Alzheimer’s disease (AD) has a substantial genetic component, but a large proportion of its heritability remains unexplained to date. Genetic association and interaction analyses of time-to-event outcomes of AD hold great promises to uncover the missing heritability, since time-to-event outcomes are more informative than the corresponding binary responses (event occurrence), which have been commonly used in genome-wide association studies. Also, predicting AD-related survival outcomes is of benefit to the prevention and treatment of the disease, and leveraging genetic information can greatly enhance the AD risk prediction accuracy. The proposed research will develop novel statistical learning models for identifying genetic factors and gene-gene/gene-environment interactions for AD-related time-to-event outcomes and predicting AD risk over time using genetic information. The developed methods will be applied to UK Biobank data to identify genes as well as G-G and G-E interactions associated with time to Alzheimer’s disease and build AD genetic risk prediction models. The expected duration of the project is three years. The success of the project will advance our knowledge on the genetic component of AD etiology, which may then inform the AD prevention and treatment.
