High-performance compute cluster for biomedical computing

描述(由申请人提供)：计算生物学和生物信息学中心(C2B2)成立于2003年，其使命是为生物医学研究团体提供结构和系统生物学算法和软件工具，用于剖析分子相互作用和基于相互作用的细胞表型阐明。在过去的几年里，C2B2研究人员开发了许多新的分析方法，这些方法导致了重要的生物学发现，包括了解DNA形状在蛋白质-DNA结合特异性中的作用，以及发现与恶性表型相关的基因，包括淋巴瘤、胶质瘤和黑色素瘤。所有这些计算研究的一个共同主线是它们非常重要的计算需求。这既是算法的组合性质的结果，也是应用算法的数据集的基因组规模的结果。例如，生成基因相互作用网络需要100多万个CPU-小时，该网络用于识别驱动胶质瘤间充质表型的主要调节因子。实际上，这种调查只有在高性能计算(HPC)环境中进行时才是可行的。为了满足这一需求，C2B2在高性能IT基础设施方面进行了大量投资(利用哥伦比亚大学、纽约州和美国国立卫生研究院的赠款)，并建立了全国最大的致力于生物调查的HPC设施之一。C2B2 HPC基础设施的核心是Titan，这是一个3712核心集群，在2008年该系统首次亮相时，Top500.org将其评为世界第124位最快的计算机。这种基础设施对C2B2的成功起到了重要作用，它使中心的研究人员能够执行原本不可行的计算调查，并允许我们招聘和留住世界一流的教师。正如过去的情况一样，我们现在已经达到了一个点，我们的调查人员的计算需求超出了我们的高性能计算基础设施的限制。历史利用率数据表明，计算消耗每年都在翻一番。因此，在过去的六到八个月里，泰坦一直在以最高运力运行。我们预计，随着C2B2从中心向完整的大学系(系)过渡，对计算的需求将进一步增加。系统生物学)，批准的预算用于再招聘9名教师。通过这一应用，我们寻求将我们的计算能力翻两番，以适应NIH资助的几个项目在2-3年内的数据分析需求。鉴于计算在C2B2研究计划中的核心作用，提供足够和易于访问的高性能计算环境是成功完成这些项目的关键先决条件。这里要求的设备采购将使我们能够在可预见的未来继续服务于哥伦比亚大学的生物医学研究需求。

DESCRIPTION (provided by applicant): The Center for Computational Biology and Bioinformatics (C2B2) was established in 2003, with the mission of providing the biomedical research community with Structural and Systems Biology algorithms and software tools for the dissection of molecular interactions and for the interaction-based elucidation of cellular phenotypes. Over the last few years C2B2 investigators have developed many novel analysis methodologies which have led to important biological discoveries including understanding the role of DNA shape in protein- DNA binding specificity and the discovery of genes causally related to the presentation of malignant phenotypes, including Lymphoma, Glioma, and Melanoma. A common thread among all these computational investigations is their very significant computational requirements. This is a result of both the combinatorial nature of the algorithms as well as the genome-scale size of the data sets on which they are being applied. For example, more than 1 million CPU-hours were needed to generate the gene interaction network that was used to identify the master regulators that drive the mesenchymal phenotype in Glioma. In practical terms, such investigations are only feasible when carried out on a high performance computing (HPC) environment. To address this need, C2B2 has made significant investments in high performance IT infrastructure (utilizing grants from Columbia University, the New York State and the NIH) and has assembled one of the largest HPC facilities dedicated to biological investigation in the nation. The centerpiece of the C2B2 HPC infrastructure is Titan, a 3712 core cluster which was ranked by Top500.org as the 124th fastest computer in the world when the system debuted in 2008. This infrastructure has been instrumental to the success of C2B2, by enabling Center researchers to execute computational investigations that would have otherwise been infeasible and by allowing us to recruit and retain world class faculty. As has been the case in the past, we have now reached a point where the computational needs of our investigators are exceeding the limits of our HPC infrastructure. Historical utilization data demonstrate that computation consumption has been doubling annually. As a result, Titan has been running at its peak capacity for the past six to eight months. We anticipate that demand for computation will increase even further as C2B2 is in the process of transitioning from a Center to a full University Department (Dept. of Systems Biology) with approved budget for 9 more faculty hires. Through this application we seek to quadruple our computational capacity to accommodate the data analysis needs of several NIH funded projects over a period of 2-3 years. Given the central role of computing in the research program of C2B2, the availability of an adequate and easily accessible HPC environment is a critical prerequisite for the successful completion of these projects. The equipment purchase requested here will allow us to continue serving the needs of biomedical research at Columbia University for the foreseeable future.