Integrating Stochasticity into Biomolecular Mechanisms: A New Direction for Biomolecular Modeling

将随机性纳入生物分子机制：生物分子生物学的新方向 生物分子建模 摘要 动力学选择在生物学中的重要作用越来越明显。然而，我们才刚刚开始 有必要的工具来量化，表征和理解它。对于生物分子过程， 多个稀有事件转换，规范假设是机制遵循一致的 转换顺序（遵循单一路径）。然而，越来越多的证据表明， 实验和生物物理测量表明，多种途径不仅是可能的，而且是必要的。 本研究的目的是发展一个实验导向的随机模拟 绘制出机制异质性的框架。作为应用程序，我将首先集中讨论辅助 ClC Cl-/H+反向转运蛋白的主动转运和几种AAA+中ATP水解驱动的易位 ATP酶，两个涉及化学反应的过程，因此需要多尺度方法， 量子到经典领域的转变 所提出的多尺度动力学建模方法集中于多步生物分子转化， 这使得它对于已建立的动力学建模的许多其它领域是独特的。因此，新方法将 其他领域的最佳做法也将加以调整。它结合了自下而上的利率计算 多尺度模拟中动力学相关转换的系数，自上而下的参数细化 基于实验数据。创新性地提出了用贝叶斯参数精化动力学解空间的方法 估计、全局敏感性分析、不确定性量化、反应路径分析和机器学习 方法.这些方法将用于更好地表征ClC-ec 1中的Cl-/H+交换机制 与Merritt Maduke（斯坦福大学）合作的反向转运蛋白。野生型系统的动力学景观将是 研究了途径异质性的作用，非整数2.2：1 Cl-：H+化学计量的起源， 以及交替访问机制的相关性。 与次级主动转运类似，ATP驱动的过程固有地涉及多个速率影响步骤 (ATP结合、水解、Pi释放、ADP释放和所有相关的构象变化）。一 多尺度反应分子动力学方法将被发展来描述ATP水解。此外，本发明的目的是， 增强的自由能采样将用于表征其他过渡，多尺度动力学模型将 被开发来探测动力学选择性的作用，并测试有争议的随机与顺序 与Chris Hill（犹他州大学）合作提出了AAA+ ATP酶中的机制。

Integrating Stochasticity into Biomolecular Mechanisms: A New Direction for Biomolecular Modeling Abstract It is increasingly apparent that kinetic selection plays an important role in biology. However, we are just beginning to have the tools necessary to quantify, characterize and understand it. For biomolecular processes involving multiple rare-event transitions, the canonical assumption is that mechanisms proceed following a consistent order of transitions (following a single-pathway). However, increasing evidence from single molecule experiments and biophysical measurements suggests that multiple pathways are not only possible, but essential. The goal of the proposed research is to develop an experimentally-directed stochastic simulation framework for mapping out mechanistic heterogeneity. As applications, I will focus, first, on secondary active transport in the ClC Cl-/H+ antiporter and ATP hydrolysis driven translocation in several AAA+ ATPases, two processes involving chemical reactions and thus requiring multiscale methods that bridge the quantum to classical realms. The proposed approach to multiscale kinetic modeling is focused on multistep biomolecular transformations, which makes it unique to many other domains of established kinetic modeling. Thus, new methods will be developed and best practices from other domains will be adapted. It combines a bottom-up calculation of rate coefficients for kinetically relevant transitions from multiscale simulations, with a top-down parameter refinement based on experimental data. Innovation is proposed to refine the kinetic solution space with Bayesian parameter estimation, global sensitivity analysis, uncertainty quantification, reaction path analysis and machine learning methods. These methods will be used to better characterize the Cl-/H+ exchange mechanism in the ClC-ec1 antiporter in collaboration with Merritt Maduke (Stanford). The kinetic landscape for the wildtype system will be studied to address the role of pathway heterogeneity, the origin of the non-integral 2.2:1 Cl-:H+ stoichiometry, and the relevance of the alternating access mechanism. Similar to secondary active transport, ATP-driven processes inherently involve multiple rate-influencing steps (ATP binding, hydrolysis, Pi release, ADP release, and all of the associated conformational changes). A multiscale reactive molecular dynamics method will be developed to describe ATP hydrolysis. Additionally, enhanced free energy sampling will be used to characterize other transitions and multiscale kinetic models will be developed to probe the role of kinetic selectivity and to test the controversial stochastic versus sequential proposed mechanisms in AAA+ ATPases in collaboration with Chris Hill (University of Utah).