Computational and Experimental Studies of Protein Structure and Design

项目简介：这是一个适用于设备的行政补充应用程序，以购买支持GPU的 计算机集群。对于父母的资助，3D蛋白质结构的确定不仅对于揭示 疾病过程的分子机制，也可用于基于结构的药物设计。随之而来的是计算性 蛋白质设计的进展可能会给治疗带来革命性的变化。随着这些进步，蛋白质和其他分子 可以作用于今天无法用药的蛋白质或明天的抗药性疾病。我的Mira研究项目 专注于蛋白质结构和设计(PS&D)的计算和实验研究。相互关联的目标是 (A)确定具有生物医学重要性的系统中的蛋白质结构和动力学；以及(B)设计蛋白质、抑制剂、 以及它们的分子相互作用，特别是预测和克服抗药性。我们在结构上开发了新的算法 分子生物学。为了克服其中的挑战，我们的算法利用组合优化、计算 几何和拓扑，并集成先进的机器学习技术。我们认为PS&D的软件必须是I) 开源和ii)自由软件。这是鱼鹰的目标。因此，我们将(C)继续开发自由、开放源码 算法和软件不仅适用于蛋白质及其相互作用设计中的挑战性问题，而且还适用于 确定困难的蛋白质结构并描述其动力学特性。 我们将使用结构数据和计算模型来理解分子机制和治疗的基础 干预，并在体外和体内进行详细的实验测量，以确认、重复和改进两者 我们对蛋白质结构和分子设计的理解。由此产生的蛋白质结构和动力学模型， 与我们新颖的设计方法一起，将阐明生物化学和药理学意义的目标。我们 还将通过在算法和建模方面的进步来推进PS&D。我们将通过以下方式测试我们的方法和预测 创建设计的蛋白质和抑制剂结构，求解经验结构，并进行体外实验 测量提纯组件的增强生物物理特性，并进行细胞内实验以测量生物功效。 我们将把我们的PS&D算法应用到生物医学的几个重要领域。我们将在以下条件下解决系统结构 我们的研究进一步发展了蛋白质结构为连续概率分布的范式.一套 提出了协同研究的推动力，其中，例如，我们将(1)预测未来蛋白质的抗药性突变 新药的靶点，(2)设计蛋白质-蛋白质相互作用(PPI)抑制剂，以“无法下药的”蛋白质为靶标，以及(3) 使用我们的PS&D方法来表征和设计抗体：抗原结构，最终目标是创造 针对病毒靶标的泛中和抗体。我们在开发新的计算方法方面的持续计划 针对早期先导准确预测潜在的药物靶点突变应该会推动更多药物的设计 坚韧耐用的第一代候选药物。

Project summary: This is an administrative supplement application for equipment, to purchase a GPU-enabled computer cluster. For the parent grant, the determination of 3D protein structures is essential not only for revealing molecular mechanism of disease processes, but also for structure-based drug design. Concomitantly, computational advances in protein design could revolutionize therapeutic treatment. With these advances, proteins and other molecules can be designed to act on today’s undruggable proteins or tomorrow’s drug-resistant diseases. My MIRA research project focuses on computational and experimental studies of protein structure and design (PS&D). The interlocking goals are to (A) determine protein structure and dynamics in systems of biomedical importance; and (B) design proteins, inhibitors, and their molecular interactions, especially to predict and overcome resistance. We develop novel algorithms in structural molecular biology. To surmount the challenges therein, our algorithms exploit combinatorial optimization, computational geometry and topology, and integrate advanced machine learning techniques. We believe software for PS&D must be I) Open-Source and II) Free software. This is the goal of OSPREY. Thus, we will (C) continue to develop free, open-source algorithms and software not only for challenging problems in the design of proteins and their interactions, but also to determine difficult protein structures and characterize their dynamics. We will use structural data and computational models to understand molecular mechanism and the basis of therapeutic interventions, and perform detailed experimental measurements in vitro and in vivo to confirm, iterate, and improve both our understanding of protein structure and molecular designs. The resulting models of protein structures and dynamics, together with our novel design methodology, will illuminate targets of biochemical and pharmacological significance. We will also advance PS&D by making algorithmic and modeling advances. We will test our methods and predictions by creating designed protein and inhibitor constructs, solving empirical structures, and performing in vitro experiments to measure enhanced biophysical properties on purified components, and in-cell experiments to measure biological efficacy. We will apply our PS&D algorithms to several areas of biomedical importance. We will solve structures of systems under our investigation and further develop the paradigm of protein structure as a continuous probability distribution. A set of synergistic research thrusts is proposed, in which, for example, we will (1) predict future resistance mutations in protein targets of novel drugs, (2) design protein-protein interaction (PPI) inhibitors that target “undruggable” proteins, and (3) use our PS&D methodology to characterize and design antibody:antigen constructs, with the ultimate goal of creating pan-neutralizing antibodies for viral targets. Our sustained program in developing novel computational methods to accurately predict potential drug target mutations in response to early-stage leads should drive the design of more resilient and durable first-generation drug candidates.