Representation of the External Exposome in Ontologies: A Scoping Review (Preprint)

Abstract

BACKGROUND The external exposome, or the totality of all external environmental factors that affect our health, remains an area of biomedicine and epidemiology that is poorly understood, especially in its relationship with various health outcomes. Research on the relationship between the external exposome and health typically relies on large, heterogeneous datasets from multiple sources. This has led researchers within the exposome community to call for improvements in data sharing and integration, which could help reduce the amount of time and effort needed to conduct exposome studies. One resource that has been identified as being well-suited for addressing this issue are biomedical ontologies, which provide a human-readable, computable semantics that can be used to create standardized, structured, linked datasets. However, the domain that the external exposome encompasses is far-reaching and includes things as varied as level of income to exposure to an arid climate. Thus, it is likely that many ontologies exist that represent some aspect of the external exposome without representing all of it comprehensively. OBJECTIVE The aim of this scoping review was to identify and evaluate all publicly available ontologies that represent some aspect of the external exposome and those that aim to represent the external exposome or external environmental exposures specifically. METHODS We conducted a scoping review of ontologies that represent some aspect of the external exposome using online literature databases and Web services for viewing and downloading ontologies and terminologies in May and August 2021, respectively. Ontologies were included if they were publicly-accessible for download, used English annotations, are currently maintained, and contain at least one class that represents some aspect of the exposome. We evaluated each identified ontology for quality with respect to usability and reusability and highlighted those that were best fit for reuse using 10 defined criteria. This review followed the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) checklist. RESULTS A total of 134 ontologies were identified from the literature and ontology Web service searches. Twenty-one (15.67%) ontologies had a domain that falls within some part of the exposome, while four ontologies aimed to represent environmental exposures, broadly. As we expected, each of these ontologies have gaps in their class coverage and varying degrees of quality-related issues. We further identified 23 ontologies that we determined to be best fit for reuse in other ontologies and have made available all of the quality evaluation data online. CONCLUSIONS As predicted, no ontology comprehensively covers the external exposome, although many aim to represent it broadly or happen to represent some aspect of it. More work is needed to make up for gaps in representation and to represent exposures related to access to resources or events that indirectly affect one’s health.