Marc Portier

The availability of accurate and up-to-date information on marine species—ranging from nomenclature and taxonomy to species occurrences—is fundamental for several stakeholders, from marine scientists to policymakers. Today, this information is provided by well-known authoritative systems, including the World Register of Marine Species (WoRMS*1), a comprehensive register of names of marine organisms and related information; Marine Regions*2, a register of georeferenced information on marine place names and areas; and the European Ocean Biogeographic Information System (EurOBIS*3), where species occurrences in European waters are recorded. These platforms ensure marine biodiversity data is accessible and interoperable, supporting various applications and use cases for fostering advancements in taxonomy, marine biology, ecology and environmental science. The MAREGRAPH project*4 has the ambition to further increase the findability, accessibility, interoperability and reusability of these foundational high-value datasets, by creating and publishing an open knowledge graph (KG) on marine biodiversity data through the semantic uplifting of the involved datasets, relying on the Semantic Web stack and its standards (Fig. 1). This requires: the definition and publication of a network of formal, reusable and extensible ontologies and controlled vocabularies for representing and linking together marine biodiversity information, as a basis for semantic interoperability; the set up of a cost-effective, flexible and scalable data architecture to publish the data as Linked Open Data and provide access through standard APIs, from a SPARQL endpoint to Linked Data Event Streams (LDES), enabling technical interoperability and serving various use cases and data access needs. the definition and publication of a network of formal, reusable and extensible ontologies and controlled vocabularies for representing and linking together marine biodiversity information, as a basis for semantic interoperability; the set up of a cost-effective, flexible and scalable data architecture to publish the data as Linked Open Data and provide access through standard APIs, from a SPARQL endpoint to Linked Data Event Streams (LDES), enabling technical interoperability and serving various use cases and data access needs. The methodology adopted for defining and publishing the KG combines the consolidated principles of the Open Standards for Linked Organisations (OSLO) framework—which defines the governance structure and a open process for developing semantic data standards—with the well-established agile and collaborative ontology development workflow of the eXtreme Design methodology (Presutti 2009), based on ontology design patterns (ODPs). Also leveraging the experience gained in defining the Marine Regions Ontology and publishing the Marine Regions gazetteer as LDES (Lonneville 2021), we are iteratively defining the ontologies for representing: marine species, whose description is rooted in the definition of taxa and their scientific names, and extends to geographic distribution, scientific literature, ecological traits and other biological information as represented in WoRMS; species observations, describing the occurrence of species in space and time, with associated descriptive data as well as biological and environmental measurements, as recorded in EurOBIS. marine species, whose description is rooted in the definition of taxa and their scientific names, and extends to geographic distribution, scientific literature, ecological traits and other biological information as represented in WoRMS; species observations, describing the occurrence of species in space and time, with associated descriptive data as well as biological and environmental measurements, as recorded in EurOBIS. Our approach to ontology design is driven by the identification of use cases and competency questions, as well as by existing models and data, also involving the community through workshops, webinars and co-creation sessions. Existing domain ontologies and reference data models, such as Biodiversity Information Standards' (TDWG) Darwin Core, the Taxonomic Concept Schema, Bioschema's profiles for taxa and taxon names, the OpenBioDiv ontology, and the Catalogue of Life Data Package (CoLDP), were considered for the identification of ODPs and reuse, particularly through ontology semantic alignments. As fostering consensus on data semantics is a critical step toward broad acceptance and adoption, semantic assets are incrementally published on a dedicated GitHub repository*5, involving again the community through public review processes. MAREGRAPH will then enable the production of linked open datasets where the data from WoRMS, Marine Regions and EurOBIS are seamlessly integrated and interlinked in a unified KG that can be further enriched and linked with data from other initiatives where marine biodiversity is the focus.

This Science Project (SP) contributes to the estimation of the impacts of the invasive species on the European Biodiversity and Ecosystems. This topic is important for European Green Deal and the new European Biodiversity Strategy. The SP is also linked with the socio-economic issues because of the NIS implications to the local ecosystems and their services, and their societal goods and services. Since many of the above impacts may be of local scale, they may alter common practices in circular economies. The SP is implemented by: (a) Combining different sources of data and information; (b) Using a dual workflow to analyse the data; (c) Integrating its resources with core EOSC services and potentially horizontal services available; (d) Engaging the relevant scientific communities. The users will be able to: (a) Analyse distribution patterns of invasive species from different sources of data; (b) Compare the above patterns; (c) Provide managerial suggestions to relevant authorities; (d) Build on the existing infrastructure to address more complex questions (e.g. future scenarios).

The current knowledge on the risk of climate change for biodiversity and ecosystems needs to be improved by seeking evidence from cross-domain analyses. As a demonstration case, this study analyses how to investigate and monitor the rapid increase of Non-Indigenous Invasive Species (NIS) in European ecosystems. These species may not only replace indigenous ones but also alter habitats, interacting with the changing environment and eventually severely influence established socio-economic regimes. The challenge is to adopt a comprehensive approach by considering the bulk of the biotic and abiotic variables and their interactions, which may be even more important for the distribution of the NIS than the occurrence of the NIS. Such approaches require access to big datasets (from genomics to in situ and satellite borne environmental data) and high computational power, especially for those models with iterative algorithms. This study aims to: integrate data from different scientific disciplines in the marine subdomain (e.g. chemistry, physics, biodiversity, ecosystems, genomics, socio-economics) into an analytical framework in order to advance our knowledge on the impact of NIS on European marine biodiversity and ecosystems; to connect the analytical framework and federate access to relevant data infrastructures at the EOSC portal in order to mobilise and empower a larger community of researchers and potential data providers; and to demonstrate and promote the benefits and potential of web-based science using EOSC.A break-through Technical Composability Layer (Tesseract, which includes an additional option with Jupyter Notebook) is used in order to achieve the horizontal composability of the services. Only FAIR-compliant datasets are used in this study. However, because the nature of the project is primarily multidisciplinary and cross-domain, the only way to guarantee that the results deriving by the different disciplines/domains are comparable is to FAIR-ify the web services used, too. This way, it is ensured that both the quality and process potential in the different disciplines and domains are comparable and therefore so are their results.This paper brings together scientists making basic research on biodiversity and ecosystems, computer engineers, including software and web developers, in order to create a FAIR-compliant virtual research environment (VRE) to achieve both the scientific goals and the community engaged.

RO-Crate (Soiland-Reyes et al. 2022) is a lightweight method to package research outputs along with their metadata, based on Linked Data principles (Bizer et al. 2009) and W3C standards. RO-Crate provides a flexible mechanism for researchers archiving and publishing rich data packages (or any other research outcome) by capturing their dependencies and context. However, additional measures should be taken to ensure that a crate is also following the FAIR principles (Wilkinson 2016), including consistent use of persistent identifiers, provenance, community standards, clear machine/human-readable licensing for metadata and data, and Web publication of RO-Crates. The FAIR Digital Object (FDO) approach (De Smedt et al. 2020) gives a set of recommendations that aims to improve findability, accessibility, interoperability and reproducibility for any digital object, allowing implementation through different protocols or standards. Here we present how we have followed the FDO recommendations and turned research outcomes into FDOs by publishing RO-Crates on the Web using HTTP, following best practices for Linked Data. We highlight challenges and advantages of the FDO approach, and reflect on what is required for an FDO profile to achieve FAIR RO-Crates. The implementation allows for a broad range of use cases, across scientific domains. A minimal RO-Crate may be represented as a persistent URI resolving to a summary website describing the outputs in a scientific investigation (e.g. https://w3id.org/dgarijo/ro/sepln2022 with links to the used datasets along with software). One of the advantages of RO-Crates is flexibility, particularly regarding the metadata accompanying the actual research outcome. RO-Crate extends schema.org, a popular vocabulary for describing resources on the Web (Guha et al. 2016). A generic RO-Crate is not required to be typed beyond Dataset*1. In practice, RO-Crates declare conformance to particular profiles, allowing processing based on the specific needs and assumptions of a community or usage scenario. This, effectively, makes RO-Crates typed and thus machine-actionable. RO-Crate profiles serve as metadata templates, making it easier for communities to agree and build upon their own metadata needs. RO-Crates have been combined with machine-actionable Data Management Plans (maDMPs) to automate and facilitate management of research data (Miksa et al. 2020). This mapping allows RO-Crates to be generated out of maDMPs and vice versa. The ELIXIR Software Management Plans (Alves et al. 2021) is planning to move their questionnaire to a machine-actionable format with RO-Crate. ELIXIR Biohackathon 2022 will explore integration of RO-Crate and the Data Stewardship Wizard (Pergl et al. 2019) with Galaxy, which can automate FDO creation that also follows data management plans. A tailored RO-Crate profile has been defined to represent Electronic Lab Notebooks (ELN) protocols bundled together with metadata and related datasets. Schröder et al. (2022) uses RO-Crates to encode provenance information at different levels, including researchers, manufacturers, biological and chemical resources, activities, measurements, and resulting research data. The use of RO-Crates makes it easier to programmatically question-answer information related to the protocols, for instance activities, resources and equipment used to create data. Another example is WorkflowHub (Goble et al. 2021) which defines the Workflow RO-Crate profile (Bacall et al. 2022), imposing additional constraints such as the presence of a main workflow and a license. It also specifies which entity types and properties must be used to provide such information, implicitly defining a set of operations (e.g., get the main workflow and its language) that are valid on all complying crates. The workflow system Galaxy (The Galaxy Community 2022) retrieves such Workflow Crates using GA4GH TRS API. The workflow profile has been further extended (with OOP-like inheritance) in Workflow Testing RO-Crate, adding formal workflow testing components: this adds operations such as getting remote test instances and test definitions, used by the LifeMonitor service to keep track of the health status of multiple published workflows. While RO-Crates use Web technologies, they are also self-contained, moving data along with their metadata. This is a powerful construct for interoperability across FAIR repositories, but this raises some challenges with regards to mutability and persistence of crates. To illustrate how such challenges can be handled, we detail how the WorkflowHub repository follows several FDO principles: Workflow entries must be frozen for editing and have complete kernel metadata (title, authors, license, description) [FDOF4] before they can be assigned a persistent identifier, e.g. https://doi.org/10.48546/workflowhub.workflow.255.1 [FDOF1] Computational workflows can be composed of multiple files used as a whole, e.g. CWL files in a GitHub repository. These are snapshotted as a single RO-Crate ZIP, indicating the main workflow. [FDOF11] PID resolution can content-negotiate to Datacite’s PID metadata [FDOF2] or use FAIR Signposting to find an RO-Crate containing the workflow [FDOF3] and richer JSON-LD metadata resources [FDOF5,FDOF8], see Fig. 1 Metadata uses schema.org [FDOF7] following the community-developed Bioschemas ComputationalWorkflow profile [FDOF10]. Workflows are discovered using the GA4GH TRS API [FDOF5,FDOF6,FDOF11] and created/modified using CRUD operations [FDOF6] The RO-Crate profile, effectively the FDO Type [FDOF7], is declared as https://w3id.org/workflowhub/workflow-ro-crate/1.0; the workflow language (e.g. https://w3id.org/workflowhub/workflow-ro-crate#galaxy) is defined in metadata of the main workflow. Workflow entries must be frozen for editing and have complete kernel metadata (title, authors, license, description) [FDOF4] before they can be assigned a persistent identifier, e.g. https://doi.org/10.48546/workflowhub.workflow.255.1 [FDOF1] Computational workflows can be composed of multiple files used as a whole, e.g. CWL files in a GitHub repository. These are snapshotted as a single RO-Crate ZIP, indicating the main workflow. [FDOF11] PID resolution can content-negotiate to Datacite’s PID metadata [FDOF2] or use FAIR Signposting to find an RO-Crate containing the workflow [FDOF3] and richer JSON-LD metadata resources [FDOF5,FDOF8], see Fig. 1 Metadata uses schema.org [FDOF7] following the community-developed Bioschemas ComputationalWorkflow profile [FDOF10]. Workflows are discovered using the GA4GH TRS API [FDOF5,FDOF6,FDOF11] and created/modified using CRUD operations [FDOF6] The RO-Crate profile, effectively the FDO Type [FDOF7], is declared as https://w3id.org/workflowhub/workflow-ro-crate/1.0; the workflow language (e.g. https://w3id.org/workflowhub/workflow-ro-crate#galaxy) is defined in metadata of the main workflow. 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. RO-Crate has shown a broad adoption by communities across many scientific disciplines, providing a lightweight, and therefore easy to adopt, approach to generating FAIR Digital Objects. It is rapidly becoming an integral part of the interoperability fabric between the different components as demonstrated here for WorkflowHub, contributing to building the European Open Science Cloud.

Background The FAIR principles (Wilkinson et al. 2016) are fundamental for data discovery, sharing, consumption and reuse; however their broad interpretation and many ways to implement can lead to inconsistencies and incompatibility (Jacobsen et al. 2020). The European Open Science Cloud (EOSC) has been instrumental in maturing and encouraging FAIR practices across a wide range of research areas. Linked Data in the form of RDF (Resource Description Framework) is the common way to implement machine-readability in FAIR, however the principles do not prescribe RDF or any particular technology (Mons et al. 2017). FAIR Digital Object FAIR Digital Object (FDO) (Schultes and Wittenburg 2019) has been proposed to improve researcher’s access to digital objects through formalising their metadata, types, identifiers and exposing their computational operations, making them actionable FAIR objects rather than passive data sources. FDO is a set of principles (Bonino et al. 2019), implementable in multiple ways. Current realisations mostly use Digital Object Interface Protocol (DOIPv2) (DONA Foundation 2018), with the main implementation CORDRA. We can consider DOIPv2 as a simplified combination of object-oriented (CORBA, SOAP) and document-based (HTTP, FTP) approaches. More recently, the FDO Forum has prepared detailed recommendations, currently open for comments, including a DOIP endorsement and updated FDO requirements. These point out Linked Data as another possible technology stack, which is the focus of this work. Linked Data Linked Data standards (LD), based on the Web architecture, are commonplace in sciences like bioinformatics, chemistry and medical informatics – in particular to publish Open Data as machine-readable resources. LD has become ubiquitous on the general Web, the schema.org vocabulary is used by over 10 million sites for indexing by search engines – 43% of all websites use JSON-LD. Although LD practices align to FAIR (Hasnain and Rebholz-Schuhmann 2018), they do not fully encompass active aspects of FDOs. The HTTP protocol is used heavily for applications (e.g. mobile apps and cloud services), with REST APIs of customised JSON structures. Approaches that merge the LD and REST worlds include Linked Data Platform (LDP), Hydra and Web Payments. Meeting FDO principles using Linked Data standards Considering the potential of FDOs when combined with the mature technology stack of LD, here we briefly discuss how FDO principles in Bonino et al. (2019) can be achieved using existing standards. The general principles (G1–G9) apply well: Open standards with HTTP being stable for 30 years, JSON-LD is widely used, FAIR practitioners mainly use RDF, and a clear abstraction between the RDF model with stable bindings available in multiple serialisations. However, when considering the specific principles (FDOF1–FDOF12) we find that additional constraints and best practices need to be established – arbitrary LD resources cannot be assumed to follow FDO principles. This is equivalent to how existing use of DOIP is not FDO-compliant without additional constraints. Namely, persistent identifiers (PIDs) (McMurry et al. 2017) (FDOF1) are common in LD world (e.g. using http://purl.org/ or https://w3id.org/), however they don’t always have a declared type (FDOF2), or the PID may not even appear in the metadata. URL-based PIDs are resolvable (FDOF3), typically over HTTP using redirections and content-negotiation. One great advantage of RDF is that all attributes are defined semantic artefacts with PIDs (FDOF4), and attributes can be reused across vocabularies. While CRUD operations (FDOF6) are supported by native HTTP operations (GET/PUT/POST/DELETE) as in LDP , there is little consistency on how to define operation interfaces in LD (FDOF5). Existing REST approaches like OpenAPI and URI templates are mature and good candidates, and should be related to defined types to support machine-actionable composition (FDOF7). HTTP error code 410 Gone is used in tombstone pages for removed resources (FDOF12), although more frequent is 404 Not Found. Metadata is resolved to HTTP documents with their own URIs, but these frequently don’t have their own PID (FDOF8). RDF-Star and nanopublications (Kuhn et al. 2021) give ways to identify and trace provenance of individual assertions. Different metadata levels (FDOF9) are frequently developed for LD vocabularies across different communities (FDOF10), such as FHIR for health data, Bioschemas for bioinformatics and >1000 more specific bioontologies. Increased declaration and navigation of profiles is therefore essential for machine-actionability and consistent consumption across FAIR endpoints. Several standards exist for rich collections (FDOF11), e.g. OAI-ORE, DCAT, RO-Crate, LDP. These are used and extended heterogeneously across the Web, but consistent machine-actionable FDOs will need specific choices of core standards and vocabularies. Another challenge is when multiple PIDs refer to “almost the same” concept in different collections – significant effort have created manual and automated semantic mappings (Baker et al. 2013, de Mello et al. 2022). Currently the FDO Forum has suggested the use of LDP as a possible alternative for implementing FAIR Digital Objects (Bonino da Silva Santos 2021), which proposes a novel approach of content-negotiation with custom media types. Discussion The Linked Data stack provides a set of specifications, tools and guidelines in order to help the FDO principles become a reality. This mature approach can accelerate uptake of FDO by scholars and existing research infrastructures such as the European Open Science Cloud (EOSC). However, the amount of standards and existing metadata vocabularies poses a potential threat for adoption and interoperability. Yet, the challenges for agreeing on usage profiles apply equally to DOIP as LD approaches. We have worked with different scientific communities to define RO-Crate (Soiland-Reyes et al. 2022), a lightweight method to package research outputs along with their metadata. While RO-Crate’s use of schema.org shows just one possible metadata model, it's powerful enough to be able to express FDOs, and familiar to web developers. We have also used FAIR Signposting (Van de Sompel et al. 2022) with HTTP Link: headers as a way to support navigation to the individual core properties of an FDO (PID, type, metadata, licence, bytestream) that does not require heuristics of content-negotiation and is agnostic to particular metadata vocabularies and serialisations. We believe that by adopting Linked Data principles, we can accelerate FDO today – and even start building practical ways to assist scientists in efficiently answering topical questions based on knowledge graphs.

The World Register of Marine Species (WoRMS) is an authoritative classification and catalogue of marine names. The WoRMS portal and available web-services are a gateway to access a treasure-chest of information, not only on taxon names themselves, but also on their mutual relations (e.g., original names, accepted versus unaccepted names, taxonomic classification), and related information such as ecological traits, distributions and linked literature. Over its fifteen years of existence, WoRMS has not only been growing in content and quality, thanks to the voluntary efforts of more than 300 experts worldwide, it has also kept a technical trajectory that involves adapting to new standards and technologies. Although WoRMS has always been relatively easily accessible through its portal and web services, and applied the basic data-sharing principles of FAIR (Findable, Accessible, Interoperable, and Reusable), there is still room for improvement. The recently growing call for globally uniform identifiers coming from the application of the FAIR data sharing principles, and the growing investment into globally open and interlinked "digital twin" representations of our oceans and the organisms found in them, have introduced the fundamentals for growing a marine knowledge graph, and as a consequence, has directed some technical attention towards applying semantic web technologies. WoRMS plays a key role in the field of (marine) biodiversity, as this research field strongly relies on the correct usage of species names, and understanding the taxonomic relationships between taxon names. As WoRMS is regarded as the authoritative resource for marine names, it is also heavily used as a quality-control tool for the correct usage of taxon names within various European and global initiatives. WoRMS provides support to global databases and infrastructures that use (or are in need of) a marine taxonomic backbone, such as the LifeWatch Species Information Backbone, the Ocean Biodiversity Information System (OBIS) and the Global Ocean Observing System (GOOS), as well as improves the content and strengthens relationships with environment-independent initiatives and infrastructures such as the Catalogue of Life (COL), the Barcode of Life Data System (BoLD) & GenBank. In addition to the taxonomic value of WoRMS, it is also highly valued for its available information on species traits, which form a critical component in ecological marine research. In its role as a marine taxonomic backbone, along with being linked to numerous other environment-independent initiatives and infrastructures, WoRMS has always required an adaptability towards the challenging new ways specific applications and concrete research have been choosing to apply the identifiers affixed by WoRMS. It is in this tradition we now announce and describe our approach to publish the content of the register as fully linked open data, using semantic web technologies. We describe in some detail the choices made to select and apply specific vocabularies that already exist for the description and interconnected linking of taxon names, to address the tension between hanging on to an historic (URN) persistent identifier and providing a dereferenceable URI that supports the appreciated "follow your nose" property, to link to other relevant registries, to design meaningful predicates for inbound links to the register and, more technically to divide the available content into meaningful sub-sections for retrieval of optional detail, and to provide a roadmap for meaningful fragmentations of the full register to allow for an effective consumption of the relevant (e.g., newly updated) parts into specific data-consumption scenarios. to select and apply specific vocabularies that already exist for the description and interconnected linking of taxon names, to address the tension between hanging on to an historic (URN) persistent identifier and providing a dereferenceable URI that supports the appreciated "follow your nose" property, to link to other relevant registries, to design meaningful predicates for inbound links to the register and, more technically to divide the available content into meaningful sub-sections for retrieval of optional detail, and to provide a roadmap for meaningful fragmentations of the full register to allow for an effective consumption of the relevant (e.g., newly updated) parts into specific data-consumption scenarios. We believe this work to be an important step towards achieving some future goals. It should further facilitate the production of automated, managed or hybrid crosswalks between various taxonomic registers and classifications. To be especially considered here, is helping to make omics taxonomic references more meaningfully comparable with WoRMS. In the process of others linking their digital objects (e.g., services, datasets, publications, experts) to taxon IDs in WoRMS, they are effectively also linking to each other, which opens the doors between apparently unconnected bodies. Its continuing use as a global standard and trustworthy reference of community-accepted names for biological taxa becomes the essential glue connecting all sorts of services, initiatives, communities.

Biodiversity Observation Networks (BONs) are important sources of information about the health and wealth of biodiversity in our world. By observing the presence or absence of species in different environments and by coupling with measurements of environmental parameters, they contribute important information to our knowledge of the natural world and to our ability to model and predict the effects of climate change. Knowledge gained from BONs can also be used to assist wider society (political, social, and economic bodies) in making biodiversity-sensitive decisions. For the data outputs from BONs to be useful to a wide audience, management of those data is crucial: the data should be published openly and they must be Findable, Accessible, Interoperable, and Reusable (FAIR). It is important that the data can be found and understood by anyone who could use them. For BON data to lead to new science, it is necessary that the data can be accessed programmatically, that full provenance is provided, and that data from different sources are interoperable. Semantic and technical interoperability should be coupled with user-friendly data archiving, managing, discovery, and User Interfaces (UIs) for those creating and for those using the data. The importance of lowering the hurdle to creating and sharing data in a FAIR way is not to be underestimated; the hardest part of the FAIR journey is often its uptake rather than its technology. The importance of eDNA (environmental DNA) to BONs, highlights the particular challenge to the management of BON data because of the complexity of the practical workflow (many parties involved doing the actual sampling, samples being shipped as well as biobanked, highly technical procedures involved at various stages), the complexity of the analysis of the DNA to measure “occurrences”, and the rapid evolution of the field of biotechnology. Version management to accommodate updates to biotechnology pipelines and reference databases, the use of IDs to allow data to be linked to evolving knowledge rather than to static information, semantic annotation aimed at multiple audiences (omics-expert and less-expert), and linking multiple-distributed datasets (omics, image, environmental, etc.) are all crucial aspects to FAIR management of omics data. As data standards evolve much more slowly than the scientific possibilities, it is also important to archive, annotate, and link all the data in a flexible way so it can be exported into today’s and tomorrow’s data formats. All of the above expectations are added as extra weight onto the field scientists and lab workers dealing with the actual samples and producing their digital outcomes. At the Vlaams Instituut voor de Zee (VLIZ) we are managing data from several BONs, and for this we are developing a data management approach that is based on Research Object Crate (RO-Crate) data packaging including the following elements: a GitHub entry point for holding the RO-Crates, uploaded data (co-located files or links to data held or published elsewhere), a straight-forward GitHub-Action workflow to publish and uplift the contents as linked open data, and UIs for data upload, search, and select. BON contributors and data managers will be able to upload the BON data (e.g., sampling log sheets, ENA (European Nucleotide Archive) accession codes, SOPs (Standard Operating Procedures), etc.), interfacing directly with GitHub or through our assisting tool to upload and create the RO-Crate descriptions, and to semantically describe their data. Scientists who analyse data will also be able to upload their workflow outputs and provenance metadata descriptions and link those to the BON data that their analysis is based on. By collecting the data and metadata in one place, the BON outputs can be rearranged into whatever published formats are required. The advantages of this approach are multiple: It relocates the authority on "what can and should be said about the data" back to the authors, who are not then limited by the data formats of the publishers. The application profile and required fields they themselves agree upon allow for expressing all the nuance and targeted information that they want, and thus create an overall stronger affinity between the produced content and its maintainers. Embracing the semantic web techniques marries a high-level uniform processing and indexing with a diversity of added detail that can be retrieved by drilling down. Similarly, clear cross-references to managed terms in managed vocabularies allows further serendipitous connections to apparently unconnected domains. Providing digital assistance allows automated provenance tracking, easy linking to agreed managed vocabularies, as well as hiding cumbersome low-level technical details (such as Git operations, format specifications, etc.). By doing so as early as possible in the processing of the information, it avoids lost memory details, late quality checking, and lengthy round trips that all add to the cost of publishing, interpreting and reusing the produced data. Finally, the available open semantics in the datasets themselves allow for advanced additional post-processing that is not limited to that provided by centralised aggregation services, but also allows for providing fragmentations, indexes and navigational indicators that support a more scalable, distributed and federated real-time pathway to selecting and analysing data in the context of specific research questions. It relocates the authority on "what can and should be said about the data" back to the authors, who are not then limited by the data formats of the publishers. The application profile and required fields they themselves agree upon allow for expressing all the nuance and targeted information that they want, and thus create an overall stronger affinity between the produced content and its maintainers. Embracing the semantic web techniques marries a high-level uniform processing and indexing with a diversity of added detail that can be retrieved by drilling down. Similarly, clear cross-references to managed terms in managed vocabularies allows further serendipitous connections to apparently unconnected domains. Providing digital assistance allows automated provenance tracking, easy linking to agreed managed vocabularies, as well as hiding cumbersome low-level technical details (such as Git operations, format specifications, etc.). By doing so as early as possible in the processing of the information, it avoids lost memory details, late quality checking, and lengthy round trips that all add to the cost of publishing, interpreting and reusing the produced data. Finally, the available open semantics in the datasets themselves allow for advanced additional post-processing that is not limited to that provided by centralised aggregation services, but also allows for providing fragmentations, indexes and navigational indicators that support a more scalable, distributed and federated real-time pathway to selecting and analysing data in the context of specific research questions.

An increasing number of researchers support reproducibility by including pointers to and descriptions of datasets, software and methods in their publications. However, scientific articles may be ambiguous, incomplete and difficult to process by automated systems. In this paper we introduce RO-Crate, an open, community-driven, and lightweight approach to packaging research artefacts along with their metadata in a machine readable manner. RO-Crate is based on Schema.org annotations in JSON-LD, aiming to establish best practices to formally describe metadata in an accessible and practical way for their use in a wide variety of situations. An RO-Crate is a structured archive of all the items that contributed to a research outcome, including their identifiers, provenance, relations and annotations. As a general purpose packaging approach for data and their metadata, RO-Crate is used across multiple areas, including bioinformatics, digital humanities and regulatory sciences. By applying “just enough” Linked Data standards, RO-Crate simplifies the process of making research outputs FAIR while also enhancing research reproducibility. An RO-Crate for this article ^{1 1} https://w3id.org/ro/doi/10.5281/zenodo.5146227 is archived at https://doi.org/10.5281/zenodo.5146227 .

Genomic Observatories (GOs) are an increasingly important resource to study the effect of climate change on marine populations. The data gathered by GOs allow one to map and track how marine populations change with time and location, and how those changes relate to the local and global conditions. Such data may be used to calculate Essential Biodiversity Variables (EBVs) and can provide important information for predictive modelling of marine biodiversity. GOs are sites that are subject to long-term scientific research, including (but not limited to) the sustained study of genomic biodiversity from single-celled microbes to multicellular organisms. We are involved in a number of GO projects, including: Ocean Sampling Day (OSD): yearly, standardised water sampling from selected sites located all over the world. The data collected consist of (a)biotic parameters, DNA sequences extracted from the water samples, and species occurrences derived from analysis of the sequences. Marine Biodiversity Observation Network of Autonomous Reef Monitoring Structures (ARMS-MBON): yearly placement of sets of stacked plates along the harbours and coasts of Europe, deployed in place for months to allow species to settle on them. The data collected consist of images of the communities that settled on the plates, DNA sequences extracted from scrapings of the plates and the surrounding water, visual observations, and species occurrences derived from analysis of the sequences and the images. Ocean Sampling Day (OSD): yearly, standardised water sampling from selected sites located all over the world. The data collected consist of (a)biotic parameters, DNA sequences extracted from the water samples, and species occurrences derived from analysis of the sequences. Marine Biodiversity Observation Network of Autonomous Reef Monitoring Structures (ARMS-MBON): yearly placement of sets of stacked plates along the harbours and coasts of Europe, deployed in place for months to allow species to settle on them. The data collected consist of images of the communities that settled on the plates, DNA sequences extracted from scrapings of the plates and the surrounding water, visual observations, and species occurrences derived from analysis of the sequences and the images. The data collected from these GOs are not particularly complicated: sampling and sequencing protocols are well established and the images are taken with standard cameras. However, complexity arises when measurements need to be linked between multiple samples extracted from each event for each location and date, to all the data from all sampling events, and the results—species occurrences and abundances—obtained from the images and the sequences for each sample, must be linked together within that sampling event and between all sampling events of the GO project. measurements need to be linked between multiple samples extracted from each event for each location and date, to all the data from all sampling events, and the results—species occurrences and abundances—obtained from the images and the sequences for each sample, must be linked together within that sampling event and between all sampling events of the GO project. We also want to be able to compare data between different GO projects, and to be able to incorporate measurements from nearby monitoring stations, as this will allow for an enhanced analysis of the evolution of marine benthic populations in light of climate change. Our aim is to adopt the Darwin Core Archive (DwC-A) OBIS-ENV-DATA format for the data from ARMS-MBON and OSD, including the linkages to the images and sequences. ARMS-MBON presents a challenge because of the wide range of data collected. Our use-case for the data format we require contains the following elements: The ARMS-MBON data that are collected are the sequences obtained from the samples and the images taken of the communities on each ARMS plate. Species occurrences and added (a)biotic parameters only come later. However, we would like to adopt the DwC-A format already being widely used from the very beginning of our data management, with the species occurrences being added to the same DwC-A files when they are determined. The ARMS plate images will go through a few stages of processing: the raw images taken by the field scientists, and later annotated images created by the image-analysis software. We would like to link the raw and processed images to each other within the DwC-A files, so the user can obtain the processed image a species was identified in, and the raw image for its re-use. This same requirement applies to the sequence data. The ARMS plate images from each sampling event number many dozens. These are archived as ZIP files in the Marine Data Archive, and the links to them are made public via the metadata record in a catalogue. We would like to be able to add links in the DwC-A files to the individual images in the ZIP file that each species occurrence came from, and from which region(s) on the images they were found; so individual images can be referenced but the space saving advantage of ZIP can be used. In addition, we would like to explore the option of adding the images within the DwC-A (ZIP) file itself, rather than referencing them as URLs. For each sampling event we collect sequences and images, and species (occurrences) will be obtained from analysis of both types of data. It is necessary that these multiple and overlapping results can be linked to each other in a clear and consistent way. This means indicating clearly where sequences and images are of the same or of different material from any plate. The ARMS-MBON data that are collected are the sequences obtained from the samples and the images taken of the communities on each ARMS plate. Species occurrences and added (a)biotic parameters only come later. However, we would like to adopt the DwC-A format already being widely used from the very beginning of our data management, with the species occurrences being added to the same DwC-A files when they are determined. The ARMS plate images will go through a few stages of processing: the raw images taken by the field scientists, and later annotated images created by the image-analysis software. We would like to link the raw and processed images to each other within the DwC-A files, so the user can obtain the processed image a species was identified in, and the raw image for its re-use. This same requirement applies to the sequence data. The ARMS plate images from each sampling event number many dozens. These are archived as ZIP files in the Marine Data Archive, and the links to them are made public via the metadata record in a catalogue. We would like to be able to add links in the DwC-A files to the individual images in the ZIP file that each species occurrence came from, and from which region(s) on the images they were found; so individual images can be referenced but the space saving advantage of ZIP can be used. In addition, we would like to explore the option of adding the images within the DwC-A (ZIP) file itself, rather than referencing them as URLs. For each sampling event we collect sequences and images, and species (occurrences) will be obtained from analysis of both types of data. It is necessary that these multiple and overlapping results can be linked to each other in a clear and consistent way. This means indicating clearly where sequences and images are of the same or of different material from any plate. Clearly, a high degree of data standardisation is necessary to enable harmonisation of the data, to accommodate the multiple streams of linked data values, which will be added by multiple creators and users of the data.