CORE 3 : Modeling Core

建模核心 总结 为了提高我们对传染病的认识，宿主病原体地图倡议（HPMI）已经启动， 系统地分析病原体和宿主之间相互作用的分子网络。而 这项工作的一部分是实验性的，这个建模核心提出了中央计算框架。 第一个计算主题涉及确定人类亚细胞亚结构和组分的方法 以及它们是如何被致病因子改变的。目标1专注于创建3D模型， 宿主-病原体蛋白复合物。它将应用已建立的综合结构生物学方法， HPMI项目，包括冷冻电子显微镜（cryo-EM），蛋白质组学（PPI，PTM和APEX）和 基于CRISPR/Cas9的基因研究。最初的工作将集中在Nsp 2-Rap 1Gds 1，ORF 8-IL 17 RA， 涉及SARS-CoV-2和人蛋白的ORF 9 B-MARK 2复合物，以及来自分枝杆菌的Pks 13 结核病，所有这些都是HPMI在以前的工作中确定的。目标2侧重于绘制宿主细胞组分， 在蛋白质复合物上和上面的鳞片，延伸到更大的隔室和细胞器。它将扩大我们的 最近令人信服的概念验证，用于创建一个公正的亚细胞成分的分层图， 人体细胞将分析这些全细胞图谱，以揭示靶向的特定亚细胞组分。 病原体和/或在病原体或宿主中处于突变选择下。 第二个计算主题涉及将宿主-病原体细胞图谱与功能分析相结合的方法 和预测医学。目标3使用这些地图构建可解释的深度学习系统来预测 包括疾病严重程度在内的传染病结局。这一目标借鉴了我们以前的工作， “可见的”学习模型（DCell和DrugCell），它们不是黑盒，但具有内部组织 由生物结构的先验知识决定。我们将从HPMI细胞图构建这样的模型， 结合了我们第一代飞行员的关键改进。最后，Aim 4将使用可视深度学习 系统以及其他机器学习模型，以设计和评估控制 M.结核病和SARS-CoV 2感染。 通过这些目标，我们最大限度地提高了我们对大分子结构和功能的基础知识 这对感染至关重要，同时将这些知识嵌入到精准医疗的智能系统中。

MODELING CORE SUMMARY To improve our understanding of infectious diseases, the Host Pathogen Map Initiative (HPMI) has launched systematic efforts to analyze the networks of molecular interactions that interconnect pathogens and hosts. While parts of this work is experimental, this Modeling Core presents the central computational framework. A first computational theme concerns methods to determine human subcellular substructures and components as well as how they are altered by interactions with pathogenic factors. Aim 1 focuses on creating 3D models of host-pathogen protein complexes. It will apply established methods of integrative structural biology to data from the HPMI projects, including cryo-electron microscopy (cryo-EM), proteomics (PPI, PTM, and APEX) and CRISPR/Cas9-based genetic studies. Initial efforts will focus on the Nsp2-Rap1Gds1, ORF8-IL17RA, and ORF9B-MARK2 complexes involving SARS-CoV-2 and human proteins, as well as Pks13 from Mycobacterium tuberculosis, all identified in previous work by the HPMI. Aim 2 focuses on mapping host cell components at scales at and above the protein complex, extending to larger compartments and organelles. It will expand on our recent compelling proof-of-concept for creating an unbiased hierarchical map of subcellular components in human cells. These whole-cell maps will be analyzed to reveal specific subcellular components targeted by pathogens and/or that are under mutational selection in either pathogen or host. A second computational theme concerns methods to integrate host-pathogen cell maps with functional analysis and predictive medicine. Aim 3 uses the maps to build interpretable deep learning systems for prediction of infectious disease outcomes including disease severity. This aim draws from our previous work to establish “visible” learning models (DCell and DrugCell), which are not black boxes but have internal organization determined by prior knowledge of biological structure. We will construct such models from HPMI cell maps, incorporating key improvements over our first-generation pilots. Finally, Aim 4 will use visible deep learning systems alongside other machine learning models to design and evaluate combinatorial biomarkers that govern M. tuberculosis and SARS-CoV2 infection in clinical settings. Through these aims, we advance our basic knowledge of the macromolecular structures and functions most crucial for infection, while embedding this knowledge within intelligent systems for precision medicine.