"Linkage Disequilibrium-Informed Deep Learning Framework for Alzheimer's Disease"

This study introduces Deep-Block, a novel multi-stage deep learning framework that incorporates biological knowledge for analyzing large-scale genomic data in Alzheimer's disease. Applied to the Alzheimer's Disease Sequencing Project (ADSP) dataset of 7,416 participants, the framework successfully identified both known AD-associated variants (including APOE) and novel genetic loci. The findings were validated through brain tissue-specific eQTL analysis and cross-referenced with established AD genetics databases.

Key Innovation: Deep-Block's three-stage approach combines linkage disequilibrium patterns, sparse attention mechanisms, and advanced machine learning algorithms to effectively process large-scale genomic data while preserving critical biological relationships.  [LINK]