Ultrafast homomorphic encryption models enable secure outsourcing of genotype imputation

Kim, Miran; Harmanci, Arif; Bossuat, Jean-Philippe; Carpov, Sergiu; Cheon, Jung Hee; Chillotti, Ilaria; Cho, Wonhee; Froelicher, David; Gama, Nicolas; Georgieva, Mariya; Hong, Seungwan; Hubaux, Jean-Pierre; Kim, Duhyeong; Lauter, Kristin; Ma, Yuanyuan; Ohno-Machado, Lucila; Sofia, Heidi J.; Son, Yongha; Song, Yongsoo; Troncoso-Pastoriza, Juan Ramón; Jiang, Xiaoqian

doi:10.1016/j.cels.2021.07.010

Cited by 49 publications

(39 citation statements)

References 60 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The copyright holder for this preprint this version posted October 1, 2021. ; https://doi.org/10.1101/2021.09.30.462262 doi: bioRxiv preprint Genotype data is encrypted once at the client and submitted to the server, which securely imputes the untyped variants without decrypting the genotypes. The imputation is performed by secure evaluation of machine learning models that we recently developed [7]. In this manuscript, we build on top of the models and develop a simple metric that can be used to estimate genotype imputation accuracy (Allelic R2) using reference panel and imputed genotypes.…”

Section: Contact Arifoharmanci@uthtmcedumentioning

confidence: 99%

Open Imputation Server provides secure Imputation services with provable genomic privacy

Kim

Wang

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

As DNA sequencing data is available for personal use, genomic privacy is becoming a major challenge. Nevertheless, high-throughput genomic data analysis outsourcing is performed using pipelines that tend to overlook these challenges. Results: We present a client-server-based outsourcing framework for genotype imputation, an important step in genomic data analyses. Genotype data is encrypted by the client and encrypted data are used by the server that never observes the data in plain. Cloud-based framework can benefit from virtually unlimited computational resources while providing provable confidentiality. Availability: Server is publicly available at https://www.secureomics.org/OpenImpute. Users can anonymously test and use imputation server without registration.

show abstract

Section: Contact Arifoharmanci@uthtmcedumentioning

confidence: 99%

Open Imputation Server provides secure Imputation services with provable genomic privacy

Kim

Wang

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…As a solution of privacy-preserving ML for genetic data, ML over encrypted data with HE is drawing attention with the annual iDASH competition [ 9 , 10 , 28 – 35 ]. Kim et al introduced a privacy-preserving logistic regression model over HE [ 28 ].…”

Section: Introductionmentioning

confidence: 99%

Secure tumor classification by shallow neural network using homomorphic encryption

et al. 2022

Self Cite

View full text Add to dashboard Cite

Background Disclosure of patients’ genetic information in the process of applying machine learning techniques for tumor classification hinders the privacy of personal information. Homomorphic Encryption (HE), which supports operations between encrypted data, can be used as one of the tools to perform such computation without information leakage, but it brings great challenges for directly applying general machine learning algorithms due to the limitations of operations supported by HE. In particular, non-polynomial activation functions, including softmax functions, are difficult to implement with HE and require a suitable approximation method to minimize the loss of accuracy. In the secure genome analysis competition called iDASH 2020, it is presented as a competition task that a multi-label tumor classification method that predicts the class of samples based on genetic information using HE. Methods We develop a secure multi-label tumor classification method using HE to ensure privacy during all the computations of the model inference process. Our solution is based on a 1-layer neural network with the softmax activation function model and uses the approximate HE scheme. We present an approximation method that enables softmax activation in the model using HE and a technique for efficiently encoding data to reduce computational costs. In addition, we propose a HE-friendly data filtering method to reduce the size of large-scale genetic data. Results We aim to analyze the dataset from The Cancer Genome Atlas (TCGA) dataset, which consists of 3,622 samples from 11 types of cancers, genetic features from 25,128 genes. Our preprocessing method reduces the number of genes to 4,096 or less and achieves a microAUC value of 0.9882 (85% accuracy) with a 1-layer shallow neural network. Using our model, we successfully compute the tumor classification inference steps on the encrypted test data in 3.75 minutes. As a result of exceptionally high microAUC values, our solution was awarded co-first place in iDASH 2020 Track 1: “Secure multi-label Tumor classification using Homomorphic Encryption”. Conclusions Our solution is the first result of implementing a neural network model with softmax activation using HE. Also, HE optimization methods presented in this work enable machine learning implementation using HE or other challenging HE applications.

show abstract

“…Hence, allowing outsourcing of computations without having to grant access to the data. In the last decade, advances in FHE schemes have propelled FHE from a primarily theoretical breakthrough to a practical solution for a wide range of applications [1][2][3]. In recent years, we have seen FHE emerging in real-world applications, e.g., Microsoft's Edge browser uses FHE to realize its privacy-preserving password monitor [1].…”

Section: Introductionmentioning

confidence: 99%

HECO: Fully Homomorphic Encryption Compiler

Viand¹,

Jattke²,

Haller³

et al. 2022

Preprint

View full text Add to dashboard Cite

In recent years, Fully Homomorphic Encryption (FHE) has undergone several breakthroughs and advancements leading to a leap in performance. Today, performance is no longer a major barrier to adoption. Instead, it is the complexity of developing an efficient FHE application that currently limits deploying FHE in practice and at scale. Several FHE compilers have emerged recently to ease FHE development. However, none of these answer how to automatically transform imperative programs to secure and efficient FHE implementations. This is a fundamental issue that needs to be addressed before we can realistically expect broader use of FHE. Automating these transformations is challenging because the restrictive set of operations in FHE and their non-intuitive performance characteristics require programs to be drastically transformed to achieve efficiency. In addition, existing tools are monolithic and focus on individual optimizations. Therefore, they fail to fully address the needs of end-to-end FHE development. In this paper, we present HECO, a new end-to-end design for FHE compilers that takes high-level imperative programs and emits efficient and secure FHE implementations. In our design, we take a broader view of FHE development, extending the scope of optimizations beyond the cryptographic challenges existing tools focus on.

show abstract

Ultrafast homomorphic encryption models enable secure outsourcing of genotype imputation

Abstract: Highlights d Fast homomorphic encryption enables secure and practical genotype imputations d Secure methods require comparable resources as nonsecure methods d Accuracy of secure methods can be improved using population specific panels

Cited by 49 publications

References 60 publications

Open Imputation Server provides secure Imputation services with provable genomic privacy

Open Imputation Server provides secure Imputation services with provable genomic privacy

Secure tumor classification by shallow neural network using homomorphic encryption

HECO: Fully Homomorphic Encryption Compiler

Contact Info

Product

Resources

About