Pilot 3 - Reproducible Workflows
Computational experiments are essential for modern research, yet their complexity often hinders reproducibility. The TIER2 Pilot 3 publication “A Virtual Laboratory for Managing Computational Experiments” introduces SCHEMA lab, an open-source virtual environment that enables the design, execution, and tracking of containerised experiments with full provenance. By capturing configurations, datasets, software environments, and performance metrics, SCHEMA lab enhances reproducibility and transparency across disciplines. It supports individual researchers and research infrastructures in organising, comparing, and reusing computational workflows, fostering credible, reusable, and FAIR digital science practices.
Stakeholders: Life scientists, computer scientists
Objectives: The main goal was to customise and evaluate tools/practices for reproducible workflows in life and computer sciences, with underlying objectives of extending the SCHEMA open-source platform to support reproducibility in both fields by leveraging software containerisation, workflow description languages (CWL, Snakemake), and experiment packaging specifications (RO-crate), particularly emphasising machine learning in computer science.