Secure computation outsourcing of genome-wide association studies using homomorphic encryption

One of the major problems in genomic research is the delivery of necessary computational power to process extremely large genomic datasets. While cloud computing provides the computing resources needed for such a scenario, the outsourcing of data processing raises a broad range of security and privacy issues as genomic data contain highly sensitive information. In our work, we explored the use of Paillier scheme, a partially homomorphic encryption algorithm to perform several GWAS computations, the Hardy-Weinberg equilibrium and χ ² test, remotely in a third-party server. We also evaluated the performance of Paillier scheme in doing these GWAS computations.