2.1 Data Sharing Policies

Data sharing is widely regarded as vital for ensuring reproducibility and has been adopted by the majority of ecological researchers (Herold 2015). The current standard practice for eDNA datasets is that upon publication of an associated manuscript, they are stored in digital repositories where they can be accessed by other researchers (Abarenkov et al. 2023; Leigh et al. 2024; Shea et al. 2023), but there is little oversight into which data are archived and how accessible they are. The establishment of public data archiving policies and requirements by scientific journals has been a key contributor to advancing data sharing practices (Sholler et al. 2019; Thrall et al. 2023; Whitlock 2011; Whitlock et al. 2010). Open-access journals such as GigaScience and GigaByte further support these efforts by providing dedicated data repositories (GigaDB) to ensure that publications are accompanied by supporting data that uphold principles of transparency, reproducibility, and FAIR standards (Armit et al. 2022). Additionally, funding agencies have also effectively implemented data sharing policies at national and international levels in various countries (“Australian Research Council,” 2021; “Open Research Europe | Open Access Publishing Platform,” 2025; “Proposal & Award Policies & Procedures Guide (PAPPG) (NSF 24-1) | NSF – National Science Foundation,” 2024; Hutchison et al. 2024).

These proactive approaches have led to a broader adoption of open data practices and a substantial increase in the proportion of published articles sharing their data. For instance, when journals mandate data archiving, there is a nearly 1000-fold increase in the ability to find data online compared to data in journals without a policy (Vines et al. 2013). Despite this progress, challenges to full FAIR principles compliance remain as the process is not yet fully standardized or unambiguously defined. For example, many datasets fall short due to inconsistent formatting and sparse metadata, limiting their interoperability and reusability (Hassenrück et al. 2021; Roche et al. 2015; Tedersoo et al. 2021). This may be due to a lack of awareness and familiarity around data standardization, existing guidelines, and the benefits of FAIR data sharing. However, existing FAIR guidelines for standardized formatting and sharing can be dissonant. Moreover, current metadata checklists (i.e., lists of data terms required and recommended for describing methods and datasets) implemented in standards are not always suitable for documenting the entire eDNA workflow. Nevertheless, each dataset published without meeting the FAIR principles can present a lost opportunity to preserve rich biodiversity data for reuse in research and decision making.

2.2 Data Standards

The establishment and adoption of standardized data formats, vocabularies, and comprehensive metadata checklists provide a robust foundation for ensuring that archived data are machine readable, interoperable, and suitable for reuse (Vines et al. 2013). Darwin Core (DwC, dwc.tdwg.org), developed by Biodiversity Information Standards (TDWG, tdwg.org) and Minimum Information about any (x) Sequences (MIxS, genomicsstandardsconsortium.github.io/mixs), developed by the Genomic Standards Consortium (GSC, gensc.org), are data standards to facilitate FAIR practices for biodiversity and DNA sequence data (Wieczorek et al. 2012; Yilmaz et al. 2011). Widely adopted by major data infrastructures like the INSDC (www.insdc.org) (Arita et al. 2021) and the Global Biodiversity Information Facility (GBIF, www.gbif.org), these standards can be powerful enablers of FAIR data practices. Furthermore, interoperability between DwC and MIxS has been established by a mapping of MIxS terms to a DwC extension (Meyer et al. 2023). The DNA-derived data extension to DwC was developed to accommodate biodiversity data in DwC-based biodiversity platforms like GBIF. The extension is currently maintained by GBIF (http://rs.gbif.org/terms/1.0/DNADerivedData) and consists of terms from MIxS, the Global Genome Biodiversity Network (GGBN) standard (https://www.ggbn.org/ggbn_portal), the Minimum Information for publication of Quantitative real-time PCR Experiments (MIQE) Guidelines (Bustin et al. 2009), and a few custom terms (e.g., atomised versions of the primer terms on MIxS) (Abarenkov et al. 2023).

Despite dynamic community-driven development (Abarenkov et al. 2023; Bustin et al. 2009; Meyer et al. 2023; Wieczorek et al. 2012; Yilmaz et al. 2011), these existing data standards lack unambiguous and comprehensive terms, vocabularies, and guidance to fully capture some of the distinctive features of eDNA data and workflows. For instance, qPCR-based eDNA workflows, first introduced around 2010 (e.g., Ficetola et al. 2008; Thomsen et al. 2012), have been increasingly used for targeted assay detection (Takahashi et al. 2023). The most comprehensive qPCR checklist (i.e., the MIQE Guidelines) was, however, developed prior to the prevalence of eDNA studies and may need to be revised and updated to align with FAIR principles for qPCR-based eDNA reporting. Furthermore, procedures for monitoring contamination and excluding nontarget taxa, and the parameters used for quality filtering and species detection vary greatly between studies, depending on the project scope and the associated financial and ecological costs of false positives and false negatives (Ficetola et al. 2015; Takahashi et al. 2023). Making detailed workflow descriptions FAIR can enhance future studies' ability to reuse data and ensure high confidence levels in species detection and taxonomic assignment. Increasing efforts have been made to establish minimum reporting requirements to validate eDNA study methods and data (Abbott et al. 2021; Borisenko et al. 2024; Gagné et al. 2021; Kimble et al. 2022; Klymus et al. 2020; Thalinger et al. 2021). As a next step, these requirements could be translated into data standards and formats to enhance human and machine readability, interoperability, and reusability.

Like other machine-derived data, eDNA datasets are comprised of a variety of data components representing degrees of processing and refinement. Therefore, defining the essential data components to be shared is crucial to ensure that shared data are standardized, complete, and reusable. Components of eDNA datasets can be broadly grouped into three categories: (1) raw data, including amplification data (i.e., Ct/Cq values from targeted assay detection studies) and raw DNA sequences (i.e., FASTQ files from metabarcoding studies), (2) interpreted data, including estimated DNA copy numbers from standard curves and detections (1/0) of targeted taxa, raw and curated ASV/OTU tables, and taxonomic assignment of each ASV/OTU, and (3) metadata describing data content, provenance, methods, and associated environmental data for each project and sample. Each data type serves a different purpose and possesses distinct qualities. For example, raw data are essential for documentation and reproducibility, as well as for enabling future re-analysis with different or updated bioinformatic pipelines. In contrast, interpreted data and metadata provide immediate value for broader biodiversity research and data-driven decision-making. Sharing all these data components in standardized formats is crucial to making data FAIR.

2.3 Cyberinfrastructure

There are a range of digital repositories available for depositing eDNA datasets, eliminating the need for a new or single repository. Instead, fostering collaborations among databases could ensure that shared eDNA datasets remain interoperable and reusable. These repositories can broadly be categorized into sequence databases and biodiversity databases. The INSDC sequence databases, which are a collaboration between the US National Center for Biotechnology Information (NCBI, Sayers et al. 2024), the European Nucleotide Archive (ENA, www.ebi.ac.uk/ena) (Yuan et al. 2024), and the DNA Data Bank of Japan (DDBJ, www.ddbj.nig.ac.jp) (Kodama et al. 2024), aim to compile and disseminate DNA sequence data and associated metadata generated worldwide (Arita et al. 2021). Data archived in NCBI, ENA, and DDBJ are routinely exchanged and synchronized among these databases. Biodiversity databases such as GBIF, on the other hand, serve as an open-access online platform for global biodiversity data, with a primary focus on species occurrences. The types of data available in GBIF include digitized specimens from natural history collections, observations from citizen science platforms (e.g., eBird (ebird.org), iNaturalist (inaturalist.org)), and DNA-associated records (e.g., barcoded specimens from the BOLD database (Ratnasingham and Hebert 2007) and records from INSDC (Arita et al. 2021)), as well as metabarcoding and metagenomic datasets from individual studies (e.g., Shea and Boehm 2024) and initiatives such as MGnify (Richardson et al. 2023) (refer to Abarenkov et al. 2023). Other biodiversity databases based on the Darwin Core standard for biodiversity data, such as the Ocean Biodiversity Information System (OBIS, obis.org) and the Atlas of Living Australia (ALA, www.ala.org.au) focus on distinct ecosystems or regions and routinely exchange datasets. Given that eDNA data encompass both biodiversity information (e.g., inferred species occurrences) and molecular sequence data (e.g., raw reads), an optimal data standard would allow for describing eDNA datasets intended to be shared through both biodiversity and nucleotide sequence databases.

Publishing eDNA data in established databases can be a simple and powerful step toward achieving FAIR eDNA, as these platforms provide persistent sample and sequence identifiers across datasets, and offer robust, indexed search functions based on user-defined parameters, such as geographical regions, ecosystem types, taxa and gene regions, and making data highly findable and accessible. Moreover, eDNA data shared in biodiversity databases like GBIF and OBIS contribute to the global biodiversity records by adding DNA-derived species occurrences alongside data from other sources. These databases do not accept all three primary eDNA data components (metadata, raw data and interpreted data) as a whole, and therefore, require linkage if data are to be accessible for re-use. For example, GBIF focuses on inferred biodiversity data, such as detected species or taxa with spatiotemporal metadata, and representative curated DNA sequences. It, however, does not host raw DNA sequencing files (i.e., FASTQ files), but may link to the corresponding INSDC records. Conversely, INSDC databases host raw sequence data, but neither the inferred biodiversity data in an interoperable format nor eDNA-derived species occurrence data without DNA sequences, such as qPCR-based detections. Collaboration efforts between GBIF and INSDC (through ENA) are in progress to ensure cross-linking of related data resources and increase interoperability between the two data infrastructures (“Cross-infrastructure collaboration with ENA improves processing, quality of DNA-derived occurrences” 2021).

In contrast, eDNA practitioners can deposit their eDNA data as a whole in open data repositories, such as Dryad (https://datadryad.org), Zenodo (https://zenodo.org), or FigShare (https://figshare.com), which accept unrestricted data types. Data in these repositories, however, are not fully FAIR, as they are not standardized by default and are not indexed to a degree where they are functionally findable and interoperable. Hence, data reusers are required to manually access and reprocess the data before it can be re-analyzed and/or combined with other datasets. As a consequence, eDNA data published in these repositories are at risk of getting lost and becoming “dark data” (Heidorn 2008).

More recently, various infrastructures have been developed specifically to present eDNA data from particular regions, institutions, or taxa, for example, All Nippon eDNA Monitoring Network (ANEMONE, db.anemone.bio), Wilderlab (www.wilderlab.co.nz/explore), eDNAtlas (www.fs.usda.gov/research/rmrs/projects/ednatlas, (Young et al. 2018)), eDNA Explorer (www.ednaexplorer.org), the Australian Microbiome (www.australianmicrobiome.com), Swedish ASV portal (https://asv-portal.biodiversitydata.se), ASV Board (https://asv.bolgermany.de/metabarcoding), Nonindigenous Aquatic Species (NAS) database (https://nas.er.usgs.gov/eDNA), and the global microbiome database under the National Microbiome Database Collaborative (NMDC, data.microbiomedata.org). These databases utilize custom data structures to serve their particular focus and functionalities, sometimes in combination with terms from more widely established data standards like MIxS and DwC. Publishing data from such databases to global databases (e.g., INSDC and GBIF) requires mapping custom data terms to an open data standard, which can be a lossy process but serves the purposes of discoverability and inclusion in the default global data pools.

While the aforementioned databases and infrastructures share the goal to promote open and data-driven science, they also cater to specific focuses and requirements. Consequently, a spectrum of repositories exists, ranging from those accepting any research data in unrestricted formats (e.g., Dryad, Zenodo, FigShare) to those exclusively processing eDNA-derived data from a single source using standardized protocols (e.g., ANEMONE and Wilderlab). Because of the numerous repositories, there is likely not a need for a new or single repository. Instead, fostering collaborations among eDNA communities, databases, data standard organizations, and journal editors and publishers could ensure that shared eDNA datasets are comprehensive, interoperable, and reusable. This involves integrating eDNA data into existing databases and standards and mapping it across past, current, and future standards. This approach retains flexibility and interoperability with existing repositories while supporting the formatting of shared datasets into common standards. This, in turn, facilitates data mobilization, collation, and reuse.

4.1 Data Formats and Components

4.2 Consistent Identifiers and File Naming

To ensure machine readability and long-lasting reference to a digital resource, consistent, persistent sample and sequence identifiers (samp_name, lib_id and seq_run_id) across datasets (Damerow et al. 2021; McMurry et al. 2017) must be used. Similarly, each file must have a clear, unambiguous name, consisting of project_id, assay_name, and seq_run_id (refer to Table 2 and example datasets) to maintain clear links between files. For example, a curated ASV/OTU table should be named as otuFinal_<project_id>_<assay_name>_<seq_run_id> (e.g., otuFinal_gbr2022_MiFish_lib20230922.csv). If multiple data components are stored within a single spreadsheet workbook, the file name should be the project_id (e.g., gbr2022.xlsx). In this case, each worksheet name should follow the format in Table 2 without the project_id as it already appears in the main file name (e.g., otuFinal_MiFish_lib20230922).

4.3 Metadata Checklist

We developed a FAIR eDNA (FAIRe) metadata checklist, providing a comprehensive vocabulary for describing different data components and methodologies (Appendix S1). Thorough documentation promotes improved transparency and reproducibility of studies, as well as enabling the evaluation of data suitability for specific reuse cases. The FAIRe checklist consists of 337 data terms (38 mandatory, 51 highly recommended, 128 recommended and 120 optional terms), organized into workflow sections (e.g., sample collection, PCR, bioinformatics). Most of the mandatory, highly recommended, and recommended information is generated naturally during the course of any project. As a result, data meeting the minimum reporting requirements can be readily made available and submitted to public repositories.

4.4 Data Submission/Publication

Publishing eDNA datasets to databases such as GBIF, OBIS, and INSDC is an essential step in ensuring FAIR data practices. These platforms ensure data findability and accessibility for scientific research and decision making that rely on open data, either through application programming interfaces (APIs) or sophisticated web-browser interfaces. They offer data validation procedures during submission and provide additional standardization to ensure interoperability. Persistent sample and sequence identifiers are provided upon the submission of data to these databases, which are essential for machine-readability and long-lasting reference to the data. For example, in GBIF, dataset authors and publishers receive credit through unique dataset Digital Object Identifiers (DOIs) that enable citation tracking and automatic usage reporting.

The guidelines provided here result in data formatted slightly different from those accepted by nucleotide and biodiversity databases. Hence, minor reformatting is required. One example is the necessary data format when multiple assays were applied in a single project. GBIF recently launched the Metabarcoding Data Toolkit (MDT), a user-friendly web application (https://www.gbif.org/metabarcoding) that reshapes tabular metabarcoding data—similar to the format proposed here—and publishes it to GBIF and OBIS complying with the guidelines in Abarenkov et al. (2023). The GBIF MDT currently requires a separate input dataset for each assay (hence information from project and sample metadata are repeated), whereas the FAIRe checklist allows the metadata from multiple assays to be combined (Figure 5). An R-script (FAIRe2MDT) has been developed (R Core Team 2024) to convert FAIRe data templates for GBIF and OBIS publication via MDT, bridging the differences in data formats between FAIRe and GBIF standards (refer to section ‘Available example datasets, scripts and tools’). Using these tools in combination streamlines the publication of eDNA data to GBIF and OBIS. Overall, there is currently a strong effort by biodiversity data platforms to improve the suitability of depositing eDNA-based data.

4.5 Available Example Datasets, Scripts and Tools

5.1 Raising Awareness

Our next step focuses on raising awareness of FAIR principles, existing guidelines, and available tools to promote their adoption with the eDNA communities. To achieve this, the lead author will host workshops with eDNA practitioners over the coming year, offering hands-on opportunities to introduce FAIR data principles and provide practical guidance on implementing standardized formatting protocols. The first workshop took place at the eDNA conference in Wellington, New Zealand, in February 2025. More than a training exercise, these workshops enable establishing direct communications, identifying communities' needs, and addressing challenges to adopting FAIR data practices within their workflows. This step is particularly important as open and FAIR data can only be achieved through eDNA practitioners' collaborative efforts and relies on their commitment to making data accessible for broader use.

Several studies have explored Open Science behaviors and identified the barriers among scientists (e.g., Norris and O'Connor 2019; Tenopir et al. 2015). These barriers include technical and resource limitations, such as lack of standards, tools, time and skills to navigate required data management systems (Tedersoo et al. 2021; Tenopir et al. 2015). Additionally, multiple perceived barriers exist, including unawareness of the value of data for others, fear of scrutiny due to potential mistakes, resistance to openly sharing data given large efforts invested to secure funding in the face of limited resource for scientific research and resistance to changing existing practices (Norris and O'Connor 2019; Tenopir et al. 2015).

To address these challenges, communicating the diverse benefits of data sharing beyond reproducibility and documentation is important. In addition to contributing to the public good, data sharing leads to various benefits to data providers. These include increased visibility and citation of associated work (Colavizza et al. 2020; Piwowar et al. 2007; Wood-Charlson et al. 2022), direct citation metrics of data accessed through databases (e.g., GBIF), and enhanced collaborations and co-authorships, which all lead to improved professional stature (Bethlehem et al. 2022; McKiernan et al. 2016; Piwowar and Vision 2013; Whitlock 2011). Many eDNA practitioners, however, may not yet be fully aware of these benefits. Establishing open and direct communications with them through workshops will help us identify their unique barriers and needs, work collaboratively to overcome them, and foster a culture of FAIR data practices in the community.

5.2 Revising, Updating and Integrating the Guidelines

As eDNA science is relatively new and rapidly evolving, new methods and technologies will emerge. To improve and adjust the FAIRe metadata checklist and formatting guidelines and keep them relevant, regular revisions and updates are essential. Input from data providers and reusers is vital to this process, and we welcome feedback at any time via https://github.com/FAIR-eDNA/FAIR-eDNA.github.io/issues. Updated guidelines will be available at https://fair-edna.github.io.

Leveraging the network of the co-authors (e.g., GSC, TDWG, ENA, GBIF, OBIS, and ALA), we aim to continuously revise and co-develop the FAIRe metadata checklist to integrate with established data standards and databases (Figure 6). Implementation and alignment of FAIRe procedures and formats will be supported through contributions to the existing guidelines for publishing DNA-derived biodiversity data (Abarenkov et al. 2023), and development of scripts, like the FAIRe-ator and FAIRe2GBIF, and tools like the GBIF MDT. Integration of the FAIRe checklist into the GSC suite (i.e., MIxS), that will facilitate its adoption within the INSDC system, will be pursued through collaboration with the GSC MIxS Compliance and Interoperability Working Group (CIG) and with input from TDWG (Figure 6).

In the scope of existing agreements between TDWG and GSC to collaborate on the development of specifications for sequence-based biodiversity data, we plan to support the integration of the FAIRe checklist into the DwC DNA Derived Data extension. Continued revision, testing, and alignment with these organizations' objectives will help ensure successful implementation. Important milestones include the publication of updated guidelines for publishing DNA-derived biodiversity data (Abarenkov et al. 2023), the publication of the eDNA MIxS checklist, and the publication of the FAIRe-aligned DwC DNA Derived Data extension. Achieving these milestones will serve as evidence that the FAIRe checklist is adequate and accepted by the eDNA community. Long-term maintenance is planned to be managed by these consortia and organizations, but will depend on sustained engagement and interest from the eDNA science community in the FAIRe checklist.

We have developed the current guidelines with great interest to integrate them into journal publication requirements. We have collaborated with Wiley and Environmental DNA to develop a roadmap for integrating FAIR data principles into their publication procedures. This roadmap involves three key steps: (1) strongly recommending adherence to the FAIRe guidelines upon journal article publication, (2) making compliance mandatory in due course, and (3) extending these requirements to additional journals including Molecular Ecology and Evolution. While most journals mandate data sharing, submitting data in a FAIR manner is not yet a widespread practice (Roche et al. 2015). Implementation of the FAIR data guidelines could be an effective strategy given the large number of eDNA studies published each year with the tremendous volume of genetic and species occurrence data they generate (Takahashi et al. 2023).

5.3 Harmonizing FAIR and CARE

The CARE principles, which stand for Collective benefit, Authority to control, Responsibility, and Ethics, were developed by the Global Indigenous Data Alliance to protect the Indigenous rights and interests in Indigenous data, which include information, data, and traditional knowledge about their resources and environments (Carroll et al. 2021, 2020). The people- and purpose-oriented CARE principles are designed to complement the data-centric FAIR principles by ensuring that Indigenous data remain FAIR while centering Indigenous sovereignty (Mc Cartney et al. 2023). Implementing the CARE principles into FAIR data initiatives can ensure that data use respects Indigenous rights, serves a meaningful purpose, and promotes the wellbeing of Indigenous Peoples and Local Communities (IPLCs) (Carroll et al. 2021; O'Brien et al. 2024).

Mc Cartney et al. (2023) established a roadmap for the lifecycle of eukaryotic biodiversity sequencing data following the CARE principles. It identifies actions required from researchers throughout each study to build sustainable partnerships with IPLCs. In the context of FAIR initiatives, it is important to develop culturally aware linked metadata. For example, generalizing or withholding culturally sensitive data, such as geographical coordinates of culturally important sites, is respected and accepted, which can be mapped under the terms dataGeneralization and informationWithheld in the FAIRe metadata checklist and other standards and databases (Chapman 2020). Information on access, use permission, and the person or organization owning and managing data rights, should be defined at the start of each project, and recorded under the terms accessRights and rightsHolder. Further works are needed in contextual metadata to document Traditional Knowledge Labels and Notices (Liggins et al. 2021; “Local Contexts – Grounding Indigenous Rights” 2023), provenance about place and people (Mc Cartney et al. 2023), and the cultural importance of species (Reyes-García et al. 2023). Meetings, workshops and collaboration across various stakeholders including IPLCs and eDNA communities can elucidate the benefits of eDNA data sharing in various aspects, exchange more information and raise discussions. All these efforts can harmonize FAIR and CARE, establish and retain the link between people and nature, and serve as a vehicle for knowledge transfer, capacity development, two-way science and biodiversity conservation.