Need Help?

Enhancing Open Data Sharing for Functional Genomics Experiments: Measures to Quantify Genomic Information Leakage and File Formats for Privacy Preservation

This study was designed to develop methods that address the privacy risk associated with genetic variants obtained from genomic sequencing experiments. The preponderance of readily available genetic sequencing experiments has enabled the generation of datasets whose target purposes range from the genotyping of individuals for medical and forensic analyses, to the acquisition of molecular information for biological model development. However, by virtue of the experimental procedure, the raw sequences are tagged with small bits of patients' private variant information. In addition, depending on the protocol, these experiments may identify sequences from exogenous species. Raw human and microbial sequences can be analyzed as a whole to learn both general facts about biology, but can also be used to determine sensitive information about an individual. This presents a privacy conundrum for data sharing.

To address privacy concerns, we have developed several mathematical and computational approaches to quantify the risk of disclosure of sensitive information from genetic sequence data, and to sanitize the associated reads in a manner that optimally maintains the utility of the data. In addition, we have generated environmental samples containing genetic information from consented individuals to evaluate the risk posed by such data. These samples were then used to assess the efficacy of the computational methods. Human-derived and -associated samples that are damaged or obscured by environmental factors represent the scope of samples used for experimentation in this study. Samples include, but were not limited to, swabs of contacted objects and glass slides coated in saliva from consented individuals. Blood samples were also acquired from the same consented individuals to create gold-standard comparison datasets.