The Hoon Cho Lab recently presented new research in genome privacy at the Research in Computational Molecular Biology Conference (RECOMB) in Cambridge, MA. A proceedings paper from the lab, co-authored by PhD student Matthew M. Hong, received RECOMB’s Best Student Paper award.
Hyunghoon (Hoon) Cho, PhD, assistant professor of biomedical informatics and data science, leads a computational biology group at Yale School of Medicine’s new department of biomedical informatics and data science. The lab’s research, primarily funded by Cho’s NIH grant on Computational Methods for Enhancing Privacy in Biomedical Data Sharing, focuses on building algorithmic solutions to tackle a variety of computational challenges introduced by the large-scale, heterogeneous, and private aspects of genomic and health-related data.
Finding relatives within a study cohort is a necessary step in many genomic studies, according to Hong, Cho and collaborators. But this process is complicated when the study cohort’s data are distributed across multiple entities subject to different data-sharing restrictions. In “Secure Discovery of Genetic Relatives across Large-Scale and Distributed Genomic Datasets,” the research team introduced a secure federated algorithm to estimate kinship between pairs of individuals without sharing any private data. The team used data from UK Biobank and All of Us to test their algorithm, which detected 94.9% of third-degree relatives and 99.9% of second-degree relatives within 15 hours of runtime.
“This work showcases how a combination of new algorithmic ideas, insights into the structure of genetic sequences, and state-of-the-art cryptographic tools can provide practical solutions for jointly analyzing sensitive data across organizations with privacy guarantees,” said Cho.
Members of Cho’s lab also participated in four out of eight presentations at the RECOMB-PRIEQ Satellite Workshop, which focuses on challenges in privacy, security, bias, and fairness in biomedical research. Anupama Nandi, PhD, postdoctoral associate of biomedical informatics and data science, was first author of “Phenotype Randomization Mechanisms for Private Release of Genome-Wide Association Statistics.” Natnatee (Ko) Dokmai, PhD, postdoctoral associate of biomedical informatics and data science, was first author of “TX-Phase: Secure Haplotype Phasing in a Trusted Execution Environment” and co-author of “Learned Sketches Offer Improved Performance for Privacy Preserving and Secure GWAS.” Matthew Mosca, MSc student, also presented “Reconstruction of Private Genomes Through Reference-Based Genotype Imputation.”
Cho joined Yale School of Medicine in 2023.