Secure Sequence Similarity Search on Encrypted Genomic Data

摘要

Genomic data is being produced rapidly by both individuals and enterprises and needs to be outsourced from local machines to a cloud for better flexibility. Outsourcing also eliminates the local storage management problem for data owners. However, sensitive data must be encrypted by data owners before outsourcing to protect data privacy and security in the cloud. As genome data is huge in volume, it is challenging to execute researchers' query securely and efficiently. In this paper, we present a prefix tree based indexing algorithm for supporting similar sequence search query. We support Hamming distance as similarity measure. The proposed method adopts semi-honest adversary model for the cloud server. The security of the shared data is guaranteed through encryption while making the overall computation fast and scalable enough for real-life biomedical applications. We evaluated the efficiency of our proposed model on a database of Single-Nucleotide Polymorphism (SNP) sequences and experimental results demonstrate that a query of hamming distance k = 2 in a database of 10000 records, where each record contains 500 nucleotides, takes approximately 4 minutes.