Characterization of Alternative Polyadenylation in Alzheimer's Disease

摘要 阿尔茨海默病（Alzheimer's disease，AD）是一种以认知功能下降为特征的缓慢进行性脑功能障碍，不可逆 记忆丧失定向障碍和语言障碍基因组技术的最新进展和 与疾病相关的爆炸性基因组信息加速了发现科学与 临床医学我们的目标是利用计算生物学、RNA生物学和生物学领域的尖端技术， 系统生物学，以确定新的预后和诊断生物标志物，并开发创新的治疗 AD的策略。我们将建立一个全面的人类多聚腺苷酸化位点档案， 各种阿帕数据库。我们将训练一个可靠的深度神经网络（DNN）模型， transfactors，然后应用该DNN预测模型来表征AD样本中的阿帕事件， 几 AD （目标1.1）。我们将开发高效和准确的方法， 学习鉴定apaQTL，以最大限度地利用基因分型数据来了解功能性 遗传变异对AD的影响我们将与其他组织产生的多组学数据进行综合分析。 项目，以了解监管网络，旨在提供更多的证据， AD中apaQTL的解释（目的1.2）。我们将根据我们建立的严格的 计算方法，以确定与AD性状相关的阿帕事件，以确定新的预后 和AD的诊断生物标志物（目标2.1）。为方便广大市民使用大型数据， 生物医学界，我们将开发一个全面的数据资源，以提供一个计算框架， 这使得用户友好的交互式探索和阿帕事件的生物医学意义的可视化 (Aim 2.2）。我们希望建立一个关键的基础，以证明阿帕事件代表了新类型的 生物标志物，并作为有前途的治疗目标，以改善患者的结果。我们提议的研究 可以为帮助精准医疗铺平创新之路，因为我们将开发高度创新的 基于深度学习的计算框架，以识别阿帕事件并进行apaQTL分析， 鉴定一类新的基于APA的生物标志物和治疗靶点。该研究具有较高的 重要性，因为它将从根本上推进我们对AD分子基础的认识， 有助于更广泛地了解AD的整体复杂性。

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.