Name
Title | ||
|---|---|---|
A Framework for Privacy-Preserving Genomic Data Analysis Using Trusted Execution Environments |
Abstract | ||
|---|---|---|
As the quantity and quality of genomic records continue to grow, geneticists at disparate institutions are increasingly motivated to integrate and analyze genomic data collections. While these endeavors are often hampered by various privacy concerns, recently developed techniques can address privacy concerns via trusted execution environments (TEEs) (e.g., hardware-enabled secure enclaves). TEEs provide confidentiality for sensitive data and means to attest to programs' correctness and their corresponding inputs. However, due to the profound consequences of private genomic data leakage for individuals, integrating TEEs for genomic analytics requires extra assurances to ensure that the program's execution does not leak participants' private information in a direct or indirect form. One major drawback of state-of-the-art TEEs is related to their shortcomings in hiding memory access patterns due to performance ramifications. As a result, the TEE's computation must be data oblivious, i.e., independent of how it accesses the sensitive private data through its course. In this paper, we present our data oblivious and privacy-preserving genomic data analysis framework using two popular genomic tasks. The first application involves securely performing a multi-institution test for the most significant single nucleotide polymorphism (SNP) in an association study. The second application focuses on the growing trend in utilizing Deep Learning (DL) based inference for genomics data. In this application, we assume that genomic data and the model are private. In both of these applications, we show that our framework can enable efficient processing and analysis of encrypted genomic data without disclosing sensitive memory access patterns. We also discuss the future considerations needed to use TEE based privacy-preserving genomic data analysis in practice. |
Year | DOI | Venue |
|---|---|---|
2020 | 10.1109/TPS-ISA50397.2020.00028 | 2020 Second IEEE International Conference on Trust, Privacy and Security in Intelligent Systems and Applications (TPS-ISA) |
Keywords | DocType | ISBN |
Privacy,Security,TEE,Intel SGX,Genomics | Conference | 978-1-7281-8544-6 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
3 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Aref Asvadishirehjini | 1 | 0 | 0.68 |
| Murat Kantarcioglu | 2 | 2470 | 168.03 |
| Bradley Malin | 3 | 1302 | 113.97 |