A Novel Three-Stage Framework for Association Analysis Between SNPs and Brain Regions

Area of Science:

• Genetics
• Neuroscience
• Computational Biology

Background:

• Current methods for analyzing single nucleotide polymorphisms (SNPs) and regions of interest (ROIs) in Alzheimer's disease (AD) lack statistical power and biological interpretability.
• Existing approaches often suffer from high regression errors and fail to provide meaningful biological insights into AD pathogenesis.

Purpose of the Study:

• To develop a novel, robust framework for SNP-ROI correlation analysis to address limitations of existing methods.
• To enhance the accuracy and biological relevance of association analyses in Alzheimer's disease research.

Main Methods:

• A three-stage framework integrating clustering, group sparse modeling, and support vector machine regression.
• Clustering is used to handle linkage disequilibrium structures between SNPs.
• Group sparse modeling incorporates prior biological information (gene structure, linkage disequilibrium) for feature SNP selection.

Main Results:

• The proposed method demonstrates superior accuracy in predicting ROIs phenotype values compared to existing methods.
• Identified feature SNPs with associated weight vectors provide insights into their importance and biological relevance.
• The framework effectively selects significant SNPs and improves predictive performance.

Conclusions:

• The novel three-stage framework offers a more accurate and biologically meaningful approach to SNP-ROI correlation analysis for Alzheimer's disease.
• This method advances the understanding of genetic contributions to AD by improving feature selection and prediction accuracy.