Content area
Full Text
This journal and Scientific Data are calling for submissions containing linked open data models that embody and extend the FAIR principles: that data should be findable, accessible, interoperable and reusable by both humans and machines. These principles are achievable with existing resources, languages and vocabularies to enable computers to combine and reanalyze data sets automatically and lead humans to new discoveries.
This May, individuals who are experts in a range of scholarly disciplines-but relatively unfamiliar with computational data modeling-learned how to make the Internet of Data and Services unearth the full implications of their research. First, in a workshop at the Lorentz Centre in Leiden, the Netherlands (https://www.lorentzcenter.nl/), data from library collections, museum accessions, African mobile healthcare records, tomato breeders and virologists were all subjected to data stewardship according to the FAIR principles (Sci. Data 15, 160018; 2016; https://www.dtls.nl/fair-data/fair-principles-explained/). Then, in a similar but accelerated event at the fifteenth-anniversary conference of Bio-IT World in Boston (http://www.bio-itworldexpo.com/), data sets dealing with pediatric oncology, personal genomic SNP variants and curated human mutations (https://www.ncbi.nlm.nih.gov/clinvar/) were given the FAIRification treatment.
What we learned from these exercises is that it currently takes about two days working with a dedicated repository and a data engineer to add and link the minimum metadata a computer needs to make a data set work with others. We also learned that, despite the motivational potential of the FAIR acronym, there is no un-FAIR....