Poseidon2LOD (NFDI4Objects Community Meeting 2025)
thor Straten, Mattis; Thiery, Florian; Schiffels, Stephan; Renz, Matthias
- Publisher
- Zenodo
- Language
- en
Abstract
Integrating complex, interdisciplinary datasets remains a central challenge in digital archaeology and cultural heritage research. In particular, archaeogenetic datasets require standardised storage and formats, as well as interoperable semantic integration with archaeological and natural science data. The Poseidon Framework meets this need by providing a community-maintained archive for archaeogenetic human genotype data. It implements the FAIR principles through a structured package format, command-line tools and open submission workflows. The Poseidon Community Archive (PCA) enables the reproducible reuse of data by linking genomic information with archaeological and laboratory metadata, with transparent provenance and validation mechanisms in place to support this. To facilitate integration of Poseidon data with other archaeo-natural datasets and the NFDI4Objects Knowledge Graph (N4O KG), the presented workflow converts the PCA data into interoperable RDF data. This workflow operates on two levels: (i) data-level standards are established by parsing PCA metadata into RDF triples and enriching them with external identifiers; and (ii) schema-level standards are ensured through the Archaeo-Natural Ontology (ArNO). Grounded in CIDOC CRM, but extended with classes from complementary ontologies, ArNO provides a shared semantic foundation for representing archaeo-natural entities and relations. It enables the consistent mapping of heterogeneous data sources and facilitates their integration into the CIDOC-CRM–based N4O KG. The RDF transformation process consists of several steps. First, averaged coordinate information is derived for sites and places from their associated discovery sites. These entities are then linked to external gazetteers, such as GeoNames and Wikidata, using label similarity and spatial matching. Next, bibliographic metadata are harmonised via the CrossRef API to ensure persistent identifiers, such as DOIs and ORCIDs. Finally, the extracted archaeological, genetic and bibliographic information is translated into ArNO-compliant RDF triples. A demo dataset comprising ancient DNA samples illustrates this approach. This work demonstrates how Poseidon data can be converted into Linked Open Data, ensuring interoperability across domains. By considering the ArNO ontology and leveraging external authority services, this paves the way for the integration of the archaeo-natural data in NFDI4Objects Task Area 3 using graph structures, such as Heterogeneous Information Networks or Knowledge Graphs. This enhances the potential for cross-dataset synthesis and reproducible research, and prepares the ground for advanced analytical methods. Ultimately, Poseidon2LOD contributes to a broader FAIR data ecosystem within NFDI4Objects Task Area 3, supporting the interdisciplinary exploration of the human past through semantically enriched, computationally accessible datasets.
Full text
Poseidon2LOD Transforming the Poseidon Community Archive into interoperable RDF data Linked Open Data Workflow The PCA data (.janno & .bib files) is converted into RDF triples to ensure datalevel standards in order to create FAIR Linked Open Data (LOD). The workflow consists of: •Combining all the PCA data •Obtaining spatial representatives for Sites and Places by aggregating the coordinate information of related Discovery Sites •Mapping Countries,Sites and Places to Geonames and Wikidata (exact match on label, similar coordinate information) •Obtaining consistent Literature information, including DOIs and ORCIDs, using the CrossRef API •Creating ArNO-consistent RDF triples A demo of the Poseidon RDF data containing the first 50 of a total of 18,929 aDNA samples together with all related information can be explored at [6] (see Figures 3 and 4). Mattis thor Straten (Kiel University) Florian Thiery (LEIZA) Stephan Schiffels (MPI EVA) Matthias Renz (Kiel University) Archaeo-Natural Ontology In order to integrate standards at the schema level of HINs, the Archaeo-Natural Ontology (ArNO) [3] was created (thor Straten et al., 2025b).It ensures that different archaeonatural data can be mapped to well-defined, reusable node and edge types (see Figure 1). Though ArNO is currently primarily designed to represent the PCA data, the incorporation of standardised reference models and ontologies enables links to be made with other archaeo-natural data repositories, including Arbodat [4, 5]. ArNO is based on CIDOC-CRM (Doerr, 2005) in order to align with the N4O KG. However, CIDOC CRM alone is insufficient to represent archaeo-natural data. Therefore, classes from other ontologies have also been integrated (see Figure 2). Try yourself [6]: Förderkennzeichen: 501836407 Imprint: NFDI4Objects Community Meeting 2025 30.09 - 02.10.2025 Links: •[1]: https://www.poseidon-adna.org/ •[2]: https://github.com/poseidon-framework/community-archive/ •[3]: https://archaeonatural-cloud.github.io/archaeonaturalontology/ •[4]: https://arbodat.info/ •[5]: https://www.pangaea.de/?q=ARBODAT •[6]: https://research-squirrel-engineers.github.io/arno-demo-1/ Bibliography: •Doerr, Martin (2005). The CIDOC CRM, an Ontological Approach to Schema Heterogeneity. https://doi.org/10.4230/DagSemProc.04391.22 •Schmid, Clemens et al. (2024). Poseidon–A framework for archaeogenetic human genotype data management. https://doi.org/10.7554/eLife.98317.1.sa2 •Shi, Chuan et al. (2017). A survey on heterogeneous information network analysis. https://doi.org/10.1109/TKDE.2016.2598561 •thor Straten, Mattis et al. (2025a). Data-Driven Community Standards for Interdisciplinary Heterogeneous Information Networks. https://doi.org/10.5281/zenodo.15040308 •thor Straten, Mattis et al. (2025b). Archaeo-Natural Ontology (ArNO). https://doi.org/10.5281/zenodo.15095412 Figure 1: Representation of the ArNO ontology as used to represent Poseidon data. Coloured ellipses represent ArNO classes, white squares represent external identifiers, white rectangles represent attributes. Poseidon Community Archive Poseidon [1] is an open framework for human genotype data in archaeogenetics that implements the FAIR principles via a package format that is readable by both humans and machines, interoperable command-line tools, and public, community-maintained archives. Packages couple genotypes with rich archaeological and laboratory context, and are versioned, making datasets: •Findable through a web API and browser interface •Accessible as downloadable packages •Interoperable across standard formats •Reusable through transparent provenance and validation. The Poseidon Community Archive (PCA) [2] curates author-submitted, publication-wise and reprocessed datasets. This enables reproducible synthesis and re-analysis across the field. (Schmid et al., 2024.) Finding Standards Heterogeneous Information Networks (HINs) are multi-typed graph models that can represent rich semantic information by modelling semantic relationships between multiple types of entities. Various machine learning techniques can be used to identify relevant connections between similar entities and discover semantic patterns in HINs (e.g., Shi et al., 2017). The motivation behind HINs is to find crossdataset dependencies in the archaeo-natural data of NFDI4Objects Task Area 3 (TA3), i.e., data that combines archaeological and natural aspects. However, mapping the different datasets to a common format that further ensures interoperability with the NFDI4Objects Knowledge Graph (N4O KG) is not straight-forward. Therefore, the current focus is on interoperability of the archaeonatural data in TA3at schema and data levels (thor Straten et al., 2025a) prior to analysis. Figure 2: Ontological foundation of ArNO (thor Straten et al., 2025a). Figure 4: The spatial Discovery Site associated with the aDNA Sample. Figure 3: An aDNA Sample (Ötzi) along with its related information.
