Bayesian methods for integrative structural biology: validation, sampling and modeling with EM data

综合结构生物学的贝叶斯方法：使用 EM 数据进行验证、采样和建模

• 批准号: 427880355
• 负责人: Professor Dr. Michael Habeck
• 金额: --
• 依托单位: Arbeitsgruppe Mikroskopische Bildanalyse
• 依托单位国家: 德国
• 项目类别: Research Grants
• 财政年份: 2019
• 资助国家: 德国
• 起止时间: 2018-12-31 至 2022-12-31
• 项目状态: 已结题
• 来源: https://gepris.dfg.de/gepris/projekt/427880355?language=en
• 关键词: Bayesian; methods; integrative; structural; biology

Hybrid methods provide insights into the 3D structures of large macromolecular complexes. Although the number of hybrid structures continues to grow, the community still lacks a generally accepted quality measure for structures obtained by integrating data from multiple experimental sources. Statistical methods, in particular those based on Bayesian inference, enable us to make sound statements about the quality of a hybrid structure. The first aim of this project is to develop new quality measures for the validation of hybrid structures by applying concepts and techniques from Bayesian inference. A prerequisite for statistical model assessment is that we sample conformational space exhaustively. Therefore, an important aspect is to also improve conformational sampling techniques.The second aim of this proposal is to enhance Bayesian modeling with cryo-electron microscopy (cryo-EM) data. Cryo-EM has emerged as a powerful method to characterize the structure of large macromolecular assemblies and can reach atomic or near-atomic resolution. To improve structural modeling with cryo-EM maps, we will build on our Inferential Structural Determination (ISD) software, which currently can be used for rigid and flexible fitting into low- to medium-resolution density maps. Our goal is to also support modeling with high-resolution maps such that structure modeling with ISD spans the entire range from high to low resolution. To this end, we will develop new probabilistic models for high-resolution maps, efficient algorithms for enhanced conformational sampling and methods for de novo modeling with cryo-EM maps.

混合方法提供了对大分子复合体的3D结构的洞察。尽管混合结构的数量持续增长，但对于通过整合来自多个实验来源的数据而获得的结构，社区仍然缺乏一个普遍接受的质量衡量标准。统计方法，特别是那些基于贝叶斯推理的方法，使我们能够对混合结构的质量做出合理的陈述。该项目的第一个目标是通过应用贝叶斯推理的概念和技术，为混合结构的验证开发新的质量衡量标准。统计模型评估的一个前提是我们对构象空间进行了详尽的抽样。因此，一个重要的方面就是改进构象采样技术。这项建议的第二个目标是利用冷冻电子显微镜(Cryo-EM)数据来增强贝叶斯建模。低温电子显微镜已经成为表征大分子集合体结构的一种强有力的方法，并且可以达到原子或近原子的分辨率。为了改进低温EM图的结构建模，我们将建立我们的推断结构确定(ISD)软件，该软件目前可以用于刚性和弹性拟合低到中分辨率的密度图。我们的目标是还支持使用高分辨率地图进行建模，以便使用ISD进行结构建模可以覆盖从高分辨率到低分辨率的整个范围。为此，我们将开发用于高分辨率地图的新的概率模型、用于增强构象采样的高效算法以及利用低温EM地图进行从头建模的方法。