Summarizing and visualizing structural changes during the evolution of biomedical ontologies using a Diff Abstraction Network

Abstract

Biomedical ontologies are a critical component in biomedical research and practice. As an ontology evolves, its structure and content change in response to additions, deletions and updates. When editing a biomedical ontology, small local updates may affect large portions of the ontology, leading to unintended and potentially erroneous changes. Such unwanted side effects often go unnoticed since biomedical ontologies are large and complex knowledge structures. Abstraction networks, which provide compact summaries of an ontology’s content and structure, have been used to uncover structural irregularities, inconsistencies and errors in ontologies. In this paper, we introduce Diff Abstraction Networks (“Diff AbNs”), compact networks that summarize and visualize global structural changes due to ontology editing operations that result in a new ontology release. A Diff AbN can be used to support curators in identifying unintended and unwanted ontology changes. The derivation of two Diff AbNs, the Diff Area Taxonomy and the Diff Partial-area Taxonomy, is explained and Diff Partial-area Taxonomies are derived and analyzed for the Ontology of Clinical Research, Sleep Domain Ontology, and Eagle-I Research Resource Ontology. Diff Taxonomy usage for identifying unintended erroneous consequences of quality assurance and ontology merging are demonstrated.

A unified software framework for deriving, visualizing, and exploring abstraction networks for ontologies

“…Introduced by Ochs et al [5] as a standalone process, difference (diff) partial-area taxonomies summarize the structural differences between two ontology releases. This module will support the derivation of diff partial-area taxonomies as part of the OAF.…”

“…Even with well-established ontology editing tools such as Protégé, the size and complexity of many ontologies makes their maintenance difficult. In previous work, we introduced different kinds of abstraction networks [4], compact summaries of structure and content of an ontology, to support ontology maintenance [5], evolution tracking [6,7], and quality assurance (QA) [8–10], among other use cases. These abstraction networks are smaller and easier to comprehend than the original ontologies (for an example, see Fig.…”

“…This is a worthwhile endeavor, as we have successfully used partial-area taxonomy abstraction networks (described in detail in Section 2.1) to support ontology quality assurance [8–10,18,23], evolution tracking [5–7], and other use cases. Thus, the OAF is expected to support unified abstraction-network-based quality assurance methodologies, where the same quality assurance techniques will be applicable to all or most of the ontologies in the same family.…”

Emerging semantics to link phenotype and environment

“…A "top-down" approach to ontology development, in which classes that constitute the top levels of a new ontology come from an existing domain or upper-level ontology (e.g., CARO, UBERON, PO, BFO-Grenon & Smith, 2004;Haendel et al, 2008;Mungall et al, 2012;Cooper et al, 2013), can result in a shared structure and homogenized development across ontologies, although more specific classes will still require alignment. Aligning ontologies manually is a large task and it is difficult to know the full consequences of an alignment without testing (Ochs et al, 2015). The ability to support the provenance of alignments and re-alignments can translate into trust and continued investment.…”

“…is a co-evolution between natural language and ontologies, which can complicate the recording of provenance and reduce backwards-compatibility (Seppälä, Smith & Ceusters, 2014;Ochs et al, 2015). Thus, as it stands, a researcher wishing to perform a meta-analysis has to manually integrate data sets, which often requires discussions with data providers to clarify meaning.…”