Alzheimer's Disease Diagnostics Using miRNA Biomarkers and Machine Learning

Area of Science:

• Biochemistry
• Genomics
• Computational Biology

Background:

• Current Alzheimer's disease (AD) diagnosis methods, such as memory tests and brain scans, are often imprecise, invasive, or costly.
• MicroRNA (miRNA) dysregulation in blood shows promise as a non-invasive biomarker for AD diagnosis and potential treatment.
• Identifying reliable biomarkers is crucial for improving early and accurate detection of AD.

Purpose of the Study:

• To discover novel miRNA biomarkers for Alzheimer's disease (AD).
• To develop and validate a machine learning (ML) model for accurate AD diagnosis using miRNA profiles.
• To explore the diagnostic potential of miRNA expression patterns in blood.

Main Methods:

• Utilized pathways and target gene networks associated with confirmed AD miRNA biomarkers.
• Developed multiple diagnostic models based on significant miRNA expression differences in serum and plasma.
• Employed machine learning algorithms trained on filtered, disease-specific miRNA datasets.

Main Results:

• The best serum-based ML model achieved 92.0% accuracy in identifying miRNA biomarkers for AD.
• The best plasma-based ML model achieved 90.9% accuracy in identifying miRNA biomarkers for AD.
• Generated thousands of descriptors from AD-implicated miRNA, target genes, and pathways to enhance biomarker discovery.

Conclusions:

• A machine learning model incorporating miRNA, genomic, and pathway descriptors enables accurate prediction of Alzheimer's disease.
• This approach offers a promising, highly accurate, and potentially non-invasive method for AD diagnosis.
• The developed models and identified biomarkers can strengthen future diagnostic strategies for AD.