The paper reports on the ongoing effort to define the scope and practical cases for the use of persistent identifiers in research organizations that operate large-scale facilities.
Facilities science or “lab science” is a branch of research
performed by visitor scientists on large-scale scientific
instruments: synchrotron radiation and neutron sources,
powerful lasers, telescopes, or supercomputers. Facilities
science business model and research lifecycle are similar
across different instruments and geographical locations,
and involve extensive management of data and other
information artefacts [
The progress of information technology makes it
possible to mint and manage many types of persistent
identifiers for a variety of uses beyond traditional Digital
Object Identifiers (DOIs) for research papers. The use of
persistent identifiers is often considered in the context of
Open Science as a practice in support of FAIR principles
[
This paper outlines an ongoing effort to define the
scope and practical cases for the use of persistent
identifiers in facilities research as a contribution to
FREYA project [
The most well established use of persistent identifiers (PIDs) in facilities research context is assigning them to research papers published by visitor scientists and/or personnel who support and operate facilities instruments. Digital Object Identifier (DOI) is possibly the most popular type of a PID for research papers and it can be
Many types of these PIDs can be used for both
peerreviewed papers and for all kinds of “grey literature”. No
type of PID, including DOI (e.g. minted by Zenodo [
Investigation is a generalization of a notion of an experiment performed on a large-scale research facility. Investigation may include one or more experiments, with some of the experiments potentially used for instrument calibration and other experiments used for the purposeful data acquisition.
Investigation time granted by facilities can be considered a non-monetary form of a research grant. Facilities typically assign unique identifiers to investigations; these IDs are not universal but facilityspecific. If we extend the analogy of the investigation time award to research grant, an investigation ID is then similar to a grant ID assigned by funding agencies.
Some facilities such as STFC ISIS neutron and
muon source [
Each DOI assigned to an investigation is resolved in a landing Web page supported by the facility on its own Web server. The DOI can be assigned and the landing page for it created right after the time slot is granted to a visitor scientist and before the conduct of the actual experiments. The landing page is then populated with metadata collected from the research proposal managed by a facility-specific proposal system. Once the investigation is actually conducted and experimental data is collected, the landing page is supplied with a link to the data holdings, which may be restricted for an embargo period to the scientists who actually performed the experiments.
Systematic assignment of DOIs to investigations, as
well as having other structured metadata for them, make
the circulation of investigations in research discourse in
many respects similar to the circulation of research
papers [
Data collected by Facilities is "raw" experimental data
with, typically, no persistent identifiers assigned to it.
Having this said, the data can be sometimes indirectly
accessed (and indirectly cited) via the respective landing
page and the DOI that resolves in that landing page. As
an example, ISIS neutron and muon facility [
Some other facilities publish datasets on the Web or
in the FTP archives [
Facilities’ practice when there is no directly
resolvable persistent identifiers for data, or they are
URLs at best, is not unique. Even in cases when a
dedicated service for data citation such as DataCite [
ORCID [
Facilities do mint their own identifiers for visitor
scientists but these identifiers circulate only within a
local proposal system. Some facilities, such as Oak
Ridge National Laboratory (ORNL) [
facilities, and opportunities for the advanced uses of existing PID types
papers and the associated research data. To better cater for impact measurements, the instrument PIDs and the facility PIDs can be included in a common vocabulary that allows a certain level of machine reasoning (semantic inference). This will allow citation-based measuring of facilities impact when researchers cite not a facility as a whole but a particular instrument of it.
ORCIDs proliferation in facilities research, which is a multi-disciplinary research by its nature, is uneven across different disciplines. ORCIDs are promoted by libraries and IT departments of the organizations that operate facilities, yet key stakeholders for the wider ORCIDs adoption are facilities’ user offices that manage research proposals from visitor investigators.
The aforementioned ORCID User Facilities and
Publications Working Group [
Another opportunity for better ORCIDs adoption is
doing some work on behalf of researchers, e.g. the
autopopulation of institutional repositories, such as
ePubs [
As mentioned above in Section 2.3, persistent identifiers that presumably designate data can in fact be resolved in different information artefacts, not necessarily data per se. This presents difficulties for data staging to computation and visualization by virtue of the respective PIDs resolution.
For the purposes of data staging, PID resolution can
be seen as an API call with one parameter [
The protocol for data staging via PIDs can be
implemented using content negotiation mechanisms
offered by PID management services such as DataCite
[
Bibliographic records for research papers as well as other
records of science such as landing pages for
investigations or records of research awards (grants)
have a potential for their enrichment with references to
persistent identifiers such as PIDs for researchers,
organizations, or well-established systems of identifiers
for chemical substances and molecular structures [
The records enrichment, in order to be scalable and maintainable, should involve methods of textual analysis and machine learning. This presents excellent opportunities for data scientists and data engineers to showcase their methods and tools that can be used for matching textual names of researchers, organizations or chemical structures with commonly used identifiers, and then for building graphs that interlink these uncovered identities. 3.5 New discourse
representations of facilities research For facilities, the potential for better – well-structured and open – representation of research supported by them is in part related to the aforementioned opportunities for records enrichment (see Section 3.4). This can include better identification of instruments and researchers (see Sections 3.1 and 3.2).
Another opportunity for novel representations of
facilities research is the further promotion of
investigations (structured descriptions of them) as true
components of research discourse across different
facilities that are ready to accept common practices of
minting PIDs for investigations and encouraging visitor
scientists to use these PIDs as references in their research
papers. In fact, not only a well-formed investigation
record can be cited from a research paper, but
investigations can cite previous research, too, such as
previous investigations and research papers that
contribute to research behind a proposed investigation.
This allows mixing up research papers and investigations
citation in a common citation graph where anything can
refer to anything [
PID-rich descriptions of facilities research may
benefit from emerging Open Access information services
such as Open Citations [
New information-rich representations of facilities
research discourse may become possible only with
support of the proper governance, data curation and
social practices; technology alone is just a facilitator of
change. This understanding of the importance of
practices and attitudes towards PIDs use is getting
traction in projects such as FREYA [
Wider and well-curated use of persistent identifiers can support FAIR principles for research data management and stewardship. Multidisciplinary research supported by large-scale facilities presents good opportunities for the application of information technology for PIDs management, as well as for the development of best practices and communities of practice around PIDs use cases that we briefly describe in this paper.
The practical implementation of these use cases can be supported by research facilities (in particular, their user offices), by information and IT departments of organizations where facilities operate, as well as through collaborative projects and volunteer work in international forums such as Research Data Alliance. Acknowledgments. This work is supported by funding from the Horizon 2020 FREYA project, Grant Agreement number 777523. The views expressed are the views of the authors and not necessarily of the project or the funding agency.