Introduction: African Americans (AA) are widely underrepresented in plasma biomarker studies for Alzheimer's disease (AD) and current diagnostic biomarker candidates do not reflect the heterogeneity of AD.
Methods: Untargeted proteome measurements were obtained using the SomaScan 7k platform to identify novel plasma biomarkers for AD in a cohort of AA clinically diagnosed as AD dementia (n=183) or cognitively unimpaired (CU, n=145). Machine learning approaches were implemented to identify the set of plasma proteins that yields the best classification accuracy.
Results: A plasma protein panel achieved an area under the curve (AUC) of 0.91 to classify AD dementia vs CU. The reproducibility of this finding was observed in the ANMerge plasma and AMP-AD Diversity brain datasets (AUC=0.83; AUC=0.94).
Discussion: This study demonstrates the potential of biomarker discovery through untargeted plasma proteomics and machine learning approaches. Our findings also highlight the potential importance of the matrisome and cerebrovascular dysfunction in AD pathophysiology.
Keywords: African American; Alzheimer’s dementia; Alzheimer’s disease; Machine learning; SomaScan; biomarkers; brain; cerebrovascular; matrisome; network analysis; plasma.