The FAIR acronym in the context of biomedical data refers to the need for any data produced with public funds to be Findable, Accessible, Interoperable and Reusable. Let’s analyse each of the words that compose the acronym.
- Findable: Because so much data is produced, be it via Next Generation Sequencing or other means (e.g., high throughput technology), the way experiments are now carried out is data heavy. This data is expected to be deposited in a public archive and referenced by the appropriate publication that has results about this data. But to be findable is not easy. Data needs to be properly described and exposed to the researcher who is searching to use it for his/her research.
- Accessible: Even if the data is found, that does not mean that it is readily available for download. This is not only for technical reasons. There may be consent frameworks that protect the data because it proceeds from patients whose identity and privacy need to be respected. In many cases data access may be restricted and have a high barrier for access.
- Interoperable: I have found that if any of these words are hot at the moment, Interoperability is probably the winner. Today, as I attended the ELIXIR-UK Small and Medium Enterprise Forum on ‘Enabling Discoverability in
Bio-Data Innovation‘ at Churchill College in Cambridge, Interoperability was among the top buzzwords. I heard Professor Carole Goble mentioning today at the conference that Interoperability has to be there for a purpose. It cannot be done for the sake of it. There needs to be a problem, usually integration of datasets or concatenation of softwares in a pipeline.
Today I had the privilege to chair the session at the ELIXIR-UK SME event about ‘Challenges in FAIR principles implementation’. The Hyve CEO Keen van Bochove said in his presentation I chaired that in order to operationalise FAIR principles the appropriate tools are needed, which we right now do not have yet. Other challenges mentioned by Kees include the fact that the data landscape for patients is sparse and diverse. Moreover, Pete McQuilton (AKA Drosophilic), during the same session gave us an overview of the FAIR landscape in ELIXIR. He made clear that in order to assess whether data shared is FAIR we need agreed metrics, which at the moment we do not have. His team, led by Susanna Sansone at Oxford University, conducted a survey where they identified some of these potential metrics, but more work is needed.
At the end of the session we had an interactive panel discussion where we asked the audience what their thoughts are regarding the single biggest challenge to adoption of FAIR values widely. It is not political or technical factors what the audience think the biggest barrier is cultural.
The overwhelming majority of the ELIXIR Innovation and SME Forum, approximately 90%, thought that the single biggest barrier to adoption of FAIR values is the current culture in the biomedical research arena.
This, indeed, is one of the top conversations that the bioinformatics community is thinking about today. This ‘gut feeling’ the community has remains to be tested more widely and if so we need to have success stories where we clearly quantify how much society benefits by making data FAIR. This indeed will be one of the solutions to change the culture: to find and communicate success stories and cases where we as a community prove that FAIR principles are imperative to embrace, regardless of any circumstance. End of day 1 of #ELIXIR4Innovation.