A framework for reproducibly managing coupled research software and data assets based on shared transformation...

In computational science, coupling research software and data into workflows is essential for addressing complex research questions and ensuring reproducibility. However, current metadata schemas rarely provide sufficient information about data models in software interfaces and datasets, hindering their effective integration into workflows. Additionally, most workflow management tools are not designed to handle metadata for process reproduction. This article refines the DataDesc metadata schema to annotate data models not only in software interfaces but also in datasets and presents an extension to the DataDesc framework for automatically comparing data models and identifying transformation requirements. The ioProc workflow manager is introduced to bundle shared transformation functions into adapter workflows and ensure transparent process documentation. Two use cases from energy systems analysis demonstrate the workflow design and implementation approaches. Overall, this article bridges the gap between abstract guidelines, such as the findability, accessibility, interoperability, and reusability (FAIR) principles, and researchers' daily data- and software-driven analyses, promoting reusability, reproducibility, and transparency in science.