The Mathematical and Computational Modelling of Cytokine Networks

免疫系统呈现出一个复杂的多尺度网络，具有细胞、细胞间和细胞内的组成部分，并作为多尺度连接和信号传导的典范；从细胞到细胞，从细胞因子到细胞，从化学物质到细胞内的化学物质。特别是细胞因子是在免疫系统中具有根本重要性的扩散信号分子，协调不同免疫细胞之间的通讯。细胞因子相互作用可以用复杂网络来概括，与动态系统重叠，因此通常表示为非线性常微分方程的耦合系统。该博士项目的目标之一是检查和理解从实验数据推断细胞因子网络时模型选择和参数估计的网络不确定性问题，以及使用动力系统理论来简化和研究所得模型。例如，这项工作的进一步目标是研究此类网络上动力系统在存在时间和空间扰动的情况下的稳定性。最终目标是将这种理解应用于质疑模型研究，在针对自身免疫和炎症性疾病的细胞因子靶向治疗的背景下，研究患者反应的时间动态以及潜在生物标志物的行为。初步研究将检查与炎症性肠病相关细胞因子谱的扰动研究中推断的网络相关的动力系统，这些细胞因子谱是在来自健康人类捐赠者的体外单核细胞中观察到的，这些捐赠者是通过牛津胃肠道生物库招募的（11/YH/0020 和 16/YH/0247）[1]。可以考虑基于来自该平台的数据或来自牛津大学肯尼迪风湿病研究所正在运行的关节炎治疗加速计划的类似数据集进行进一步的网络推理研究。该研究的影响将在于开发利用多维扰动研究合理开发细胞因子相互作用的计算机模型的方法，进而为系统地告知干预自身免疫性疾病细胞因子治疗的潜在目标提供范围。这项研究的新颖性在于系统地使用网络、动力系统和贝叶斯推理的理论来研究大型细胞因子系统，并与数学生物学、非线性系统以及统计学和应用概率的 EPSRC 领域保持一致。最后，该项目将涉及与葛兰素史克研究人员的广泛互动，他们通过 EPSRC iCASE 奖项为学生提供部分资助。

The immune system presents a complex multi-scale network, with cellular, inter- and intra-cellular components and serves as an exemplar of multiscale connectivity and signalling; from cell to cell, cytokine to cell and chemical to chemical within cells. In particular the cytokines are diffusing signaling molecules of fundamental importance in the immune system that orchestrate the communication among diverse immunological cells. Cytokine interactions may be summarised in terms of complex networks, overlaid with a dynamical system and thus often represented as coupled systems of nonlinear ordinary differential equations. One aim of this doctoral project will be to examine and understand the problem of network uncertainty for model selection and parameter estimation in inferring cytokine networks from experimental data, as well as the use of dynamical system theory to simplify and investigate the resulting models. For instance, a further aim of this work will be to investigate the stability of dynamical systems on such networks in the presence of temporal and spatial perturbations. A final aim will be to apply such understanding to interrogating modelling studies investigating the temporal dynamics of patient responses, and the behaviour of potential biomarkers, in the context of cytokine-targeting treatments for autoimmune and inflammatory diseases.The initial studies will examine the dynamical system associated with the inferred network from a perturbation study of the cytokine profiles relevant to inflammatory bowel disease that are observed within in-vitro monocytes sourced from healthy human donors, who had been recruited via the Oxford gastrointestinal biobank (11/YH/0020 and 16/YH/0247) [1]. Further network inference studies based on this data from either this platform may be considered or similar datasets from the Arthritis Therapy Acceleration Programme, that is being run within the Kennedy Institute of Rheumatology at the University of Oxford. The study impact will be in developing methodologies for the rational development of in silico models for cytokine interactions using multidimensional perturbation studies, in turn offering the scope for systematically informing potential targets for intervention in cytokine treatments for auto-immune disorders. The novelty of this study concerns the systematic use of the theory underlying networks, dynamical systems and Bayesian inference for the study of large systems of cytokines and aligns with the EPSRC areas of mathematical biology, nonlinear systems together with statistics and applied probability. Finally the project will involve extensive interaction with researchers from GlaxoSmithKline, who are part funding the studentship via an EPSRC iCASE award.