Immune profiling of multi-parameter flow cytometry using computational statistics

描述（由申请人提供）：本申请涉及广泛的挑战领域（04）临床研究和特定的挑战主题04-AI- 102：人类对感染和免疫的免疫反应-通过现代免疫学方法和系统生物学进行分析。定量监测复杂免疫应答的能力对于开发有效疫苗和发现用于临床试验的诊断或预后生物标志物至关重要。多参数流式细胞术（FCM）可以用单个外周血（PB）样本测量多个免疫参数（细胞表型、活化或成熟状态、细胞内细胞因子或其他效应分子浓度），并提供理想的免疫应答详细快照。该项目的目的是开发和验证客观的统计方法来分析FCM数据，并应用这些方法来发现基于FCM的免疫相关性的疗效，在充分表征的艾滋病毒和晚期黑色素瘤队列。多色FCM技术的最新进展允许在单细胞水平上同时测量多达20种荧光标记物，并且这些最先进的测定显示出对感染和疫苗接种的免疫应答的巨大前景。然而，用于分析FCM数据的软件尚未跟上步伐，仍然依赖于串行2D门控方法，这些方法对于多维数据集的分析来说是次优的。因此，FCM的结果在不同的机构之间可能存在很大差异。直接处理多维数据的系统生物学方法需要设计能够跟上FCM技术创新的快速步伐的软件。我们建议开发计算统计模型，使用多参数FCM来表征免疫应答谱，并实现用于自动FCM分析和发现预测性免疫特征的高效软件。我们的具体目标是：1）开发多变量计算统计方法，以在多个样本中一致地表征FCM数据; 2）在广泛的FCM样本集上验证自动化细胞亚群鉴定; 3）基于FCM数据的统计模型鉴定免疫特征，预测感染或疫苗接种结果。该研究将大大扩展FCM分析的实用性，提供有效的自动统计方法和工具，用于从FCM数据中识别异质细胞亚群和免疫特征。这将使使用多参数FCM的任何人受益，特别是对疫苗开发，临床诊断和免疫治疗的影响，因为他们共同需要客观的FCM软件进行免疫分析。因此，我们的提案直接解决了挑战主题04-AI-102的目标，代表了对安全有效疫苗的合理设计和开发具有重大潜在影响的方法学进展。拟议的应用旨在开发自动化细胞亚群识别和预测免疫分析的感染和疫苗接种结果从多参数流式细胞术数据。该项目的成功将产生统计方法和软件，这些方法和软件将产生更准确、可重复的流式细胞术分析，并确定感染控制和疫苗效力的免疫相关性。

DESCRIPTION (provided by applicant): This application addresses broad Challenge Area (04) Clinical Research and specific Challenge Topic 04-AI- 102: The human immune response to infection and immunization - Profiling via modern immunological methods and systems biology. The ability to monitor complex immune responses quantitatively is essential for the development of effective vaccines and the discovery of diagnostic or prognostic biomarkers for clinical trials. Multi-parameter flow cytometry (FCM) can measure multiple immune parameters (cell phenotype, activation or maturation status, intracellular cytokine or other effector molecule concentrations) with a single peripheral blood (PB) sample, and provides a detailed snapshot of the immune response that is ideal for profiling. The purpose of this project is to develop and validate objective statistical methods to profile FCM data, and apply these methods to discover FCM-based immune correlates of efficacy in well characterized HIV and advanced melanoma cohorts. Recent advances in polychromatic FCM technology allow the simultaneous measurement of up to 20 fluorescent markers at the single cell level, and these state-of-the-art assays show tremendous promise for profiling the immune responses to infection and vaccination. However, software for analysis of FCM data has not kept pace and still relies on serial 2D gating methods that are sub-optimal for analysis of multi-dimensional data sets. As a result, FCM results can be highly variable across different institutions. Systems biological approaches that handle multi-dimensional data directly are needed to design software that can keep up with the rapid pace of FCM technological innovations. We propose to develop computational statistical models to characterize immune response profiles using multi- parameter FCM, and to implement efficient software for automated FCM analysis and discovery of predictive immune signatures. Our specific aims are to 1) develop multivariate computational statistical methods to characterize FCM data consistently across multiple samples; 2) validate automated cell subset identification on a broad set of FCM samples; and 3) identify immune signatures based on statistical models of FCM data that predict infection or vaccination outcome. The research will substantially extend the utility of FCM analysis with effective, automated statistical methods and tools for identifying heterogeneous cell subsets and immune signatures from FCM data. This will benefit anyone using multi-parameter FCM, with particular impact on vaccine development, clinical diagnostics and immune therapeutics, given their common need for objective FCM software for immune profiling. Hence, our proposal directly addresses the objectives of Challenge Topic 04-AI-102, and represents methodological advances with major potential impact on the rational design and development of safe and effective vaccines. The proposed application seeks to develop automated cell subset identification and predictive immune profiling of infection and vaccination outcomes from multi-parameter flow cytometry data. Success in this project will result in statistical methodology and software that will produce more accurate, reproducible flow cytometry analysis, as well as identify immune correlates of infection control and vaccine efficacy.