RGD) to standardize quantitative rat phenotype data in order to integrate results from multiple studies into the PhenoMiner database and data mining tool. These ontologies provide the framework for presenting what was measured, how it was measured, and under what conditions it was measured.

Results: There has been a continuing expansion of subdomains in each ontology with a parallel 2-3 fold increase in the total number of terms, substantially increasing the size and improving the scope of the ontologies. The proportion of terms with textual definitions has increased from ~60% to over 80% with greater synchronization of format and content throughout the three ontologies. Representation of definition source Uniform Resource Identifiers (URI) has been standardized, including the removal of all non-URI characters, and systematic versioning of all ontology files has been implemented. The continued expansion and success of these ontologies has facilitated the integration of more than 60,000 records into the RGD PhenoMiner database. In addition, new applications of these ontologies, such as annotation of Quantitative Trait Loci (QTL), have been added at the sites actively using them, including RGD and the Animal QTL Database.

Conclusions: The improvements to these three ontologies have been substantial, and development is ongoing. New terms and expansions to the ontologies continue to be added as a result of active curation efforts at RGD and the Animal QTL database. Use of these vocabularies to standardize data representation for quantitative phenotypes and quantitative trait loci across databases for multiple species has demonstrated their utility for integrating diverse data types from multiple sources. These ontologies are freely available for download and use from the NCBO BioPortal website at http://bioportal.bioontology.org/ontologies/1583 (CMO), http://bioportal.bioontology.org/ontologies/1584 (MMO), and http://bioportal.bioontology.org/ontologies/1585 (XCO), or from the RGD ftp site at ftp://rgd.mcw.edu/pub/ontology/.

Figure 1

The clinical measurement ontology 2012 vs. 2013. A. Additions and improvements to the CMO have resulted in an expansion of both the number of terms and the scope of the ontology. In May of 2012, there were 13 direct child terms under the root “clinical measurement”. As of July 2013, this had increased to 26. The vertical arrows point to the level in the display which corresponds to the vocabulary nodes directly under the root. B. Adjustments to the branch for “body morphological measurement” and addition of a new branch for “organ measurement” clarified the morphological terms and allowed for addition of organ-specific physiological terms.

Figure 2

The measurement method ontology 2012 vs. 2013. Addition of new terms such as “chromatography” necessitated the creation of a “molecular separation method” branch under “ex vivo method”. The term “gel electrophoresis”, as a type of molecular separation method, was moved from being a direct child of “ex vivo method” into the new branch.

Figure 3

Access to strain-specific quantitative phenotype data from RGD strain report pages. All available quantitative phenotype data for a strain is accessible from the RGD strain report page’s phenotype profile. In the section labeled “Phenotype Values via PhenoMiner”, select a CMO term to view values for that strain.

Figure 4

Use of the CMO, MMO, XCO, VT, and RS ontologies to annotate RGD QTLs. The RS Ontology (A) and the CMO, MMO, XCO, and VT Ontologies (B) are used at the Rat Genome Database to standardize the presentation of the rat strains crossed, the specific measurement that was made, the method that was used to make that measurement, the conditions under which the measurement was made, and the specific trait that was measured, respectively. Annotations are assigned an evidence code of “IED” or “inferred from experimental data” to indicate the type of evidence (i.e., experimental) which supports the use of these terms.