Genetic Architecture of Pure Alzheimer's Disease and Mixed Pathology

点击翻译按钮获取中文摘要

Co-occurring neuropathologies such as Lewy bodies, hippocampal sclerosis, and microinfarcts likely influence heterogeneity in genetic studies of Alzheimer’s disease (AD). The genetic architecture of these co-occurring neuropathologies is not clear. Most AD genetic analyses use clinical AD dementia phenotypes and are limited by phenotype misclassification: many clinical AD “controls” harbor AD pathology while clinical AD “cases” have no or minimal AD pathology. Only 20% of pathology confirmed AD dementia cases have only AD pathology (“pure AD”); 80% have co-occurring neuropathologies (“mixed AD”). A complementary strategy is to perform genome-wide association studies (GWAS) specifically distinguishing between pure and mixed AD. The overarching goal of this proposal is to further scientific understanding of pure AD and mixed AD by integrating neuropathology together with advanced statistical approaches and extensive multi-omics data. Aim 1: Distinguish the genetic architecture of pure AD vs. mixed AD vs. pathology-free controls and determine how associations are driven by particular neuropathological patterns. 1a. Perform GWAS of pathology-confirmed AD controls vs. pure AD vs. mixed AD using a multinomial regression framework. We will also examine how associations are driven by particular neuropathological features furthering our understanding of underlying mechanisms. 1b. Infer possible mechanisms by integrating biological knowledge. We will perform transcriptome- wide analyses and gene-network analyses of pure AD and mixed AD, integrating human PPI data. Aim 2: Identify factors specific to pure AD or mixed AD. 2a. Use machine learning and harmonized clinical data to isolate factors associated with pure or mixed AD. Identified clinical risk factors could be used for future clinical trial stratification. 2b. Perform genetic correlation analysis and Mendelian randomization to estimate correlated genetic effects and potential causal effects between candidate risk factor traits and pure or mixed AD. Genetic correlation analysis will gain novel insights into the shared genetic basis between biobank-scale GWAS traits and pure AD and mixed AD. We will follow-up significant correlations with local genetic correlation and Mendelian randomization (MR) to identify specific genomic regions contributing to the correlation and quantify causal effects of these traits on pure AD and mixed AD. Aim 3. Characterize the cellular and molecular consequences of pure AD vs mixed AD. We will harness the data generated by the SEA-AD and the ROS-MAP studies to characterize the molecular changes occurring for our phenotypes at cell type resolution. We will perform abundance analyses (3a), differential gene expression (3b), and differential chromatin accessibility (3c) in each cell type across multiple brain regions. We will investigate loci from recent GWASs of AD dementia and candidates from Aim 1.