Paper Title
Blood RNA and Autoantibodies as Early Parkinson's Disease Biomarkers: A Validation of Present Problems and a Discovery of Two Biosignatures

Abstract
Parkinson's disease (PD) is the second most prevalent neurodegenerative disease globally, inflicting numerous families and creating a substantial financial burden to the health care system. An early diagnosis of the disease will mitigate its effects by introducing patients to appropriate medical options earlier, thus slowing the disease progression. Blood test based on discovered blood-based biomarkers is a promising, cheap, and effective method for early PD diagnosis. However, problems in some recent PD blood-based biomarker studies render the biomarkers those study discovered less competent in diagnosing early, pre-symptomatic PD. To delineate and address those problems, this work validates some sources of variability in PD blood biomarker studies and generates two stringent sets of blood biomarkers to diagnose early PD accurately. A novel Machine learning (ML) platform is used in this study to train datasets created from GEO data. The adaptation of this platform ensures that the most suitable algorithm can train any dataset included in this study. This work validates the PD stage as a significant confounding factor for early PD's biomarkers discovery. Biological sex is found to confound advanced PD biomarker studies potentially. Two biosignatures with potent predictive power for early PD are also defined. My study provides some valuable advice for the experimental design of future studies aiming to discover PD biomarkers. This study also generates two sets of biomarkers with potential medical applications. Keywords - Parkinson’s Disease, Machine Learning, Biomarker, Blood