Updates from the MADICES (Machine-Actionable Data Interoperability for the Chemical Sciences) Workshop

https://www.psdi.ac.uk/ Updates from the MADICES (Machine-Actionable Data Interoperability for the Chemical Sciences) Workshop Ontologies4Chem 2025 11th November 2025 Dr Samantha Pearman-Kanza University of Southampton Introducing MADICES Core Themes Year Focus Area Key Themes 2022 Foundations • Historical challenges in chemical data archiving • Need for machine-actionable formats and FAIR principles • Barriers: lack of standards, metadata issues, long-term storage • The community needs clear standards, easy to use tools and integrations into existing tool chains • Report: https://madices.github.io/docs/2022/report 2024 Interoperability & Semantic Annotation • Platform-agnostic data exchange using JSON-LD and RO-Crate • Semantic annotation tools and ontology integration • Handling proprietary data formats and streaming • Early-stage annotation and vendor collaboration • Semantic meaning is KEY for reproducibility and open science acceleration • Report: https://madices.github.io/docs/2024/report 2025 Scaling & AI Integration • Education and outreach for semantic data • Best practices for Semantic Annotation • ELN Interoperability & Semantic Metadata • AI/LLM integration in lab workflows • FAIR instrument control and workflow schemas • Report: Coming Soon ELN Interoperability Introduction What is ELN Interoperability? The ability of Electronic Laboratory Notebooks (ELNs) to exchange, interpret, and reuse data across platforms and workflows. Supports FAIR principles: Findable, Accessible, Interoperable, Reusable. Community Need Enables collaboration across institutions. Enhances reproducibility and long-term data stewardship. Reduces vendor lock-in and supports integration with broader research infrastructure. MADICES 2025 Goals: Define and formalize ELN interoperability use cases. Evaluate mechanisms like the .eln File Format and RO-Crate Schema Plus. Recommend future directions for semantic interoperability. “Electronic Lab Notebooks are great, but not on vacation” Cartoon by Phil Johnson for MIT. ELN Interoperability Use Cases Data Preservation Singular ELN Backup & Restore: Recover user data within the same ELN system. Whole ELN Backup & Restore: Restore entire ELN ecosystem after migration or failure. Data Migration Singular ELN Migration: Transfer user data to a different ELN with semantic mapping. Multiple ELN Migration: Institutional migration with complex structures and templates. Collaboration & Publishing Cross-platform ELN Collaboration: Share experiments across different ELNs. ELN Archiving for Publication: Preserve and publish ELN data alongside research outputs. Integration with External Systems Exchange ELN Data with Other Systems: Use ELN data in external workflows and archives. Instrument Workflow Integration: Automate data ingestion from instruments into ELNs. CC BY-ND 4.0 Errant Science - https://errantscience.com / Barriers & Challenges Diverse Data Models: ELNs vary in how they structure experiments, samples, and metadata, making consistent data exchange difficult. Semantic Annotation Gaps: Not all ELNs support ontologies or semantic schemas, limiting the ability to interpret and reuse data. Export/Import Limitations: Round-trip data exchange is often incomplete or lossy, especially when metadata is not preserved. Link Rot and External Dependencies: Referencing external ontologies or resources can lead to broken links and outdated references. However, incorporating entire ontology files can lead to bloated oversized files. Complex Data Structures: Nested and interlinked data objects are difficult to flatten and represent in standardised formats. Instrument Integration: Manual data handling from instruments is inefficient and error-prone, and integration often requires custom solutions. Ownership and Access Control: Managing permissions and ownership during data export/import adds complexity, especially across institutional boundaries. Scalability and Performance: Large ontologies and datasets can slow down processing and validation, requiring optimisation strategies. User Burden: Without automation and intuitive tools, users may struggle to curate, annotate, and share data effectively. Mechanisms & Approaches RO-Crate & RO-Crate Schema Plus: JSON-LD format for packaging research data with metadata. ELN File Format (Developed by the ELN Consortium to support standardized data packaging and exchange across eln platforms uses RO-Crate) Ontologies & SHACL: Modular ontologies and SHACL shapes enable validation and semantic consistency. Modularisation improves usability. API-Based Exchange: ELNs like elabFTW use APIs to resolve linked data. Machine-actionable links and self-describing data are key. Instrument Integration: SILA enables direct instrument-ELN interaction. FINALES acts as a broker for coordinated communication in large-scale projects. Cartoon created by ErrantScience.com for AI4SD: licensed under CC-BY-NC ELN Interoperability Recommendations Standardize Export Formats: Promote adoption of RO-Crate, RO-Crate Plus, and .eln profiles across ELNs. Develop Semantic Mappings: Agree on shared ontologies and schemas for metadata. Avoid passing internal ELN-specific data that lacks external meaning. Automate Instrument Workflows: Use standards like SILA to streamline data flow. Instruments should be able to send RO-Crates directly to ELNs with minimal manual steps. Create Demonstrators: Build proof-of-concept integrations to validate approaches. Examples include ingesting RO-Crates into platforms like OpenBIS [23] or DataLab [24]. Support API-Driven Interoperability: Ensure ELNs expose sufficient API endpoints for data access and manipulation. Consider access levels and ownership logic at import/export stages. Enable Self-Describing Data: Use semantic annotations and indexing algorithms to help systems follow relevant links and resolve references automatically. Cartoon created by ErrantScience.com for AI4SD: licensed under CC-BY-NC Semantic Annotation What is Semantic Annotation? Linking data to ontological concepts for machinereadable metadata. Supports FAIR principles: Findable, Accessible, Interoperable, Reusable. Community Need Enhances data interoperability, discoverability, and reuse. Addresses challenges in scientific data integration and automation. MADICES 2025 Goals Provide guidance and best practices for semantic annotation. Identify and address key challenges in implementation and adoption. CC BY-ND 3.0 Dataedo - https://dataedo.com/cartoon/