COMPUTATIONAL INTEL CLUSTERS

计算英特尔集群

• 批准号: 6121969
• 负责人: THOMAS F COLEMAN
• 金额: $ 1.8万
• 依托单位: CORNELL UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 1998
• 资助国家: 美国
• 起止时间: 1998-12-01 至 1999-11-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/6121969
• 关键词: biological products; biomedical resource; computers; proteins; structural biology; technology /technique

One approach to the prediction of protein structure from sequence defines the native structure as the global energy minimum. Because the search for the global energy minimum of even a relatively small protein is impractical in an all-atom approach, during the past few years we have developed a simplified united-residue representation of the polypeptide chain. Once the global energy minimum of the simplified chain has been found, the chain can be converted to an all-atom chain and the search completed with an all-atom force field, such as ECEPP/3 used in this work. During this year, we focused our attention on the final development of the force field to include multibody or correlation terms that are necessary for the stabilization of regular secondary structures and developing and implementing the methods of global optimization to search the conformational space of the simplified chain. Based on the distance-scaling method of global optimization, we developed an efficient and reliable procedure for global optimization that we call the Basin-to-Deformed-Basin-Coupling method that is based on locating the regions of significantly low energy by gradually deforming the energy surface and locating the minima corresponding to the original potential in these regions by reversing the deformation. At present, the method is capable of locating the global energy minima of model polyalanine chains of up to 30 residues in length.

一种从序列预测蛋白质结构的方法 将固有结构定义为全局能量最小值。因为 寻找全球最小的能源，即使是相对较小的 在过去的几年里，蛋白质在全原子方法中是不切实际的 多年来，我们已经开发出一个简化的联合剩余表示法 多肽链。曾经全球能源的最低限度 已找到简化链，可以将该链转换为 全原子链和用全原子力场完成的搜索， 例如本工作中使用的ECEPP/3。在这一年里，我们专注于 注意力场的最终发展，包括 对象所必需的多体或相关项 规则二级结构的稳定化和发展 实现全局优化的方法来搜索 简化链的构象空间。基于 针对全局优化的距离尺度方法，提出了一种新的全局优化算法 高效可靠的全局优化过程，我们称之为 基于定位的盆地-变形盆地耦合方法 能量显著较低的区域通过逐渐变形 能量面和寻找与原始能量面对应的极小值 通过反转形变，这些区域的潜力。目前， 该方法能够定位模型的全局能量极小值 长达30个残基的聚丙氨酸链。