Detection of Association Features Based on Gene Eigenvalues and MRI Imaging Using Genetic Weighted Random Forest

Area of Science:

• Neuroscience
• Genetics
• Biomarker Discovery

Background:

• Alzheimer's disease (AD) research benefits from integrating imaging and genetic data to find biomarkers.
• Understanding AD progression requires identifying key biomarkers across disease stages: healthy controls (HC), early mild cognitive impairment (EMCI), late mild cognitive impairment (LMCI), and AD.

Purpose of the Study:

• To develop a novel data fusion and genetic weighted random forest method for identifying significant Alzheimer's disease biomarkers.
• To enhance the differentiation between HC, EMCI, LMCI, and AD stages using integrated genetic and imaging data features.

Main Methods:

• A novel data fusion technique was employed, amplifying differences between cognitive states using gene p-value matrix eigenvalues.
• A genetic weighted random forest model was constructed using fused features, incorporating genetic evolution for decision tree diversity and weighting.
• The model was analyzed post-training to detect significant fused features and identify important biomarkers.

Main Results:

• The proposed model demonstrated high performance and generalization in validation experiments.
• Significant genes including CSMD1, CDH13, PTPRD, MACROD2, and WWOX were identified as potential AD biomarkers.
• Key biological pathways implicated in AD were identified: calcium signaling, arrhythmogenic right ventricular cardiomyopathy, and glutamatergic synapse.

Conclusions:

• The developed model provides an accurate and efficient approach for identifying significant biomarkers in Alzheimer's disease.
• The findings offer valuable insights into the mechanisms of AD progression and potential therapeutic targets.