Democratizing clinical-genomic data: How federated platforms can promote benefits sharing in genomics

Abstract

Since the first sequencing of the human genome, associated sequencing costs have dramatically lowered, leading to an explosion of genomic data. This valuable data should in theory be of huge benefit to the global community, although unfortunately the benefits of these advances have not been widely distributed. Much of today’s clinical-genomic data is siloed and inaccessible in adherence with strict governance and privacy policies, with more than 97% of hospital data going unused, according to one reference. Despite these challenges, there are promising efforts to make clinical-genomic data accessible and useful without compromising security. Specifically, federated data platforms are emerging as key resources to facilitate secure data sharing without having to physically move the data from outside of its organizational or jurisdictional boundaries. In this perspective, we summarize the overarching progress in establishing federated data platforms, and highlight critical considerations on how they should be managed to ensure patient and public trust. These platforms are enabling global collaboration and improving representation of underrepresented groups, since sequencing efforts have not prioritized diverse population representation until recently. Federated data platforms, when combined with advances in no-code technology, can be accessible to the diverse end-users that make up the genomics workforce, and we discuss potential strategies to develop sustainable business models so that the platforms can continue to enable research long term. Although these platforms must be carefully managed to ensure appropriate and ethical use, they are democratizing access and insights to clinical-genomic data that will progress research and enable impactful therapeutic findings.

Legal aspects of privacy-enhancing technologies in genome-wide association studies and their impact on performance and feasibility

“…HE and SMPC profit in return from reduced communication and computation overhead [ 36 ]. Figure 4 shows which concrete actions can be taken when projects fall within the scope of the GDPR in the following six areas: data collection [ 87 ], data storage [ 39 , 88 ], quality control [ 89 – 91 ], genotype imputation [ 45 , 52 ], SNV association tests and follow-up analysis [ 21 ] as well as visualisation in distributed GWAS analysis.

Fig.

…”

Efficacy of federated learning on genomic data: a study on the UK Biobank and the 1000 Genomes Project

“…Federated learning, on the other hand, is designed to operate on isolated data silos. Therefore, there is a growing demand for federated learning infrastructures for healthcare data (Nik-Zainal et al, 2022 ; Alvarellos et al, 2023 ). Researchers have applied federated learning to a variety of healthcare data, including electronic health records (Vaid et al, 2021 ), medical images (Li et al, 2020 ) and wearables (Chen et al, 2020 ).…”