FedGMMAT

FedGMMAT utilizes a federated statistical testing approach to perform efficient genetic association tests that correct for confounding fixed and additive polygenic random effects while ensuring sensitive genetic data remains local to each collaborating site.

Key Features:

• Federated Analysis: Genetic data never leaves local sites, with intermediate statistics protected by encryption.
• Confounding Correction: Corrects for confounding fixed effects and additive polygenic random effects across heterogeneous sample sizes.
• Privacy Preservation: Operates under a privacy-preserving framework compliant with regulations such as HIPAA.
• Statistical Equivalence: Achieves virtually identical results to pooled analysis with practical resource requirements.

Scientific Applications:

• Collaborative Genetic Studies: Enables high-powered association studies across institutions without centralizing Protected Health Information or genetic data.
• Disease Genetics Research: Supports identification of genetic determinants of diseases through federated analysis of large-scale genetic and phenotypic data.

Methodology:

FedGMMAT implements federated generalized linear mixed model association tests using simulated and real datasets for validation.