FAIR Data Principles in the EU: Compliance and Implementation

Researchers funded by the European Commission or the US National Institutes of Health (NIH) must now treat data management as a core part of the research lifecycle. The FAIR principles, which stand for findable, accessible, interoperable, and reusable, define the minimum standard for how research data should be prepared, described, and shared. For bioinformatics professionals, compliance affects grant eligibility, publication requirements, and the long-term utility of generated datasets. This article explains what the FAIR principles require, how European and US funding mandates differ, and how to implement compliant data practices in bioinformatics projects.

Understanding the FAIR Principles

The FAIR Guiding Principles were published in 2016 in Scientific Data and have since become the reference framework for research data management across disciplines. The principles were designed to support the reuse of data and other digital research objects, including algorithms, tools, and workflows that produced the data. Each letter in the FAIR acronym represents a cluster of requirements that together make data maximally useful for both humans and machines.

Findability

Findable data can be located by both people and automated systems. This requires that data and metadata are assigned persistent identifiers, that metadata clearly describe the data, and that the data or metadata are registered or indexed in a searchable resource. In bioinformatics, this typically means depositing datasets in recognized repositories such as those maintained by the National Center for Biotechnology Information (NCBI) or the European Bioinformatics Institute (EMBL-EBI). The NCBI Data Resources and EMBL-EBI Training portals provide access to the major repositories where bioinformatics data should be deposited.

Accessibility

Accessibility means that data can be retrieved using their identifiers, and that the protocol for retrieval is standardized and openly available. Importantly, accessibility does not require data to be open to everyone. Data can be accessible under authentication or authorization procedures, such as controlled access for human genomic data. The key requirement is that the conditions of access are clearly stated and that the retrieval mechanism works reliably.

Interoperability

Interoperable data can be integrated with other datasets and processed by different systems. This requires the use of formal, shared vocabularies and ontologies, and the inclusion of qualified references to other data. In bioinformatics, interoperability often means using community standards for data formats and metadata schemas. Community-developed minimum information checklists are designed to drive rich and consistent reporting of metadata, underpinning the reproducibility and reuse of data. These reporting guidelines are usually in the form of narratives intended for human consumption, but modular and reusable machine-readable versions are also needed to provide quantitative and verifiable measures of the degree to which metadata descriptors meet community requirements.

Reusability

Reusable data are richly described with accurate and relevant attributes, released with clear usage licenses, and accompanied by detailed provenance information about how they were generated. Reusability is the ultimate goal of the FAIR principles because it enables new research questions to be answered with existing data. The FAIRification process aims to translate raw datasets into FAIR datasets in a general way, addressing data curation, validation, deidentification, versioning, and indexing.

Funding Mandates and Policy Context

European Commission Requirements

The European Commission requires FAIR data management for EU-funded projects. This requirement applies to projects funded under Horizon 2020 and Horizon Europe, and it extends to experiments conducted at EU-funded research infrastructures. The requirement is part of the broader European Open Science Cloud initiative, which aims to create a trusted environment for sharing and reusing research data across Europe.

For bioinformatics researchers, the practical implication is that data management plans must be submitted at the project proposal stage and updated throughout the project lifecycle. These plans must describe how data will be made FAIR, which repositories will be used, and what access conditions will apply. The European Commission also requires that research data be deposited in recognized repositories, with the expectation that data will be as open as possible and as closed as necessary.

NIH Genomic Data Sharing Policy

The NIH Genomic Data Sharing Policy establishes expectations for the sharing of genomic data generated with NIH funding. The policy applies to large-scale genomic data, including genome-wide association studies, whole-genome sequencing, and transcriptomic data. Under this policy, researchers must deposit data in designated data repositories, submit data to the appropriate NIH data sharing repository, and ensure that data are shared in a timely manner.

The NIH policy emphasizes the importance of informed consent for data sharing, and it requires that data be shared in a manner consistent with the consent under which they were collected. For bioinformatics researchers working with human genomic data, this means that data use limitations must be respected and that controlled access mechanisms must be implemented where required.

Alignment Between EU and NIH Requirements

While the EU and NIH mandates differ in their specific requirements, they share a common foundation in the FAIR principles. Both funders expect data to be deposited in recognized repositories, described with rich metadata, and made available for reuse under clear conditions. For researchers funded by both organizations, the practical approach is to design data management practices that satisfy the stricter of the two requirements for each data type.

At a Glance: FAIR Compliance Overview

The FAIRification Workflow

The FAIRification process transforms raw datasets into FAIR-compliant datasets through a structured series of steps. The GO FAIR initiative defined a seven-step FAIRification process focusing on data, but also indicating the required work for metadata. This process aims to address the translation of raw datasets into FAIR datasets in a general way, without considering specific requirements and challenges that may arise when dealing with particular types of data.

Step 1: Analyze Data Characteristics

The first step is to understand the nature of the data being processed. Bioinformatics data can include raw sequencing reads, processed alignments, variant calls, expression matrices, and clinical annotations. Each data type has different requirements for metadata, format, and repository selection. For health research data, additional considerations include ethical implications and legal frameworks that affect how data can be shared.

Step 2: Define the Target Metadata Model

Metadata models describe the attributes that must be captured to make data interpretable. Community-developed minimum information checklists provide the foundation for metadata models in many bioinformatics domains. These checklists are designed to drive rich and consistent reporting of metadata, underpinning the reproducibility and reuse of data. Machine-readable versions of these checklists provide quantitative and verifiable measures of the degree to which metadata descriptors meet community requirements.

Step 3: Transform Data Formats

Raw data often exist in proprietary or project-specific formats that are not interoperable. The FAIRification process requires transforming data into standard formats that can be processed by different tools. For health data, the Health Level Seven International Fast Healthcare Interoperability Resources standard provides a framework for structuring clinical data. The use of HL7 FHIR resources supports the revised FAIRification workflow for health research data management.

Step 4: Create Rich Metadata

Metadata creation involves extracting information from the documentation and transforming it to a knowledge graph using semantic web standards. This process requires careful attention to the vocabulary used, with terms drawn from recognized ontologies and controlled vocabularies. The metadata must describe both the data themselves and the experimental conditions under which they were generated.

Step 5: Assign Persistent Identifiers

Persistent identifiers ensure that data can be found and cited over the long term. For bioinformatics data, this typically means obtaining accession numbers from repositories such as NCBI or EMBL-EBI, or DOIs from services like Zenodo. The identifiers must be registered in a way that ensures they remain resolvable over time.

Step 6: Register in a Searchable Repository

Depositing data in a recognized repository ensures that it can be discovered by other researchers. The choice of repository depends on the data type and the requirements of the funding agency. For genomic data, NCBI and EMBL-EBI repositories are the standard choices. For other types of bioinformatics data, general-purpose repositories like Zenodo may be appropriate.

Step 7: Validate and Release

The final step is to validate that the FAIRified dataset meets the requirements of the target metadata model and to release it with appropriate access conditions and licenses. Validation should include checks for metadata completeness, format compliance, and identifier resolution.

Practical Implementation for Bioinformatics Projects

Repository Selection

The choice of repository is one of the most consequential decisions in FAIR compliance. For bioinformatics data, the major options include the NCBI Data Resources and the EMBL-EBI Training repositories. NCBI maintains databases for sequence data, genomic data, and biomedical literature. EMBL-EBI provides similar services for European researchers and maintains cross-references with NCBI databases.

For datasets that do not fit the standard bioinformatics repository categories, general-purpose repositories like Zenodo provide an alternative. Zenodo was developed with support from the European Union and accepts data from any discipline. The repository assigns DOIs, supports versioning, and provides access to the full research community.

Metadata Standards Selection

The selection of metadata standards should be driven by the data type and the requirements of the target repository. For plant phenotyping data, the MIAPPE standard provides a framework for describing experimental conditions. For clinical data, HL7 FHIR resources provide the structure for interoperability. For materials science data, community discussions have emphasized the need for FAIR-compliant metadata schemas that address the specific challenges of computational and experimental data.

The key consideration is that metadata standards must be machine-readable to support automated processing and validation. Narrative descriptions intended for human consumption are insufficient for FAIR compliance because they cannot be processed by automated tools.

Data Format Decisions

Data formats determine whether data can be processed by standard bioinformatics tools. The FAIR principles require that data be represented in formats that are widely used and openly documented. For sequence data, this means using formats like FASTQ and BAM. For variant data, VCF is the standard format. For expression data, formats like TSV or HDF5 may be appropriate depending on the analysis pipeline.

The transformation of data into standard formats is often the most time-consuming part of the FAIRification process. Automated tools can assist with this transformation, but manual curation is often required to ensure that all relevant information is captured.

Versioning and Provenance

Data versioning is essential for reproducibility and for tracking the evolution of datasets over time. The FAIRification workflow for health research data identified versioning as a gap in the original GO FAIR process. Each version of a dataset should have its own persistent identifier, and the relationship between versions should be documented.

Provenance information describes how data were generated, including the instruments, software, and parameters used. This information is essential for reusability because it allows other researchers to understand the limitations and appropriate uses of the data.

Assessment and Measurement of FAIR Compliance

Automated Assessment Tools

Several automated tools have been developed to assess FAIR compliance. The F-UJI tool evaluates datasets against the FAIR principles and provides a compliance score. FAIR-Checker is another automated tool that assesses dataset quality. These tools evaluate aspects such as the presence of persistent identifiers, the completeness of metadata, and the accessibility of data through standard protocols.

A study of 384 datasets in the energy research domain used FAIR-Checker and F-UJI to evaluate FAIR compliance. The results showed strengths and weaknesses in FAIR principle compliance within the energy domain, providing insights for research data management efforts. Similar assessments in the addiction research field found that none of the datasets evaluated reached an average of 30 percent FAIR compliance, with findability scoring highest and interoperability and reusability scoring lowest.

Manual Assessment Approaches

Automated tools cannot capture all aspects of FAIR compliance. Manual assessment is needed to evaluate the quality of metadata, the appropriateness of vocabularies, and the clarity of usage licenses. Manual assessment should be conducted by someone familiar with the data domain and the requirements of the target community.

Common Assessment Findings

Studies of FAIR compliance across multiple domains have identified consistent patterns. Findability is typically the best-scored principle because most repositories assign persistent identifiers and register data in searchable databases. Accessibility also tends to score reasonably well because repository download mechanisms are generally reliable. Interoperability and reusability consistently score lowest because they require more effort in metadata creation and format standardization.

The implication for bioinformatics researchers is that FAIR compliance efforts should focus on the areas where compliance is weakest. This means investing in metadata creation, adopting community standards, and documenting provenance information.

Common Failure Patterns in FAIR Implementation

Incomplete Metadata

The most common failure in FAIR implementation is incomplete metadata. Researchers often deposit data with minimal descriptions that are insufficient for other researchers to understand what the data contain or how they were generated. This failure undermines both findability and reusability because the data cannot be discovered through search and cannot be interpreted by potential users.

Inappropriate Repository Selection

Choosing a repository that does not match the data type or the requirements of the funding agency is another common failure. Some repositories have specific requirements for data formats and metadata that must be met before deposit. Researchers who do not check these requirements may find that their data cannot be deposited or that the deposit does not satisfy funding requirements.

Ignoring Machine Readability

Metadata that is only readable by humans does not support automated processing. The FAIR principles require that metadata be machine-readable so that automated tools can discover and process data. Researchers who create narrative descriptions without structured metadata fail to meet this requirement.

Inadequate Versioning

Data that is not versioned cannot be cited reliably, and the relationship between different versions of a dataset cannot be traced. This failure undermines reproducibility and makes it difficult for other researchers to understand which version of a dataset was used in a particular analysis.

Neglecting Provenance

Provenance information is essential for reusability because it allows other researchers to understand the limitations and appropriate uses of the data. Researchers who do not document the experimental conditions, software versions, and analysis parameters used to generate data limit the potential for data reuse.

Data Sharing Platforms and Their Limitations

Platform Diversity and Fragmentation

The landscape of data sharing platforms is diverse and fragmented. A scoping review and cross-sectional survey of COVID-19 data sharing platforms and registries identified 44 registries and 20 platforms. These resources were concentrated in high-income countries and siloed by comorbidity, body system, and data type. Resources for harmonizing and sharing clinical data were less likely to implement FAIR principles than those sharing omics or imaging data.

The finding that more data sharing does not equate to better data sharing highlights the need for semantic and technical interoperability of platforms and registries. For bioinformatics researchers, this means that depositing data in a repository is not sufficient. The repository must support the integration of data with other resources and the application of standard vocabularies.

Federated Approaches

Federated machine learning architectures built upon FAIR health data enable collaborative model building among data owners without sharing their datasets. In a real-world application, five healthcare organizations transformed their datasets into HL7 FHIR resources via a common FAIRification workflow and software set, generating FAIR datasets. Each organization deployed a federated learning agent within its secure network, connected to a cloud-based manager. The federated model achieved an accuracy rate of 87 percent in predicting 30-day readmission risk for chronic obstructive pulmonary disease patients.

This approach demonstrates the value of FAIR health data in machine learning when utilized in a federated manner that ensures privacy protection without sharing data. For bioinformatics researchers working with sensitive data, federated approaches may offer a path to compliance that respects privacy requirements.

Domain-Specific Considerations

Health Research Data

Health research data present unique challenges for FAIR compliance. The FAIRification workflow for health research data must address technical barriers, ethical implications, and legal frameworks. The analysis of the FAIRification workflow from these perspectives identified gaps in data curation, validation, deidentification, versioning, and indexing.

Deidentification is a critical requirement for health data. The process of removing personally identifiable information must be documented and validated to ensure that privacy is protected. The legal framework for health data sharing varies by jurisdiction, and researchers must ensure that their data sharing practices comply with applicable regulations.

Plant Phenotyping Data

Plant phenotyping experiments are conducted under a variety of experimental parameters and settings for diverse purposes. The data they produce is heterogeneous, complicated, often poorly documented, and difficult to reuse. Meeting societal needs for nutrition, crop adaptation, and stability requires more efficient methods toward data integration and reuse.

A case study examining the reuse of plant phenotyping data found that making data FAIR involves significant effort. The process of discovering a phenotypic dataset on a FAIR data point, verifying the existence of related datasets with environmental data, acquiring both, and integrating them presented challenges. The use of metadata standards such as MIAPPE was essential for reusability and reproducibility.

Computational Materials Science

The expansive production of data in materials science, their widespread sharing, and repurposing requires educated support and stewardship. The implementation of the FAIR data principles must not be too narrow. The wider materials science community ought to agree on strategies to tackle the challenges that are specific to its data, both from computations and experiments.

For computational materials science data, the FAIRification of metadata related to ground-state and excited-states calculations, potential-energy sampling, and generalized workflows requires a constructive approach. Challenges with the FAIRification of experimental data and materials science ontologies remain to be addressed.

Photon and Neutron Research Infrastructures

Users of photon and neutron research infrastructures employ a wide range of scattering, imaging, and spectroscopic methods investigating the behavior of matter with a broad scientific base across physics, chemistry, and biology. They produce large data volumes of up to 1 petabyte per day in some cases. To ensure all these data are FAIR requires an enormous effort from the research infrastructures.

The effort requires the expansion of data storage capacity, the development and deployment of software for effective data storage, metadata schemes, and implementation of effective data pipelines at each individual experiment. FAIR data also affect the carbon footprint related to large amounts of data and raise questions related to user authentication, rights of access, and cyber security. The research infrastructures alone cannot achieve such a transformational process. Cooperation of the user communities is essential as they need to create and utilize existing tools to deliver FAIR data.

Records and Documentation Requirements

Data Management Plans

Data management plans are the primary documentation tool for FAIR compliance. These plans must describe the types of data that will be generated, the standards that will be applied, the repositories that will be used, and the access conditions that will apply. For EU-funded projects, data management plans must be submitted at the proposal stage and updated at regular intervals.

Metadata Records

Metadata records must be maintained for each dataset. These records should include the persistent identifier, the metadata schema used, the vocabularies applied, and the date of deposit. The metadata should be sufficient for another researcher to understand what the data contain and how they were generated.

Provenance Logs

Provenance logs document the history of data from generation through processing to final deposit. These logs should include information about the instruments used, the software versions applied, the parameters set, and the individuals responsible for each step. Provenance logs are essential for reproducibility and for diagnosing problems in data processing.

Version Histories

Version histories track the evolution of datasets over time. Each version should have its own persistent identifier, and the relationship between versions should be documented. Version histories are important for citation and for understanding which version of a dataset was used in a particular analysis.

Quality Controls and Validation

Metadata Validation

Metadata validation checks that metadata records meet the requirements of the target metadata model. This includes checking for required fields, validating vocabulary terms, and ensuring that identifiers are properly formatted. Automated validation tools can perform many of these checks, but manual review is often needed to assess the quality of narrative descriptions.

Format Validation

Format validation checks that data files conform to the requirements of the target format. This includes checking file structure, encoding, and the presence of required headers. Format validation is particularly important for data that will be processed by automated pipelines.

Identifier Resolution

Identifier resolution checks that persistent identifiers resolve to the correct data or metadata. This validation should be performed at the time of deposit and periodically thereafter to ensure that identifiers remain functional.

Accessibility Testing

Accessibility testing verifies that data can be retrieved using the documented protocols. This includes testing download mechanisms, authentication procedures, and the behavior of the repository under different conditions.

Limitations and Interpretation Constraints

Automated Assessment Limitations

Automated FAIR assessment tools have limitations that must be understood when interpreting their results. These tools can check for the presence of identifiers, the completeness of metadata fields, and the accessibility of data through standard protocols. They cannot assess the quality of metadata content, the appropriateness of vocabulary choices, or the clarity of usage licenses.

Domain-Specific Constraints

FAIR compliance must be interpreted in the context of the specific data domain. The requirements for genomic data differ from those for clinical data, which differ from those for materials science data. Researchers should not apply a one-size-fits-all approach to FAIR compliance but should instead follow the standards and practices of their specific community.

Privacy and Ethical Constraints

The FAIR principles do not require data to be open. Data can be accessible under controlled conditions, and privacy and ethical considerations may require such controls. Researchers must balance the goal of data sharing with the obligation to protect research participants and respect data use limitations.

Professional Escalation Criteria

When to Seek Expert Assistance

Researchers should seek expert assistance when they encounter challenges that exceed their expertise. This includes situations where the data type is novel, where the metadata standards are unclear, or where the repository requirements are complex. Data stewards and bioinformatics support staff can provide guidance on FAIR implementation.

When to Consult Legal Counsel

Legal counsel should be consulted when data sharing raises legal questions. This includes situations involving data use agreements, intellectual property, or cross-border data transfers. Legal review is particularly important for health data and for data that may be subject to export controls.

When to Escalate Technical Issues

Technical issues that cannot be resolved through standard troubleshooting should be escalated to the repository support team or to the developers of the tools being used. This includes issues with identifier resolution, metadata validation, or data format transformation.

Frequently Asked Questions

What is the difference between FAIR data and open data?

FAIR data and open data are related but distinct concepts. Open data refers to data that anyone can access, use, and share without restrictions. FAIR data refers to data that are findable, accessible, interoperable, and reusable. FAIR data can be open, but it can also be accessible under controlled conditions. The FAIR principles emphasize the conditions of access and the quality of metadata instead of the absence of restrictions.

Which repositories satisfy EU FAIR data requirements?

The European Commission does not maintain a single list of approved repositories. Instead, it requires that data be deposited in repositories that support FAIR principles. For bioinformatics data, the NCBI Data Resources and EMBL-EBI Training repositories are the standard choices. General-purpose repositories like Zenodo also satisfy EU requirements for many data types.

How does the NIH Genomic Data Sharing Policy relate to FAIR principles?

The NIH Genomic Data Sharing Policy requires that genomic data generated with NIH funding be shared in a timely manner through designated data repositories. The policy aligns with FAIR principles by requiring that data be deposited in repositories that assign persistent identifiers, support metadata creation, and provide access controls. The policy adds specific requirements for informed consent and data use limitations that go beyond the general FAIR framework.

What are the most common reasons for low FAIR compliance scores?

Studies of FAIR compliance across multiple domains have found that interoperability and reusability consistently score lowest. Common reasons include incomplete metadata, the use of non-standard vocabularies, the absence of machine-readable metadata, and the lack of clear usage licenses. Findability and accessibility tend to score higher because most repositories assign persistent identifiers and provide reliable download mechanisms.

How long does the FAIRification process take for a typical bioinformatics dataset?

The time required for FAIRification depends on the data type, the quality of the existing documentation, and the availability of automated tools. Simple datasets with good documentation can be FAIRified in a few days. Complex datasets with poor documentation may require weeks of effort. The FAIRification workflow for health research data identified data curation, validation, deidentification, versioning, and indexing as the most time-consuming steps.

Can FAIR principles be applied to data that cannot be shared openly?

Yes, FAIR principles can be applied to data that cannot be shared openly. The accessibility principle requires that data be retrievable using their identifiers, but it does not require that access be granted to everyone. Data can be accessible under controlled conditions, such as through a data access committee or a data use agreement. The key requirement is that the conditions of access are clearly stated and that the retrieval mechanism works reliably.

What is the role of metadata standards in FAIR compliance?

Metadata standards define the attributes that must be captured to make data interpretable. Community-developed minimum information checklists provide the foundation for metadata models in many bioinformatics domains. Machine-readable versions of these checklists provide quantitative and verifiable measures of the degree to which metadata descriptors meet community requirements. The use of recognized metadata standards is essential for interoperability and reusability.

How should researchers handle versioning of FAIR datasets?

Researchers should assign a new persistent identifier to each version of a dataset and document the relationship between versions. The metadata for each version should describe what changed and why. Versioning is essential for reproducibility because it allows other researchers to cite the specific version of a dataset used in their analysis. The FAIRification workflow for health research data identified versioning as a gap in the original GO FAIR process.

Related Bioinformatics Guides

• GWAS QC Steps: Structural Analysis and Computational Methodologies in Bioinformatics
• The European Bioinformatics Institute (EMBL-EBI): A Comprehensive Reference for Veterinary Computational Biology
• The Protein Data Bank (PDB): Archival Standards, Structural Validation Metrics, and Bioinformatics Integration Protocols
• The History and Evolution of Bioinformatics
• Bioinformatics in Agriculture and Crop Improvement

References and Further Reading

• EMBL-EBI Training. European Bioinformatics Institute.
• NCBI Data Resources. National Center for Biotechnology Information.
• Genomic Data Sharing Policy. National Institutes of Health.
• The FAIR Guiding Principles. Scientific Data.
• From Raw Data to FAIR Data: The FAIRification Workflow for Health Research.. Methods of information in medicine, 2020.
• FAIR, ethical, and coordinated data sharing for COVID-19 response: a scoping review and cross-sectional survey of COVID-19 data sharing platforms and registries.. The Lancet. Digital health, 2023.
• Machine actionable metadata models.. Scientific data, 2022.
• The benefits and struggles of FAIR data: the case of reusing plant phenotyping data.. Scientific data, 2023.
• FAIR data for prehistoric mining archaeology.. International journal on digital libraries, 2021.
• Shared metadata for data-centric materials science.. Scientific data, 2023.
• Privacy-preserving federated machine learning on FAIR health data: A real-world application.. Computational and structural biotechnology journal, 2024.
• FAIR data - the photon and neutron communities move together towards open science.. IUCrJ, 2025.
• Impact of medicaid waivers on medication for opioid use disorders in residential facilities: Evidence from twenty-five states.. 2026.
• Provision of recommended treatment for stimulant use disorder in United States substance use treatment facilities.. 2026.
• How do services offered within opioid treatment programs vary based on state methadone policies?. 2026.
• Integrating evidence-based methamphetamine interventions into methadone clinics: Structural barriers and implementation opportunities.. 2026.
• Assessing the utility of health access data and social determinants of health in ecological suicide prediction models.. 2026.
• A cross-sectional analysis of county-level determinants of pandemic-era policy implementation.. 2026.
• Feasibility and Preliminary Effectiveness of a Data-to-Care Intervention for HIV-Positive People Who Inject Drugs in Ukraine: A Pilot Cluster-Randomized Trial.. 2026.
• fair_data.py: implementing FAIR data compliance in Tribchem. 2025.
• Measuring FAIR Data Compliance in German Energy Research. Open Conference Proceedings, 2026.
• Investigating FAIR data principles compliance in horizon 2020 funded Agri-food and rural development multi-actor projects. Agricultural Systems, 2024.
• The compliance to FAIR principles of shared data in addiction research. Scientometrics, 2025.
• Towards FAIR Data Workflows for Multidisciplinary Science: Ongoing Endeavors and Future Perspectives in Plasma Technology. International Conference on Data Technologies and Applications, 2024.
• Compliance-by-Design: Validating Decentralised Architectures for Digital Product Passports under the EU Data Governance Act. Proceedings of the 2025 21st International Conference on Network and Service Management AI and Sustainability in the Future of Network and Service Management Cnsm 2025, 2025.
• DICONDE: Enabling FAIR Data and Communication in NDE 4.0. Handbook of Nondestructive Evaluation 4 0, 2025.
• Ontology Network for the Standardization of the EU CBAM report data. Ceur Workshop Proceedings, 2024.

This article is educational and does not replace validated analysis plans, institutional policy, clinical interpretation, or specialist review.