Protein Structure Prediction Using First Principles

描述（由申请人提供）：本项目的总体目标是提高我们对蛋白质第一原理结构预测的理解。基于我们早期对Astro-Fold的研究工作（Klepeis和Floudas，2003 c）（见C.2节中的图1），拟议的研究旨在开发一个增强的蛋白质结构预测框架。该框架包括（a）用于预测螺旋片段的自由能计算，（B）用于预测β链和β折叠拓扑结构的组合优化方法，（c）用于预测螺旋间三级接触的新提出的方法，（d）用于预测一般螺旋结构的新提出的方法。 残基-残基接触预测，（e）从（a-d）中的预测导出改进的距离约束，以及新提出的基于优化的迭代方法，（e）通过aBB确定性全局优化方法和扭转角动力学产生低能三级蛋白质结构的系综，以及（f）新提出的用于选择预测的三级蛋白质结构的方法，从低能量蛋白质结构的系综，使用基于高和/或中分辨率诱饵开发的距离依赖力场。我们提出了以下四个具体目标： 具体目标1：研究和开发一种新的方法，用于预测a-螺旋蛋白和a/B蛋白的螺旋间三级接触。 具体目标二：研究并开发一种新的优化方法，用于预测α螺旋，α/β和β蛋白中的残基-残基接触。 具体目标3：研究一种新的基于迭代优化的方法，用于生成用于三级结构预测的附加和改进的距离约束。 具体目标4：研究和开发一个强大的力场，它将能够区分从三级蛋白质结构搜索产生的高分辨率和/或中等分辨率诱饵之间的折叠结构。研究了力保持展开的鲁棒优化方法。调查，测试和验证蛋白质结构预测的整体增强框架。 公共卫生相关性使用第一性原理的蛋白质结构预测不仅对计算生物学和化学具有根本重要性，而且对于促进对药物发现有直接影响的蛋白质-蛋白质和蛋白质-肽相互作用的理解也具有重要意义。对球状蛋白和膜蛋白结构的预测方法的改进将导致新药物的开发，这将有利于公众健康。

DESCRIPTION (provided by applicant): The overall aim of this project is to improve our understanding of first principles structure prediction of proteins. Based upon our earlier research work on the Astro-Fold (Klepeis and Floudas, 2003c) (see Figure 1 in section C.2), the proposed research is directed towards the development of an enhanced framework for protein structure prediction. This framework consists of (a) free energy calculations for helical segment prediction, (b) a combinatorial optimization approach for the prediction of beta strands and beta sheet topology, (c) a new proposed approach for predicting inter-helical tertiary contacts, (d) a new proposed approach for general residue-residue contact prediction, (e) the derivation of improved distance restraints from predictions in (a-d), and a new proposed optimization-based iterative approach, (e) the generation of an ensemble of low energy tertiary protein structures via the aBB deterministic global optimization approach and torsional angle dynamics, and (f) a new proposed approach for the selection of the predicted tertiary protein structure, from the ensemble of low energy protein structures, using a distance dependent force field developed based on high and/or medium resolution decoys. We put forward the following four specific aims: Specific Aim 1: Investigate and develop a novel approach for the prediction of inter-helical tertiary contacts for a-helical proteins, and a/b proteins. Specific Aim 2: Investigate and develop a novel optimization method for the prediction of residue-residue contacts in alpha helical, alpha/beta and beta proteins. Specific Aim 3: Investigate a new iterative optimization-based approach for the generation of additional and improved distance restraints for the tertiary structure prediction. Specific Aim 4: Study and develop a powerful force field which will be able to discriminate the folded structure among high resolution and/or medium resolution decoys generated from the search of tertiary protein structures. Study a robust optimization approach for the force held development. Investigate, test, and validate the overall proposed enhanced framework for protein structure prediction. PUBLIC HEALTH RELEVANCE The protein structure prediction using first principles is not only of fundamental importance to computational biology and chemistry, but also of major importance for advancing the understanding of protein-protein and protein-peptide interactions which have a direct impact on drug discovery. Improvements of predictive methods for the elucidation of protein structures for globular and membrane proteins will lead into development of novel drugs which will benefit the public health.