CompBio: Integration of Bioinformatics Modeling and Coarse Grain Dynamics: A New Computational Approach to Solve Complex Biological Problems

在新兴计算模型和技术计划(EMT)的支持下，费城科学大学(USP)的研究人员计划开发一种新的计算方法来研究复杂的生物系统。在这项建议中，我们计划研究趋化作用(对化学物质的反应)所使用的信号转导途径的一小部分。我们的方法结合了生物信息学辅助的分子建模技术和一种新的粗粒(CG)分子动力学方法。拟开展的工作包括三大部分。首先，将开发新的CG蛋白质模型和参数，以补充最近创建的CG动力学程序和CG脂质双层模型。其次，将利用初级序列分析、同源建模、对接和随机环构建等多种生物信息学和分子建模技术，构建跨膜信号转导系统的复杂结构框架-细菌趋化蛋白复合体。第三，复杂的结构框架将映射到蛋白质和脂肪双层的CG表示。将在毫秒尺度上进行CG模拟，以表征细菌趋化系统的结构、构象和动力学性质。CG动力学模拟的主要优点是计算量比传统的全原子分子动力学少四个数量级，但牺牲了一些原子细节。因此，它将使生物跨膜系统的动力学模拟首次达到毫秒级的时间尺度。由于许多感兴趣的生物事件(如信号转导)都发生在这个时间尺度上，所提出的模拟将显著增强我们解决复杂生物问题的计算能力。通过模拟，作为细胞质结构域信号传递过程的一部分的特定构象变化将被阐明，这一过程目前尚不清楚。虽然结果本身很重要，但它们将有助于展示这些技术的力量，并允许扩展到其他有时间和空间限制的领域。这些成果将加速膜蛋白的研究，进而对计算生物学、化学、生物化学、物理化学等众多研究领域以及医学、生物和物理等相关领域产生广泛而重大的影响。拟议的工作将由来自USP计算生物学和化学不同领域的一组专家合作进行。来自不同学科(生物信息学、化学、计算机科学)的研究生将接受培训，并进行许多拟议的研究，因此，赠款将促进学生的教学和培训以及跨科学学科的学习。此外，研究结果将在同行评议的期刊、国家和国际会议、网站和课堂上广泛传播。生成的计算机程序以及开发的参数将通过互联网和公共许可免费共享，以利于学术和非商业研究。

With support from the Emerging Models and Technologies for Computation Program (EMT), the researchers at University of the Sciences in Philadelphia (USP) plan to develop a novel computational approach for studying complex biological systems. In this proposal we plan on studying a small part of the signal transduction pathway used in chemotaxis (the movement in response to a chemical). Our approach combines bioinformatics-assisted molecular modeling techniques with a novel coarse grain (CG) molecular dynamics method. The proposed work includes three major parts. First, new CG protein models and parameters will be developed to compliment the recently created CG dynamics program and CG lipid bilayer models. Second, a complex structural framework of a transmembrane signal transduction system, the bacterial chemotaxis protein complexes, will be constructed by utilizing a variety of bioinformatics and molecular modeling techniques such as primary sequence analysis, homology modeling, docking and random loop building. Third, the complex structural framework will be mapped into the CG representation of the protein and lipid bilayer. The CG simulations at the millisecond scale will be carried out to characterize structural, conformational and dynamic properties of the bacterial chemotaxis system. The major advantage of CG dynamics simulation is that the computational demands are approximately four orders of magnitudes less than the conventional all-atom molecular dynamics, at the expense of some atomistic details. It will thus enable dynamics simulations of a biological transmembrane system to the milliseconds timescale for the first time. Since many biological events of interest (e.g. signal transduction) happen on this time scale, the proposed simulations will significantly enhance our computational capabilities in solving complex biological problems. Through the simulation, specific conformational changes as part of the signaling process in the cytoplasmic domain, which remains unknown, will be elucidated. While the results are important in themselves, they will serve to demonstrate the power of these techniques, and allow for extensions into other fields with temporal and spatial limitations. The results will accelerate the research on membrane proteins and by extension, have a broad and significant impact in a wide number of research areas, including computational biology, chemistry, biochemistry, physical chemistry, as well as related fields such as medicine, biology and physics. The proposed work will be carried out by collaboration of a group of experts from different areas of computational biology and chemistry at USP. Graduate students from different disciplines (bioinformatics, chemistry, computer science) will be trained and perform many of the proposed studies, thus the grant will promote teaching and training of students and learning across scientific disciplines. Additionally, the results of the research will be broadly disseminated by publication in peer-reviewed journals, at national and international conferences, on web sites and in the classrooms. The computer programs generated as well as parameters developed will be shared freely to benefit academic and non-commercial researches, via internet and public licensing.