Authors:
Ramnath Sai SagarBLOOMINGTON, IN – Inpher —in collaboration with CEA and KU Leuven— won 1st place at the 2019 iDash Competition Genome Privacy & Security Competition, hosted by the Indiana University Luddy School of Informatics, Computing, and Engineering. Inpher Chief Computer Scientist Dr. Nicolas Gama and Director of Security Innovation Dr. Mariya Georgieva led the team to first prize in the competition’s Track 2, which focused on secure genotype imputation using fully homomorphic encryption.
Inpher wins iDash 2019 for its work in fully homomorphic encryption
The iDash Competition is an annual international competition for developing privacy-preserving techniques on bioinformatic datasets. The primary goal of the competition is to evaluate the performance of state-of-the-art privacy-preserving methods —such as fully homomorphic encryption and secure multiparty computation— that “ensure rigorous data confidentiality during data analyses in a cloud environment.” This year, Track 2 challenged participants to develop an efficient and accurate homomorphic encryption-based method for genotype imputation. The team —Dr. Georgieva and Dr. Gama from Inpher, Dr. Carpov from CEA (France), and Dr. Chillotti from KU Leuven (Belgium)— won first prize in the competition with the fastest and most accurate solution.
The winning solution was based on TFHE library and the hybrid framework Chimera, which were both developed in part by the Inpher team. The homomorphic evaluation of the prediction ran in less than a second for the prediction of 500 missing single nucleotide polymorphisms (SNPs) for 500 individuals, which means that, for the first time, the concrete cost of fully homomorphic encryption has become negligible compared to only reading the input data. The significant performance improvements on fully homomorphic encryption achieved in the 2019 iDash Competition proves that the technology is ready to be applied in real world applications — contact us today to learn how!