JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Confidential computing for population-scale genome-wide association studies with SECRET-GWAS

  • 0Computer Science and Engineering, University of Michigan, Ann Arbor, MI, USA. jonaher@umich.edu.
Nature Computational Science +

|

September 12, 2025

|

September 12, 2025

View abstract on PubMed

Summary

This summary is machine-generated.

Genomic data diversity is limited. SECRET-GWAS uses confidential computing for privacy-preserving, collaborative genome-wide association studies (GWAS), enabling faster analysis of large datasets.

Area Of Science

  • Genomics
  • Computational Biology
  • Privacy-Preserving Technologies

Background

  • Genomic datasets from single institutions lack global diversity, hindering the study of rare variants and diseases.
  • Collaborative genome-wide association studies (GWAS) are crucial for comprehensive genomic research but face privacy and accuracy challenges.
  • Existing privacy-preserving solutions struggle with the performance demands of large-scale regression analyses.

Purpose Of The Study

  • To develop a rapid, privacy-preserving tool for population-scale collaborative GWAS.
  • To enable efficient execution of linear and logistic regression on distributed genomic data without compromising privacy.
  • To address the limitations of previous secure computing approaches in supporting complex GWAS methods.

Main Methods

  • Implementation of SECRET-GWAS, a tool leveraging confidential computing on an Intel SGX cloud platform.
  • Application of system optimizations including streaming, batching, and data parallelization to enhance performance.
  • Integration with the Hail genomic analysis framework and protection against hardware side-channel attacks.

Main Results

  • SECRET-GWAS efficiently scales linear and logistic regression to over a thousand processor cores.
  • The tool enables multivariate linear and logistic regression GWAS queries on population-scale datasets from ten sources.
  • Analysis completion times were significantly reduced: 4.5 minutes for linear regression and 29 minutes for logistic regression.

Conclusions

  • SECRET-GWAS provides a rapid and privacy-preserving solution for collaborative, population-scale GWAS.
  • The tool overcomes performance limitations of previous methods, supporting widely used regression techniques.
  • SECRET-GWAS enhances global genomic data diversity representation and facilitates the study of rare genetic factors.