Privacy Risks and Anonymization of Microbiome Data

Abstract: “The microbial communities of the human body are subject to extensive research. While individual variations in the microbiome reveal valuable information about health and diseases, they also allow for the identification of individuals among populations of hundreds. In 2015, Franzosa et al. presented a method for the unique characterization of individuals via codes constructed from their microbiome samples. The temporal stability of these codes allows to correctly match different samples from the same individual, thus posing substantial privacy risks. The true-positive rate was particularly high on the widely studied microbiome of the gastrointestinal tract.

In this talk, we discuss a distance-based method for personal microbiome identification. Together with a routine for avoiding spurious matches, it further improves upon the performance of Franzosa et al.'s technique and underlines the need for solutions to protect the microbiome as personal and sensitive medical data. We will briefly outline one such solution, namely an adaptation of the well-known concept of k-anonymity.”