FAIRification of CIP legacy datasets

The International Potato Center (CIP) has undertaken a structured FAIRification process to improve the
quality, interoperability, and reuse of its legacy and ongoing research datasets hosted in CIP’s
Dataverse. Aligned with CGIAR FAIR and AI-ready data principles, this effort focuses on the adoption
of Core Sets of Variables validated at the CGIAR system level, embedding standardized data practices
across key research domains. As an initial priority, CIP has FAIRified legacy datasets in crop breeding,
agronomy, plant health (late blight), and in-situ crop diversity, cleaning and harmonizing existing
records and assigning persistent DOIs to ensure traceability and long-term access. In total, dozens of
high-value legacy datasets—comprising hundreds of field books and multiple years of observations—
have been standardized and made interoperable for downstream analyses, meta-studies, decision support
systems, and AI-enabled applications. In parallel, CIP proposes a minimum Core Set of Variables for
biodiversity to address the specific requirements of in-situ crop diversity datasets, complementing
existing agronomy and breeding standards. For plant health and disease epidemiology, the FAIRification
process builds on the agronomy core while incorporating domain-specific extensions suitable for late
blight monitoring, modeling, and decision support. Looking forward, CIP commits to embedding the
Core Set of Variables by default in all future breeding datasets from 2026 onward, ensuring FAIR
compliance at the data collection and management stages rather than through retrospective curation.
This FAIR-by-design approach strengthens data quality, reduces future curation costs, and positions
CIP’s data assets as AI-ready, reusable resources that directly support CGIAR’s system-wide data
governance and digital transformation objectives.