International Association
for Cryptologic Research

Following recent advances in high throughput sequencing, it can be expected that, in the near future, more and more individuals will have their whole genome extracted, stored and analyzed in a routine fashion. Although this perspective is full of promises in terms of personalized preventive and curative medicine as well as medical research, it should also be acknowledged that the genome of an individual is in essence extremely sensitive data from (at least) a privacy standpoint. Thus, for such personalized medicine-oriented platforms to reliably exist, it is necessary to develop specific counter-measures providing intrinsic protection of the genomic data when manipulated by an IT infrastructure.

In this context, a very promising approach is grounded in homomorphic encryption as a means of computing directly over encrypted data.

The purpose of the present postdoctoral offer is thus to investigate the practical relevance of using homomorphic encryption techniques for privacy-preserving genetic data processing. The main use case will consist in performing requests on a database of genomes represented by their variants. Several scenarios will be investigated in particular with respect to the privacy of the request itself on top of the privacy of the genetic data. In this various scenarios, the candidate is expected to identify the most suitable homomorphic encryption techniques ranging from additive-only (e.g. suitable for private requests on unencrypted data) and multiplicative-only (e.g. suitable for disjunctive public requests on encrypted genetic data) homomorphic encryption systems to the use of the more recent (and more costly) fully homomorphic encryption techniques. The candidate will also be expected to build prototypes for one or more of the above scenarios in order to experimentally demonstrate the practical viability of the solutions, in particular with respect to performances.