Characterization of Alternative Polyadenylation in Alzheimer's Disease

点击翻译按钮获取中文摘要

Abstract Alzheimer's disease (AD) is a slowly progressive brain disorder characterized by cognitive decline, irreversible memory loss, disorientation, and language impairment. Recent advances in genomic technologies and the explosive genomic information related to disease have accelerated the convergence of discovery science with clinical medicine. We aim to utilize cutting-edge techniques in computational biology, RNA biology, and systems biology to identify novel prognostic and diagnostic biomarkers and to develop innovative therapeutic strategies for AD. We will establish a comprehensive archive of human polyadenylation sites by combining various APA databases. We will train a reliable deep neural network (DNN) model by considering both cis ad trans factors, and then apply this DNN prediction model to characterize APA events in AD samples across several AD consortia (Aim 1.1). We will develop highly efficient and accurate approaches based on deep learning to identify apaQTLs in order to maximize the utility of genotyping data to understand the functional effects of genetic variants in AD. We will perform integrative analysis with multi-omics data generated by other projects to understand the regulatory network, aiming to provide additional evidence for functional interpretation of apaQTLs in AD (Aim 1.2). We will perform integrative analysis with our established rigorous computational approaches to identify APA events associated with AD traits, in order to identify novel prognostic and diagnostic biomarkers for AD (Aim 2.1). To facilitate the utilization of large-scale data by the broad biomedical community, we will develop a comprehensive data resource to provide a computational framework that enables user-friendly interactive exploration and visualization of the biomedical significance of APA events (Aim 2.2). We expect to build a critical foundation to demonstrate that APA events represent novel types of biomarkers and serve as promising therapeutic targets to improve patient outcomes. Our proposed research could pave the innovative way for aiding precision medicine because we will develop highly innovative computational framework based on deep learning to identify APA events and perform apaQTL analysis to identify a novel class of APA-based biomarkers and therapeutic targets. The proposed research is of high significance because it will fundamentally advance our knowledge about the molecular basis of AD and contribute to a broader understanding of the overall complexity of AD.