FRG: Metalloproteins: Computational Challenges, Methods, and Tools

技术摘要：材料研究司和网络基础设施办公室出资支持这项工作。这一奖项授予一个跨学科的重点研究小组，以支持计算机研究和教育，以调查与选定金属蛋白的结构和功能有关的关键问题。由于这必然涉及多方法方法，作为其计划的一部分，该小组将开发一套用于大规模生物分子模拟的计算工具。研究小组将在先前NSF支持下的成果的基础上，通过向信息技术研究征集方案颁发奖项来实现。取得的一些成就包括：开发了高度并行的粒子网格Ewald分子动力学(PMEMD)程序，该程序目前已并入Amber；快速实现了高阶静电多极，这是对Amber中的阿米巴力场进行有效编码的基础；分子静电势的新划分；用于量子蒙特卡罗模拟的pfaffian波函数的开发；从头算分子动力学的耦合；以及新的量子蒙特卡罗开源程序包QWALK的建立。该小组将专注于研究：(I)信号转导RAS蛋白；(Ii)依赖维生素K的凝血蛋白；(Iii)选择DNA修复酶，以及(Iv)金属多肽和金属蛋白结构域的折叠/展开。方法的发展包括：(I)新的量子蒙特卡罗/分子动力学方法和基于配对和pfaffian的新型变分波函数，它们允许精确处理具有复杂多参考关联的量子过程；(Ii)基于高斯基组和分布多极的经典区域的精确静电描述；(Iii)进一步发展量子力学和分子力学界面；以及(Iv)发展基于自适应偏置势能的新的自由能方法。开发的方法将被整合到上述包中，并在包括新一代Petaflop机器在内的多种架构上进行测试。该奖项支持高性能仿真中心在计算方法和科学方面的持续课程开发，该中心是在北卡罗来纳州立大学成立的，是前一个资助周期的一部分。该小组还将开发一门新的本科计算物理课程。非技术摘要：材料研究部和网络基础设施办公室出资支持这项工作。这一奖项授予一个跨学科的重点研究小组，以支持计算机研究和教育，以调查与选定蛋白质的结构和功能相关的关键问题，这些蛋白质由一个金属原子或簇组成，这些原子或簇结合到一组蛋白质上，这些蛋白质通过蛋白质-蛋白质相互作用结合在一起。在这一过程中，该团队将开发一套计算工具，用于对生命系统中发生的大分子进行大规模计算机模拟。对复杂大分子进行可靠的计算机模拟极具挑战性。该团队将专注于金属蛋白和金属酶？它们由一个金属原子或原子簇组成，这些原子结合到一组蛋白质上，这些蛋白质通过蛋白质相互作用而结合在一起。金属原子控制着分子的形状，从而控制着它的生物功能。它还可以实现特定的生化反应，而不是直接参与其中。这些都是生命中一类重要的蛋白质--S过程。它们的故障是许多致命疾病的根本原因。该团队将开发计算机模拟技术，捕捉在不同长度尺度上主导的物理现象，使复杂的模拟能够理解它们在对生命重要的各种过程中的作用，例如蛋白质是如何合成的，DNA是如何修复的。由此产生的网络工具将通过众所周知的网络工具套件提供给更广泛的科学界。该奖项支持高性能模拟中心在计算方法和科学方面的持续课程开发，该中心是在北卡罗来纳州立大学成立的，作为前一个资助周期的一部分。该小组还将开发一门新的本科计算物理课程。

TECHNICAL SUMMARY:The Division of Materials Research and the Office of Cyberinfrastructure contribute funds to support this work. This award to an interdisciplinary focused research group supports computational research and education to investigate key issues related to the structure and function of selected metalloprotiens. As this of necessity involves a multi-method approach, the group will develop a set of computational tools for large-scale biomolecular simulations as part of its program.The research team will build on the achievements under prior NSF support through an award made on a proposal to the Information Technology Research solicitation. Some achievements include: the development of the highly parallel Particle Mesh Ewald Molecular Dynamics, PMEMD, code currently incorporated into AMBER; a fast implementation of higher-order electrostatic multipoles that is the basis of an efficient coding of the AMOEBA force fields in AMBER; a new partitioning of the molecular electrostatic potential; the development of pfaffian wavefunctions for quantum Monte Carlo simulations; the coupling of ab initio molecular dynamics; and the establishment of the new quantum Monte Carlo open source package QWalk.The team will focus on investigating: (i) signal transduction Ras proteins; (ii) Vitamin K-dependent coagulation proteins; (iii) select DNA repair enzymes, and (iv) the folding/unfolding of metallopeptides and metalloprotein domains. Methods development thrusts include: (i) new quantum Monte Carlo /molecular dynamics methods and new types of variational wave functions based on pairings and pfaffians, which allow for accurate treatment of quantum processes with complicated multi-reference correlations; (ii) the accurate description of electrostatics for the classical region based on Gaussian basis sets and distributed multipoles; (iii) further development of quantum mechanics and molecular mechanics interfaces; and (iv) developing new free-energy methods based on adaptively biasing potentials. The developed methods will be incorporated into the above packages and tested on multiple architectures including the new generation of petaflop machines. This award supports continuing curricula development in computational methods and science at the Center for High Performance Simulation, which was founded at NC State as part of the previous funding cycle. The group will also develop a new undergraduate Computational Physics course.NON-TECHNICAL SUMMARY:The Division of Materials Research and the Office of Cyberinfrastructure contribute funds to support this work. This award to an interdisciplinary focused research group supports computational research and education to investigate key issues related to the structure and function of selected protiens that consist of a metal atom or cluster bound to a group of proteins that are held together by protein-protein interactions. In the process, the team will develop a set of computational tools for large-scale computer simulations of large molecules that occur in living systems.Reliable computer simulations of large complicated molecules are extremely challenging. The team will focus on metalloproteins and metalloenzymes ? which consist of a metal atom or cluster of atoms bound to a group of proteins that are held together by protein-interactions. The metal atom controls the shape of the molecule and so its biological function. It may also enable particular biochemical reactions without directly participating in them. These are an important class of proteins in life?s processes. Their malfunction is the root cause of many fatal diseases. The team will develop computer simulation techniques that will capture physical phenomena that dominate on different length scales to make the complicated simulations to understand their role in various processes important to life, for example how proteins are synthesized and how DNA is repaired. The resulting cybertools will be made available to the broader scientific community through well known suites of cybertools.This award supports continuing curricula development in computational methods and science at the Center for High Performance Simulation, which was founded at NC State as part of the previous funding cycle. The group will also develop a new undergraduate Computational Physics course.