Data is the cornerstone of medical research and one of the world’s most valuable assets. Why then, are datasets rarely standardized, exchanged, or reused? In the academic world, most attention is given to the publication of the research results and the FAIR Data principles get left behind. The lion’s share of academic institutes and journals do not have incentives or policies in place for sharing data after publication. And even if institutes have such policies in place, it is quite difficult for researchers to comply with them. After publication, the underlying datasets therefore typically end up in a drawer and are forgotten. Used datasets are not often looked at again by the same researcher, let alone reused by other researchers. In fact, a majority of medical research data is not reusable. This leads to a tremendous missed opportunity, as datasets could be reused to test for replicability, meta-analyses, or to test completely new hypotheses.
In recent years there has been increased discussion about replicability of research findings and reducing waste in clinical science. Then, in 2016, an international group of researchers came together in the Netherlands to discuss how to enable the sharing and reuse of datasets. The result was a highly influential paper with guidelines on how to share data based on the so-called FAIR Data Principles (Findable, Accessible, Interoperable and Reusable). The paper was picked up by the European Commission and the G20. Researchers receiving H2020 funding are now encouraged to make their data FAIR. Sharing data based on the FAIR principles means that the data is standardized and described in such a way that it can be easily reused by both humans and machines. In this article, we will describe 7 initiatives promoting FAIR data sharing that you should know about.
Source: Scientific Data
FAIR Data Initiative #1: GO-FAIR
GO-FAIR is an international organization established to promote and implement FAIR data sharing and aims to implement the vision of the European Open Science Cloud. The organization consists of three pillars:
- Go Change aims to create a paradigm shift in academic culture toward FAIR data sharing. For this to happen there needs to be proper incentives and rewards for sharing research data. The group works on promoting policies, but also offers a research starter kit on how to make your research data FAIR.
- Go Train aims to train researchers to become FAIR data stewards. Because of the enormous increase in the amount of data produced, there is an increasing need by institutes for professional data stewards. Go Train offers a training program to become a professional data steward and provides certificates. Interested in a job as a data steward? Check out their curriculum.
- Go Build envisions an internet of FAIR data and services. It supports communities with the technical implementation of creating FAIR data infrastructure.
The approach of GO-FAIR is to change the academic culture from the bottom-up. Inspired? Local initiatives are encouraged to get involved.
Source: The propeller of academic research. GO-FAIR
FAIR Data Initiative #2: Institute for Data Science, Maastricht University
The Institute for Data Science is a new institute at Maastricht University, founded by one of the main contributors of FAIR, Dr. Michel Dumontier. The institute’s mission is to train scientists and create a platform for collaborative and reproducible science across disciplines. One of its research fields is the development of Artificial Intelligence (AI) coupled with FAIR data. Other research groups are also doing a great job to promote FAIR, such as the Human Genetics group at the Leiden University Medical Center.
Source: Dr. Dumontier: “What’s not to like about big, open, FAIR data?” IDS
FAIR Data Initiative #3: FAIR Metrics
The GO FAIR Metrics Group is a collaboration by some of the original authors of the FAIR Data Principles paper working on a framework for evaluating FAIRness of data quantitatively. Research communities can propose their own metrics based on this framework. The framework allows research communities to design and submit their own metrics of FAIRness in their field. The FAIR Metrics framework will lead to standards of ‘FAIRness’ in research communities. This project is still work in progress. You can contribute to this initiative by sharing the metrics of your own research group.
FAIR Initiative #4: Fairsharing.org
FAIRsharing.org is a searchable registry of standards, databases and data policies following FAIR Data Principles. It is maintained by the University of Oxford E-research Center. As data resources are often fragmented, FAIRsharing.org connects the landscape of the different standards, databases and data policies of a specific field. If you are looking for interoperable datasets following similar taxonomy and standards this is a good place to start. If you are looking for an overview on which standards (actively maintained ontologies) you could use for your research project this is also a good resource.
FAIR Initiative #5: Personal Health Train
Findability, accessibility, interoperability and reusability are principles that are not just applicable to research. FAIR Data Principles could also have great potential in data exchange between hospitals, physicians and caregivers. The issue with transferring clinical data is that it is patient privacy sensitive. It therefore cannot be openly shared in the cloud. The Dutch Technology Lifesciences (DTL) institute is working on an infrastructure that allows data control to remain within the organization with only authorized packages of data to be transferred between institutes. DTL collaborates on this project with a large number of partners, including Castor EDC.
Source: The Personal Health Train concept video
FAIR Data Initiative #6: DANS
FAIR data sharing has its roots in medical research, but is also being adopted in other disciplines. This process is accelerated by institutes such as DANS (Data Archiving and Networked Services). DANS is a Netherlands based institute that offers an online platform with research datasets. Their platform currently contains 240,000+ datasets in a wide range of disciplines. The DANS institute is currently working on implementing FAIR metrics scoring system to their resources. If you are looking for public datasets for your research, this is a great place to look.
Source: (F+A+I)/3=R, according to DANS
FAIR Data Initiative #7: ELIXIR Rare Diseases Community
One field for which data sharing is of particular importance is that of orphan diseases. As it is difficult to recruit patients, adequately powered studies are rare. By sharing data across institutes in a standardized format, it will be possible to combine datasets to create better insights into these diseases. For this reason, European life-science data organization ELIXIR created a community aimed at rare diseases. The community plans to build an infrastructure that will enable researchers to discover, access and analyze different rare disease repositories across Europe. As the portal is still in a pilot phase, Castor EDC went ahead and implemented fairification of the registry of vascular anomalies (VASCA).
Castor EDC is committed to FAIR Data
Castor EDC believes that the FAIR Data Principles will improve healthcare in the long run by helping to ensure medical guidelines are based on higher quality evidence. By incorporating the FAIR Data Principles in Castor, we eliminate these barriers as we allow researchers to expose their data in a way that research data can be shared easily between research projects worldwide.