Parkinson's disease is one of the most common neurodegenerative diseases. Numerous glitches in the genome each make a small contribution to the onset and progression of Parkinson's. This information from many common sites of DNA variation can be integrated into a single genome-wide polygenic score. We hypothesize that such genetic scores will powerfully predict the risk of progression from Parkinson's disease to dementia. Dementia is one of the most debilitating complications of Parkinson's progression that burdens patients and their caregivers. The pace of deterioration varies dramatically between patients for reasons that are poorly understood. In Specific Aim 1, we will develop and characterize genome-wide polygenic scores in thousands of patients followed for up to 12 years with tens of thousands of longitudinal visits. We will use recently developed computational algorithms and large longitudinal cohorts to derive and evaluate robust polygenic predictors of disease progression. These genome-wide polygenic scores will integrate genetic variants at genome-scale into a single quantitative measure of inherited prognosis. In Specific Aim 2, we will replicate the scores in five independent longitudinal cohorts and evaluate their predictive performance for early time points. A genome-wide polygenic score has the potential to transform clinical trial design and accelerate the development of novel therapies. If successful, the study will develop a polygenic score that can be used to enrich clinical trials with a homogeneous target population and thereby improve our ability to identify novel disease-modifying therapeutics. Moreover, the genome-wide progression score is poised to reveal new genetic network mechanisms and to advance precision medicine.