Paper 2019/1305

Privacy-Preserving Computation over Genetic Data: HLA Matching and so on

Jinming Cui, Huaping Li, and Meng Yang

Abstract

Genetic data is an indispensable part of big data, promoting the advancement of life science and biomedicine. Yet, highly private genetic data also brings concerns about privacy risks in data shar- ing. In our work, we adopt the cryptographic prim- itive Secure Function Evaluation (SFE) to address this problem. A secure SFE scheme allows insti- tutions and hospitals to compute a function while preserving the privacy of their input data, and each participant knows nothing but their own input and the final result. In our work, we present privacy-preserving solutions for Human Leukocyte Antigen (HLA) matching and two popular biostatistics tests: Chi-squared test and odds ratio test. We also show that our protocols are compatible with multiple databases simultaneously and could feasibly han- dle larger-scale data up to genome-wide level. This approach may serve as a new way to jointly analyze distributed and restricted genetic data among insti- tutions and hospitals. Meanwhile, it can potentially be extended to other genetic analysis algorithms, allowing individuals to analyze their own genomes without endangering data privacy.

Metadata
Available format(s)
PDF
Category
Applications
Publication info
Published elsewhere. 1st Federated Machine Learning workshop (2019) in conjection with IJCAI2019
Keywords
MPCmultiparty computation
Contact author(s)
cuijinming @ genomics cn
jamie cui @ outlook com
History
2019-11-11: received
Short URL
https://ia.cr/2019/1305
License
Creative Commons Attribution
CC BY

BibTeX

@misc{cryptoeprint:2019/1305,
      author = {Jinming Cui and Huaping Li and Meng Yang},
      title = {Privacy-Preserving Computation over Genetic Data: HLA Matching and so on},
      howpublished = {Cryptology ePrint Archive, Paper 2019/1305},
      year = {2019},
      note = {\url{https://eprint.iacr.org/2019/1305}},
      url = {https://eprint.iacr.org/2019/1305}
}
Note: In order to protect the privacy of readers, eprint.iacr.org does not use cookies or embedded third party content.