Hierarchical Methods for Large BioMolecular Complexes

描述（由申请人提供）：大生物分子复合物（LBC）形成负责大多数生物过程的机制，并与理解许多疾病（如癌症和代谢紊乱）相关。这些结构的知识不仅可以提供大分子如何在组装中起作用的机械描述，还可以为开发与疾病相关的治疗干预提供线索。今天，利用冷冻电子显微镜（CryoEM），电子断层扫描（ET）和X射线晶体学（X射线）或核磁共振光谱（NMR）的LBC的混合实验方法，需要巧妙地补充更快，更准确的计算处理，以尽可能最好的分辨率水平对LBC进行最终的超微结构解析。该建议解决了一个增强的和自动化的计算处理管道的发展，一旦体积CryoEM地图的LBC已被重建。特别是，我们提出了分层计算表示，算法和软件，自动化的LBC的结构特征的确定，以及加快LBC和相关的蛋白质和/或核酸之间的匹配和拟合技术的发展。 更确切地说，我们的具体目标是：目的1：开发算法，用于确定结构特征的LBC从冷冻-EM地图在三个不同的形态尺度。目标2：开发用于生成分层的、体积样条近似的LBC的确定的结构特征的算法，以促进基于快速傅立叶的相关搜索方法。目标3：使用我们的体积样条近似开发快速相关搜索方法，包括结构特征识别，结构拟合，LBC和蛋白质/RNA对接。目标4：实现，测试，包装和免费分发我们的结构特征确定和快速搜索/拟合技术。

DESCRIPTION (provided by applicant): Large Biomolecular Complexes (LBCs) form the machinery responsible for most biological processes and are relevant to understanding many diseases such as cancer and metabolic disorders. Knowledge of these structures would provide not only the mechanistic descriptions for how macromolecules act in an assembly but also clues in developing therapeutic interventions related to disease. Today, hybrid experimental approaches for LBCs utilizing cryo-electron microscopy (CryoEM), electron tomography (ET) and X-ray crystallography (Xray) or nuclear magnetic resonance spectroscopy (NMR), need to be ably complimented with faster and more accurate computational processing for final ultrastructure elucidation of LBCs at the best level of resolution that is possible. This proposal addresses the development of an enhanced and automated computational processing pipeline, once a volumetric CryoEM map of an LBC has been reconstructed. In particular we propose the development of hierarchical computational representations, algorithms and software, which automates structural feature determination of LBCs as well as speeds up match and fitting techniques between LBCs and relevant proteins and/or nucleic acids. More precisely, our specific aims are: AIM 1: To develop algorithms for determining structural features of LBCs from Cryo-EM maps at three different morphological scales. AIM 2: To develop algorithms for generating hierarchical, volumetric spline approximations of the determined structural features of LBCs to facilitate fast Fourier based correlation search methods. AIM 3: To develop fast correlation search methods using our volumetric spline approximations for LBCs, including structural feature identification, structure fitting, LBC and protein/RNA docking. AIM 4: To implement, test, package and freely distribute our structural feature determination and fast search/fitting techniques.