Biodiversity research is entering a new era in which molecular data form a continuous spectrum rather than a set of isolated approaches. At one end lies DNA barcoding (Hebert et al. 2003), which uses short, standardised genetic markers to identify species and quantify community composition through metabarcoding. At the other end lies reference genome sequencing (Supple and Shapiro 2018), which aims to reconstruct full chromosomal assemblies for selected taxa. Although these two approaches differ in scope and scale, their convergence is increasingly evident in how data are managed, analysed, validated and shared. Together, they form a complementary strategy for documenting life, from local inventories to complete genomic blueprints.

DNA barcoding, spearheaded by initiatives such as the International Barcode of Life (iBOL, International Barcode of Life consortium (2025)), has transformed species identification by enabling scalable, sequence-based taxonomy. Meanwhile, large-scale genome sequencing efforts, such as the European Reference Genome Atlas (ERGA, Mazzoni et al. (2023)), have redefined standards for completeness, accuracy and contextual metadata. Both depend on the same principles: transparent provenance, persistent identifiers and interoperable repositories. These principles create fertile ground for alignment. This convergence represents a philosophical shift towards treating biodiversity data as a shared digital ecosystem.

The key commonality lies in data handling. Whether assembling gigabase-scale genomes or processing short barcode reads, both fields require consistent metadata capture, version-controlled analytical workflows and reliable publication in trusted repositories such as the European Nucleotide Archive (ENA, O’Cathail et al. (2025)) and the Barcode of Life Data Systems (BOLD, Ratnasingham and Hebert (2007)). The adoption of community-driven tools and standards, such as Galaxy (Goecks et al. 2010) for workflow execution, WorkflowHub (da Silva et al. 2020) for sharing and documenting pipelines, PlutoF (Abarenkov et al. 2010) and COPO (Shaw et al. 2024) for data management and metadata negotiation and systems such as UNITE (Abarenkov et al. 2024) and BOLD for taxonomically anchored sequence data, creates bridges between communities that once operated in parallel. Through these connections, molecular data can flow seamlessly between generation, processing, validation, publication and application.

This shared informatics foundation enables coordinated scaling of biodiversity genomics. Automated validation pipelines, harmonised metadata vocabularies and FAIR (Findable, Accessible, Interoperable, Reusable, Wilkinson et al. (2016)) data practices ensure that sequence data from diverse origins can be reused across the biodiversity research domain. Crucially, convergence in data handling strengthens the link between specimens, sequences and analyses, preserving provenance even as datasets grow in complexity and size. The result is a knowledge system in which short barcode markers can guide genome sequencing priorities and reference genomes can refine barcode-based identification, with each informing the other through shared digital infrastructure.

This report describes how these conceptual alignments translate into practical workflows and shared infrastructure. They show that the union of barcoding and genome sequencing comprises a synchronisation of data logic: a shift from separate production lines to a common architecture for biodiversity knowledge.

The convergence between DNA barcoding and genome sequencing becomes tangible in how the data are handled after laboratory work ends. Despite their differences in scope, with barcodes distilling diversity into compact diagnostic sequences, while genomes capture it at full chromosomal resolution, both approaches pass through strikingly similar stages of digital stewardship. What unites them is not the size of their data, but their logic: structured data processing and validation, transparent provenance and deposition into shared repositories where the information can be discovered, reused and connected across research domains.

Both barcoding and genome sequencing workflows begin with an initial round of pre-processing quality control to verify that sequencing reads meet baseline standards of completeness and accuracy. From that point, they follow parallel analytical paths that differ mainly in technical detail. Barcoding workflows assemble reads into high-fidelity consensus markers for species identification and community profiling, while genome workflows assemble larger contigs and scaffolds. Then, a consensus is computed, which is annotated. The result is validated and submitted (see Fig. 1). In both cases, every computational step is scripted, version-controlled and logged to ensure reproducibility and to maintain an unbroken record of data provenance.

The key to this shared logic lies in metadata management. Each dataset is accompanied by structured information about its origin, processing and validation, ensuring that specimens, sequences and analyses remain linked throughout their lifecycle. Tools such as COPO, a metadata brokering system that supports FAIR-compliant submission to public repositories and PlutoF, which manages collection of metadata for barcoding and metabarcoding data, illustrate how distinct communities can align around compatible standards. Once validated, data and metadata are deposited in trusted archives, such as ENA for raw sequences and assembled genomes, UNITE for fungal metabarcoding data and BOLD for specimen barcodes, forming an interconnected landscape of biodiversity genomics.

Beyond these repositories, a growing ecosystem of workflow platforms reinforces convergence in analytical practice. The Galaxy environment enables researchers to run, share and reproduce pipelines transparently, while WorkflowHub provides a registry for documenting, crediting or attributing and versioning these pipelines as community assets. Both barcoding and genome sequencing initiatives now adopt these tools to promote openness and interoperability. Increasingly, workflow outputs, metadata and provenance records are bundled into Research Object Crates (RO-Crates, Soiland-Reyes et al. (2022)), i.e. lightweight digital containers that capture every element of a computational process, from raw data to final results, in a machine-readable format. This approach ensures that datasets can travel intact between systems and over time without loss of context.

When viewed through the lens of data handling, the two approaches to biodiversity genomics reveal a single conceptual architecture. Each begins with sequence validation, proceeds through structured assembly and validation pipelines and ends in the publication of self-contained research objects that embody the principles of transparency, reproducibility and reuse. These parallels do more than simplify logistics: they enable interlinking barcode-based surveys with genome assemblies, allowing the same digital infrastructure to support biodiversity discovery, ecological monitoring and evolutionary research. What once were separate data production lines are now strands of the same informatic fabric, woven together by shared standards and a commitment to open, reproducible science.

Translating the shared logic of biodiversity genomics data handling into practice requires infrastructure that supports both the high-throughput demands of DNA barcoding and the high-complexity demands of genome sequencing. The challenge is not only to process large volumes of sequence data, but to do so in a way that preserves provenance, harmonises metadata and ensures interoperability across platforms and repositories. Across European biodiversity genomics efforts, this principle has been realised through coordinated implementation of data systems, validation pipelines and tracking tools that operate on a common digital foundation.

The joint activities, including whole-genome sequencing and barcoding, have delivered several key outcomes that have enhanced data management and integration. Central to these achievements are the data flows for both the (meta)barcoding and genome sequencing efforts, which trace data from initial collection through processing to final publication. These process maps represent the first comprehensive documentation of how data move through large-scale biodiversity genomics initiatives in Europe.

Throughout the implementation of the activities, several significant challenges emerged that are likely to be relevant for future biodiversity genomics initiatives. Many of these challenges stemmed from the inherent complexity of coordinating data management across different scientific communities and diverse technical approaches. The challenges underscore the critical importance of considering the FAIR principles.

The current landscape of biodiversity data management reveals significant fragmentation, with multiple specialised repositories, each serving distinct purposes, making data discovery and linking complex (Fig. 2). This is exemplified by The BioImage Archive, Genomes on a Tree (GoaT), the ERGA (ERGA 2025b) and Darwin Tree of Life (DToL, The Darwin Tree of Life Project Consortium (2022)) data portals, part of the broader EBI Biodiversity Portal, BOLD, GGBN and GBIF. Whilst these platforms are often interconnected, for example, ENA is linked to both BOLD and GBIF, the lack of a unified data model and standardisation impedes seamless data integration.

A fundamental challenge has been reconciling metadata standards between the barcoding and genome sequencing areas. While both communities require similar core information about specimens and sampling events, they have historically used different terms and structures to represent this information. The barcoding and observational communities have typically used their own standards (e.g. BOLD) or DarwinCore (e.g. GBIF and OBIS) and the genomics community has used GSC MIxS (e.g. INSDC’s ENA). Efforts to align these standards, for example, at the 2024 BGE Hackathon (Vos 2025), revealed subtle, but important differences in how sampling events, specimens and derived samples are conceptualised and documented across communities. Bridging these differences required careful negotiation to ensure that neither community's specific needs were compromised, while still achieving sufficient standardisation for interoperability.

Efforts are ongoing to address these challenges by promoting standards like DwC (Wieczorek et al. 2012) and GSC MIxS (Yilmaz et al. 2011), creating mappings between existing standards and ensuring open access and comprehensive documentation to enhance data interoperability and reusability across scientific communities. Both metadata concepts and valid vocabulary ‘values and ranges’ need to be aligned for easier interoperability. Fortunately, some fundamental values, for example, date annotation, are handled consistently across most standards conforming to ISO 8602, whereas, for example, country nomenclature can vary.

We note and encourage the gradual increase in the adoption of ontologies rather than flat value lists, as this allows granular mapping. Organism taxonomy is of the highest importance, with many different species taxonomies; for example, the NCBI taxonomy is widely used in the general genomics community (INSDC), while WoRMS is more common in marine observations and the barcoding community publishes data anchored in the BOLD taxonomy.

Even over a few months, standards evolve, as new and emerging concepts crystallise, so mapping is not static. The FAIR-IMPACT project (FAIR-IMPACT 2025) is currently developing guidelines for mapping processes and for generating, storing, sharing and updating them over time (Juty et al. 2025). Even when the same concepts are present in the linked repositories, the requirements for which metadata concepts are mandatory or optional may differ, affecting data reuse opportunities. Efforts, such as Bioschemas, play a role in recommending cardinality and marginality for a core set of metadata properties, assisting users in providing sufficient metadata for reuse; the BGE genome-related activities are involved in the refinement of a ‘BioSample’ (a task within the ‘Sample’ group Bioschemas (2025)) type and profile, in collaboration with a wider community membership. The efforts to increase FAIRness also make it easier to incorporate diverse data, for example, satellite oceanographic data or previous observations from other groups in the same forest.

Tracking samples, data, metadata and outputs across distributed teams and institutions proved challenging. The complexity of modern genomics processes, involving multiple processing steps at different facilities, demanded sophisticated tracking systems. However, different institutions often had existing systems and practices that needed to be accommodated. Balancing the need for standardised tracking with institutional autonomy required flexible solutions, such as the GTC, that could integrate with various local systems, while maintaining consistent project-wide monitoring.

Quality control processes presented another significant challenge, particularly in combining automated and manual assessment steps. For barcode data, we needed to implement automated validation of sequence quality and metadata completeness, while still allowing for expert review of taxonomic assignments. Similarly, genome assembly quality assessment requires automated computation of standard metrics alongside manual inspection and correction of assembly characteristics. The same holds true for the accompanying sample metadata, such as sampling location, sampling methods etc., as demonstrated by the EAR reporting and review system. Coordinating these hybrid automated and manual processes across distributed teams required careful attention to process design and documentation.

The publication of complete data packages - encompassing raw data, metadata and analysis results - posed challenges that existing databases were not fully equipped to handle. While established repositories exist for sequence data and certain metadata types, there were gaps in the infrastructure for publishing pipeline configurations, intermediate analysis results and links between different data types. This led to the challenge of deciding what auxiliary data to preserve and how to maintain persistent connections between related data objects stored in disparate systems.

Traditionally, this role has been filled by a peer-reviewed publication; however, as the number of biodiversity genomics projects scales into the thousands, this quickly becomes infeasible. Utilising “catch-all” research objects, such as RO-Crate, which collect information on the links and provenance of all data, metadata, individuals and protocols used in each step in a machine-readable format, facilitates monitoring and publishing the full pipeline and details at each step for each project. Assigning credit and authorship for data products proved challenging due to the collaborative nature of biodiversity genomics work. Traditional academic credit systems are not well-suited for recognising the diverse contributions involved in generating a genome assembly or (meta)barcode dataset.

Particular challenges were faced in appropriately crediting field collection work, laboratory processing, bioinformatic analyses and expert curation, especially when these contributions were distributed across multiple institutions. Since this is a general challenge in biodiversity, it was systematically addressed, starting with requirements gathering from the wider biodiversity community. This was initiated at the ELIXIR BioHackathon (Brown et al. 2024) with a survey of key ‘biodiversity’ properties and continued at the German BioHackathon (Juty and Gaignard 2025), focusing on identifying core properties with Biodiversity community representatives (including Biosamples/ENA). This work also brings in a broader community from the Bioschemas working group, which is leading the development of ‘Sample’ and ‘BioSample’ types and profiles, where the ‘BioSample’ type (Bioschemas 2019) is currently being actively revised.

Using those Bioschemas/BioSample-type properties, work is ongoing to build on the basic set to develop an RO-Crate profile (described above) that better captures metadata properties common to the conceptual workflows of both BGE streams. This includes appropriate recognition or accreditation of the various actors involved in sampling and analysis processes, for example, sampling authors, preservation authors, biobanking authors etc.

Encapsulation and implementation of diverse data elements, including sample and specimen descriptions and associated resources (such as permits and material transfer agreements), tied to external links (e.g. images, taxonomic links, publications) and the software/scripts used in processing are essential for ensuring data usability and provenance. Achieving this requires an understanding of the different components and the ability to abstract and link granular provenance information as needed. Such an approach facilitates robust data integration and reuse, while maintaining transparency. Using frameworks like RO-Crate to encapsulate and link these components significantly enhances data interoperability and accessibility.

These challenges have provided valuable lessons for the genomics community and have driven the development of novel solutions, some of which are still evolving. They highlight the importance of flexible, interoperable systems and the need for community-driven standards development. Future biodiversity genomics initiatives would benefit from considering these challenges early in their planning phases.

Based on the described experiences, several key recommendations for implementing best data management practices in distributed biodiversity genomics projects have been developed. These recommendations focus on practical approaches that have proven effective in coordinating complex data processing flows across multiple institutions and research communities.

Utilising existing infrastructure, such as GBIF, BOLD, PlutoF, ENA, GGBN and WorkflowHub, where possible, is strongly recommended for distributed biodiversity genomics projects rather than developing new solutions. This applies particularly to data repositories, pipeline platforms and metadata catalogues. The key is identifying gaps where new development is necessary and focusing resources there. It was found that extensive use of APIs to link existing systems was more sustainable than creating new, monolithic platforms. When integrating project management tools with data infrastructure, clear guidelines should be established to enable users to perform the required activities and ensure smooth handover between teams and systems.

The following critical risk areas that require careful management were identified:

Loss of links between physical specimens and digital data (Groom et al. 2021);

Projects should establish monitoring and mitigation strategies for these risks from the outset.

The presented experience has highlighted the value of using RO-Crate as a standardised packaging format for biodiversity genomics data. It is recommended to develop specific RO-Crate profiles for different data types and pipelines within any given project and related work has begun (see above). Relevant profiles should align with emerging community standards, such as the Bioschemas BioSample profile (cf. Bioschemas Sample Group, Bioschemas (2025)), while maintaining the flexibility to accommodate project-specific requirements. The RO-Crate approach has proven valuable for maintaining links between different data objects and preserving pipeline provenance information and these objects are already consumable within Galaxy and WorkflowHub. The next step after refining the object itself is developing methods for integration into existing infrastructures, such as those listed above.

From a technical infrastructure perspective, it is recommended:

Implementing automated validation pipelines for both data and metadata;

These recommendations are based on practical experience rather than theoretical ideals and recognise that their implementation may need to be adapted to specific project contexts. However, they provide a foundation for effective data management in large-scale biodiversity genomics initiatives.

As biodiversity genomics continue to scale globally, several key areas emerge in which further development of data management approaches will be crucial. These opportunities span technical, organisational and social aspects of genomics data handling.

The rapid evolution of sequencing technologies presents both opportunities and challenges for data management. Emerging technologies promise higher throughput, longer reads and novel data types, requiring flexible data handling systems that can adapt to these advances. Future infrastructure development should anticipate these changes by emphasising modularity and extensibility in data processing pipelines. Additionally, the increasing volume of genomic data demands more sophisticated approaches to data compression, transfer and storage.

User interfaces for data access and analysis need significant attention. While current systems serve the needs of specialist users, broader adoption of biodiversity genomics approaches requires more accessible interfaces. Future developments should focus on creating intuitive, well-documented interfaces that lower the barriers to data access while maintaining scientific rigour. This is particularly important for enabling broader participation in biodiversity research and monitoring.

A critical area for future work is the further harmonisation of metadata and data standards across the biodiversity genomics community. While BGE has made progress in aligning barcoding and genome-sequencing standards, significant work remains to create genuinely interoperable data ecosystems. This includes developing standard protocols for new data types and ensuring compatibility with emerging biodiversity data standards. Indeed, other programmes globally have identified similar issues, so it is a shared effort.

An ambitious future goal is the comprehensive integration of BGE outputs through an overarching data model. We envision a system where each component - from field collections to final analyses - has a unique identifier and can be seamlessly referenced through standardised protocols. The proposed approach of using RO-Crate as an integration layer shows promise, allowing individual components to remain in their native repositories, while maintaining machine-readable connections between them.

The role of data management in advancing biodiversity research continues to evolve. Future developments should focus on enabling novel analyses across multiple data types and supporting automated hypothesis generation through integrated data mining. This requires technical infrastructure and new approaches to data modelling and knowledge representation. AI will likely play an increasing role in aiding the capture of metadata, for example, in the lab. The data and metadata management recommendations, when implemented, increasingly facilitate the machine accessibility and interoperability of the biodiversity data and, thus, the provision of “AI-ready” data.

Credit and attribution systems need substantial development to better reflect the collaborative nature of biodiversity genomics work. Future systems should provide finer-grained recognition of contributions, including sampling work, bioinformatic analysis and data curation. Integration with persistent identifiers for people (e.g. ORCID) and institutions should be expanded and new metrics for tracking data reuse and impact should be developed.

Finally, better systems for understanding the impact and utilisation of biodiversity genomics data are needed. This includes tracking not only academic reuse, but also applications in conservation, environmental monitoring and policy development. Understanding these patterns of data use will be crucial for directing future infrastructure development and ensuring the long-term relevance of biodiversity genomics data.

These future directions represent significant challenges, but addressing them is essential to realising the full potential of biodiversity genomics for understanding, monitoring and preserving global biodiversity. The foundations laid by BGE provide a starting point for these developments, but continued coordination and investment will be needed to achieve these ambitious goals.

AI - Artificial Intelligence