High Performance Computational System to Support LCMS/Proteomics Analysis

描述（由申请人提供）：布法罗大学（UB）和区域合作伙伴罗斯韦尔公园癌症研究所（RPCI）和豪普特曼-伍德-沃德研究所（HWI）加入了基础设施的主要持续投资，以支持面向基础，应用和临床生物医学应用的蛋白质组学研究。分布式区域仪器网络提供最先进的液相色谱/质谱（LC/MS）仪器。为我们的研究人员联盟生成蛋白质组学数据的关键仪器包括沃茨LC-QTOF Premier（RPCI）、配备电子转移解离的Thermo LTQ-XL线性离子阱和多维纳米流LC（UB Pharmaceutical Sciences Instrumentation Core），以及UB新的纽约州生物信息学和生命科学卓越中心（CBLS）的Thermo LTQ Orbitrap离子色谱分离和纳米流LC，物理上与RPCI园区相邻。额外的LC/MS仪器支持正在进行的蛋白质组学研究，并提供增强的定量能力。其中包括UB Pharmaceutical Sciences Core的4台LC/三重四极杆MS（2台API 3000和2台Thermo Quantum Ultra EMR），RPCI的1台API 3000和Thermo Quantum Ultra。对于这个由NIH支持的实验室组成的多机构联盟来说，从基础和临床研究样本中获得的大型蛋白质组学数据集的计算分析是一个严重的瓶颈。由于研究的性质，使用SEQUEST分析实验数据可能需要在标准高端工作站上进行数小时至数天的计算，这取决于基质复杂性和预期的肽修饰，使用ICAT或iTRAQ等试剂定量相对表达，以及研究中的蛋白质组。为了缓解这一瓶颈，我们建议获得一个专门设计用于加速LC/MS蛋白质组数据分析的专用高性能计算（HPC）集群。支持自动化的软件已在手或承诺收购，包括SEQUEST，X！Tandem、Mascot和Scaffold。拟议的网格支持的计算系统将被安置在UB计算研究中心（CCR），一个建立的HPC研究中心位于靠近LC/MS蛋白质组学设施。拟议的高性能计算集群将纳入充足的大容量存储和备份设施。在数据采集地点战略性地放置工作站，将为调查人员提供对HPC系统的分散访问。蛋白质组学软件的基准测试表明，MS数据分析的显著加速可以通过收购拟议的系统来实现，该系统利用了四核处理器技术的进步，并利用了非常大量的现有CCR基础设施。鉴于我们的研究人员在基础和临床科学方面的重点，拟议的设备将对威胁生命的疾病的新疗法的发展以及我们对基本生化和生理过程的理解产生直接，重大的影响。公共卫生相关性：在蛋白质组范围内分析蛋白质表达和翻译后修饰（PTM）是一项重要的新兴技术，将有助于我们了解基本的生化过程，细胞反应网络以及疾病过程对它们的影响。这些信息也有助于开发新的机械靶向药物。当前和悬而未决的NIH支持的研究项目受到蛋白质组学研究中遇到的大型数据集和/或复杂PTM的计算分析瓶颈的阻碍。提出的网格使能高性能计算集群和分布式处理网络将有效地解决我们的区域蛋白质组学研究联盟的这一计算瓶颈，从而直接推进许多NIH支持的项目，旨在改善艾滋病、心血管疾病和癌症等严重疾病的治疗。

DESCRIPTION (provided by applicant): The University at Buffalo (UB) and regional partners Roswell Park Cancer Inst. (RPCI) and Hauptman-Wood- ward Research Inst. (HWI) are joined in a major continuing investment in infrastructure to support proteomic research directed toward basic, applied, and clinical biomedical applications. A distributed regional network of instrumentation provides access to state-of-the-art liquid chromatography/mass spectrometry (LC/MS) instrumentation. Key instruments generating proteomics data for our consortium of investigators include a Waters LC-QTOF Premier (RPCI), a Thermo LTQ-XL linear ion trap equipped with Electron Transfer Dissociation and multidimensional nano-flow LC (UB Pharmaceutical Sciences Instrumentation Core), and a Thermo LTQ Orbitrap with ion chromatography fractionation and nano-flow LC in UB's new NY State Center of Excellence in Bioinformatics & Life Sciences (CBLS), located physically contiguous to the RPCI campus. Additional LC/MS instruments support the ongoing proteomics research and provide enhanced quantitative capabilities. These include four LC/triple-quadrupole MS (2 API3000 and two Thermo Quantum Ultra EMR) in the UB Pharmaceutical Sciences Core, an API3000 and Thermo Quantum Ultra at RPCI. For this multi-institutional consortium of NIH-supported laboratories, computational analysis of the large proteomics data sets acquired from basic and clinical research samples presents a severe bottleneck. Because of the nature of the research, analyses of experimental data using SEQUEST can require hours to days of computation on standard high-end workstations, depending on the matrix complexity and peptide modifications expected, the use of reagents such as ICAT or iTRAQ to quantify relative expression, and the proteome under study. To alleviate this bottleneck, we propose to acquire a dedicated high-performance computing (HPC) cluster specifically designed to accelerate LC/MS proteomic data analysis. Cluster-enabled software is in hand or committed for acquisition, and includes SEQUEST, X!Tandem, Mascot, and Scaffold. The proposed grid- enabled computing system will be housed in the UB Center for Computational Research (CCR), an established HPC research center located in close proximity to the LC/MS proteomics facilities. Adequate mass storage and backup facilities will be integrated into the proposed HPC cluster. Workstations placed strategically at the sites of data acquisition will provide decentralized access to the HPC system for investigators. Benchmarking of the proteomics software reveals that a considerable acceleration of MS data analysis can be achieved by acquisition of the proposed system, which exploits advances in quad-core processor technology and leverages very substantial existing CCR infrastructure. Given the emphasis of our investigators in both basic and clinical sciences, the proposed equipment will have direct, major impact on the development of new therapies for life-threatening diseases, as well as on our understanding of basic biochemical and physiological processes. PUBLIC HEALTH RELEVANCE: The analysis of protein expression and post-translational modifications (PTMs) on a proteome-wide scale represents an important emerging technique that will assist in our understanding of fundamental biochemical processes, networks of cellular responses, and the impact of disease processes upon them. This information can also contribute to the development of new, mechanistically-targeted drugs. Current and pending NIH-supported research projects are hampered by bottlenecks in the computational analysis of the large data sets and/or complex PTMs that are encountered in proteomics research. The proposed grid-enable high-performance computing cluster and distributed processing network will effectively address this computational bottleneck for our regional proteomics research consortium, and thereby directly advance the numerous NIH- supported projects aimed at improving the therapy of serious diseases such as AIDS, cardiovascular disease, and cancer.