Precision Medicine Digital Twins for Alzheimer’s Target and Drug Discovery and Longevity - PROJECT SUMMARY Alzheimer’s disease (AD) is a devastating neurodegenerative disease and it is lack of effective disease-modifying treatments. Medical digital twins are computational disease models for target identification and drug discovery. However, how to organize and prioritize drug targets and candidate AD treatments in digital twins at drugome- wide and genome-wide scales are challenging. Our team developed AlzGPS, a genome-wide positioning systems platform to catalyze multi-omics for AD drug discovery. We also created The Alzheimer’s Cell Atlas (TACA), a single-cell transcriptomics and network pathobiology map for target identification and drug repurposing at brain cellulome-wide scales. We demonstrated that systematic identification and characterization of underlying pathogenesis and disease progression at cellulome- and genome-wide scales, will serve as a foundation for identifying and validating disease-modifying targets and treatments in AD or even longevity. We hypothesize that the digital twins tools for coordinated acquisition and seamless curation of multimodal data will be transferrable to any aging therapeutic development domains and will be applicable beyond digital twins, to expand artificial intelligence (AI) and machine learning (AI ML) workflows in AD target and drug discovery. We thus posit that a drugome-wide and genome-wide, precision medicine digital twins platform that identifies likely causal AD genes and networks from human genome sequencing and multi-omics findings, enables a more complete mechanistic understanding of AD pathobiology and the rapid development of disease-modifying targets and treatments with great success. Our goal is to ethically acquire and responsibly disseminate comprehensive patient-specific multimodal data sets, which will form the basis for scientific, technological, and translational studies to design and evaluate digital twins, and explore their integration to AD target and drug discovery. Aim 1 will develop and test an interpretable mechanistic deep learning framework to identify disease-modifying targets and networks for AD and longevity. We will develop a human protein-protein interactome network topology-based deep learning framework (R21 phase) and identify putative drug targets for AD and longevity through integrating multimodal data (genetics, genomics, transcriptomics, proteomics, and clinical) from AD sequencing project (ADSP), the AD knowledge portal, Longevity Consortium, and the Accelerating Medicines Partnership-AD (R33 phase). Aim 2 will develop and apply AI ML technologies for collaborative end-to-end analyses of single-cell multi-ome data. We will develop and implement a graph embedded gaussian mixture variational autoencoder network algorithm (R21 phase) and identify AD cell type-specific genes/targets, regulatory networks, and ligand- receptor interactions (R33 phase). Aim 3 will implement and test precision medicine Digital Twins for drug repurposing in AD and AD-related dementia (R33 phase). All Digital Twins codes, toolbox packages, and data developed will be shared through the ADSP and the AD knowledge portal based on the FAIR principles. This project is highly feasible and potentially transformative for both Alzheimer’s data science and precision medicine.
