What is it?
A pilot project designed to encourage authors to provide unambiguous identification of 3 types of key research resources (antibodies, model organisms, and software tools) used in the course of a scientific investigation. Authors will be asked to supply a minimal set of metadata, e.g., catalog number and vendor, and an accession number, similar to a Genbank ID.
Why is it necessary?
Accurate identification and tracking of research resources used to produce the findings of a scientific study are needed to promote reproducible science, to design and troubleshoot experiments, to measure the impact of individual research resources, and to data mine the scientific literature. As shown in the study by Vasilevsky et al. (2013), authors are not providing sufficient detail in their materials and methods such that the resources used can be conclusively identified by another human being. But, perhaps of more importance, we now have to start developing citation practices that make it easier for machines to extract this information. Unique identifiers ensure that the research resource is uniquely identified and can be recognized by computer algorithms. The RII thus addresses both good research practice and 21st century publishing requirements. More information can be found on the RII Working Group page.
What research resource types can I find in this portal?
In the initial pilot phase we provide three types of resources:
- Antibodies: primarily commercial antibody reagents, plus a smaller number of lab-sourced ones.
- Organisms: mouse (MGI and IMSR), zebrafish (ZFIN and ZIRC), worm (CGC and Wormbase), fruit fly (BDSC), and rat (RDG) model organisms.
- Tools: software, databases, and services aggregated from a wide range of providers.
How can we participate?
The RII is a voluntary project designed to test whether authors can accurately identify their research resources and supply the appropriate metadata and identifiers. We are asking that all interested journals and publishers sign up at: https://www.force11.org/f11_tool/?A=L25vZGUvNDg1MA==&, and to display the logo for the project on their website. Journals will then be asked to solicit this information during the pilot project phage, currently underway and continuing through 2014. Individuals and other interested parties, e.g., organizations and libraries, are welcome to join the initiative.
Who provides the accession numbers for these types of research resources?
Identifiers are provided by the authoritative source for each type of research resource. The Neuroscience Information Framework (http://neuinfo.org), an initiative of the US National Institutes of Health Blueprint Consortium, provides the identifiers for software and antibodies, through their Research Resource Catalog and the Antibody Registry. Model organism identifiers are provided by the appropriate model organism community database, e.g., ZFIN or the Mouse Genome Informatics database. The RII project aggregates these identifiers and makes them available from a single portal (http://scicrunch.com/resources).
Where should this information be provided?
Authors should add this information to the Materials and Methods section of their paper, just as they would a regular citation or a Genbank Accession Number. However, as the Materials and Methods is often behind a pay wall, we are asking all participants to make the identifiers available outside of the pay wall and that journals adopt uniform standards. For the pilot project, we are proposing using the keyword field to report the identifiers, but we will accept other suggestions.
What would the citation look like?
The citation would include typical information, e.g., the vendor or source, plus the catalog number, if one is available, and the RRID. At this time, we aren't asking for specific xml tags or links to be embedded, although in the future, we think that such machine actionable information will be highly desirable.
Antibody: "Sections were stained with a rabbit polyclonal antibody against ERK1 (Abgent Cat# AP7251E, RRID:AB_2140114)."
Software tool: "Images were analyzed using CellProfiler Image Analysis Software, V2.0, ( http://www.cellprofiler.org/; RRID:nif-0000-00280)
Model organism: "Subjects in this study were Fgf9Eks/Fgf9+ mice ( RRID: MGI:3840442)"
Do we need to modify our manuscript submission system?
No, not for the pilot project.
Who will review the information for accuracy?
Volunteer curators for the Resource Identification Initiative from the working group hosted at FORCE11.
How should this information be solicited?
We are asking that an email correspondence be sent to the author asking for their participation. The email will make it clear that participation is voluntary, i.e., not a condition of acceptance. Based on our preliminary discussions with editors and publishers, we are leaving it up to the individual journal to decide at which point in the publication process they wish to solicit the information, e.g., at time of submission, at time of review, or at time of acceptance. We then ask that a follow-up survey be sent to the authors to see how easy or difficult it was to comply with the request.
What are the benefits to authors?
Our own experience as research scientists tell us that authors spend a lot of time looking for reagents and tools that work under certain circumstances or trying to trouble shoot experiments. At the most basic level, the RII will help with this, because authors will be able to find all papers that used a particular antibody, animal or software tool. They will also be able to search by metadata, e.g., the species in which an antibody is raised, that may or may not appear in the paper, but which does occur in the specification of the antibody. However, we believe that the RII, if broadly implemented beyond the pilot project, will provide many additional benefits down the road to science in general, and by extension, to scientists. For example, we currently have no way of alerting readers of a paper that a problem has been found with a particular antibody or software tool. The RII will provide the basis for developing such an alerting service. We believe that it will provide the basis for data mining, by allowing researchers to aggregate and compare all findings on a particular animal model, which right now are scattered throughout the literature. Finally, funders and resource providers will be able to track the utilization of their resources within papers in a way that our current citation system does not allow. We want to be able to track actual USE of the research resource, rather than a mention of the paper that might describe it for some other purpose.
What happens after the pilot project is completed?
We will be setting up some demonstration projects that will show authors, researchers, resource providers, and funders how they can use this information to improve their science and track usage. We will make the data set available to all participants to use to enhance their online offerings. We hope that this pilot project will lead to long-lasting changes in the way we report research resources within published works, leading to a set of standards appropriate for 21st century publishing. Our goal is to make all reporting of research resources:
- Machine processable and machine actionable
- Available outside of paywalls
- Uniform across all publications
Examples of the RRIDs currently being used can be found in Google Scholar or PubMed.
Can you guarantee that the ID sources will not go away?
No, we can't. Nothing in life is permanent. What we can guarantee is that they will be persistent and globally unique. But the beauty of including the RRID citations in the literature is that even if the source databases go away, the RRID's will be forever linked with the individual research resource inside of the paper. Thus, the literature itself, with its current mechanisms for persistence, is a redundant source of information about RRID's.