Deriving genomic diagnoses without revealing patient genomes

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Interpreting patient genomes requires comparison with other genomes, but sharing raises privacy concerns.
• Numerous monogenic diseases have identified genomic diagnoses and potential therapeutic targets.

Purpose of the Study:

• To demonstrate a privacy-preserving method for genomic diagnosis using secure multiparty computation.
• To enable accurate identification of disease-causing genetic variants while maintaining participant confidentiality.

Main Methods:

• Implementation of secure multiparty computation (SMC) for distributed genomic data analysis.
• Application of SMC in diverse clinical scenarios, including small cohorts, trio analyses, and multi-institutional collaborations.

Main Results:

• Successful identification of causal variants for monogenic diseases in real-world scenarios.
• Discovery of previously unrecognized disease genes and variants.
• Achieved up to 99.7% privacy for participants' sensitive genomic data.

Conclusions:

• Secure multiparty computation offers a viable solution for privacy-preserving genomic diagnosis.
• This approach facilitates collaborative research and accelerates the discovery of genetic disease mechanisms.
• The method balances the need for large-scale genomic data analysis with stringent participant privacy requirements.