A diagnostic plasma omics-biomarker for Alzheimer's disease informed by microglial single-cell transcriptomics: A...

Area of Science:

• Neuroscience
• Genomics
• Biomarker Discovery

Background:

• Current Alzheimer's disease (AD) diagnosis relies on neuropathology, limiting early detection before neurodegeneration.
• Transcriptomic data offers a promising avenue for developing early-stage, blood-based diagnostic biomarkers for AD.

Purpose of the Study:

• To develop and validate an omic-informed framework for blood-based Alzheimer's disease biomarkers.
• To identify gene panels from microglial transcriptomic data predictive of AD and applicable to blood samples.

Main Methods:

• Analyzed microglial gene expression using single-nucleus RNA sequencing (snRNA-seq) and six statistical methods.
• Identified and evaluated 78 gene panels (30-2000 genes) for AD patient vs. control classification.
• Mapped top gene panels (300, 50, 30 genes) to blood (monocyte) transcriptomic data using graph-based approach and optimal transport.

Main Results:

• The 300-gene panel achieved an AUC of 0.7 and 75% accuracy for AD classification, with lower accuracy (53%) for cognitively normal individuals.
• Reduced gene panels (50 and 30 genes) showed similar AUCs but improved the balance between AD and normal control prediction.
• The 50-gene panel yielded an AUC of 0.7, with 65% AD accuracy and 71% normal accuracy.

Conclusions:

• Integrating multiomics datasets enhances precision and comprehensiveness in Alzheimer's disease biomarker discovery.
• The developed omic-informed framework shows potential for improved early-stage Alzheimer's disease diagnosis using blood samples.