Introduction to the FAIRplus project
The FAIRplus project ran from January 2019 to December 2022 and produced a reusable FAIRification Framework for life science data. The framework includes tools and guidelines for making life science data FAIR (Findable, Accessible, Interoperable, Reusable). See About the project.
FAIRification Framework
Determine the goals for FAIRification in terms of desired usability of data that isn't currently possible.
- If you already have a goal in mind for your dataset, start with the FAIR wizard.
- If you have a dataset but do not have a clear idea of how it could be improved, start with the DataSet Maturity tool.
- If you are not sure which of these routes is best for you, have a look at the recipe about the FAIRification framework in the FAIR Cookbook.
Examine in detail a dataset's current and expected data requirements as well as available resources and expertise.
Before you can put together a workplan to implement the FAIRification goal, you need to make sure that you fully understand all the data requirements, project capabilities and resources for your project.
For example, if the FAIRification goal targets the annotation of data with open terminologies, it is only implementable if the data is sufficiently well understood to identify suitable controlled vocabularies and ontologies, and if expertise is available to perform the annotation to a sufficiently high standard.
Data requirements
At this stage, you need to identify the characteristics your FAIR dataset should exhibit based on the previously defined FAIRification goal, such as conforming to a specific community standard, creating identifiers, etc. Then you can define the tasks required to fulfil these characteristics.
Every task should be linked to a FAIR indicator, which will specify the actions to be taken but also help identify the expertise needed to fulfil it. In addition, these tasks are expected to have varying levels of complexity depending on the maturity level targeted for the dataset.
Capabilities and resources
In addition to identifying data-related requirements, realising a FAIRification goal also depends on services, capabilities and resources that FAIR data management and hosting environments should exhibit to enable and support the realization of a FAIR dataset.
These tasks are categorized as ”FAIRification capabilities and resources” and include considerations such as data access, data hosting, ontology services and data sharing amongst others, but also the human resources available to perform FAIRification. As with data requirements these capabilities are expected to vary depending on the level of maturity achieved or targeted.
FAIRification backlog
Now that you have the FAIRification goal, you can now go ahead and create a list of tasks that you would like to do. These tasks do not include implementation considerations yet. This is best done in an agile manner where for each task you assign a priority based on what needs to be completed first and have a person responsible for taking care of the task completion.
Iteratively design and implement FAIRification tasks to reach the original goals
The practical part of the process centres around the FAIRification cycle, which consists of three separate stages: assessment, design and implementation.
This phase typically consists of multiple FAIRification cycles applied in an iterative fashion, with each FAIRification cycle focusing on a single FAIRification goal. The optimal cycle time can vary depending on the project and on the FAIRification goal of that cycle.
An assessment step sits both at the start and the end of each implementation cycle, with the output assessment of one cycle usually serving as input to the next one, where appropriate.
For the first cycle, the assessment should have been completed as part of the project examination phase. The assessment is ideally done by one or more reviewers who are not part of the project team performing the FAIR implementation tasks.
Based on the prioritised tasks defined in the FAIRification backlog (see the "Examine your project" section above), one or more related tasks are picked up for the current FAIRification cycle.
During the design stage, concrete actions from the FAIRification template are identified to achieve the FAIRification goal identified for this cycle. These actions form the FAIRification work plan to be realised during the implementation stage.
Note down all the potential tasks you can think of irrespective of which ones can be done. Each task should be assigned a priority level in terms of where the biggest impact (to FAIR level) can be achieved most easily and efficiently (available resources).
| Task | Priority | Task moderator |
|---|---|---|
| Map all anatomy-related terminology to UBERON | Medium | John Doe |
| Cross-check metadata records against MINSEQE and identify missing data elements | High | Jane Smith |
| Propose DUO-based data reuse conditions and circulate to all project partners for approval | Low | Peter Parker |
At this stage, one will have a complete list of all possible tasks that have been identified, that could contribute towards the defined FAIR goal. In addition, other tasks that have been discovered, which may be implementable at a later stage, should also have been documented. These will form part of the final feedback on future tasks, which can be used to improve FAIR status of datasets at a later time, if desired.
From the list of all possible tasks, it is necessary to define those that could be:
- Achievable in the time frame defined for the cycle (eg 3 months)
- Accomplishable with the team composition available
This refined list of tasks represents the FAIRification work plan for the cycle. The work plan should include a task status to indicate which tasks are completed, in progress and still to do.
See the FAIR Cookbook for practical guidance on common FAIRification tasks.
| Task | Priority | Task moderator | Task status |
|---|---|---|---|
| Map all anatomy-related terminology to UBERON | Medium | John Doe | Done |
| Cross-check metadata records against MINSEQE and identify missing data elements | High | Jane Smith | Ongoing |
| Propose DUO-based data reuse conditions and circulate to all project partners for approval | Low | Peter Parker | To do |
Review outcomes and assess the success against the original goals.
In this final phase, the cumulative outputs of all the FAIRification processes should be compared to the initial project goals, to assess the overall success of the process. This should include another assessment of your dataset with the FAIR DSM tool.
This review phase should include feedback from one or more reviewers not directly involved in the practical implementation work but familiar with the overall data and scientific goals of the project. It may show that the desired degree of FAIRness has not yet been achieved, in which case a renewed cycle of assessment and FAIRification may be considered. If your goals have been achieved, then congratulations on your progress!
Remember - FAIR is a journey!