RDFBones: A framework for the standardisation of osteological data

This is the text version of podium presentation by Felix Engel and Stefan Schlager that Felix Engel delivered at the 2019 annual meeting of the Swiss Society for Anthropology (SGA) on 16 November 2019 in Basel.

Abstract

Standardisation of research data helps to increase their comprehensibility, traceability and reproducibility. It supports aggregation of large databases as the compatibility of data items can be assessed and the reliability of information scrutinised. Standardised data are a prerequisite for the analysis of large bodies of data contributed by multiple researchers, enabling investigations at a level that, otherwise, would not be possible. Recently, the incentive for data standardisation is increasing as funding agencies start to demand measures for research data management as part of proposed research projects. At the same time, funding is provided for the build-up of research data infrastructures. Future developments might bring increased publication of primary data and mechanisms for their citeability.

Several approaches to the standardisation of osteological data have been developed in Biological anthropology. Most of these assume the creation of a (relational) database to be subsequently filled with information. In such scenarios, changes to the database structure are problematic in multiple respects causing traditional data standards to be unchangeable and not providing for variation or extension. On the other hand, the rigidity of data standards might prevent their broad adoption in research in the face of biological anthropology's diversity of research topics and methods.

In other life sciences, researchers have adopted methods of semantic data modelling from information science. These aim at using machines to support not just the processing but also the understanding of data. Knowledge domains are modelled as network graphs, referred to as ontologies. Data items from existing databases can be mapped onto ontology concepts enabling aggregation of disparate datasets. This approach follows recent trends in information technology where conceptual data models are separated from the logics of data storage. Semantic data modelling is routinely applied in various disciplines and has helped to build up large bodies of research data.

RDFBones (Engel & Schlager 2019) is an ontology modelling research data from osteological investigations. It draws on several existing ontologies, including the Ontology for Biomedical Investigations (OBI), the Foundational Model of Anatomy (FMA) and the CIDOC Reference Model (CIDOC CRM). These resources provide most of the concepts needed to convey the meaning and inner coherence of osteological datasets and need only few additions covering specific requirements of research in biological anthropology.

RDFBones is designed to be continuously extended in order to cover a variety of research topics and methods. To this end, the RDFBones core ontology supports custom extensions written by researchers. Extensions are comparatively easy to write and provide a formalised description of scientific concepts and methods.

Advantages of this approach include transparent aggregation of disparate datasets and flexible database models catering to the needs of specific investigations. Additional benefits are access to existing knowledge graphs for data analysis (e. g. by drawing on the FMA to put observations into an anatomical context) and opening up osteological data to a variety of contextual information from other knowledge domains.

REFERENCE Engel, F. & Schlager, S (2019). RDFBones – making research explicit: an extensible digital standard for research data. Anthropologischer Anzeiger, 76(3), 245–257.