Chan, F. M., Badawi, A. Q. A. A., Sim, J. J., Tan, B. H. M., Sheng, F. C., & Aung, K. M. M. (2021). Genotype Imputation with Homomorphic Encryption. 2021 6th International Conference on Biomedical Signal and Image Processing. doi:10.1145/3484424.3484426
Genotype imputation is a technique used to determine unobserved genomic markers when sequencing genomic data. This is a cost effective method for sequencing a genome. Due to the large amount of personal identifiable information involved in genomic imputation, there is a rising concern for analysis of such nature to be secure and private.
We describe a method using homomorphic encryption (HE) to perform genotype imputation in a secure and private setting. Our solution first involves training a logistic regression model and performing the imputation in the encrypted domain.
We have implemented our solution over using the open sourced Homomorphic Encryption library, SEAL. We are able to impute 500 SNPs within 5 minutes, with an accuracy of 97.3%.
This research / project is supported by the A*STAR - RIE2020 Advanced Manufacturing and Engineering (AME) Programmatic Program
Grant Reference no. : A19E3b0099