Biomarker Identification for Alzheimer's Disease Using a Multi-Filter Gene Selection Approach

Area of Science:

• Neuroscience
• Genetics
• Biomarker Discovery

Background:

• Alzheimer's disease (AD) lacks effective therapies, necessitating research into reliable biomarkers and therapeutic targets.
• Dementia and cognitive decline associated with AD underscore the urgent need for advanced research strategies.

Purpose of the Study:

• To develop and validate an aggregative multi-filter gene selection approach for identifying robust Alzheimer's disease biomarkers.
• To uncover potential therapeutic targets by analyzing differentially expressed genes in AD.

Main Methods:

• Integrated hub gene ranking (degree, bottleneck) with feature selection (Random Forest, Double Input Symmetrical Relevance) and ranking aggregation.
• Analyzed five AD-related microarray datasets (GSE48350, GSE36980, GSE132903, GSE118553, GSE5281) and validated findings on an independent dataset (GSE109887).
• Utilized logistic regression to assess the predictive value of identified genes, achieving an AUC of 86.8 on the validation set.

Main Results:

• Identified 803 overlapping differentially expressed genes from 464 AD and 492 normal cases across diverse brain regions.
• Prioritized 50 genes with significant predictive value for Alzheimer's disease.
• Pathway analysis indicated involvement of these genes in synaptic vesicle cycles, neurodegeneration, and cognitive function.

Conclusions:

• The developed gene selection approach effectively identifies robust AD biomarkers.
• The 50 prioritized genes offer promising therapeutic targets for Alzheimer's disease.
• Findings provide critical insights into the biological mechanisms underlying AD, paving the way for novel treatments.