Prediction and analysis of Alzheimer’s disease using blood microbiome

Abstract

INTRODUCTION: The human microbiome plays a role in human health and disease. Specifically, the gut microbiota is known to contribute to Alzheimer’s disease (AD); however, the blood microbiome remains less studied. Here, we investigated the link between the blood microbiome and AD because it is more convenient to obtain blood samples than gut samples. METHODS: We analyzed microbial reads from blood RNA sequencing data of 65 amyloid-β (Aβ)-positive and 56 with Aβ-negative subjects. We investigated differences in microbial signatures in blood between Aβ-positive and negative patients. Finally, we accessed the value of these data for predicting Aβ status using machine-learning models. RESULTS: We found significant changes in the blood microbiome diversity and differentially abundant microbial biomarkers as a function of Aβ status. The prediction performance of Aβ status was larger than the area under the curve value of 0.7. DISCUSSION: Microbiome-based AD diagnostics could aid in early diagnosis of AD.