Machine Learning-Driven Identification of Alzheimer’s Disease-Specific Targets

The objective of this study is to prioritize Alzheimer’s disease (AD)-associated genetic targets using computational and machine learning approaches applied to genome-wide association study (GWAS) summary statistics derived from the NIAGADS dataset (NG00075). Specifically, we will analyze variant-level association signals from the International Genomics of Alzheimer’s Project (IGAP) meta-analysis (Kunkle et al., 2019) and integrate these signals into gene-level representations to identify genes that may be associated with AD pathology and potential molecular targets. Using statistical aggregation and machine learning modeling approaches, we will evaluate patterns of genetic association across variants mapped to genes and genomic regions to prioritize candidate genes that may be linked to AD biological mechanisms. These analyses will leverage association statistics and allele frequency information available in the controlled-access IGAP summary statistics. Allele frequency information provides population-level context for variant-disease associations and can support prioritization of candidate genes whose associated variants may reflect biological processes relevant to AD across broader segments of the population represented in the GWAS. This study is a secondary computational analysis of existing, de-identified GWAS summary statistics. The dataset contains variant-level association metrics describing the statistical relationship between genetic variants and AD. No individual-level genotype data, clinical records, or personally identifiable information will be accessed. All analyses will be conducted using secure institutional computing systems in compliance with NIAGADS data use policies. Computational analyses will be performed on a HIPAA-compliant High-Performance Computing cluster operated by Research Computing at West Virginia University that provides a secure institutional environment for data storage and analysis. The phenotypic characteristic evaluated in association with genetic variants is AD case-control status as defined in the original IGAP meta-analysis, and variants will be analyzed in relation to their reported statistical association with AD.

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder that impacts millions of people worldwide. Previous large genetic studies have identified many differences in people’s DNA that are associated with an increased risk of developing AD. By studying these genetic differences across large groups of individuals, researchers can identify genes that may be connected to the biological changes that occur in the brain during the disease. In this project, we will analyze existing genetic research data to identify genes that are strongly associated with AD. The long-term goal of this work is to help identify molecular targets that could inform future therapeutic delivery strategies aimed at improving the ability of treatments to localize to and remain in affected brain tissue.