Development of Generalized-Ensemble Algorithms and their Application in Protein Studies

密歇根理工大学的乌尔里希·汉斯曼（Ulrich Hansmann）获得了理论与计算化学项目的奖励，以进一步开发用于蛋白质折叠模拟的广义集成算法。该奖项的部分资金由分子和细胞生物学部门的分子生物物理学项目提供。在成功破译整个基因组之后，一个新的挑战出现了：对于大多数新发现的序列，我们不知道相应蛋白质的功能。可靠的计算工具允许人们从氨基酸序列中研究蛋白质的结构和功能的出现，因此可以更深入地了解细胞基本过程的分子基础。对于详细的蛋白质模型，在计算机模拟中探索这种关系是极其困难的，并且在简单的室温模拟中，精确计算物理量的计算工作量随着残基的数量呈指数增长。为了克服这一困难，PI率先采用了广义集合方法，并证明了其模拟多达40-50个氨基酸的肽和蛋白质的可行性。通过进一步推进该方法并在其高度可扩展的软件包SMMP中实现该方法，PI的目标是推导出描述蛋白质（通常由50-200个氨基酸组成）中稳定结构域的折叠和相互作用的模型。将算法优化与统计力学的基本思想结合起来，PI正在探索75个残基载脂蛋白calbinding D9K和84个残基四聚体BBTA2的折叠和组装；以及铜离子与酵母铜硫蛋白37残基截断形式的结合。他的目标是仅仅从蛋白质内部原子之间以及蛋白质与周围环境之间的物理相互作用来理解这些过程。因此，这项工作对生物学领域产生了广泛的影响，同时也通过包括一些美国印第安学生在内的学生参与这项研究。

Ulrich Hansmann of Michigan Technological University is supported by an award from the Theoretical and Computational Chemistry program for work to further develop generalized-ensemble algorithms for use in protein folding simulations. Partial funding for this award has been provided by the Molecular Biophysics program within the Division of Molecular and Cellular Biology.After the successful deciphering of whole genomes a new challenge has emerged: for most of the newly found sequences we do not know the function of the corresponding proteins. Reliable computational tools that allow one to study the emergence of structure and function of proteins from their sequence of amino acids would therefore lead to deeper insight in the molecular foundation of the fundamental processes in cells. Exploring this relationship in computer simulations is extremely difficult for detailed protein models, and the computational effort to calculate accurately physical quantities increases exponentially with the number of residues in simple room-temperature simulations. In order to overcome this difficulty, the PI has pioneered the generalized ensemble approach and demonstrated its feasibility for simulation of peptides and proteins of up to 40-50 amino acids. By further advancing this approach and implementing it in his highly scalable software package SMMP the PI aims at deriving models that will describe folding and interaction of stable domains in proteins (usually consisting of 50-200 amino acids). Connecting optimization of algorithms to fundamental ideas of statistical mechanics the PI is probing folding and assembly of the 75-residue apo calbindin D9K and the 84-residue tetramer BBTA2; and binding of copper ions to the 37-residue truncated form of yeast copper thionein. His goal is to understand these processes solely from the physical interactions between the atoms within a protein, and between the protein and the surrounding environment. The work is, thus, having a broad impact on the field of biology as well as through the involvement of students, including some American Indian students, in this research.