Minutes of the ODEX4ALL project meeting (via Skype)
Attendees: Luiz Olavo Bonino, Arnold Kuzniar, Anand Gavai, Kees Burger & Mark Thompson
Date : 26-06-2015
- FAIR Data Point
- Luiz proposed to start with a light-weight software layer called FAIRport and to postpone the work on the FAIRifier
- N.B.: In fact, Arnold/Anand are already looking into how to FAIRify the data from Richard Finker's (WUR) BreeDB database of tomato germplasms collection, which serves as a use case
- hence we will assume FAIR Data Resource, which provides FAIR compliant data i.e. in one of the RDF serialization formats which support named graphs (supporting nanopublications): RDF/TriG, RDF/N-Quads, RDF/JSON
- N.B.: Users should be able to exchange between different RDF serializations (including Turtle etc.); however, the exchange/conversion from named graphs to unnamed graphs might result in loss of information
- the aim is to expose user/data specific FAIR Profiles to the data consumers via Web GUI and APIs (REST?)
- write the libraries in different languages (e.g. R, Python, Perl, Java) for a particular version of the Data FAIRport specification (Mark Wilkinson also implemented a prototype in Perl); the idea is to support a broad community of users
- specification/software layer needed to monitor usage patterns (matrix) of FAIR Data Points.
2. Triple annotator of nanopublications is work in progress by Kees and Mark
- the tools is part of the BRAIN platform hosted by Euretos
3. Potential collaboration with the group of Morris Swertz on Molgenis.
- similar goals as the FAIR Data Initiative
- they have developed a pipeline that takes RDF data as input and exposes FAIR Profiles through the web, both GUI and API
Questions by Arnold/Anand:
- After data are made FAIR, does Molgenis platform take these data as input to generate API and web interface?
Is Molgenis another Data FAIR point, similar to BreeDB, in the FAIR ecosystem?
Luiz: Currently, Molgenis is not FAIR compliant. The goal is to, later, Molgenis automatically generates FAIR Data Points for the datasets submitted to it.
The first step is for us to develop a first FAIR Data Point (none exists now) for the (selected parts of) BreeDB. Then we can try this initial reference implementation with other datasets such as the NanoPub Store. Once we reach a point where the reference implementation of the FAIR Data Point satisfies some quality criteria, we can start the collaboration with the Molgenis team to adjust Molgenis in such a way that it will generate the FAIR Data Points.
- Are nanopublications stored as named RDF graphs only?
Luiz: Yes. It is a requirement of the model for the graph to have a name. A nanopublication has 4 graphs. See http://nanopub.org/guidelines/working_draft/
- Has anyone of you (besides Mark Wilkinson) understanding of the Data FAIRport Specification (v0.1) document? We would appreciate your input on this to get started. Thanks.
Luiz: The specification of the FAIR Profiles are not yet in a stable form. So, we should start the work with the assumption that the specification .
- We will use the JIRA/Connfluence collaboration tool (Agile Platform) to share project-related documents, use cases, software requirements and specifications. Luiz already created an account for Arnold and Anand (https://dtl-fair.atlassian.net)
- One could translate data from a relational database to RDF graph using one of the mapping languages: R2RML and D2RQ.
- Jesse van Dam (WUR) is also interested in contributing to the FAIR developments.
- Availability of people involved during summer:
- Kees on holidays: July 25 - Aug 9
- Luiz on holidays: July 3 - July 26
- Mark on holidays : Jul 8 - Aug 4
- Arnold and Anand will look into Mark Wilkinson's Perl implementation of the Data FAIRport specification (v0.1) and Molgenis API.
- Arnold and Anand will continue working on the BreeDB use case.
- Next week Thursday we will have a Skype meeting with Luiz???