Federico Mazzone 

Genomic data is at the center of many innovations and studies in modern healthcare, from

personalized treatments to early detection of genetic diseases. However, analyzing this data

typically involves sharing it with external entities, like research collaborators, cloud services, or

third-party providers. This raises serious privacy concerns. Unlike other personal data, a person’s

genome is a permanent identifier and its leakage could affect not only the individual, but also

their relatives, potentially exposing disease predispositions and physical traits. This project

develops methods to analyze genomic data securely, using cryptographic techniques that allow

computations to be performed directly on encrypted inputs. This means that data can remain

private, even to those performing the analysis, while still enabling meaningful results. To make

this feasible at scale, the project combines algorithmic optimizations with hardware acceleration

(e.g., using GPUs and FPGAs) to reduce the cost of encrypted computation. The outcome will

support secure collaboration across institutions, allowing researchers to work with sensitive data

without exposing it, and enabling secure biomedical research across institutions.