DuAL-Net: A Hybrid Framework for Alzheimer's Disease Prediction from Whole-Genome Sequencing via Local SNP Windows and Global Annotations

TL;DR

DuAL-Net is a hybrid deep learning framework that predicts Alzheimer's disease risk from whole-genome sequencing data by integrating local SNP window analysis and global functional annotations, achieving improved accuracy and interpretability.

Contribution

The paper introduces DuAL-Net, a novel hybrid deep learning approach combining local and global genomic features for AD prediction from WGS data, with enhanced performance and interpretability.

Findings

Achieved an AUC of 0.671 in AD prediction.

Outperformed models using bottom-ranked and random SNPs.

Identified known and novel AD-associated SNPs.

Abstract

Alzheimer's disease (AD) dementia is the most common form of dementia. With the emergence of disease-modifying therapies, predicting disease risk before symptom onset has become critical. We introduce DuAL-Net, a hybrid deep learning framework for AD dementia prediction using whole genome sequencing (WGS) data. DuAL-Net integrates two components: local probability modeling, which segments the genome into non-overlapping windows, and global annotation-based modeling, which annotates SNPs and reorganizes WGS input to capture long-range functional relationships. Both employ out-of-fold stacking with TabNet and Random Forest classifiers. Final predictions combine local and global probabilities using an optimized weighting parameter alpha. We analyzed WGS data from 1,050 individuals (443 cognitively normal, 607 AD dementia) using five-fold cross-validation. DuAL-Net achieved an AUC of 0.671 using top-ranked SNPs, representing 35.0% and 20.3% higher performance than bottom-ranked and randomly selected SNPs, respectively. ROC analysis demonstrated strong positive correlation between SNP prioritization rank and predictive power. The model identified known AD-associated SNPs as top contributors alongside potentially novel variants. DuAL-Net presents a promising framework improving both predictive accuracy and biological interpretability. The framework and web implementation offer an accessible platform for broader research applications.