Explicit ions in implicit solvent: fast and accurate.

项目摘要。本提案响应PAR-17-046“技术探索性研究 发展（R21）"。该提案的目标是设计和测试一个新的“GB- 离子”模型，将允许快速和准确的生物相关的动力学原子模拟 结构，如蛋白质和DNA在隐式水与明确的离子。 现代生物分子科学的进展，从结构生物学到基于结构的药物设计， 原子级建模和模拟方法大大加速了这一进程，这些方法弥合了理论与实际之间的差距， 和实验所谓的隐式溶剂化模型提供了速度和通用性的关键优势 通过表现水的影响--通常是这种模拟中计算成本最高的部分-- 以近似的方式。由此带来的建模工作的加速在许多领域都至关重要，从 从人类生物学的基础研究到新药的设计;快速隐式溶剂法可以 使得在传统的显式溶剂中进行昂贵的模拟成为可能 approach.然而，最适合原子模拟的方法论版本-所谓的 广义玻恩（GB）模型-在其基础上有一个严重的缺陷，妨碍了它在系统上的使用， 需要明确处理生物相关离子的问题。事实上，大多数生物医学 由于这个原因，相关的应用程序是目前GB无法达到的-这些是大多数系统，其中多个 一价离子如Mg 2+或Ca 2+起着关键作用，或者在单价离子与特定位点的结合是 重要.核酸和染色质的离子转运或致密化只是一长串例子中的两个。 GB模型的这种严重局限性将以一种新颖的、系统的方式来解决; 一个新的隐式溶剂化模型将通过一个试点实施和生物测试证明 相关结构。我们将开发一个新的模型，GB-ION，类似的精神广义玻恩，治疗 离子明确，在同一水平的准确性和效率作为目前的快速分析GB模型。 具体来说，GB将扩展到多个断开的电介质边界，超出 当前模型假设的单连接球形拓扑。新的原型模型将在 对于单价、二价和三价离子的代表性实例进行参数化。我们将在几个模型上进行测试。 生物相关的结构和过程，并在开源软件包中实现其原型，广泛 使用（AmberTools或/和OpenMM。） 结果将有利于整个生物分子建模社区建立有效的方法 为了在需要明确处理离子的情况下进行快速隐式溶剂原子模拟， 这是大多数生物医学相关的模拟。

Project Summary. This proposal responds to PAR-17-046 “Exploratory Research for Technology Development (R21)”. The goal of the proposal is to design and test in a pilot implementation of a novel “GB- ION” model that will allow fast and accurate atomistic simulations of dynamics of biologically relevant structures such as proteins and DNA in implicit water with explicit ions. Progress in modern bio-molecular sciences, from structural biology to structure-based drug design, is greatly accelerated by methods of atomic-level modeling and simulations that bridge the gap between theory and experiment. The so-called implicit solvation model provides critical advantages of speed and versatility through representing the effects of water – often the most computationally expensive part of such simulations – in an approximate manner. The resulting speed-up of modeling efforts is critical in many areas, from fundamental research in human biology to design of novel medicines; fast implicit solvent methodology can make possible simulations that are otherwise prohibitively expensive within the traditional explicit solvent approach. However, the version of the methodology best suited for atomistic simulations – the so-called generalized Born (GB) model – has a critical flaw in its foundation that precludes its use on systems and problems where explicit treatment of biologically relevant ions is needed. In fact, the majority of bio-medically relevant applications is out of reach to current GB for this reason – these are most of systems where multi- valent ions such as Mg2+ or Ca2+ play a critical role, or where binding of mono-valent ions to specific sites is important. Ion transport or compaction of nucleic acids and chromatin are just two examples out of a long list. This serious limitation of the GB model will be addressed in a novel, systematic way; advantages of the new implicit solvation model will be demonstrated through a pilot implementation and testing on biologically relevant structures. We will develop a novel model, GB-ION, similar in spirit to the generalized Born, that treats ions explicitly, at the same level of accuracy and efficiency as the current fast analytical GB models. Specifically, the GB will be extended to work for multiple, disconnected dielectric boundaries, beyond the singly-connected spherical topology that the current model assumes. The new prototype model will be parametrized for representative examples of mono-, di-, and tri-valent ions. We will test the model on several biologically relevant structures and processes, and implement its prototype in an open source package, widely used (AmberTools or/and OpenMM.) Results will benefit the entire biomolecular modeling community by establishing validity of an approach to carry out fast implicit solvent atomistic simulations in situations where explicit treatment of ions is necessary, which is the majority of bio-medically relevant simulations.