PARALLEL PROCESSING RESOURCE FOR BIOMEDICAL SCIENTISTS

生物医学科学中存在着许多问题，这些问题太复杂了，无法 通过实验或大型机电源来解决。其中， 具有重要生物学意义的大分子的结构是一种卓越的结构。 明显的例子是干扰素，它是除一个原核以外所有原核的产物- 癌基因，以及荷尔蒙和其他生物信使的所有受体。 计算建模填补了实验方法留下的空白，并产生了 对机制和结构的洞察。计算资源 在这项建议中描述的唯一适合于检查生物 这些分子的影响。更具体地说，康奈尔大学的目标是 NIH理论中心的资源包括： O加强理论中心的优势，促进协作 生物学家、化学家、物理学家、计算机之间的研究努力 科学家和数值数学家，推动算法发展 和软件的进步，这将为进一步的进步奠定基础 生物医学； O为生物医学研究人员提供高度并行计算的途径 有可能实现非常高水平的 灵活性和性价比达到新水平的计算能力，以及 在研究人员的实验室里很容易复制； O在蛋白质折叠和其他研究方面取得科学进展 具有生物活性的大分子区域 资源的计算能力； O开发写作方法并评估现有的写作方法 使并行系统之间的可移植性成为可能的程序，从而 更新、更强大的机器可以立即投入使用； O广泛宣传两项就 其内的资源和技术发展； O利用从开发生物医学分子中获得的知识 为高度并行的系统构建应用程序以进一步增强 并行技术。

A multitude of problems exist in biomedical science that are too complex to be solved by experimentation or mainframe power. Among these, the structure of biologically important macromolecules is a preeminent one. Obvious examples are the interferons, the products of all but one proto- oncogene, and all receptors for hormones and other biological messengers. Computation modeling fills the gap left by experimental methods and yields insights into mechanisms and structures. The computational resources described in this proposal are uniquely suited to examine the biological effects of these molecules. More specifically, the aims of the Cornell Theory Center NIH Resource are: o To build on the strengths of the Theory Center to promote collaborative research endeavors among biologists, chemists, physicists, computer scientists, and numerical mathematicians, pushing algorithm developments and software advances that will lay the basis for further advances in biomedical science; o To provide to biomedical researchers access to highly parallel computing environments that have the potential to achieve very high levels of computing power at new levels of flexibility and cost/performance and which are easily duplicated in researchers' laboratories; o To achieve scientific advances in protein folding and the study of other biologically active macromolecules areas that would be impossible without the computing capability of the Resource; o To develop methodologies and assess existing methodologies for writing programs to make portability between parallel systems possible, so that newer and more powerful machines can immediately be put to use; o To broadly disseminate the results both of research conducted on the Resource and technological developments within it; o To use the knowledge gained from developing biomedical molecular structure applications for highly parallel systems to further enhance parallel technology.