Hands-on exploration of SPHN Semantic Interoperability Framework for FAIR knowledge graphs

Overview

SWAT4HCLS (Semantic Web Applications and Tools for Health Care and Life Sciences) is an annual conference at the intersection of semantic web technologies and biomedical research. SWAT4HCLS 2025 was held in Barcelona, Spain.

I, alongside my colleague Dr. Vasundra Touré, organized and delivered a hands-on tutorial introducing the SPHN Semantic Interoperability Framework to an international audience of researchers and data engineers.

The tutorial was designed as an end-to-end walkthrough: starting from defining health data concepts, generating semantic artifacts, and finishing with transforming mock data into validated, SPHN-compliant RDF. Participants worked through the full pipeline using two tools developed within SPHN: SchemaForge and Connector.

The SPHN Semantic Interoperability Framework

The Swiss Personalized Health Network (SPHN) is a national research infrastructure initiative that facilitates the exchange of health-related data in a FAIR manner within Switzerland. One key part of the SPHN is its Semantic Interoperability Framework that provides tools and resources for defining semantics for clinical data representation and exchange.

The framework adopts W3C standards with:

• RDF as the core data model for representing health data and its relationships
• RDFS and OWL for formally capturing the semantics of health concepts
• SHACL for defining and enforcing structural and semantic constraints on the data
• SPARQL for querying the resulting knowledge graphs

Together, these standards ensure that data produced within the SPHN network is interoperable, enabling discovery and reuse of health-related data across Switzerland.

Tutorial Structure

Part 1: SPHN SchemaForge

The first part of the tutorial introduced SPHN SchemaForge, a web service that takes a structured Excel file containing concept definitions, properties, and relationships as input and automatically generates a suite of semantic artifacts:

• RDF Schema: a formal representation of the defined concepts
• SHACL shapes: validation constraints that enforce the schema's rules on instance data
• SPARQL queries: pre-built queries for exploring and extracting data from a knowledge graph
• HTML documentation: human-readable reference documentation for data providers and researchers

Participants worked through the process of defining health data concepts in the Excel template, submitting it to SPHN SchemaForge, and inspecting the generated artifacts — gaining a concrete understanding of how semantic modeling decisions translate into machine-readable outputs.

Part 2: SPHN Connector

The second part focused on SPHN Connector, an data transformation pipeline that transforms mapped and structured source data (in a variety of formats) into RDF that is validated and compliant with SPHN semantics.

Participants mapped sample mock data against the schema generated in Part 1, ran the ingestion pipeline, and inspected the resulting RDF output. The validation step, using the SHACL shapes produced by SchemaForge, ensured that the ingested data conformed to the defined constraints before being loaded into the knowledge graph.