IGSN ID Implementation Exemplars: CSIRO and AuScope

Through the partnership between DataCite and IGSN e.V., DataCite services can be used to register International Generic Sample Numbers (IGSN IDs) for material samples. The blog series 'IGSN ID Implementation Exemplars' showcases sample management workflows developed by the community that incorporate IGSN ID registration. In each post, we offer practical guidance on how to work alongside disciplinary sample experts to register IGSN IDs within DataCite services.

Following the IGSN–DataCite partnership, the Commonwealth Scientific and Industrial Research Organisation (CSIRO) and AuScope in Australia have been working in parallel to develop sample management systems that include IGSN ID registration. The CSIRO and AuScope systems are based on different repository platforms: Specify versus CKAN. The systems also have different audiences/remits: an internal sample management for national collections versus an open researcher catalogue. Here, we present both to enable the community to compare and contrast these systems.

CSIRO Mineral Resources (CMR) Specify Sample Catalogue

CSIRO is Australia's national science agency—a governmental corporate entity. It has existed for over a century, working with industry, government, and the research community to translate scientific and technological advancements into solutions to address Australia’s most significant challenges. CSIRO is one of the largest and most prominent mineral research and development organisations in the world, and it manages several national collections for the benefit of research and industry.

CSIRO is one of the founding members of IGSN e.V., adopting the IGSN ID early for its CMR holdings, and developing an internal system to manage the samples and the IGSN ID registration process. Unfortunately, a combination of limited maintenance and technology obsolescence of the CMR legacy system has, in recent years, led to digital curation challenges, and it has become increasingly difficult recently to find and access mineral samples collected from CSIRO projects. These difficulties have been combined with physical curation challenges, which are handled on a project basis and do not always fully comply with workplace safety requirements.

Since its mineral samples are both essential and expensive research assets, in 2023, CMR decided to streamline its digital and physical sample curation processes. Through its Discovery Program, CMR funded the one-year End-to-End Sample Curation Project, which has produced the following deliverables:

• A set of guidelines for curating physical samples, including a sample curation workflow.
• A catalogue to manage and access samples (CMR Specify Samples Catalogue; Fig. 1).
• Engagement with and training for CSIRO researchers around sample management best practices.

To exemplify the applicability and value of the new guidelines and catalogue, they were demonstrated on two CMR collections, one for rocks and another for soils, jointly comprising around 1,500 samples. The catalogue now contains over 34,000 records of samples collected from various mineral exploration projects in CMR, including microanalysis samples and the CMR rockstore legacy collection. The curation of a further ~3,500 rock and mineral samples is in progress and will be added to the catalogue in the coming months.

As has already been alluded to, the CMR Sample Catalogue has been configured based on Specify 7 (Fig. 2). The project team comprising engineers and sample curators were careful in their choice of Specify, ultimately deciding that it has not only the needed functionality for managing and accessing material samples but also several other benefits for CMR over similar repository platforms:

• Specify is used for data management by the CSIRO National Research Collections, so internal expertise and integration opportunities already exist. 
• It is collaborative, open-source software.
• It is machine and human friendly. 
• It has a well-established consortium with solid community support.

With the CMR Specify Sample Catalogue in place, the next stage is to integrate it with relevant FAIR-enabling services (Fig. 3). In this regard, the catalogue is now becoming a norm for the management of samples by CMR researchers and curators, and it is in the process of being added to CMRs' Data Portal to enable users to discover and access samples. The catalogue has also been interfaced with the Mindat.org API—the largest open database of minerals, rocks, meteorites, and their localities—and several internal CSIRO Enterprise Services. Further integrations with the Fieldmark electronic field notebook and CSIRO's existing DOI/IGSN ID registration services will finally realise the complete end-to-end tracking capability envisioned by the project. By connecting with relevant FAIR-enabling services, the ultimate goal will be for the system to enable CSIRO to quickly discover all its research assets.

This work is supported through the CSIRO Mineral Resources (Discovery) Strategy Project OD-233946 – End-to-End Sample Tracking. We acknowledge the contribution of the project members: Anusuriya Devaraju, Anusree Ramachandran Menon, Tenten Pinchand, Tina Shelton, Kirsten Fenselau, and Jacob Walmsley.

AuScope Sample Repository

AuScope is funded by the Australian Government under the National Collaborative Research Infrastructure Strategy. It is Australia's provider of research infrastructure to the national Geosciences community, working on both fundamental questions and grand challenges. AuScope helps researchers in academia, government, and industry to address such questions and challenges through shared and free use of its infrastructure, supporting national leadership and collaboration across groups. As part of its infrastructural offerings to the Australian Geosciences community, AuScope has recently launched its Sample Repository (Fig. 4).

The main objective for building the Sample Repository is to store, disseminate, and publish sample metadata produced by AuScope-funded projects and instruments, ensuring that AuScope is in line with best practices in infrastructure development both nationally and internationally. In particular, it was developed as part of a multi-year strategy for the AuScope Downward-Looking Telescope that is aligned with Australia's National Research Infrastructure implementation plan.

The Sample Repository enables the Australian Geosciences researcher community to share and make available its metadata for others. To assist with this, the repository provides a public catalogue with spatial search capabilities to facilitate discovery and online indexing (Fig. 5). Special attention has also been given to the design and provisioning of an API for integration with other applications and services. Samples seldom come individually, and thus, it is essential to provide an API that can be used to register collections of samples in the catalogue as a bulk upload.

Other features of the Sample Repository include:

• Authentication via the Australian Access Federation
• Batch sample metadata submission
• Semantic linking of samples and sample collections
• Custom sample metadata schema and controlled vocabularies
• Simplified curation process for approved data depositors

The Samples Repository was developed alongside the AuScope Data Repository. These repositories have been designed to follow AuScope's core principles of providing open-source, sustainable, and secure infrastructure that enables researchers and research organisations to be attributed for their outputs, increases the FAIRness of those outputs, and ensures the evolving CARE Principles are adhered to, providing flexibility for future improvements in technology and processes (Fig. 6).

Although the AuScope Data Repository and the AuScope Sample Repository have different underlying models, they are both based on the CKAN repository platform. They are also integrated with the DataCite DOI/IGSN ID registration service operated by the Australian Research Data Commons. The AuScope repositories, therefore, share a common underlying system and functionality, putting the curation of research data and sample collections side-by-side, and enabling the seamless identification, citation, and linkage of samples and collections to research data and publications.

The decision to use CKAN as the repository's underlying platform came after performing a comprehensive technology review of similar repository platforms. CKAN and The Dataverse Project (Dataverse) were the two primary candidates—both open-source and with a strong community, but with their own advantages and disadvantages. Even though Dataverse better matched the technical system requirements, CKAN's modular architecture offered more flexibility for extensibility and ease of adding missing features. Also, at the time of conducting the review, Dataverse was mainly focused on providing a robust data repository functionality and had little provision for handling sample metadata, which was critical for the AuScope Sample Repository project. Another key factor was the parallel development of the AuScope Data Repository, and sharing technology and engineering capability improved the efficiency of both project teams.

The project team behind the AuScope Sample Repository plans to organize a series of workshops and activities to promote its public release and use by the national Geosciences community. Once the repository’s adoption increases, it will be registered with re3data, and integration with other services and portals (e.g., the AuScope Discovery Portal) will be explored.

Conclusion

The main lesson learnt from the development of the CMR and AuScope systems is that effective sample curation and management is more than simply providing technology and infrastructure, it requires buy-in from all of the relevant stakeholders, as well as a shared understanding and the widespread implementation of best practices. It is also clear that the physical and digital sample curation processes are complementary and must be implemented in parallel. Digital adoption is essential for system uptake, and digital solutions should be applied at the earliest stages of the sample lifecycle.