Transformative Computational Infrastructures for Cell-Based Biomarker Diagnostics

项目摘要 患者样本中异常细胞群的存在是对各种人类疾病的诊断， 尤其是白血病和淋巴瘤。用于基于细胞的诊断评估的主要技术之一 是流式细胞术，它使用荧光试剂来测量细胞群体的分子特征 在复杂的混合物中。而血液病的诊断通常采用细胞计数法。 对于恶性肿瘤，它可以更广泛地应用于其他疾病的诊断(如哮喘、过敏和 自身免疫)，如果它可以重复用于解释更高的复杂性染色面板和识别 更细微的细胞群体差异。流式细胞术分析也被广泛用于单细胞表型分析 在转译研究中，探索正常和异常生物过程的机制。 最近，质量细胞术的发展有望进一步增加单细胞的应用。 细胞学评估，以了解广泛的生理、病理和治疗过程。 当前的细胞仪数据分析实践依赖于二维数据图的“手动选通” 识别复杂混合物中的细胞子集。然而，这个过程是主观的、劳动密集型的，并且 不可复制，难以部署在多中心转化性研究或临床试验中 议定书的标准化和协调是至关重要的。该项目的目标是开发、验证和 传播用户友好的基础设施，用于计算分析两种诊断方法的细胞仪数据 和发现应用程序，可以帮助克服当前手动分析的限制，并提供 通过以下目标更有效、更客观和更准确地进行分析：具体目标1--实施一项新的 计算基础设施-FlowGate-用于细胞学数据分析，包括视觉分析和 机器学习；特定目标2--评估FlowGate在细胞群体特征方面的效用 机械论翻译研究(T1)；具体目标3--评估 FlowGate临床诊断与当前诊断标准分析的比较 细胞学数据(T2)；具体目标4--开发培训和教育资源并指导实施 开展外联活动，鼓励采用和使用由此产生的FlowGate网络基础设施。 该项目将通过克服采用的关键障碍，在推进翻译科学方面产生重大影响 通过促进分析流水线的优化，为用户提供直观的 对接，并提供定向培训活动。已开发的计算软件的应用 改善急性髓细胞白血病和慢性淋巴细胞性白血病诊断的基础设施将有助于对准确性的新强调 通过更精确地量化患者特定的肿瘤和正常反应特征来进行医学研究 细胞群。虽然FlowGate将由加州大学圣地亚哥分校、加州大学欧文分校和斯坦福大学CTSA开发， 由此产生的计算基础设施将免费提供给整个研究界。

Project Summary The presence of abnormal cell populations in patient samples is diagnostic for a variety of human diseases, especially leukemias and lymphomas. One of the main technologies used for cell-based diagnostic evaluation is flow cytometry, which employs fluorescent reagents to measure molecular characteristics of cell populations in complex mixtures. While cytometry evaluation is routinely used for the diagnosis of blood-borne malignancies, it could be more widely applied to the diagnosis of other diseases (e.g. asthma, allergy and autoimmunity) if it could be reproducibly used to interpret higher complexity staining panels and recognize more subtle cell population differences. Flow cytometry analysis is also widely used for single cell phenotyping in translational research studies to explore the mechanisms of normal and abnormal biological processes. More recently, the development of mass cytometry promises to further increase the application of single cell cytometry evaluation to understand a wide range of physiological, pathological and therapeutic processes. The current practice for cytometry data analysis relies on “manual gating” of two-dimensional data plots to identify cell subsets in complex mixtures. However, this process is subjective, labor intensive, and irreproducible making it difficult to deploy in multicenter translational research studies or clinical trials where protocol standardization and harmonization are essential. The goal of this project is to develop, validate and disseminate a user-friendly infrastructure for the computational analysis of cytometry data for both diagnostic and discovery applications that could help overcome the current limitations of manual analysis and provide for more efficient, objective and accurate analysis, through the following aims: Specific Aim 1 – Implement a novel computational infrastructure – FlowGate – for cytometry data analysis that includes visual analytics and machine learning; Specific Aim 2 – Assess the utility of FlowGate for cell population characterization in mechanistic translational research studies (T1); Specific Aim 3 – Assess the robustness and accuracy of FlowGate for clinical diagnostics in comparison with the current standard-of-care analysis of diagnostic cytometry data (T2); Specific Aim 4 – Develop training and educational resources and conduct directed outreach activities to stimulate adoption and use of the resulting FlowGate cyberinfrastructure. The project will have a major impact in advancing translational science by overcoming key hurdles for adoption of these computational methods by facilitating analysis pipeline optimization, providing intuitive user interfacing, and delivering directed training activities. The application of the developed computational infrastructure for improved diagnostics of AML and CLL will contribute to the new emphasis on precision medicine by more precisely quantifying the patient-specific characteristics of neoplastic and normal reactive cell populations. Although FlowGate will be developed by the UC San Diego, UC Irvine, and Stanford CTSAs, the resulting computational infrastructure will be made freely available to the entire research community.