Deciphering the dynamics of physiological systems using quantitative and computational approaches

生理系统表现出不同程度的复杂性和与它们不断变化的环境相互作用的迷人方式。实验研究这些系统的困难是一个重大的科学挑战，迫使我们采用替代方法来分析它们。在我的研究计划中，我们设计了定量方法（采用非线性动力学和计算生物学领域）与实验技术并行，研究神经生理和免疫系统的丰富动力学。这包括检查大脑内神经元的电生理和激素节律，分析单细胞和群体水平上的免疫细胞反应/动力学和相互作用，以及了解亚细胞过程（例如离子通道，受体，蛋白质等）如何在集成网络中控制两个系统的超细胞行为。为了实现这一点，我们设计了计算工具（例如信号处理和机器学习算法）来分析体内和体外数据，和/或开发数学模型以及定量方法来研究它们的行为。我们根据我们对调节它们的生理过程的理解构建了计算机模型，然后通过设计基于推理的拟合技术对实验数据进行验证。后者通过进行系统的模型比较和参数优化来确定最可能的模型和参数集。利用稳态分析、慢速分析、数值技术和生物物理方法，研究这些模型，以破译生理系统在亚/超细胞和（元）网络水平上的行为，提供有关其一般和紧急行为的见解，确定其中异常的原因，测试假设并指导新的实验，做出实验可测试的预测，并开发技术工具来操纵它们。开发的模型还提出了有趣的数学问题，这些问题可以使用我们设计的计算工具进行数值处理，也可以使用非线性（随机）动力学方法进行分析。

Physiological systems exhibit varying degrees of complexity and fascinating ways of interacting with their actively-changing environment. The difficulty in investigating these systems experimentally is a major scientific challenge that compel us to employ alternative approaches to analyze them. In my research program, we devise quantitative approaches (adopted from the field of Nonlinear Dynamics and Computational Biology) in parallel with experimental techniques, to study the rich dynamics of neurophysiological and immunological systems. That includes examining electrophysiological and hormonal rhythms in neurons within the brain, analyzing immune-cell response/kinetics and interactions at the single cell and population levels and understanding how subcellular processes (e.g. ion channels, receptors, proteins, etc.) govern the supracellular behavior of both systems within an integrated network. To accomplish this, we design computational tools (e.g. signal processing and machine learning algorithms) to analyze in vivo and in vitro data, and/or develop mathematical models along with quantitative methods to study their behavior. We construct in silico models based on our understanding of the physiological processes regulating them, followed by validating them against experimental data by designing inference-based fitting techniques. The latter is done by performing systematic model comparisons and parameter optimization to determine the most likely model and parameter set(s). Using steady state analysis, slow-fast analysis, numerical techniques and biophysical methods, the models are then investigated to decipher the behavior of physiological systems at the sub/supracellular and (meta)network levels, to provide insights about their general and emergent behavior, to identify causes of abnormalities within them, to test hypotheses and guide new experiments, to make experimentally testable predictions, and to develop technological tools to manipulate them. The models developed also raise intriguing mathematical questions that are tackled either numerically using computational tools that we design, or analytically using methods of nonlinear (stochastic) dynamics.