High Performance Computing for Multiscale Modeling of Biological Systems

一、总体--摘要 我们建议更新生物医学技术和研究资源(BTRR) 生物系统多尺度建模的性能计算，以下简称 MMBioS.MMBioS是由卡内基的匹兹堡大学(匹兹堡大学；牵头机构)联合发起的 梅隆大学(CMU)、匹兹堡超级计算中心(PSC)和索尔克研究所 生物学研究(索尔克)。我们的使命是继续开发计算方法和可用的 在计算技术与生活之间的界面上推进研究和培训的软件工具 科学。我们的生物学主题仍然是：真实而高效的建模、分析和模拟 分子和细胞的结构和动力学有助于理解和预测 生物功能/功能障碍的多尺度机制，重点是突触信号和 调节事件，从而促进了对抗神经和免疫的新疗法的发现 系统紊乱。在上届奖项取得进展的基础上，开始填补这一空白 在不同尺度的结构生物学、细胞微生理学和大型- 生物图像分析，我们现在进一步扩大我们的努力，以开发更强大的工具和 一个有效实施和使用我们技术的综合平台。我们已经增加了 我们的技术研究和开发项目的范围和数量从3个增加到4个，以推进和 能够适应分子建模(tr和d1)、细胞建模(tr和d2)、(细胞)网络 建模(tr&d3)和图像衍生建模(tr&d4)方法和软件面临新的挑战。 这些都是由七个推动生物医学项目(DBP)推动的：神经递质的动力学 分子和细胞水平的转运蛋白(DBP1；美国国立卫生研究院和佛罗里达大学)；调节和结合 关于PSD-95及其与AMPAR交易的关系(DBP2；加州理工学院)，多巴胺的多尺度模拟 转运蛋白功能(DBP3；PIT)；T细胞信号的时空建模(DBP4；英国布里斯托尔)； 构建动态的转录和染色质结构空间图(DBP6；NIH)；结构和 突触的功能(DBP7；UT Austin)；以及使用大型规则集进行建模的可扩展方法 和图像(DBP8；哈佛大学)。之前关于功能连接学的DBP5(Allen Brain Institute) 已成功完成。我们将继续开展积极的培训和传播计划，并 广泛的服务项目协作(C&SP)，利用独特的体验 PSC的能力，计算系和系统生物系的优势 (PIT)和计算生物学(CMU)，以及计算神经生物学的前沿研究 索尔克实验室。

I. Overall - Abstract We propose to renew the Biomedical Technology and Research Resource (BTRR) on High Performance Computing for Multiscale Modeling of Biological Systems, hereafter referred to as MMBioS. MMBioS is a joint effort between the University of Pittsburgh (Pitt; lead institution), Carnegie Mellon University (CMU), the Pittsburgh Supercomputing Center (PSC), and the Salk Institute for Biological Studies (Salk). Our mission is to continue to develop computational methods and usable software tools to advance research and training at the interface between computing technology and life sciences. Our biological theme remains: realistic and efficient modeling, analysis and simulations of molecular and cellular structure and dynamics toward understanding and predicting the origin and mechanism of biological function/dysfunction at multiple scales, with focus on synaptic signaling and regulation events, thus facilitating the discovery of new treatments against nervous and immune systems' disorders. Building on the progress made during the past award in starting to fill the gap between modeling efforts at disparate scales of structural biology, cellular microphysiology and large- scale bioimage analysis, we now further expand our efforts toward developing more powerful tools and an integrated platform for efficient implementation and use of our technology. We have increased the scope and number of our Technology Research and Development Projects from 3 to 4, to advance and enable the adaptation of molecular modeling (TR&D1), cell modeling (TR&D2), (cellular) network modeling (TR&D3), and image-derived modeling (TR&D4) methods and software to new challenges. These are driven by seven Driving Biomedical Projects (DBPs) on: the dynamics of neurotransmitter transporters at both molecular and cellular levels (DBP1; NIH and U of Florida); regulation and binding to PSD-95 and its relation to AMPAR trafficking (DBP2; Caltech), multiscale modeling of dopamine transporter function (DBP3; Pitt); spatiotemporal modeling of T cell signaling (DBP4; Bristol, UK); constructing a dynamic, spatial map of transcription and chromatin structure (DBP6; NIH); structure and function of synapses (DBP7; UT Austin); and scalable approaches to modeling using large sets of rules and images (DBP8; Harvard). Previous DBP5 (Allen Brain Institute) on functional connectomics has been successfully completed. We will continue our vigorous training and dissemination programs, and a broad range of Collaboration of Service Projects (C&SPs), taking advantage of the unique experience and capabilities of the PSC, the strengths of the Departments of Computational and Systems Biology (Pitt) and Computational Biology (CMU), and cutting-edge research at the Computational Neurobiology Laboratory at Salk.