Duality Technologies Researchers Accelerate Privacy-Enhanced Collaboration on Genomic Data
Duality Technologies, a leading provider of Privacy-Enhancing Technologies (PETs), announced the publication of an article in PNAS (Proceedings of the National Academy of Sciences of the United States of America), on Secure large-scale genome-wide association studies using homomorphic encryption.
The paper, co-authored by Prof. Shafi Goldwasser, Dr. Marcelo Blatt, and Dr. Yuriy Polyakov of Duality Technologies, along with Dr. Alexander Gusev of the Dana-Farber Cancer Institute and Harvard Medical School, presents a solution leveraging homomorphic encryption (HE) to perform large-scale Genome-Wide Association Studies on encrypted genetic and phenotype data for a dataset of over 25,000 individuals, yielding results 30 times faster than an alternative state-of-the-art method based on Multiparty Computation (Cho et al 2018 Nat Biotechnol). This work was partially supported by a grant from the National Human Genome Research Institute of the National Institutes of Health (NIH).
Recommended AI News: Fixd Maintenance Management Software to Expand COVID-19 Industry Support Package
Genome-Wide Association Studies seek to identify genetic variants associated with a particular trait and have been crucial in understanding complex diseases. However, they depend on individual-level, highly-sensitive genomic variant data which is typically protected by privacy regulations – making it challenging to gather large volumes of data that are required for this type of research. HE technology enables collaboration on large-scale genomic and clinical studies while protecting the privacy of the individual participants. The performance breakthrough documented by the Duality researchers demonstrates that HE is a practical solution for privacy-preserving GWAS computations, with the ability to scale genome-wide analyses to hundreds of thousands of individuals.
GWAS can yield insights into a range of diseases and is currently being utilized to identify genetic variants associated with susceptibility or response to COVID-19. As health authorities and medical researchers seek to understand the true nature of the novel condition, homomorphic encryption enables researchers to pool individual-level data on a global scale to perform genomic analysis in a privacy-preserving manner. Duality’s technology demonstrated by this advanced research is currently available to the healthcare industry.
The article details a privacy-preserving framework based on advances in HE and demonstrates that Duality’s advanced solution can swiftly perform accurate GWAS analysis while keeping all individual data encrypted. The technology described in the paper is highly scalable: the researchers’ extrapolations show that this method can perform GWAS on 100,000 individuals and 500,000 Single Nucleotide Polymorphisms (SNPs) in 5.6 hours on a single server node (or in 11 minutes on 31 server nodes running in parallel). These performance results are significantly faster than prior secure computation methods with a similar level of accuracy. The advances can also be applied to other branches of medical research such as clinical trials, drug repurposing and rare disease studies.
Recommended AI News: New Jersey Sportsbooks Hit 2-Year Low While Online Casinos Boom, According to PlayNJ
“Today, the critical barrier to personalized medicine is no longer just data availability, but data privacy. With high accuracy even for very subtle associations, HE-GWAS results can be used to derive polygenic risk scores for accurate patient stratification and prediction of future disease,” said Dr. Alexander Gusev of the Dana-Farber Cancer Institute and Harvard Medical School. “Ongoing precision medicine studies can immediately benefit from these capabilities by enabling secure collaboration across clinical institutions without requiring complex data sharing agreements or compromising individual-level privacy. This technology can also empower patients to participate in research studies directly and receive personalized results knowing that their individual data will not be exposed.”
“At this time of crisis, when the world’s eyes turn towards the potential held by collaboration on sensitive individual medical data, our interdisciplinary research team of cryptographers and data scientists, in cooperation with a leading bioinformatician, demonstrated the feasibility of data privacy compliant global collaboration,” said Prof. Shafi Goldwasser, Director of the Simons Institute for the Theory of Computing at UC Berkeley, Co-Founder and Chief Scientist of Duality Technologies, and co-author of the paper. “Our implementation shows that homomorphic encryption is practical, functional, and is significantly faster than existing options for secure data analysis. Duality Technologies is proud to collaborate with Dr. Alexander Gusev to highlight the use of its market-ready analytics platforms in scientific research, demonstrating that homomorphic encryption will play a central role in protecting privacy while alleviating the current global health crisis. This has significant implications for healthcare, and a variety of other vital fields.”
“We are excited by the outcome of this research, demonstrating that complex data science computations are feasible at scale on homomorphically encrypted data,” said Dr. Marcelo Blatt, Head of Data Science at Duality Technologies. “These capabilities are now available in our SecurePlus Platform for the broader healthcare industry, enabling privacy-enhanced medical collaboration.”
“We have far exceeded what is considered state-of-the-art performance of homomorphic encryption computations,” added Dr. Yuriy Polyakov, Principal Scientist at Duality Technologies. “The secret behind our breakthrough is our unique combination of world-leading cryptographic and data science experts. It is gratifying to see these technological advances being used to address real-world challenges in a variety of regulated industries.”
Recommended AI News: Raima and MontaVista Announce Strategic Partnership
Comments are closed, but trackbacks and pingbacks are open.