Phenotype harmonisation for multi-study genomics collaborations: lessons learned from African research initiatives

Abstract The importance of phenotype harmonisation has been highlighted by the need for an increase in the number of large-scale collaborations aiming to capitalise and make the best use of existing data resources. Phenotype harmonisation, a sub-component of data harmonisation, describes a retrospective data transformation process aimed at unifying disparate datasets (consisting of phenotype, environmental exposure, and clinical data) to maximise their interoperability and reuse. In this review, we aim to provide a brief background for phenotype harmonisation and summarise our own experiences within such activities to address the lack of information about how to approach phenotype harmonisation. We raise key considerations for researchers embarking on a phenotype harmonisation effort, including: the importance of participation and collaboration; viability assessment; preparation; data management; training and knowledge sharing. As phenotype harmonisation is evolving across the world, our shared experiences and setting globally acceptable standards will accelerate knowledge discovery and unlock the potential of bringing together multiple datasets.