Title: Harnessing machine learning, chemoproteomics, and in silico drug trials to repurpose drugs for Alzheimer's disease
Authors: Mark Albers - Harvard Medical School (United States) [presenting]
Abstract: The exploration of transcriptomes and proteomes derived from brains with Alzheimers disease (AD) by powerful computational tools has the potential to identify pathways and targets involved in the initiation and/or progression of AD. Distinguishing primary disease drivers from secondary events is a key challenge. We use three integrated, complementary informatics approaches to discover and probe potential pathways in AD using FDA-approved drugs. First, we apply classical and network aware machine learning approaches to identify pathways and targets altered in AD brains at different stages of disease progression using data from Accelerating Medicines Partnership-AD. Second, we use systems pharmacology approaches to analyze RNA-seq and proteomic data collected from cultured human neural cells following exposure to potential disease drivers and/or FDA-approved drugs in order to discover the target selectivity of lead compounds. Moreover, these data are fed back into the predictive model as CNS-cell type-derived priors for further refinement. Third, we will introduce how in-silico drug trials in electronic health record data, which will be discussed in greater detail in this session, can evaluate candidate approved drugs using real world data. Together, these data packages will help to prioritize and design follow on clinical and translational studies to test causality of disease pathways using a repurposed drug, evaluated by positive biomarker and clinical outcomes.