PROJECT SUMMARY/ABSTRACT Alzheimer's disease (AD) is the most common form of dementia characterized by progressive loss of cognitive function. Unfortunately, currently there is no effective treatment for AD and clinical interventions of AD have largely failed despite enormous efforts. For the current application, we seek to develop multimodal machine learning models by leveraging the rich collection of AD-related omics data and phenotypical data recently generated from large-scale collaborative projects such as Alzheimer Disease Neuroimaging Initiative (ADNI), Accelerating Medicines Partnership-AD (AMP-AD) and the Alzheimer's Disease Sequencing Project (ADSP). Three aims will be pursued in the current application. Aim 1. We will build an expandable multimodal unsupervised machine learning framework to investigate AD heterogeneity. Given the multifactorial nature of AD, we will perform AD subtyping by harnessing the rich information across multiple spectrum of data. Aim 2. We will build an expandable multimodal supervised machine learning framework to quantify AD risk from longitudinal follow up of cognitively normal elders. The models will be built from genetic susceptibility and gene regulatory information as well as endophenotypes measured when participants were cognitive normal. Aim 3. We will build AD-related gene interaction networks in post-mortem human brain samples. We will examine the association of multiple omics data with AD in brain samples, and build tissue-specific interaction networks to understand potential molecular mechanisms underlying AD pathogenesis. The present application represents an innovative approach to identify individuals at high risk of AD from both clinical and genetic risk factors in ethnically diverse populations. The outlined strategy will provide new insights into the risk stratification and prevention strategies for AD. We also commit to share our methods through GitHub or CRAN for free access across the scientific community.