Advancing collaborative genomic research with federated learning: Overcoming data sovereignty barriers.

As genomic research expands globally, data sovereignty, privacy concerns, and regulatory restrictions present significant challenges for cross-border collaboration, particularly among low- and middle-income countries (LMICs). Traditional data-sharing models often conflict with ethics, governance policies, and local regulations, limiting the ability to conduct large-scale multiethnic genomic studies. Federated Learning (FL), a decentralised Artificial Intelligence (AI) approach, offers a promising solution by enabling collaborative analysis of distributed datasets without transferring sensitive data. This project will explore FL and similar AI-driven techniques to enhance collaborative synergy in genomic research while safeguarding data sovereignty.

• Address data sovereignty challenges in collaborative genomic research across LMICs by developing AI-driven solutions that preserve privacy and comply with local governance frameworks.

• Compare and evaluate the application of FL against other statistical models in disease risk prediction, focusing on complex diseases such diabetes, cardiovascular disease (CVD), or cancer using multiomics and clinical datasets from diverse populations.
• (Advanced) Develop a framework for a Softwareas- a-Service (SaaS) platform that facilitates secure and scalable genomic data analysis for common diseases in Southeast Asia.

This project will begin with a comprehensive review of FL applications in biomedical research, followed by testing hypotheses on disease risk prediction models using FL across multiomics and clinical datasets in different languages and healthcare settings.

The study will assess model accuracy, data security, and scalability while addressing computational and infrastructure challenges in LMICs. Based on these findings, the project will explore the design and implementation of a SaaS-based platform that integrates FL for secure, privacy-preserving genomic analysis in collaborative research environments.

This project presents an opportunity to pioneer AI-driven solutions for data sovereignty challenges ingenomic research, fostering equitable participation of disadvantaged communities in global precision medicine efforts. By leveraging FL, the research aims to unlock the potential of multi-ethnic genomic datasets while ensuring compliance with local policies and ethical guidelines. Additionally, the development of a scalable SaaS platform could lay the groundwork for future commercialisation or public-sector adoption, transforming the way insights from genomic data is analysed and shared across borders.