Differential privacy under dependent tuples—the case of genomic privacy

Abstract

Motivation The rapid progress in genome sequencing has led to high availability of genomic data. However, due to growing privacy concerns about the participant’s sensitive information, accessing results and data of genomic studies is restricted to only trusted individuals. On the other hand, paving the way to biomedical discoveries requires granting open access to genomic databases. Privacy-preserving mechanisms can be a solution for granting wider access to such data while protecting their owners. In particular, there has been growing interest in applying the concept of differential privacy (DP) while sharing summary statistics about genomic data. DP provides a mathematically rigorous approach but it does not consider the dependence between tuples in a database, which may degrade the privacy guarantees offered by the DP. Results In this work, focusing on genomic databases, we show this drawback of DP and we propose techniques to mitigate it. First, using a real-world genomic dataset, we demonstrate the feasibility of an inference attack on differentially private query results by utilizing the correlations between the tuples in the dataset. The results show that the adversary can infer sensitive genomic data about a user from the differentially private query results by exploiting correlations between genomes of family members. Second, we propose a mechanism for privacy-preserving sharing of statistics from genomic datasets to attain privacy guarantees while taking into consideration the dependence between tuples. By evaluating our mechanism on different genomic datasets, we empirically demonstrate that our proposed mechanism can achieve up to 50% better privacy than traditional DP-based solutions. Availability https://github.com/nourmadhoun/Differential-privacy-genomic-inference-attack. Supplementary information Supplementary data are available at Bioinformatics online.

Inference attacks against differentially private query results from genomic datasets including dependent tuples

“…Later, for location datasets, Liu et al (2016) showed how an adversary with the knowledge of pairwise dependencies between tuples can predict users’ locations ( Liu et al , 2016 ). Recently, Almadhoun et al (2020) analyzed the decrease in the privacy guarantees of DP-based mechanisms when there are dependent tuples in a statistical genomic dataset.…”

“…In this work, we demonstrate the scale of attribute inference attacks using differentially private results of two complex and real-life queries over statistical genomic datasets [compared to the simple sum query considered in the study by Almadhoun et al (2020) ]. As opposed to Liu et al (2016) , which only considers pairwise correlation between the tuples, we consider interdependent correlations between dataset participants.…”

“…correlations between genomes of family members in a genomic dataset), and hence their privacy guarantees degrade if such correlations are used by the adversary. An adversary can use auxiliary channels to learn about such dependencies in the dataset and exploit this vulnerability in DP mechanism as shown by Liu et al (2016) and Almadhoun et al (2020) .…”

“…For the attribute inference, we follow the same attack model by Almadhoun et al (2020) . We assume a stronger adversary than the standard DP adversary with the following assumptions: (i) the adversary has access to the membership information of all n members in the dataset.…”

“…In previous work, focusing on statistical genomic datasets, we have shown how this dependency between the tuples (i.e. dependency between the genomes of individuals from the same family) results in additional information leak from the differentially private query results ( Almadhoun et al , 2020 ). Focusing on a simple count (sum) query, we have analyzed how the privacy guarantees of DP-based solutions degrade when there are dependent tuples in a genomic dataset.…”