A 3D Data Base for Electron Microscopy and Related Fields ofUltrastructure

几种三维成像技术将彻底改变大分子组装、亚细胞结构及其相互作用的研究。三维电子显微镜和共聚焦光学显微镜就是明显的例子。这些技术产生了以各种格式存储的大量数据，只有通过与作者的特殊安排才能访问。在科学出版物中，只能捕捉到低分辨率密度分布所代表的结构的几个方面。其他研究人员在同一领域或重叠领域进行“数据挖掘”的潜力很大，这些领域目前尚未使用，因为这些数据实际上无法访问。通过对NSF数据库活动计划的这一申请，PI希望继续并扩大美国-欧洲联合倡议，以建立一个分布式数据库，该数据库将容纳通过使用各种实验方法(EM、LM、X光等)产生的任何低分辨率(非原子)3D数据。该提案建立在国际和平协会和马德里何塞·玛丽亚·卡拉佐博士共同开发的现有原型的基础上，该原型已经可以通过Mosaic/WWW公开访问。根据提议，卡拉佐博士将两次访问沃兹沃斯中心，每次为期五个月，继续与国际和平协会及其小组密切互动，设计和测试分布式数据库和相关软件工具。计算机科学专业的一名研究生将在纽约州立大学的PI和Daniel Rosenkrantz博士的指导下，设计新的工具，允许输入分段数据并在复杂的3D数据查询中使用，例如在两个低分辨率结构的对接中，或者在两个具有部分-整体关系的结构的比较中。我们计划在中期组织一次研讨会(由欧洲政府共同资助)，届时来自3D超微成像中心的科学家以及科学数据库和网络领域的专家将汇聚一堂。通过研讨会，我们寻求征求关键意见，并就如何最好地分担数据筛选中的责任以及如何确保工作的连续性展开讨论。

Several techniques of three-dimensional imaging are about to revolutionize the study of macromolecular assemblies, subcellular structures, and their interactions. 3D electron microscopy and confocal light microscopy are obvious examples. These techniques produce large amounts of data stored in various formats and accessible only through special arrangements with the authors. Only few aspects of a structure represented by a low-resolution density distribution can be captured in scientific publications. There is a large potential for "data mining" by other researchers within the same field or overlapping fields that is currently unused because the data are virtually inaccessible. With this application to the Databases Activities Program of NSF, the PI wishes to continue and expand a joint US-Europe initiative for the establishment of a distributed database that would accommodate ANY low-resolution (non-atomic) 3D data produced by using various experimental methods (EM, LM, x-ray, etc.). The proposal builds on an existing prototype that was developed jointly by the PI and Dr. Jose Maria Carazo in Madrid and that is already publicly accessible through Mosaic/WWW. According to the proposal, Dr. Carazo will visit the Wadsworth Center twice, for a period of five months each, to continue the design and testing of the distributed database and associated software tools, in close interaction with the PI and his group. A graduate student in Computer Science, who will be supervised by the PI and Dr. Daniel Rosenkrantz at SUNY, will design new tools allowing segmentation data to be entered and used in complex 3D data queries, as for instance required in docking of two low-resolution structures, or in the comparison of two structures that have a part-whole relationship. We plan to organize a workshop in mid-term (to be jointly funded by the European Government) that will bring together scientists from centers of 3D ultrastructural imaging and specialists in the areas of scienti fic databases and networking. With the workshop we seek to solicit critical input and start a discussion on how responsibilities might best be shared in the screening of data and how to ensure continuity of the endeavor.