DOI: 10.3897/arphapreprints.e157339

Authors: Soulaine Theocharides, Sam Leeflang, Wouter Addink, Sharif Islam

Abstract: Biodiversity research relies on physical specimens stored in natural science collections, which serve as enduring reservoirs of data about organisms and their environments. However, these reservoirs remain siloed. The concept of Digital Specimen addresses the challenges posed by the vast amount of disconnected digital biodiversity data available today. The existing approach involves converting analogue records into digital replicas stored in local databases, leading to isolated and fragmented datasets that are difficult to integrate and utilise efficiently. The Digital Specimen aims to overcome this by establishing an interconnected network of digital objects on the Internet. Digital Specimens are FAIR Digital Objects (FDOs), structured digital entities that adhere to the FAIR principles: Findable, Accessible, Interoperable, and Reusable. FDOs have the potential to enhance the accessibility and interoperability of data from natural science collections by providing unique identifiers, descriptive metadata, and defined operations. DiSSCo utilises the FDO framework to enhance the accessibility and interoperability of biodiversity research data from natural science collections. FDOs facilitate seamless data exchange by providing structured digital objects with unique identifiers, descriptive metadata, and defined operations. As part of making Digital Specimens FDOs, DiSSCO implemented FDO records, metadata records associated with a Persistent Identifier, which further enable machine actionability. A Digital Object repository was developed for the purposes of storing and acting upon digital specimens. Three technological pillars compose the repository: a relational database stores the latest version of the digital specimen and is used for retrieving specimens by their identifier; an indexing solution provides full search capabilities on digital specimens; and a document store holds previous versions of a digital specimen for provenance purposes. There are three ways a user may interact with the digital object repository: a REST API; a user-friendly web portal; and a DOIP server.To ingest data from multiple source systems, a harmonised data model was developed, called OpenDS. Built upon existing international standards like DarwinCore and ABCD, OpenDs accommodates complex structures necessary to capture information about multiple taxonomic identifications, events, agents, and relationships to other data sources. DiSSCo has decided to adapt the GBIF Unified Model (UM) for specimen data, ensuring interoperability and avoiding the development of potentially competing standards. By aligning with the GBIF UM, DiSSCo enhances interoperability with GBIF and promotes the establishment of a unified data modelling standard within the biodiversity community, facilitating seamless data exchange and integration with data aggregators like GBIF.

HTML

XML

PDF

DOI: 10.3897/rio.10.e126586

Authors: Donat Agosti, Laurence Bénichou, Ana Casino, Lars Nielsen, Patrick Ruch, Puneet Kishor, Lyubomir Penev, Patricia Mergen, Christos Arvanitidis

Abstract: Knowledge about biodiversity is largely embedded in a daily growing corpus of over 500 million pages of biodiversity literature that is not machine-actionable. It is thus not open to building a biodiversity knowledge graph, or facilitating the use of artificial intelligence tools. This hinders the completion of a much-needed taxonomic name reference system, prevents the discovery of the biotic interactions underpinning the prediction and understanding of global change trends and consequences, viral spillovers, annotation of genes with their respective phenotypes, and their citations in various domains dealing with biological species such as conservation, agriculture, medicine, life sciences and industry, necessary to achieve the objectives of the Green Deal and address the targets identified in the Global Biodiversity Framework. This Policy Brief highlights key actions that can liberate the scientific data published, exploit their use , promote an enhanced way to publish, and ultimately foster excellence and innovation in biodiversity science, monitoring and conservation.

HTML

XML

PDF

DOI: 10.3897/rio.10.e126588

Authors: Lyubomir Penev, Quentin Groom, Ana Casino, Boris Barov

Abstract: A new community of research infrastructures has joined forces to provide scientists with seamless access to the plethora of data, services and tools in biodiversity research. New levels of technological innovation and interoperability between infrastructures foster unprecedented access to biodiversity data across all data domains and the entire research lifecycle, thus advancing open science practices and strengthening Europe's position in the global biodiversity research landscape. This policy brief highlights the potential benefits derived from enhanced connectivity and interoperability among various types of biodiversity data, fostering innovation and advancements in biodiversity science, monitoring, conservation, and policy development.

HTML

XML

PDF

DOI: 10.3897/arphapreprints.e114920

Authors: Mathias Dillen, Andreas Plank

Abstract: In this task, the aim is to develop a workflow that should facilitate the linking process of collector name strings to PIDs for those collectors. Such a workflow should help scale up the number of links being made, make the process more efficient and should take advantage as much as possible of existing work and infrastructures, so as not to reinvent the wheel. As such, the work can be roughly split into a few subtasks:- Make existing linking workflows more easily implementable in other contexts and by other infrastructures. This includes finding ways for such workflows to produce links that can easily be published, i.e. in a standardised format compatible with existing infrastructure. The suitability of different infrastructures for making established links available should also be assessed.- Establish, document and improve the comprehensiveness, findability and interoperability of the content in PID-minting resources, in particular Wikidata as it can be edited openly.- Refine the decision making process of establishing links, by implementing and improving the methods that can be used to validate potential links.In this document, the focus lies on linking people. We will propose a workflow to 'roundtrip' links established through the Bionomia platform back to the collections holding the attributed specimens, as well as making them available for use by other BiCIKL infrastructures. We will also refine existing automated linking workflows and pilot the new functionalities on the (botanical) collections of the task partners. These refinements will be influenced by an assessment of the current state of Wikidata, investigated through shape expressions constructed from commonly used queries and from Wikidata records which have been linked in previous efforts such as the Botany Pilot, Bionomia and published specimen data to GBIF.

HTML

XML

PDF

DOI: 10.3897/arphapreprints.e107169

Authors: Wouter Addink, Niki Kyriakopoulou, Lyubomir Penev, David Fichtmueller, Ben Norton, David Shorthouse

Abstract: United and coordinated efforts of biodiversity data infrastructures are needed to bring together various data forms from many different scientific areas. Biodiversity data are considered of great importance and use when they form a network of knowledge that can be seamlessly integrated and presented to various audiences, promoting both research and education. The Biodiversity Community Integrated Knowledge Library (BiCIKL) project seeks to maximise the potential of integrated data sources by striving to connect fragmented data derived from biological, paleontological, and geological specimens and collections, as well as all derived information such as literature in the form of taxonomic treatments, research papers etc., taxonomic information and molecular sequences provided by these infrastructures, under the umbrella of common digital practices and policies in curation, data sharing and open data access over different scientific fields. One of the main goals of BiCIKL is to create bi-directional links between various data types, a process enabled by: a) the adoption of globally unique and persistent identifiers upon agreement among all stakeholders, that link to digital specimen objects, collections, taxonomic treatments, people, sequence data and taxa, and b) implementation of the best practices for the generation, management and curation of interlinked data by the host infrastructures. At the same time, infrastructures should be readily discoverable and accessible by end users, providing data that enable re-usability. In this manual we give an overview of the best practices and their associated recommendations for infrastructures on making the most out of their services and data, for establishing a network of knowledge with other infrastructures, for servicing researchers, data providers and other end users. These guidelines have been developed in collaboration with the infrastructures through Technical RI Forum meetings organised in the context of the BiCIKL project. Practices and recommendations were divided into six categories: 1) modalities of access, 2) building communities and trust, 3) technology and standards, 4) versioning of APIs and their data, 5) bi-directional linking between infrastructures and 6) API design patterns and naming conventions. A second division into three user groups (Infrastructures, Data providers, Users e.g. Researchers, Developers and Citizen scientists) is presented in Appendix I.

HTML

XML

PDF

DOI: 10.3897/arphapreprints.e107168

Authors: Wouter Addink, Sharif Islam, Mathias Dillen, Anton Güntsch, Soulaine Theocharides

Abstract: Persistent Identifier (PID) systems are the foundation for achieving the FAIR Guiding Principles (“findable, accessible, interoperable and reusable”). As FAIR data and connecting different data classes (i.e. specimens, genomics, observations, taxonomy and publications) are essential aspects of the BiCIKL project, we need a PID system at least at the European level to create and maintain identifiers for the digital representation of specimens and samples, called Digital Specimens (DS) (Hardisty et al. 2022). The PID system provides the mechanism to ensure that identifiers are globally unique, persistent and resolvable. This system should also manage associated metadata, facilitate provenance, enable discovery, manage states and the life cycle of the PID, link to other derived data and digital content, and allow content providers to enforce metadata constraints. For the successful provision of a PID system, this design document has been created to guide us during the implementation and operation phases. The document is based on an earlier milestone (MS28) that was used for discussion and evaluation with potential end-users.

HTML

XML

PDF

DOI: 10.3897/arphapreprints.e107167

Authors: Kessy Abarenkov, Allan Zirk, Guy Cochrane, Vishnukumar Kadhirvelu, Joana Pauperio, Olaf Bánki, Jerry Lanfear, Filipp Ivanov, Timo Piirmann, Raivo Pöhönen, Urmas Kõljalg

Abstract: This deliverable report includes description of the work steps towards building a web interface for the reporting of errors and gaps in sequenced material source annotations as part of the Task 8.3 of BiCIKL. Beta version of the web interface has been published and is available for the registered users of PlutoF platform.

HTML

XML

PDF

DOI: 10.3897/arphapreprints.e107166

Authors: Soulaine Theocharides, Niki Kyriakopoulou

Abstract: Work package 11 of the BiCIKL project involves developing software tools to support a FAIR experience for members of the biodiversity research community. The package overall focuses on Findability, by providing tools to search and answer questions, and Accessibility, through developing links across various biodiversity data sources and research tools. Task 11.2 specifically involves prediction of new links using machine learning. We chose to demonstrate the functionality of machine learning link prediction with plant-pollinator interactions. This type of interaction was chosen due to the wealth of data available, particularly on the Global Biotic Interactions (GloBI) database, as well as this kind of interaction’s ecological and economic significance. The result was a RESTful API capable of predicting plant-pollinator interactions among a predefined set of species. Predictions are made on-the-fly, at the time of the request. The GitHub repository for the API can be found here: https://github.com/DiSSCo/BiCIKL_Linkages_APIThe API takes either a plant or a pollinator as inputs, and outputs potential matches based on a user-defined confidence score. The API’s prediction is powered by a random forest classifier stored on disk. The classifier was trained on the taxonomic hierarchy of observed plant-pollinator pairs obtained from the GloBI database. When evaluating the likelihood of an interaction, the trained classifier looks at the taxonomic hierarchy of both the plant and pollinator and outputs a confidence score. What pairs are returned is determined by the minimum confidence score set by the user.

HTML

XML

PDF