Machine learning methods for risk prediction of late-onset Alzheimer's

 
   Lead Investigator:    Janet Cady
   Institution      :    Parabon Nanolabs
   E-Mail           :    janet@parabon.com
   Proposal ID      :    996



   Proposal Description:
   The ability to accurately predict risk for late-onset Alzheimer?s Disease (LOAD) before pathological changes manifest will allow clinical trials to recruit at-risk patients for pre-symptomatic treatments and measure treatment success more accurately. Despite recent advancements, existing genetic risk prediction models (GRPMs) for LOAD lack sufficient discrimination ability to support clinical applications. We hypothesize that much of the missing predictive power can be explained by non-additive (epistatic) interactions among genetic variants, which most methods fail to exploit. In this project we will use machine learning methods to detect high-order epistatic interactions that are predictive of LOAD pathogenesis. We will then use these epistatic interactions to build a model that will allow us to predict LOAD risk at a given age long before disease onset.