PDaSP Track 2: Confidential Genome Imputation and Analytics (CoGIA)

Understanding genetic variations and their impact on human health is essential for advancing medicine and developing new treatments. Researchers need access to comprehensive genetic databases that represent diverse populations to effectively analyze individual genetic information. Despite the growing number of large-scale genetic datasets being generated, access to these data remains limited for many researchers due to stringent data-sharing policies driven by privacy concerns. Even when access is granted, the immense size of modern genetic datasets, coupled with the growing complexity and costs of analysis, presents significant barriers for individual researchers to fully use these resources in their studies. This project addresses this challenge by developing secure methods that allow researchers to analyze their genetic data using controlled-access reference datasets without compromising the privacy of either dataset. This work serves the national interest by accelerating scientific progress and improving public health through expanded access to genetic data resources, strengthening the security infrastructure that supports genetic research, fostering public trust in genetic studies, and enhancing protection of national genetic databases. This project develops secure algorithms and deployment-ready software to support genome imputation and analytics services that preserve the confidentiality of all data involved. The research is organized into three integrated components. First, the team will develop practical algorithms and tools for genome analysis in trusted execution environments using data-oblivious techniques to prevent side-channel leakage by ensuring uniform observable behavior across all inputs. Second, the project will develop a multi-layered security framework that combines data transformation, usage monitoring, and secure testing environments to safeguard controlled-access reference datasets. Third, the research will extend these capabilities to a federated network of genomic repositories, enabling privacy-preserving, cross-institutional genome imputation and phasing through novel federated algorithms and identity management strategies. The project leverages secure hardware and related privacy technologies to enable services that allow users to analyze their genomic data using controlled-access reference datasets without compromising the confidentiality of either. These research activities will produce deployable, confidential genome analytics services and foundational techniques for future privacy-preserving workflows built on trusted execution environments. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.