From July 20 until July 22 a small group of selected experts met at Oxford University (UK) to revise the specifications of the de facto standard for capturing information about biomedical research in a structured and computer-supported way. The meeting was part of the ELIXIR UK Node activities and with the participation of members of the ISA user community, ELIXIR, EBI and NL Nodes, FAIRDOM, RO, and Nanopublications groups.
The ‘investigation, study, assay’ (ISA) standard and associated tools provide a backbone for computer-assisted research and data descriptions. These can be about complete study designs, but also about specific technologies used and the effects those have on data property descriptions, it can even be used for minute note-keeping on lab protocol execution. One of the strengths of ISA is that it allows the combination of different studies, using different assays, in one analysis. ISA thus helps researchers and data stewards to follow FAIR data principles. FAIR stands for Findable, Accessible, Interoperable and Reusable. ISA helps to make data more interoperable by allowing more explicit descriptions; based on agreed on terms from ontologies and controlled vocabularies. That of course also helps to find studies, since you can now know the terms to look for or you can use tools to translate between equivalent terms. Finding data and having them in an interoperable form are key steps in making data reusable.
The revised ISA specifications will have a clearer underlying backbone comprising a core that is used across all tools that use ISA, and specialised modules for sub-domains (e.g. imaging, RNA sequencing, metabolomics, nanomaterial descriptions). Developing the modules will be a coordinated action between core developers and selected members from the ISA community (ISA Commons). ISA has stand-alone tools for structuring and annotating the research process and its outcomes, but – perhaps more importantly – can be used to extend commonly used tools.
Apart from the core ISA team led by Susanna Sansone (Oxford e-Research Centre), participants included metabolomics repository developers from the European Bioinformatics Institute (EBI), engineers of the data publishing journal GigaScience, and Research Object designers from the University of Manchester. Marco Roos (LUMC, DTL) was invited to the meeting given his involvement in Elixir-NL and previous contributions, such as to the development of nanopublications and merging this with two other important mechanisms for study capture: Research Objects , and ISA (DOI: 10.1371/journal.pone.0127612). Other representatives of DTL/ELIXIR-NL were Chris Evelo & Egon Willighagen of Maastricht University, and Jildau Bouwman of TNO.
The Netherlands has a long-standing experience in capturing study data in the life sciences and medicine. This work has lead to further developments of the ISA model. For example, ISA has been extended to describe systems biology investigations that encompass both laboratory experiments and mathematical modelling analyses (Katy Wolstencroft, FAIRDOM project), and it has been used for complex designs in nutritional studies (Jildau Bouwman, Phenotype database). This experience will be leveraged in future developments of the ISA-TAB standard by forming a study capture DTL Interest Group that will be led by Jildau Bouwman.