Investigating Cis- and Trans-Genetic Regulation of Brain Transcriptomics and Proteomics Associated with AD/ADRD

Emory University, Atlanta GA

Investigators

Abstract

PROJECT SUMMARY Although genome-wide association studies (GWAS) have discovered ~75 significant risk genes associated with AD/ADRD, the underlying molecular mechanisms remain elusive. This gap has been partially filled by integrating different omics data together with GWAS data. For example, tools developed can integrate transcriptome-wide association study (TWAS) and proteome-wide association study (PWAS) with GWAS data to detect additional AD risk genes and inform on their mediation via transcriptome or proteome with AD dementia. Yet, recent TWAS and PWAS analyses for AD dementia have only considered cis-acting genetic regulations (Gap 1), generally only used omics data from one tissue (Gap 2) and conduct TWAS and PWAS separately (Gap 3). This study will extend and develop several new analytic tools that can be applied to genomic data from multiple tissues. This will increase the number of AD risks genes identified from current TWAS/PWAS studies and better inform on their molecular mechanisms. We will also validate AD risk genes identified in this study using both computational methods and biological experiments in a model organism. In Aim1, we will use our recently developed BGW method to estimate genome-wide cis- (nearby) and trans- acting (distal) QTL effect sizes for transcriptome-wide genes and proteome-wide proteins in DLPFC and three other tissues (Supplementary motor area (SMA), spinal cord, muscle). In Aim2, we will apply our recently developed Stacked Regression based method to leverage these estimated genome-wide xQTL weights across all four tissues, to improve the yield of separate tissue-specific TWAS and PWAS for AD dementia for the target DLPFC tissue. In Aim3a, we will first develop a novel tool to conduct joint TWAS and PWAS in the same analytic framework, which can test for genetic effects mediated by transcriptome while adjusting for PWAS effects (and vice versa), and simultaneously test for genetic effects mediated by transcriptome-proteome interactions. In Aim 3b, we will apply this novel tool to conduct joint genome-wide TWAS and PWAS for AD dementia in the target DLPFC tissue. We will also explore secondary TWAS/PWAS analyses for other tissues and combine the results by the omnibus Aggregated Cauchy Association Test (ACAT) across all four tissues to provide secondary lists of significant risk genes of AD dementia. Aim 4, will validate AD risk genes identified, by both computational approach with the probabilistic Mendelian Randomization tool (PMR-Egger, Aim 4a) and biological experiments with the established Drosophila AD model (Aim 4b). We will examine the underlying biological mechanisms of these validated significant risk genes by protein-protein interaction and pathway analyses with STRING (Aim 4c). Public Health Impact: New tools developed in this study will improve TWAS/PWAS studies to identify more risk genes for AD, better inform on their underlying mechanisms and provide new targets for drug discovery for new AD therapies. This study can strongly impact ADRD as well as genetic studies of other human diseases.