SGER: Discrete Event Simulation of Self-Assembly Kinetics

EIA-0320595Russell Schwartz卡内基梅隆大学项目摘要：自组装动力学的离散事件模拟该项目的目标是开发一种用于模拟广义自组装系统的新型计算工具。自组装系统由许多小组件或子单元组成，它们在适当的条件下自发排列成更大的结构。 在许多医学上重要的自组装系统中，有病毒蛋白壳或衣壳，它在病毒的遗传物质周围形成保护层；淀粉样蛋白，与阿尔茨海默病、亨廷顿病和朊病毒病有关的蛋白质纤维团聚体；和不规则的蛋白质聚集体。 对于所有这些系统，组装过程仅被部分了解。 此外，自组装作为构建纳米级人造设备和材料的一种手段最近引起了人们的兴趣。 由于许多自组装过程的尺寸小、速度快且复杂，事实证明它们很难进行实验分析。 因此，模拟方法已成为深入了解自组装过程的重要途径。 该项目旨在建立在该领域先前工作的基础上，创建一个足够通用的自组装过程模型，以捕获各种自组装系统，但速度足够快，可以在合理的时间内处理实际的模拟尺寸。 基本方法将涉及将先前针对该问题的建模工作中开发的技术与以前未用于自组装模拟的计算方法相结合。 模拟器将使用主要基于先前的局部规则动力学的自组装动力学模型。模型，它提供了低级亚基相互作用方面的高级自组装行为的通用表示。 它将使用称为离散事件优先级队列的计算数据结构来有效地实现。这将允许模拟器在离散状态的变化（例如相互绑定的子单元）之间进行步进，而不需要在所有时间步骤上进行显式积分。 其结果将是比以前的方法更快地模拟高度通用的自组装模型。 该模拟器将用 Java 实现，以方便开发、可扩展性和可移植性。 实施将通过不同的阶段进行，这些阶段致力于开发对象模型（指定计算机代码片段如何相互交互）、编码和测试原型模拟器，以及最终确定优化且记录良好的发布质量版本。 最终结果将是一个独立的模拟工具和一组可在其他程序中扩展和使用的计算类。 这项工作主要需要在自组装过程的数学模型和通过离散事件队列方法进行有效模拟的算法方面进行创新。 需要进一步创新，整合生物物理学、算法、软件工程和用户界面设计等领域的现有知识，以产生多功能、易于使用的图形模拟工具。 预计该项目将产生多项效益。 其影响将主要集中在自组装领域，通过提供一种通用工具，整个领域的研究人员可以使用该工具对跨尺寸和时间尺度的已知系统进行建模、开发新系统的计算原型以及对这两个领域的干预进行实验。 它还将通过开发新颖的模拟方法、在计算模拟器中的实现以及针对该问题的算法优化，为生物物理模拟的一般领域提供新的方法和经验。 该项目的跨学科性质将增强其影响力，为计算界提供生物物理系统中问题的新变化，并为生物物理学界提供可用于解决其他问题的新计算技术。 这项工作还将具有教育价值，为学生（包括两名本科生）提供跨学科研究经验，并提供可用作研究和教学工具的模拟器。

EIA-0320595Russell SchwartzCarnegie Mellon UniversityProject Summary: Discrete Event Simulation of Self-Assembly Kinetics The goal of this project is to develop a novel computational tool for simulating generalized self-assembly systems. Self-assembly systems consist of many small components, or subunits, that spontaneously arrange themselves into larger structures under appropriate conditions. Among the many medically important self-assembly systems are viral protein shells, or capsids, which form protective coats around the genetic material of viruses; amyloids, fibrous agglomerations of proteins that are implicated in Alzheimer.s disease, Huntington.s disease, and the prion diseases; and irregular protein aggregates. For all of these systems, the process of assembly is only partially understood. In addition, self-assembly has attracted recent interest as a means of constructing man-made devices and materials on the nanometer scale. Due to the small size, speed, and complexity of many self-assembly processes, they have proven difficulty to analyze experimentally. Simulation approaches have therefore emerged as a crucial avenue for gaining insight into the self-assembly process. This project seeks to build on the prior work in the area by creating a model of the self-assembly process sufficiently versatile to capture a wide variety of self-assembly systems, yet fast enough to handle realistic simulation sizes in a reasonable time. The basic methodology will involve combining techniques developed in prior modeling work on this problem with a computational method that has not previously been used for self-assembly simulation. The simulator will use a model of self-assembly dynamics based largely on the prior .local rules dynamics. model, which provided a versatile representation of high-level self-assembly behavior in terms of low-level subunit interactions. It will be efficiently implemented using a computational data structure called a .discrete event priority queue,. which will allow the simulator to step between changes in discrete state (such as subunits binding to one another) without the need for explicit integration over all time steps. The result will be faster simulation of a highly general self-assembly model than was possible with prior methods. The simulator will be implemented in Java to facilitate ease of development, extensibility, and portability. Implementation will be conducted through distinct phases devoted to developing an object model (which specifies how pieces of computer code interact with one another), coding and testing a prototype simulator, and finalizing an optimized and well documented release-quality version. The end result will be both a stand-alone simulation tool and a set of computational classes available for extension and use in other programs. This work will require innovation primarily in mathematical models of self-assembly processes and in algorithms for their efficient simulation by a discrete event queue methodology. Further innovation will be needed in the integration of existing knowledge from such areas as biophysics, algorithms, software engineering, and user interface design to produce a versatile, easy-to-use graphical simulation tool. The project can be expected to yield several benefits. Its impact will be primarily on the field of self-assembly, by providing a general tool that can be used by researchers throughout the field for modeling known systems across size and time scales, developing computational prototypes of novel systems, and experimenting with interventions in both. It will also provide new methods and experience to the general field of biophysical simulation through the development of a novel simulation methodology, its implementation in a computational simulator, and optimization of algorithms for this problem. The cross-disciplinary nature of the project will enhance its impact by providing for the computational community new variations on problems to be found in biophysical systems and providing for the biophysics community new computational techniques that can be brought to bear on other problems. The work will also have educational value by providing interdisciplinary research experience to students, including two undergraduates, and by providing a simulator that can be used as both a research and a teaching tool.