Kinetic Phenotype Discovery Informatics for Neurological Diseases

项目摘要 神经系统疾病的数量明显超过其他治疗领域的疾病， 发病率高于其他任何疾病类别。然而，制药业一直不成功， 找到有效的药物这些失败的一个重要因素是缺乏足够的模型系统， 对疾病的基本认识影响诊断和治疗。因此，有一个强大的， 对使用患者来源的细胞模型来理解致病机制的新兴趣 潜在的神经系统疾病表型。为了真正了解这些机制， 表型，有必要使用延时显微镜分析动态事件。这些工具补充了 RNA谱分析研究，能够以高通量对致病过程进行单细胞解析， 以前所未有的规模对高度多样化或大量复制的患者集进行调查。发现 预测疾病表型在一个大的代表性患者样本，一个系统的，公正的方法是 需要挖掘延时显微镜图像序列、患者临床和伴随数据。有 因此，迫切需要下一代分析工具来发现疾病预测 表型对患者变异稳健。 因此，我们建议开发一个可教的动力学信息学发现（KID）工具的基础上， 层次推理框架如果得到证实，KID工具将迅速用于转化研究 在许多疾病中使用患者来源的细胞模型。它可以促进向广泛采用的范式转变 用于治疗发现、优化、分层和诊断发现的患者细胞模型。 我们这个快速通道项目的直接目标是通过展示 它可以根据疾病和疾病特征（如发病年龄）对患者进行分类。一期 我们将开发原型KID工具和初步的患者面板。我们将通过发现 表型，然后在盲法测试中根据表型对患者进行准确评分。在第二阶段，我们将 开发完整的患者面板和预产品KID工具。我们将通过在盲测中对患者进行评分来验证， 通过与靶向转录谱分析和基因组测试的确证来验证表型。我们将 将表型减少到单个时间点，并在靶向化合物筛选中显示功效。 具体目标是：第1阶段（目标1）：开发原型KID工具和初步患者面板 掺入神经元放电报告基因。(Aim 2）：在初步患者面板中验证原型KID工具。 第二阶段（目标1）：完成KID工具的beta原型和KID工具确认的患者面板。(Aim 2）、 疾病表型的发现和验证在整个患者面板。(Aim 3）：疾病表型验证。

Project Summary Neurological disorders significantly outnumber diseases in other therapeutic areas and are growing in incidence faster than any other disease classes. However, the pharmaceutical industry has been unsuccessful in coming up with effective drugs. A big factor in these failures has been a lack of adequate model systems for fundamental disease understanding affecting both diagnosis and treatment. There is therefore a strong, emerging interest in the use of patient-derived cell models to understand the pathogenic mechanisms underlying neurological disease phenotypes. To gain a true understanding of these mechanisms and phenotypes, it is necessary to analyze dynamic events using time-lapse microscopy. These tools complement RNA profiling studies by enabling single-cell resolution of pathogenic processes at high-throughput, enabling investigation of highly diverse or largely replicative patient sets at an unprecedented scale. To discover predictive disease phenotypes across a large representative patient sample, a systematic, unbiased approach is needed to mine time-lapse microscopy image sequences, patient clinical and concomitant data. There is therefore a critical need for a next-generation analytical tool to enable the discovery of disease predictive phenotypes robust to patient variations. Thus, we propose to develop a teachable kinetic informatics discovery (KID) tool based on a hierarchical inference framework. If proven, the KID tool would be rapidly adopted for translational research using patient-derived cell models in many diseases. It could facilitate a paradigm shift towards broad adoption of patient-cell models for therapeutics discovery, optimization, stratification and diagnostic discovery. Our immediate objective for this Fast-Track project is to develop and validate the KID tool by showing that it can classify patients on the basis of disease and disease characteristics such as age-of-onset. In Phase I we will develop the prototype KID tool and preliminary patient panel. We will prove feasibility by discovering phenotypes and then accurately scoring patients based on the phenotypes in blind tests. In Phase II we will develop the full patient panel and pre-product KID tool. We will validate by scoring patients in blind tests and validate phenotypes through corroboration with targeted transcriptional profiling and genomic tests. We will reduce the phenotypes to a single time point and show efficacy in a targeted compound screen. The Specific Aims are: Phase 1, (Aim 1): Develop the prototype KID tool and preliminary patient panel incorporating neuronal firing reporter. (Aim 2): Verify the prototype KID tool in a preliminary patient panel. Phase II, (Aim 1): Complete the beta prototype KID tool and patient panel for KID tool validation. (Aim 2): Disease phenotype discovery and verification in the full patient panel. (Aim 3): Disease phenotype validation.