RUI Protein Conformation: New Methods for Global Optimization and Parameter

该项目涉及开发和改进一种新的基于凸全局低估因子(CGU)的构象搜索策略。CGU搜索策略已经在Sun模型上进行了成功的蛋白质折叠测试。该项目将(1)加速CGU方法的快速构象搜索；(2)通过参数估计来改进势能函数。将使用几个改进来加速该方法，包括(A)使用核心定向链生长方法作为前端，(B)在适用的情况下使用han和Baker二级结构规则来减少自由度数，以及(C)使用阿佩尔方法来减少所需的能量评估次数。为了改进搜索方法，使其与更好的能量函数一起工作，该项目将系统地从Sun模型转向更现实的蛋白质势函数和模型。为此，将开发一种新的计算参数估计方法，该方法可以同时调整所有能量函数参数，以减小模型中的全局最小构象与选定蛋白质集合中所有成员的已知天然构象之间的差异。

This project involves the development and improvement of a new conformational search strategy based on a convex global underestimator (CGU). The CGU search strategy has already been successfully tested on the Sun model for protein folding. This project will (1) speed up the CGU method for fast conformational searches and (2) improve the potential energy function by parameter estimation. Several enhancements will be used to speed up the method, including (a) using a core-directed chain growth method as a front end, (b) using the Han and Baker secondary structure rules, where applicable, to reduce the number of degrees of freedom, and (c) using Appel's method to reduce the number of required energy evaluations. To improve the search method so that it will work with better energy functions, this project will systematically move away from the Sun model toward more realistic protein potential functions and models. To do this, a new computational parameter estimation method that can simultaneously adjust all the energy function parameters will be developed, so as to reduce the difference between the global minimum conformation in the model and known native conformation for all members of a selected set of proteins.