Approved Research

Analyzing genotype-phenotype associations with fairness and privacy

Lay summary

The goal of this project is to conduct a comprehensive study of fairness and privacy of sharing summary statistics about genomic datasets, in light of significant advances in the field of algorithmic fairness and data privacy over the last 10 years. It is our hope that new tools can be applied within the genomic context to unlock high-fidelity sharing of the results of genomic analyses that can improve health outcomes not just for a specific population, but broadly for diverse groups of individuals of different ancestry and sociodemographic backgrounds. This has the potential to unlock greater prediction accuracy for clinical insights, collaboration, data access, and statistical power, across all studies using genomic data --- both safeguarding data sharing in a more rigorous way and unlocking future discoveries.