Abstract
AbstractParkinson’s Disease (PD) exhibits significant individual variability, and recent Artificial Intelligence advancements have identified three distinct progression subtypes, each with known clinical features but unexplored gene expression profiles. This study aimed to identify the transcriptomics characteristics of PD progression subtypes, and assess the utility of gene expression data in subtype prediction at baseline. Differentially expressed genes were subtype-specific, and not typically found in other PD studies. Pathway analysis showed distinct and shared features among subtypes. Two had opposing expression patterns for shared pathways, and the third had a more unique profile with respect to the others. Machine Learning revealed that the progression subtype with the worst prognosis can be predicted at baseline with 0.877 AUROC, yet the contribution of gene expression was marginal for the prediction of the subtypes. This study offers insights into PD subtypes transcriptomics, fostering precision medicine for improved diagnosis and prognosis.
Publisher
Cold Spring Harbor Laboratory