Open research questions in Research Data Management Practices

What the literature leaves open

• While these funds are limited to the very short term (with the PLATO project ending in June 2025, the CoDOA and the DOACH projects ending in 2026), the premise of these projects is to better understand the landscape and potential for Diamond OA in Switzerland and to build collaborative infrastructure that could sustain Diamond OA nationally in the longer term. The policy has faced resistance from several APC-based journals, and the final outcome of its implementation remains uncertain. 1 Landscape Report on Open Access Diamond publishing in Africa, Europe and Latin America SERBIA The subsidy system is insufficiently transparent, as neither the lists of funded publications, nor the amounts awarded to individual publications are publicly available. One outcome of the conference acknowledged that there is insufficient funding for Diamond OA in Switzerland. 1 Landscape Report on Open Access Diamond publishing in Africa, Europe and Latin America SWITZERLAND licensing remain insufficiently covered.

  Landscape Report on Diamond Open Access Publishing in Africa, Europe, and Latin America · 2026 · DOI

• 67. Blumberg, K., Miller, M., Ponsero, A. & Hurwitz, B. Ontology-driven analysis of marine metagenomics: what more can we learn from our data? Gigascience 12, (2022). 68. Blumberg, K. L. et al. Ontology-Enriched Specifications Enabling Findable, Accessible, Interoperable, and Reusable Marine Metagenomic Datasets in Cyberinfrastructure Systems. Front. Microbiol. 12, 765268 (2021).

  Advancing FAIR data towards comparable, organized, predictive AI-ready data for community validation · 2026 · DOI

• Toro, S. et al. Dynamic Retrieval Augmented Generation of ontologies using artificial intelligence (DRAGON-AI). J. Biomed. Semantics 15, 19 (2024). 16. Niyonkuru, E. et al. Leveraging generative AI to assist biocuration of medical actions for rare disease. Bioinform. Adv. 5, vbaf141 (2025). 17. van Reisen, M. et al. Towards the tipping point for FAIR implementation. Data Intell 2, 264–275 (2020). 42. The FAIR cookbook, containing recipes to make your data more FAIR. https://github.com/FAIRplus/the-fair-cookbook (2026). 43. Jackson, R. et al. OBO Foundry in 2021: operationalizing open data principles to evaluate ontologies. Database (Oxford) 2021, (2021). 44. Vendetti, J. et al. BioPortal: an open community resource for sharing, searching, and utilizing biomedical ontologies. Nucleic Acids Res. 53, W84–W94 (2025). 18. Simpson, A. et al. MISIP: a data standard for the reuse and 45. Wood-Charlson, E. M., Crockett, Z., Erdmann, C., Arkin, A. P. & reproducibility of any stable isotope probing-derived nucleic acid sequence and experiment. Gigascience 13, (2024). Robinson, C. B. Ten simple rules for getting and giving credit for data. PLoS Comput. Biol. 18, e1010476 (2022). 19. Meyer, F. et al. CAMI Benchmarking Portal: online evaluation and ranking of metagenomic software. Nucleic Acids Res. 53, W102–W109 (2025). 20. Gebre, S. G. et al. NASA open science data repository: open science for life in space. Nucleic Acids Res. 53, D1697–D1710 (2025). 21. The Human Microbiome Project Consortium. A framework for human microbiome research. Nature 486, (2012). 22. The Human Microbiome Project Consortium. Structure, function and diversity of the healthy human microbiome. Nature 486, (2012). 23. Van Den Bossche, T. et al. The Metaproteomics Initiative: a coordinated approach for propelling the functional characterization of microbiomes. Microbiome 9, 243 (2021). 46. Koblitz, J., Reimer, L. C., Pukall, R. & Overmann, J. Predicting bacterial phenotypic traits through improved machine learning using highquality, curated datasets. Commun. Biol. 8, 897 (2025). 47. Fadum, J. M., Borton, M. A., Daly, R. A., Wrighton, K. C. & Hall, E. K. Dominant nitrogen metabolisms of a warm, seasonally anoxic freshwater ecosystem revealed using genome resolved metatranscriptomics. mSystems 9, e0105923 (2024). 48. Thorpe, A. C. et al. River biofilm bacteria as sentinels of national-scale freshwater ecosystems. bioRxiv, (2025). 49. Stach, T. L. et al. Complex compositional and metabolic response of river sediment microbiomes to multiple anthropogenic stressors. ISME Commun, (2025). 24. Field, D. et al. The Genomic Standards Consortium. PLoS Biol 9, 50. Berman, H. M. et al. The Protein Data Bank. Nucleic Acids Res. 28, e1001088 (2011). 235–242 (2000). 25. Hedlund, B. P. et al. SeqCode: a nomenclatural code for prokaryotes described from sequence data. Nat. Microbiol. 7, 1702–1708 (2022). 26. Whitman, W. B. et al.

  Advancing FAIR data towards comparable, organized, predictive AI-ready data for community validation · 2026 · DOI

• Beyond what is described in the manuscript, and demonstrating the power of being engaged, was an additional exploratory attempt to predict pigmentation. Many strains were flagged as “pigment-producing” while the pigment name was recorded as “no pigment”, presenting an internal inconsistency that rendered the labels unsuitable for ML. The issue was reported and quickly corrected in the database. This illustrates exactly how FAIR + COPE is an iterative scientific investigation, and updating FAIR data saves time: controlled vocabularies and validation rules should enforce trait/ value consistency (Comparable/Organized), and structured feedback loops between data users and curators (Engaged) turned FAIR data records into FAIR + COPE, AI-ready datasets that are reproducible and easier to validate. Real-world demonstration of impact: Community river microbiome catalogue. Establishing FAIR + COPE processes at the onset of a multistudy cooperative project can streamline data integration and elevate the final product. The Genome Resolved Open Watershed database (GROWdb)7 started as a coordinated effort to align sampling designs, analytical workflows, and metadata reporting structures, while still allowing for individual researchers to pursue their research questions and hypotheses. By defining the minimum metadata, core environmental measurements, and quality control expectations, data comparability, and therefore aggregated analysis for statistical power, was possible beyond what is typically achieved by a single project effort. In addition, the GROWdb data were extensively labeled using standards and ontological terms that explicitly connect the microbes, as biological observations, to their environmental and ecosystem context. Sample-Data-Environment linkages enable rapid query, recombination, and direct use in modeling and ML workflows. Example predictions explored include: forecasting of microbial functional responses to changes in oxygen, nutrients, and hydrologic conditions; estimation of microbial contributions to biogeochemical fluxes across watersheds; and identification of conserved functional responses to shared environmental stressors. Throughout the collection, generation, and release of the GROWdb, community engagement was a key driver that refined both data quality and usability. For example, GROWdb leads aligned metadata practices with emerging community guidance for sample and metadata reporting34. This collaboration identified gaps, improved variable definitions, and resulted in more complete metadata, relative to what individual projects typically document. Engagement with scientists on specific projects such as lakefocused analyses that incorporated GROWdb into broader ecosystem studies (e.g., Fadum et al47.) further ensured which data structures would support diverse reuse cases.

  Advancing FAIR data towards comparable, organized, predictive AI-ready data for community validation · 2026 · DOI

• https://doi.org/10.1038/s42003-026-10694-y Fig. 2 | Persistent identifiers connect FAIR data to people, places, and other research outputs. Persistent Identifiers (PIDs) connect people, organizations, and research products, from individuals (ORCID) and organizations (ROR, Research Organization Registry) to samples (IGSN, International Geo/General Sample Number34,53). Related identi- fiers can be included in the PID metadata to establish connections in machine-readable formats. Data Management Plans (DMPs), especially those supported by the DMPTool (dmptool.org) help organize research outputs like protocols, datasets, software, and publications. These are typically given DOIs (Digital Object Identifier) and connected by relationship types, including is_part_of / has_part or cites / is_citedby. Finally, research projects can be given a Research Activity iDentifier (RAiD, https:// raid.org). Box 1 | Examples of FAIR + COPE resources recommended by the authors Australian BioCommons uses a nuanced engagement strategy (consultations, meetings, surveys) to understand and solve community-level data standards, analysis, and management challenges such as issues with data submission or access. Solutions – developed in partnership with infrastructures, institutes and repositories – include building or adopting technologies with ongoing community involvement and engagement. For example, Galaxy Australia’s Microbiology Lab (microbiology.usegalaxy.org.au) provides tools, workflows and compute for analysis, developed with the international microbial research community54 and the Australian Microbiome Analysis Community (www. biocommons.org.au/microbiomeanalysis). Bacterial and Viral Bioinformatics Resource Center (BV-BRC55) leverages a unified data model to support a web-based AI assistant and enhanced web-based visualization and analysis tools as a service to aid in data integration for computable comparisons and workflows, which are shared back to the community for better understanding of pathogen biology and infectious diseases. Department of Energy Systems Biology Knowledgebase (KBase56) has an object-based data model that enables researchers to (1) automatically convert data files into interoperable objects, (2) perform provenanced data analyses that build model predictions from complex, multiscale biological data, and then (3) immediately share with collaborators for feedback or publish them alongside a journal article. ELIXIR57 maintains a list of endorsed deposition databases for the submission of experimental data and supports communities in converging on shared standards, databases, and tools within their domain (e.g. Microbiome Community and Federated Human Data Community)58,59.

  Advancing FAIR data towards comparable, organized, predictive AI-ready data for community validation · 2026 · DOI

• metadata, and provenance minimizes the data management effort required for data release or publication. Agent-assisted tools that support the initial steps in the FAIR + COPE processes will be essential for questions that extend beyond a single domain or institution; reach across time, space, and biological scales; and aim to integrate different data modalities. They will also accelerate the generation of AI-ready data. COPE is operationalized by integrating community engagement FAIR has been embraced by the life sciences community17. However, once a FAIR dataset is released, it is rarely updated to incorporate new biological knowledge. FAIR + COPE embeds engagement into each step of Comparability, Organization, and Predictive analysis – from ontology curation and the formation of new standards18, to workflow sharing or open model validation – creating an iterative loop where data and models are continually improved (Fig. 1). Engagement in FAIR + COPE is operationalized through defined feedback structures between data producers, consumers, and the communities establishing standards. Examples of these feedback loops already exist: Data competitions, like the Critical Assessment of Metagenome Interpretation19, provide insight and user research opportunities, and community-led standards development cataloging efforts12,18,20–24 that ensure data analysis and standards evolve alongside needs of the research. New resources continue to make discovery easier. For example, the Multi-Omics Metadata Standards Integration Working Group (MOMSI) surveyed the landscape of standards and generated a set of 250 standards, universal and omics specific, released as a collection (https:// fairsharing.org/5742)12, with user rdamomsi.github.io/Dashboard. Another example of deep community engagement was the development of the proposed nomenclatural code for uncultivated microbes, SeqCode25,26, and the widely utilized Genome Taxonomy Database (GTDB27; https://gtdb.ecogenomic.org), which provides consistently rank normalized genome-based taxonomy for prokaryotic genomes. exploration available and resource Engagement has always been a key component of the scientific process, typically at the beginning (proposal review) and the end of a study (publication review). FAIR + COPE elevates engagement as an essential component throughout the data lifecycle. The challenge will be to grow the existing culture of structured participation (i.e., review panels and journal reviews) into a more iterative, interactive culture, where creating and sharing comparable and organized data, evaluating diverse data sources for cross-study integration, and testing, validation, and updates of predictions from those data, is the norm. This is going to be even more critical in the age of AI, as human-in-the-loop validation and testing will be necessary to test AI-generated predictions.

  Advancing FAIR data towards comparable, organized, predictive AI-ready data for community validation · 2026 · DOI

• Fig. 1 | FAIR + COPE improvements to data enables iterative improvements to predictions that can be iteratively validated by the community. The FAIR data principles are necessary, but being able to make data Comparable and Organized for Predictive biology that can be validated by an Engaged community (FAIR + COPE) is an iterative process. Many resources support various aspects of FAIR + COPE, but the scientific community needs to invest in expanding and connecting those resources to fully enable the vision it embodies. Action and examples are provided to demonstrate feasibility. generated through testing and validation, updates to existing Comparable and Organized data need to be made and Predictive models need to be regenerated and re-evaluated. COPE (comparable, organized, predictive, and engaged) goes beyond being a set of principles; COPE acknowledges the active, iterative scientific process required to create and maintain the predictive inferences that act as the foundation of tools, parameters, and data products used for Artificial Intelligence (AI)-driven modeling and longterm data reuse. The COPE process for data integration goes beyond the FAIR principles The FAIR principles are intentionally high-level, providing a foundation for data discoverability and access. In practice, however, FAIR compliance is not universal and does not guarantee that datasets from different sources can be directly integrated for domain-specific predictive modeling or AIdriven analyses. While FAIR enables the linking of data products by persistent identifiers, FAIR + COPE further extends that linking to additional domain-relevant metadata (e.g., sampling conditions, experimental methods, units, etc.14) for the evaluation of comparability and establishing organizational relationships. Currently, the COPE process is often manual, focused on a particular research question, and done without full transparency or reproducibility. Even if the data are FAIR, integration of domainspecific data across studies still requires harmonization. The data must be converted to comparable units and methodological parameters, and organized with consistent ontological labels and standardized terms. The resulting predictive analyses are often done on local machines, sometimes using older database or tool versions, and testing and validation is performed by close collaborators / co-authors already engaged in the project. This siloed approach also highlights how implicit predictive inference, such as gene annotation, normalization, and labeling, is often not well documented and difficult to update, therefore making data difficult to reuse. COPE, as a mostly manual process, is time consuming, scope-limited, and hard to reproduce. To accelerate FAIR + COPE, application and assessment of units, parameters, standards, tools, and database updates must be automated, documented, and available for validation. Also, predictive assertions are not limited to analysis outputs of comparable, organized data; it also includes the necessary updates to labels or models that underpin what makes data comparable and organized. FAIR + COPE does not specify which predictive models should be used, only that the labels and units be made explicit, versioned, and retain a link to the original FAIR data they were derived from. Predictive requires Comparable data to define valid inputs, Organized structures to capture and propagate inferred relationships, and Engaged communities to iteratively evaluate, validate, and revise predictions as knowledge advances. This is what enables data from different origins to be confidently integrated, updated, and made ready for modeling and AI-workflows. Examples of automation using large language models (LLMs) already exist: leveraging FAIR-aligned workflows and tools to apply standardized labels to data (comparable) and rapidly identify contextuallyrelated research outputs (organized), when labeled by standard identifiers15,16. By embedding domain-aware AI tools, FAIR can be transformed into FAIR + COPE, a more automated process that reduces effort while maximizing synergisms across diverse data sources and data origins to enable novel exploration and Predictive analysis. By leveraging comparable, organized data, metadata, and provenance, reproducible Predictive analyses can generate models of biological systems with accompanying estimates of uncertainty, or reveal integration issues and inconsistent analysis.

  Advancing FAIR data towards comparable, organized, predictive AI-ready data for community validation · 2026 · DOI

• Though the ALP specifically used the Researcher Workbench data platform, the utilization of academic libraries as a central point for outreach and training could be replicated for future use with other data science program models.

  Advancing data-driven health research from the All of Us data training and engagement program · 2026 · DOI

• Though the ALP specifically used the Researcher Workbench data platform, the utilization of academic libraries as a central point for outreach and training could be replicated for future use with other data science program models.

  Advancing data-driven health research from the All of Us data training and engagement program · 2026 · DOI

• Expanding the geographical coverage particularly in regions with limited data, will ad- dress regional disparities and improve the dataset’s compre- hensiveness. Increasing the proportion of data with detailed methodolog- ical information while reducing the percentage of unspeci- fied or poorly documented entries will enhance the GHFDB’s https://doi.

  The 2024 release of the Global Heat Flow Database (GHFDB): quality assessment, metadata standards, and a century of geothermal data · 2026 · DOI

• 3.1. Fitness for purpose Regarding the prospective design for the Fitness-for-purpose assessment, at this stage we propose using the dqv:UserQualityFeedback class, targeting as well the dataset, and whose body and contents may further use specific classes and properties to represent the required components out of the QUANTUM specification provided in annex 3). 3.2. Assessment Status and Provisional Labels The QUANTUM RDF template may also include metadata indicating whether the generated assessment is final or provisional. This is useful because some dimensions may receive a score of zero not because the dataset performs poorly, but because those dimensions have not yet been assessed or the assessment workflow has not been completed. The RDF should therefore make the status of the assessment explicit, so that users and consuming systems can distinguish between a finalized label and a temporary or incomplete assessment. At the current stage, the RDF generated by the QUANTUM tool should be understood as a raw assessment output rather than a formally certified label. Since no external certifying authority or digital signature is currently attached to the generated RDF, the template should avoid representing provisional results as certified outputs. Instead, the assessment can include a human-readable note using dcterms:description or rdfs:comment, for example indicating that the label is temporary, draft, pending validation, or not yet finalized. Where useful, temporal metadata such as dcterms:issued or dcterms:valid may also be added to record when the assessment was generated and, if applicable, the period for which it should be considered valid. 3.3. Authorship of the label generation PROV-O can be used to document how the RDF assessment was produced. The generated assessment or quality annotation may be linked through prov:wasGeneratedBy to a generation activity representing the QUANTUM labelling tool. This allows the RDF to state that the output was automatically generated by the tool, while still leaving room for future extensions where a final label may be validated, digitally signed, or certified by an authorized body. In such future cases, the RDF could include additional provenance or certification metadata identifying the validating authority and the certification mechanism used. Task 3.2 explains how Health Data Access Bodies (HDABs) can practically support the full journey of an AI medical device, from development and testing to real-world use and later updates.

  Summary Report on the RDF Template for the QUANTUM DQ&U Label · 2026 · DOI

• on Open Science https://www.budapestopenaccessinitiative.org/read/ - https://openaccess.mpg.de/Berlin-Declaration - http://viennaprinciples.org/v1/ - https://doi.org/10.54677/MNMH8546. 9 Diagram by Bianca Gualandi, Mario Marino, Giulia Caldoni UNESCO Open Science Recommendation The term Open Science covers a variety of efforts and practices focused on making scientific research more open to society, thus making it transparent, accessible, collaborative, and reproducible. The aims of Open Science practices: • make multilingual scientific knowledge • openly available, accessible and reusable for everyone, increase scientific collaborations and sharing of information for the benefits of science and society, • open the processes of scientific knowledge creation, evaluation and communication to societal actors beyond the traditional scientific community.

  2026-05-21_Data-Management-Plans-Psicostat-UNIPD-Vasone-Visentin · 2026 · DOI

• Calibration and measurement capabilities The Mutual Recognition Arrangement of the International Committee for Weights and Measures (CIPM-MRA) provides a framework through which NMIs demonstrate the equivalence of their measurement results. This equivalence is declared in the form of Calibration and Measurement Capabilities (CMCs). Currently, the integration of the RM Explorer apps with the BIPM Key Comparison database (BIPM-KCDB) enables the app to display the list of CMCs associated with the NRC. As the BIPM-KCDB evolves, the joint use of chemical identifiers and machinereadable metadata could enable further integration, allowing one to link each certified property in a reference material to a CMC that supports it.

  Establishing digital reference material documentation infrastructure for chemical metrology · 2026 · DOI

• Accurate and efficient data extraction and analysis from cloud in- cident reports are important for improving the dependability of cloud computing services. To address this challenge, we propose a 020406080100Cost0.650.700.750.800.850.90AccuracyAvg. Cost: 30.48Avg. Acc.: 0.83AWSzero-shotfew-shotGPT3.5GPT4oClaude3.5Claude4Gemini2.0Gemini2.50255075100125150175Cost0.630.650.680.700.730.750.78AccuracyAvg. Cost: 61.22Avg. Acc.: 0.72AZUREzero-shotfew-shotGPT3.5GPT4oClaude3.5Claude4Gemini2.0Gemini2.5020406080100120140Cost0.550.580.600.620.650.680.700.73AccuracyAvg. Cost: 51.89Avg. Acc.: 0.66GCPzero-shotfew-shotGPT3.5GPT4oClaude3.5Claude4Gemini2.0Gemini2.5 Leveraging LLMs for Structured Information Extraction and Analysis from Cloud Incident Reports (Work In Progress Paper) ICPE Companion ’26, May 04–08, 2026, Florence, Italy methodology that demonstrates how to leverage LLMs for struc- tured data extraction. In this work, we collect 3,000 incident reports from three cloud operators, and annotate 460 of them for evaluation. We propose five prompt components and design six strategies. Using both lightweight and advanced LLMs, we then develop data extraction pipeline to extract ten types of information from textual incident reports. After that, we evaluate and compare the accuracy, latency, and cost of six LLMs, providing insights into prompt and model selection adapting to different requirements in practical report ex- traction. Overall, we summarize 6 key findings, and provide open- source artifacts as valuable resources for system researchers, cloud engineers, and service users to better understand and improve cloud incident management. Our future work includes: (1) Optimized evaluation of prompts. Further research is required to optimize prompt design, and to evaluate different components through controlled experiments. (2) Proactive incident prediction. The extracted information can be fur- ther used to predict incident duration and root causes, which helps better demonstrating downstream utility in incident mitigation. (3) Advanced LLM techniques. More advanced LLM techniques, such as fine-tuning and Retrieval Augmented Generation (RAG), can be applied to further improve extraction accuracy in complex fields thorough historical patterns.

  Leveraging LLMs for Structured Information Extraction and Analysis from Cloud Incident Reports (Work In Progress Paper) · 2026 · DOI

• Establishing how EOSC, Copernicus, Destination Earth, EuroHPC and AI Factories can function as truly complementary layers needs clarification on integration mechanisms and operational coordination.

  Strengthening Europe’s sovereignty and interoperability in Earth observation data · 2026 · DOI

• Additional findings regarding social scientists' data-sharing behaviors include: (1) those who do share qualitative data in data repositories are more likely to share their research tools than their raw data; and (2) perceived technical support and extrinsic motivation are both strong predictors of qualitative data sharing (a previously underresearched subtype of social science data sharing).

  Surveying research data-sharing practices in US social sciences: a knowledge infrastructure-inspired conceptual framework · 2022 · DOI

• Research data management is essential for high-quality reproducible research, yet relatively little is known about how research data management is practiced by graduate students in Civil and Environmental Engineering (CEE).

  Understanding Research Data Practices of Civil and Environmental Engineering Graduate Students · 2022 · DOI

• Further work on RO-Crate profiles include to formalise links to the API operations and repositories [FDOF5,FDOF7], to include PIDs of profiles and types in the FAIR Signposting, and HTTP navigation to individual resources within the RO-Crate.

  Creating lightweight FAIR Digital Objects with RO-Crate · 2022 · DOI

• BACKGROUND AND OBJECTIVE: There is currently no standardised way to share information across disciplines about initiatives, including fields such as health, environment, basic science, manufacturing, media and international development.

  Standardised data on initiatives—STARDIT: Beta version · 2022 · DOI

• This paper first outlines challenges experienced by researchers engaged in a large-scale coding project; then highlights valuable lessons learned in large-scale coding projects; and finally discusses opportunities for further research on comparative case study analysis focusing on social-ecological systems and common pool resources.

  Challenges and opportunities in coding the commons: problems, procedures, and potential solutions in large-N comparative case studies · 2016 · DOI

• Findings are that RDA is appropriate for describing alternative publications, though expansion and improvement is warranted in documenting makers, addressing intellectual property, approaching privacy concerns, facilitating subject and genre analysis, undertaking object cataloging, applying companion standards and external vocabularies, and using RDA where boundaries between work, expression, and manifestation are blurred.

  RDA and the Description of Zines: Metadata Needs for Alternative Publications · 2014 · DOI

• The following article addresses this gap in the literature by analyzing the FRAD conceptual model, examining its applicability to an authority file for manuscripts, and proposing a way to implement and display this entity-relationship model in a local authority file.

  Applying the FRAD Conceptual Model to an Authority File for Manuscripts: Analysis of a Local Implementation · 2009 · DOI

• Findings The paper finds that significant support was expressed for the provision of bi‐directional links between source and output repositories, tempered by a limited knowledge of repositories among the survey constituency and the need for reassurance on measures for the protection of data ownership.

  Project StORe: making the connections for research · 2007 · DOI

• A conclusion attempts to draw up a balance‐sheet of CURL involvement in resource description and discovery, summarising what has been achieved and what remains to be done.

  CURL and resource description and discovery · 1999 · DOI

• The Conference agreed that action be taken during 1963-64 on the following points : 1) Development of new archives and systems of information exchange between archives : .’ It was strongly recommended that studies be carried out of practicability of establishing further archives for survey data in Europe, and that a blueprint be worked out through consultations with all : interested parties for a system of cooperation and exchange between such data archives. Steps should be taken to ensure that all new archives to be established--whether nationally or by specific fields--be urged to join such a system. It was also recommended that any European system of archives should establish close cooperation, and arrange for exchange of information and data, with all major archives outside of Europe. The hope was expressed that all archives in Europe would make their collections accessible to the scholarly community on the freest possible basis. It was further recommended that the work currently under way on methods of electronic data retrieval of survey materials be given the fullest possible support, and that efforts in this direction in Europe and the United States of America be effectively coordinated. 2) Action to promote the utilization of survey archives : The Conference welcomed the initiative taken by the University of Cologne and supported by UNESCO for a European Training Seminar in Survey Techniques to be held for the first time in 1964. The Confe- rence stressed the importance of training in comparative cross-national survey analysis, and recommended that efforts be made to organize such Seminars in a regular sequence at a number of European research centres. The Conference also welcomed the initiative taken by the University of Amsterdam to establish under the auspices of NN’apor an international periodical for current information on survey questions and response distribution, and recommended that this journal be made an organ for information on archive developments. The Conference called upon all research centres in Europe to cooperate in this venture, and to communicate the results of surveys to the journal.

  Conference on Data Archives in the Social Sciences · 1963 · DOI

Most-cited papers in Research Data Management Practices

• The PRIDE database at 20 years: 2025 update · Nucleic Acids Research · 2024 · 1,817 citations
• Datasheets for datasets · Communications of the ACM · 2021 · 1,324 citations
• The conundrum of sharing research data · Journal of the American Society for Information Science and Technology · 2012 · 597 citations
• Exploring and retrieving sequence and metadata for species across the tree of life with NCBI Datasets · Scientific Data · 2024 · 501 citations
• The NHGRI-EBI GWAS Catalog: standards for reusability, sustainability and diversity · Nucleic Acids Research · 2024 · 366 citations
• Indigenous Data Sovereignty in the Era of Big Data and Open Data · Australian Journal of Social Issues · 2020 · 161 citations
• Managing by Data: Algorithmic Categories and Organizing · Organization Studies · 2020 · 102 citations
• GenBank 2025 update · Nucleic Acids Research · 2024 · 99 citations
• Data, measurement and empirical methods in the science of science · Nature Human Behaviour · 2023 · 94 citations
• Digital Extended Specimens: Enabling an Extensible Network of Biodiversity Data Records as Integrated Digital Objects on the Internet · BioScience · 2022 · 94 citations

Most recent work

• Paper 14 - Government Structural Reference Infrastructure for AI and AGI Environments · Zenodo (CERN European Organization for Nuclear Research) · 2026
• LEE SHARKS — Knowledge Graph and Metadata Packet for AI Indexing: Canonical Author Identity, Disambiguation Protocol, and Entity Map (EA-KG-01) — Crimson Hexagonal Archive · Zenodo (CERN European Organization for Nuclear Research) · 2026
• A collaborative submission model for building high-quality data resources at scale through partnership · bioRxiv · 2026
• Integrating the biodiversity genomics continuum: harmonising data from barcodes to reference genomes · Research Ideas and Outcomes · 2026
• Evaluating the role of metadata standards in enhancing data discoverability and interoperability in academic digital repositories · Digital Library Perspectives · 2026
• Strengthening Europe’s sovereignty and interoperability in Earth observation data · 2026
• Evolve with your research: stepwise system evolution from document-driven to fact-centric research data management in materials science · Journal of Cheminformatics · 2026
• Requirement Analysis and Use Case Engineering via ethnographic methods – Negotiating FAIRness for Sensitive Interview Transcript Data · Zenodo (CERN European Organization for Nuclear Research) · 2026
• Proposed curriculum for library and information science (LIS) education to meet the emerging trends in artificial intelligence (AI) · Digital Library Perspectives · 2026
• Data Literacy for the 21st Century: Perspectives from Visualization, Cognitive Science, Artificial Intelligence, and Education · 2026

Related topics in Computer Science