Predicting phenotypes in Alzheimer's Disease using machine learning-enabled polygenic risk models Lead Investigator: Thomas Beach Institution : Vivid Genomics, Inc. E-Mail : tom@vividgenomics.com Proposal ID : 1297 Proposal Description: Alzheimer?s Disease affects 5 million people in the US, at an annual estimated cost of $250B. Alzheimer?s drug development has failed more than 150 times in the last 20 years. Despite these intensive efforts, there are no disease modifying therapeutics. This failure has been hypothesized to result from many factors but perhaps the most important of these has been unrecognized heterogeneity in clinical trial populations. Amyloid-beta plaques and tau tangles are the signature diagnostic findings at autopsy, and each are thought to independently contribute to cognitive deterioration but until recently there were no means of screening trial participants to make sure these targets were actually present. Enrolling subjects without these lesions would result in decreased effect sizes for trial agents specifically targeting them. Recently-developed PET imaging methods can now identify higher densities of these lesions in living subjects but their expense is inhibitory and early disease stages are still undetectable. Additionally, there are numerous Alzheimer?s co-morbidities that can only be identified at autopsy, including Lewy body disease, cerebral amyloid angiopathy, TDP-43 pathology and brain microinfarcts. Using postmortem phenotype data from a cohort of ~1100 human brain samples Vivid Genomics, Inc. has developed Genomic Biopsy???, a collection of three prototype genetic biomarker assays that predict the presence of amyloid, Lewy Bodies, and Cerebral Amyloid Angiopathy. The assays were generated by combining machine learning and statistical genetic analysis of DNA. The objective of this proposal is to optimize and validate these initial assays and develop additional tests to predict the presence of Lewy bodies, TDP-43 pathology and microscopic infarcts. The project aims to further develop, optimize, validate, and perform CLIA validation of a set of genetic biomarker assays for Alzheimer?s disease and its