Chapter 5: Computational Workflows

In order to investigate RQ3, and considering important parts of the FAIR principles are Reuse and provenance, this chapter examines in closer details how FAIR Digital Objects and RO-Crate can be used with Computational Workflows.

Section 5.1 proposes that tools in computational workflows, when wrapped as interoperable building blocks, can be considered as FAIR Digital Objects, with a use case from biomolecular simulation.

Sections 5.2 and 5.3 explore how FDOs and Research Objects can be constructed incrementally using computational workflows, with a use case from specimen digitization in natural history collections.

Section 5.4 presents a profile of RO-Crate to capture workflow execution provenance, with incremental granularity levels and six workflow engine implementations. Use cases include machine learning-aided tumour detection and compatibility with PROV approaches.

Supplementary materials that may assist readers of this chapter provide further details on FAIR Computational Workflows [Goble 2020], WorkflowHub [Goble 2021], Common Workflow Language [Crusoe 2022] and making a software tool workflow ready [Brack 2022a].

On the aspects of workflow provenance, recommended reading in supplementary materials covers CWLProv [Khan 2019], RO-Crate in Galaxy [De Geest 2022] and Common Provenance Model [Wittner 2020, Wittner 2023a].