Abstract: Despite recent developments, the genetics and biology of Alzheimer’s disease remain insufficiently characterized. As an important first step toward developing effective treatment strategies to slow or prevent Alzheimer’s disease onset, the identification of relevant genetic markers is crucial. In the present study, we analyzed transcriptomic and multi-omic datasets across multiple cohorts (the Alzheimer’s Disease Neuroimaging Initiative, Religious Orders Study and Rush Memory and Aging Project, Mount Sinai Brain Bank, and Mayo Clinic Alzheimer’s Disease Genetics Studies) using gene set enrichment analysis, machine learning algorithms, and polygenic risk scoring to identify gene sets relevant to Alzheimer’s disease risk and pathological features. For prioritized gene sets, we performed epigenome-wide association studies to assess DNA methylation patterns, and used multi-omic mediation analysis to characterize the causal gene regulatory networks. Overall, we identified several key gene sets relevant to Alzheimer’s disease pathology—particularly, those related to immune system function and mitochondrial dysfunction. Upregulated pathways, including neutrophil degranulation and tumor necrosis factor-α signaling pathways, correlated strongly with aspects of neuroinflammation in Alzheimer’s disease. By contrast, downregulated oxidative phosphorylation pathways further suggested mitochondrial dysfunction. Gene sets that contained mitochondrially located genes (e.g., SGK1 and LRRK1) were identified as significantly contributing to neurodegeneration. Moreover, genes such as CXCL1, TGFB2, and DUSP1 were consistently implicated in all datasets, thus emphasizing their involvement in immune modulation and mitochondrial function. The multimodal investigation outlined in the current study represents useful steps toward comprehending the genetic architecture of Alzheimer’s disease, including an expanded understanding of the spatial interactions of genes associated with disease susceptibility. Mitochondrial dysfunction and immune modulation were pathological pathways that converged on Alzheimer’s disease and future treatment novel options. Using the frameworks provided in the current comprehensive study, we present opportunities to explore targeted treatment strategies that may alter immune systems and mitochondrial function to optimize treatment outcomes for individuals at increased risk of or living with Alzheimer’s disease. 

Key words: Alzheimer’s diseasel DNA methylationl epigeneticl gene regulatory networksl machine learningl mitochondrial dysfunctionl multi-omics analysisl nerve regenerationl neuroinflammation