Study Rationale: Current diagnostic criteria and treatment options are designed for the “average” person with Parkinson’s disease (PD). This one-size-fits-all approach ignores the huge biological and genetic variability that is present in the PD population, as well as the diversity of symptoms that affect individuals’ quality of life.
Hypothesis: Here we propose to derive a new, data-driven patient screening tool that predicts individual disease evolution and treatment response.
Study Design: Using data from PD populations and prodromal individuals who have just begin to experience initial symptoms, we will develop a machine-learning model that integrates clinical scores, brain imaging, physiological assays and biosensor data. We will use the model to identify biologically informed disease “biotypes” that can be differentially targeted for therapeutic development. Finally, we will investigate how early these biotypes emerge in the prodromal population and how they persist over time in the PD population.
Impact on Diagnosis/Treatment of Parkinson’s disease: The proposed project will provide a way to objectively integrate multi-omic biological information and generate a comprehensive disease phenotype for each individual. This data-driven method will help to design optimal treatment strategies for individuals, to aid disease forecasting and to identify people who should be enrolled in clinical trials.