Bridging Metabolomics and Toxicology: Inside the ELIXIR CZ–NL Staff Exchange

In April, researchers from ELIXIR Czech Republic and ELIXIR Netherlands came together in Maastricht for the first phase of a staff exchange designed to tackle a persistent challenge in the life sciences: how to better connect findings from toxicology and metabolomics. While both domains generate rich, complementary data, they often evolve in parallel and in a disconnected setting - using different tools, identifiers, and conceptual frameworks. During this exchange, researchers from RECETOX (Brno) and Translational Genomics (Maastricht) set out to change that.

14 May 2026

No description

Why this exchange matters to life science resources from across Europe

The initiative was built around three core goals:

1.      strengthening a shared portfolio of interoperable tools and services,

2.      facilitating knowledge transfer across nodes, and

3.      improving integration between the metabolomics and toxicology communities.

At its heart lies the Adverse Outcome Pathway (AOP) framework - a structured way to represent biological mechanisms linking (molecular) events and stressors to adverse outcomes. By aligning AOP-based approaches with metabolomics workflows, the teams aim to enable more holistic analyses that connect chemical exposure, molecular changes, and biological effects.


From alignment to interoperability

Rather than starting from scratch, the exchange focused on deepening connections between existing tools. Technologies such as BridgeDb, IDSM, MSMetaEnhancer, WikiPathways, and FixID were central to discussions, with an emphasis on moving beyond basic compatibility toward full interoperability.

One particularly promising direction involves federated queries that bridge datasets and domains - for example, querying AOP knowledge bases using mass spectrometry (MS) data through integrations between IDSM and AOP-Wiki-RDF. This kind of cross-domain querying could significantly streamline how researchers link experimental metabolomics data to mechanistic toxicology knowledge.

Alongside identifier mapping and ontology alignment, participants explored how AI-driven tools can enhance data integration. Applications ranged from text mining (e.g., aoptk) to image-based knowledge extraction (e.g. PFOCR), highlighting how automation can help scale the interpretation of complex biological data and establish FAIR-compliant best practices.

Hands-on collaboration and concrete progress

The exchange was not just conceptual - it was highly practical. Working sessions focused on improving and extending key tools:

  • MSMetaEnhancer is being expanded with additional identifier conversion capabilities, integration with BridgeDb, and improved chemical name resolution, including the addition of OPSIN services and Wikidata database.
  • aoptk, a Python package for large-scale LLM-driven analysis of scientific literature to support toxicological outcome assessment, is advancing toward richer, identifier-aware outputs, integration with ontology mapping services like ZOOMA, and support for nanopublication formats. Development also includes the incorporation of AOP-Wiki data, complementing its existing integration of sources like PubMed, PubMed Central, and Europe PMC.
  • SPARQL endpoints provide access to AI-ready life science data on the fly, by leveraging the semantic web framework that also supports all internet pages. These endpoints create a Linked Open Data knowledge graph interconnecting to various databases through harmonized identifier mappings using BridgeDb.
  • Galaxy tools are being extended with SPARQL query capabilities, structural similarity calculations, and planned integration with BioDataFuse, enabling more flexible and scalable workflows.

These efforts reflect a shared priority: making tools not only interoperable, but also easier to use in real-world workflows and scalable to answer new research questions.

Toward measurable impact and a more connected ELIXIR ecosystem

The collaboration is already shaping new research directions. Planned benchmarking studies will evaluate how well different tools perform in named entity recognition (NER) for small molecules, identifier resolution, and structural similarity analysis. These efforts will provide much-needed insight into how data interpretation varies depending on tools and identifiers - a key issue for reproducibility.

What makes this exchange particularly valuable is its broader impact. By reducing fragmentation and aligning tools and standards, we help both nodes operate more efficiently and demonstrate how shared infrastructure, open collaboration, and a focus on FAIR principles can close gaps and bring communities together. The second phase of the exchange, scheduled for later this year in Brno, will build on this momentum. If the first visit is any indication, the collaboration is well on its way to turning alignment into true integration - laying the groundwork for more powerful, cross-domain life science research.

Resources

 


More articles

All articles

You are running an old browser version. We recommend updating your browser to its latest version.

More info