Scalable privacy-preserving data sharing methodology for genome-wide association studies

Area of Science:

• Genetics
• Bioinformatics
• Computer Science

Background:

• Genome-wide association studies (GWAS) generate large datasets with sensitive individual-level genetic information.
• Traditional privacy methods are insufficient for protecting GWAS data from linkage attacks.
• Differential privacy offers rigorous privacy guarantees but can impact data utility.

Purpose of the Study:

• To extend existing differentially-private methods for releasing aggregate GWAS data.
• To improve privacy protection for GWAS while maintaining data utility.
• To address the challenge of releasing privacy-preserving GWAS statistics with arbitrary case/control numbers.

Main Methods:

• Extended differentially-private methods for releasing chi-squared (χ(2))-statistics.
• Developed methods for releasing differentially-private allelic test statistics.
• Incorporated a novel interpretation assuming known control data for enhanced privacy.

Main Results:

• Proposed methods allow for flexible handling of case and control numbers in GWAS data release.
• The new methods provide enhanced privacy guarantees for aggregate GWAS data.
• Risk-utility analysis on real data demonstrated the performance of the proposed techniques.

Conclusions:

• The extended differentially-private methods offer improved privacy protection for GWAS data.
• These methods provide a valuable tool for researchers handling sensitive genetic information.
• The approach balances privacy guarantees with data utility for genetic association studies.