COMPUTER ASSISTED STRUCTURAL CHARACTERIZATION OF OLIGOSACCHARIDES

该项目的具体目标是开发计算机辅助的 寡糖的结构表征方法。 这 该项目的重点是开发一个基于万维网（WWW）的 用于分析信息处理的数据库系统， 分类称为复杂碳水化合物研究中心-神经 网络（CCRC-Net）系统。 CCRC-Net旨在使科学家 可以提交碳水化合物的光谱，以使用 CCRC的专利模式识别技术， 将光谱提交给数据库中的光谱。 从“查询 CCRC-Net制作了一个“结果页”， 这些碳水化合物相同或最接近类似的 查询谱的属性。 已经建立了联系， CCRC-Net和复杂碳水化合物结构数据库（CCSD）， CarbBank搜索程序。 CCSD/CarbBank现在位于 CCRC的网站，也将在网络上的功能集成 与其他生物序列和信息数据库，如Medline。 的 CCRC-Net为科学家提供了一组数据库及其 用于光谱分类的人工神经网络搜索引擎 和碳水化合物的化合物鉴定。 CCRC-Net是一个 独立于计算机平台的客户端-服务器系统。 唯一的 要求是用户具有图形Web浏览器（例如， Netscape、Microsoft的Internet Explorer、Mosaic）。 CCRC-Net 以“分形并行分布”的方式设计， 可以在任何级别进行扩展，而不会影响其 以一种对用户完全透明的方式来实现性能。 因此，它是一个“分布式”系统，其模块、数据或 服务可以位于世界任何地方。 CCRC-Net目前拥有 六个数据库和六个人工神经网络搜索引擎（一个 对于每个数据库）可通过WWW获得。 六个数据库包含 (i)气相色谱-电子轰击质谱联用 部分甲基化的糖醇乙酸酯（PMAA的GC-EIMS），（ii）1H-NMR 木葡聚糖的质子核磁共振谱 寡糖残基，（iii和iv）大（67个结构）和 一个小的（23个结构）组的1H-NMR谱的N-连接的 糖蛋白的寡糖侧链，（v） 来自新型隐球菌的葡糖醛酸甘露聚糖（GXM），和（vi）气体 色谱-电喷雾电离质谱（GC-EIMS） 通过还原裂解产生的部分甲基化脱水糖醇 复杂的碳水化合物。 CCRC-Net的重组图形用户 界面利用JAVA小程序进行更直观的交互， 用户. 一键点击本地目录浏览等任务 和文件传输提供了对这些功能的流线型访问。 CCRC-Net也有一个JAVA频谱查看器。 这个项目的目标是 建立CCRC网络的分形基础设施（骨干）。 新 数据库（虚拟数据模块，VDM）和搜索引擎（虚拟 搜索模块（VSM）插入主干（软件基础设施） 这反过来又提供了新模块和其他模块之间的连接， CCRC-Net模块。 一份描述这项工作的手稿正在 制备

The specific aim of this project is to develop computer-assisted methods for structural characterization of oligosaccharides. This project has focused on developing a World Wide Web (WWW)-based database system for analytical information processing and classification called the Complex Carbohydrate Research Center-Neural Network (CCRC-Net) system. CCRC-Net is designed so that scientists can submit a spectrum of a carbohydrate for identification using the CCRC's patented pattern recognition technology to compare the submitted spectrum to those in the database(s). From the "query spectrum" CCRC-Net produces a "results page" giving a "hit list" of those carbohydrates identical to or most closely resembling the attributes of the query spectrum. Links have been established between CCRC-Net and the Complex Carbohydrate Structure Database (CCSD) and its CarbBank search program. The CCSD/CarbBank now resides on the CCRC's Web site and will also be functionally integrated on the Web with other biosequence and informatic databases such as Medline. The CCRC-Net provides scientists access to a group of databases and their artificial neural network search engines for spectral classification and compound identification of carbohydrates. CCRC-Net is a client-server system that is computer platform-independent. The only requirement is that the user have a graphical Web browser (e.g., Netscape, Microsoft's Internet Explorer, Mosaic). CCRC-Net is designed in a "fractal parallel distributed" manner so that the system can accommodate expansion at any level without compromising its performance in a manner that is completely transparent to the user. Therefore, it is a "distributed" system whose modules, data, or services could reside anywhere in the world. CCRC-Net currently has six databases and six artificial neural network search engines (one for each database) available via the WWW. The six databases contain (i) combined gas chromatography-electron impact mass spectra of partially methylated alditol acetates (GC-EIMS of PMAAs), (ii) 1H-NMR (proton nuclear magnetic resonance) spectra of xyloglucan oligosaccharide residues, (iii and iv) a large (67 structures) and a small (23 structures) group of 1H-NMR spectra of the N-linked oligosaccharide side chains of glycoproteins, (v) 1H-NMR spectra of glucuronoxylomannans (GXMs) from Cryptococcus neoformans, and (vi) gas chromatography-electrospray ionization mass spectra (GC-EIMS) of partially methylated anhydroalditols generated by reductive cleavage of complex carbohydrates. CCRC-Net's restructured graphical user interface utilizes JAVA applets for more intuitive interaction for users. One-button clicks on such tasks as local directory browsing and file transfer provide streamlined acces s to these functions. CCRC-Net also has a JAVA spectrum viewer. The goal of this project is to build the fractal infrastructure (backbone) of CCRC-Net. New databases (virtual data modules, VDM) and search engines (virtual search modules, VSM) plug into the backbone (software infrastructure) which in turn provides connectivity between the new module and other CCRC-Net modules. A manuscript describing this work is being prepared.