Privacy-preserving genome-wide association studies on cloud environment using fully homomorphic encryption

Area of Science:

• Genomics
• Cryptography
• Bioinformatics

Background:

• Genomic sequencing advancements provide large-scale data for Genome-Wide Association Studies (GWAS).
• Privacy concerns regarding genomic data sharing hinder efficient medical research and disease association studies.
• Developing secure methods for handling sensitive genetic information is crucial for advancing personalized medicine.

Purpose of the Study:

• To present a cryptographic solution for secure genome-wide association studies (GWAS).
• To enable privacy-preserving analysis of large-scale genomic and phenotype data.
• To address the challenge of efficient genomic data sharing in medical research.

Main Methods:

• Utilizing a fully homomorphic encryption scheme to encrypt genotype and phenotype data.
• Performing computations on encrypted data by evaluating frequency tables.
• Employing a packing technique for efficient computation of statistical measures.

Main Results:

• The proposed solution supports secure computation of statistical tests like the chi-squared (χ2) test and linkage disequilibrium.
• Experimental results show efficient computation times: 35 ms for a χ2 test and 80 ms for linkage disequilibrium per 10,000 subjects.
• Demonstrated secure and efficient execution of GWAS algorithms in an outsourcing setting.

Conclusions:

• Cryptographic solutions based on fully homomorphic encryption are practical for secure GWAS computations.
• Appropriate encoding and packing techniques are key to the efficiency of these privacy-preserving methods.
• This approach facilitates secure genomic data analysis, promoting advancements in disease research and personalized medicine.