Related Experiment Video
Updated: Jun 1, 2025
04:41
Mapping Alzheimer's Disease Variants to Their Target Genes Using Computational Analysis of Chromatin Configuration
Published on: January 9, 2020
18.9K
LD-informed deep learning for Alzheimer's gene loci detection using WGS data
Taeho Jo1, Paula Bice1, Kwangsik Nho1,2
1Indiana Alzheimer Disease Research Center and Center for Neuroimaging, Department of Radiology and Imaging Sciences Indiana University School of Medicine Indianapolis Indiana USA.
Alzheimer'S & Dementia (New York, N. Y.)
|January 17, 2025
Summary
Deep-Block, an AI framework, identifies genetic loci for Alzheimer's disease (AD) risk using whole genome sequencing data. It finds both known and novel variants, improving AD genetic architecture understanding.
Area of Science:
- Genomics
- Bioinformatics
- Artificial Intelligence
Background:
- Exponential growth in genomic data necessitates advanced analytical tools for identifying genetic loci.
- Alzheimer's disease (AD) research requires efficient methods to analyze large-scale sequencing data for genetic risk factors.
Purpose of the Study:
- To present Deep-Block, a multi-stage deep learning framework designed to identify genetic regions associated with Alzheimer's disease (AD).
- To incorporate biological knowledge into an AI architecture for enhanced genetic analysis of high-throughput sequencing data.
Main Methods:
- Genome segmentation using linkage disequilibrium (LD) patterns.
- Sparse attention mechanisms for selecting relevant LD blocks.
- TabNet and Random Forest algorithms for single nucleotide polymorphism (SNP) feature importance quantification.
Main Results:
- Identification and ranking of 30,218 LD blocks based on AD relevance.
- Detection of novel and confirmation of known AD-associated SNPs, including APOE rs429358.
- Functional evidence for identified variants through expression Quantitative Trait Loci (eQTL) analysis across 13 brain regions.
Conclusions:
- Deep-Block effectively processes large-scale sequencing data, preserving SNP interactions and minimizing information loss.
- Identified variants show functional relevance supported by tissue-specific eQTL data.
- The framework successfully identified known and novel genetic variants, advancing the understanding of AD's genetic architecture.
