Expanding Access to Large-Scale Genomic Data While Promoting Privacy: A Game Theoretic Approach

Area of Science:

• Genomics
• Data Science
• Bioinformatics

Background:

• Large-scale genomic data repositories are emerging, necessitating privacy-preserving sharing for scientific discovery.
• Current de-identification methods are vulnerable to re-identification attacks, hindering research.
• Traditional safeguards like legal agreements and data suppression are insufficient or impede utility.

Purpose of the Study:

• To develop a novel, quantifiable framework for privacy-preserving genomic data sharing.
• To account for adversarial capabilities and tailor protections to reasonable resources.
• To balance data utility with privacy risks more effectively.

Main Methods:

• Utilized a game theoretic lens to model adversarial behavior and design data protection strategies.
• Developed the Sequence and Phenotype Integration Exchange (SPHINX) resource with genomic summary data from over 8,000 individuals.
• Conducted extensive sensitivity analyses to assess the robustness of the proposed framework.

Main Results:

• Demonstrated that a game theoretic approach provides more effective, quantifiable protections for genomic data.
• Showcased improved risk-utility balancing compared to traditional methods using the SPHINX resource.
• Validated the generalizability of the framework across various genomic data sharing initiatives.

Conclusions:

• A game theoretic framework offers a robust and adaptable solution for secure genomic data sharing.
• This approach enhances scientific discovery by enabling effective risk-utility management.
• The findings support the adoption of advanced, resource-aware privacy protections in genomics.