In Silico Screening of Medications for Slowing Alzheimer's Disease Progression
in a Clinical Trials Metadatabase 
   Lead Investigator:    Richard Kennedy
   Institution      :    University of Alabama at Birmingham
   E-Mail           :    rekenned@uab.edu
   Proposal ID      :    1211



   Proposal Description:
   The drug development pipeline for AD is costly in time and resources and prone to failure. Drug repurposing that evaluates drugs approved    for other disorders for potential efficacy in treating AD can increase the        efficiency of the development process by dramatically increasing the number of testable agents. In this proposal, we will investigate a novel in silico          approach for repurposing, identifying promising medications by screening          activity of concurrent medications on AD trajectories in two large                administrative datasets using data mining algorithms. This approach will be       guided through the following specific aims and hypotheses:                        Aim 1: Identify medications associated with reduced rate of progression for       potential drug repurposing.                                                       H1: Data mining will identify patterns of concomitant medications and             medication combinations that are associated with statistically significant        reductions in rate of progression in subjects with AD.                            H2: Subjects taking medications selected through data mining will show reduced rates of AD progression compared to subjects not taking these medications in a validation set of AD cohort data.                                                 Approach: We will utilize data from our meta-database of AD studies, which        draws from 18 Alzheimer's Dis- ease Cooperative Study (ADCS) studies,             Alzheimer's Disease Neuroimaging Initiative (ADNI), and the Coalition Against Major Diseases (CAMD) AD studies, with > 6,500 subjects and > 1,400 concurrent medications. We will use high-dimensional data mining algorithms to select promising candidates, individually and in combination, from subjects' concurrent medications based on association with reduced rate of progression on the Alzheimer's Disease Assessment Scale-cognitive (ADAS-cog). Rates of progression will be