Implementing Nestedness in Darwin Core: An epidemiology case study

• Published in: Biodiversity Information Science and Standards 2022-09, Vol.6
• Main Authors: Jaroszynska, Francesca, Body, Guillaume, Pamerlon, Sophie, Archambeau, Anne-Sophie
• Format: Article
• Language: English
• Subjects: Animal diseases; Biodiversity; Climate change; Consortia; Data models; Disease; Employment; Epidemiology; Geographical distribution; Health surveillance; Information sharing; Land use; Livestock; Pathogens; Public health; Risk assessment; Species; Transnationalism; Wildlife; Zoonoses
• ISSN: 2535-0897; 2535-0897
• DOI: 10.3897/biss.6.94355

Wildlife diseases have an impact on biodiversity, the economy, and public health. Better knowledge of disease patterns would be to the benefit of conservation measures, livestock production and, thus, ultimately human health. Nevertheless, disease surveillance systems operate mostly at a national scale, with incompatible data structures, inhibiting effective data sharing and thus rapid transnational responses to disease outbreak. As the risk of disease to biodiversity, the economy and to public health increases with climate change, land-use change and trade, the necessity for a common data standard to improve data sharing of surveillance efforts is greater than ever to enable transnational proactive and reactive measures to be taken. To address these large issues, a consortium for the European Food and Safety Authority (EFSA; the Enetwild consortium) was formed to collect existing data on wildlife host abundance and distribution in Europe. Alongside data on their associated pathogens, the consortium is attempting to develop data models to aggregate the host data according to the Darwin Core*1 standard. However, the complexity of zoonotic disease and wildlife distribution data is not easily captured in the current version of the Darwin Core standard, often comprising complex data structures with species interactions and partial information. Firstly, zoonotic disease data consist of observations of both the host and the pathogen, whereby each positive case of a disease is associated with the observation of the host species. Secondly, wildlife host distribution data frequently contain detailed information for only some individuals of a group when multiple individuals of the host species are observed simultaneously, such as life stage and sex. In order to capture these types of interactions and subsetting, we need to be able to handle the hierarchical structuring of these data. In an attempt to resolve these issues, a data model in line with the Darwin Core standard was initially developed for wildlife host population data (Enetwild Consortium et al. 2020). Here, we propose a new data model for data on the hosts’ pathogens, and use both data models to demonstrate the model efficacy for complex hierarchical data structures. The epidemiology data model is structured around the existence of a primary occurrence of the host species observation, which may consist of one or many individuals. In order to associate the presence (or absence) of the pathogen to the host