Secure large-scale genome-wide association studies using homomorphic encryption

摘要: Genome-wide association studies (GWASs) seek to identify genetic variants associated with a trait, and have been powerful approach for understanding complex diseases. A critical challenge GWASs has the dependence on individual-level data that typically strict privacy requirements, creating an urgent need methods preserve of participants. Here, we present privacy-preserving framework based several advances in homomorphic encryption demonstrate it can perform accurate GWAS analysis real dataset more than 25,000 individuals, keeping all individual encrypted requiring no user interactions. Our extrapolations show evaluate 100,000 individuals 500,000 single-nucleotide polymorphisms (SNPs) 5.6 h single server node (or 11 min 31 nodes running parallel). performance results are one order magnitude faster prior state-of-the-art using secure multiparty computation, which requires continuous interactions, accuracy both solutions being similar. also be applied other domains where large-scale statistical analyses over needed.