Preserving Genomic Privacy via Selective Sharing.

Area of Science:

• Genomics
• Privacy-Preserving Technologies
• Bioinformatics

Background:

• Genomic data sharing is crucial for medical research but poses significant privacy risks.
• Existing methods, often based on differential privacy, may not adequately protect sensitive genetic information or capture data interdependencies.

Purpose of the Study:

• To develop a privacy-preserving framework for sharing genomic data that protects sensitive information and maximizes data utility.
• To introduce a new privacy metric, epsilon-indirect privacy, suitable for selective data sharing scenarios.
• To evaluate the framework's effectiveness against traditional methods in terms of privacy and utility.

Main Methods:

• A novel framework inspired by differential privacy, focusing on selective data sharing rather than noise addition.
• Introduction of a new privacy notation, epsilon-indirect privacy, to address complex data sharing settings.
• Comparative analysis of the proposed framework against existing privacy-preserving techniques.

Main Results:

• The proposed framework effectively prevents the leakage of sensitive genomic information to potential attackers.
• The method achieves high data sharing utility by selectively sharing non-sensitive data points.
• The framework demonstrates superior performance in both privacy protection and data utility compared to previous approaches.

Conclusions:

• The developed framework offers a robust solution for secure genomic data sharing in medical research.
• Epsilon-indirect privacy provides a more accurate measure of privacy in selective data sharing contexts.
• This approach balances the need for data sharing with the imperative of individual genetic privacy.