In the two years since NSF initially funded the development of the online Microbial Germplasm Database (MGD), issues of preservation of all categories of germplasm and loss of biodiversity have become key social, political and scientific issues. There has been a concomitant intensified perception of the role of microbial germplasm in plant health, sustainable agriculture, biotechnology, biocontrol, bioremediation and genetic engineering. Clearly, and from diverse sources, there is a call for an inventory of germplasm, national and worldwide, and for its preservation. No single national collection can expect to maintain more than type species or representative isolates of the diversity of irreplaceable germplasm in the biosphere. Working culture collections, those collections, maintained by and for specific research purposes often contain thousands of isolates representing a cross-section of the genetic diversity within a single taxon. MGD has placed online a prototype database cataloging and quantifying microbial germplasm utilized in plant-related research and held in working collections at universities and research stations. We have thus far uploaded approximately 40,000 records of accession information from volunteer-researchers who have allowed their collection data to be included. This is a small fraction of the body of data available. In the project, MGD will develop a vigorous outreach program to assist researchers in developing suitable formats for data storage and in moving their collection records to a computer accessible medium. Not only text data describing germplasm, but images, charts, drawings, and diagnostic materials, as well, will be included. During the previous funding period, software was developed that allows uploading of information from disparate sources to a common area where data relating to individual strains can be browsed or searched by key words. The system is based on the Internet Gopher, a tool developed at the University of Minnesota for connecting to remote sites and allowing use of data resources at that site. The system is based on placing text files on a server in directory locations reflecting intuitive and logical associations so that a user can move to the files by simple menu choices. We have devised means to provide multiple views of records so that users may browse by any of several criteria. As a prototype, the result is quite satisfactory and has provided much useful information. We have now reached the limits of the UNIX file system, however. During this round of funding MGD will be moved to a fully relational database structure using SQL as the query language. This will necessitate modifications to the Gopher server so it can interface with SYBASE database management system installed on the MGD computer. The Gopher layer will be maintained because of its ease of use and its ubiquity among users of academic high speed networks, and because its continued evolution at other sites will bring added capability to MGD users. As part of a Masters Degree project in the Department of Computer Science, extensions to the interface will allow users to store information about queries that they find useful, define time-saving macros, establish views of the data that are unique to individual needs, and maintain other user-dependent information. Users will be afforded the benefits of a powerful query language while at the same time being shielded from its complexities. This becomes increasingly important as the volume of data on MGD grows. Because MGD users represent a considerable spectrum in terms of both computer experience and areas of interest, special attention will be devoted to identifying user requirements for the interface layers. Survey instruments will be designed and applied to a cross-section of the biological community in order to establish the prototypical conceptual models with which users approach MGD (i.e., their perceptions of the database's organization and how meaningful queries should be structured). The most common patterns of access will be identified, and database structure adapted to these patterns so that the most common user queries are easy to specify and fast to carry out. It is anticipated that an interface of this nature will have far-reaching implications for other providers of information resources. An e-mail interface, now being developed at MGD will be placed in beta testing during the next funding period and, once stable, will be placed into full production. Remote locations in the U.S. and in developing countries as well as sites served by other networks (many secondary schools, community colleges, and private companies), will be able to access data and images placed online by MGD. MGD will also make available to participants, tools that have proven useful in providing computer-supported collaborative work environments, including a prototype multiple-user workspace, which is being developed at XEROX PARC (ASTRO-VR) for the astrophysics community. The environment allows users at remote locations to enter a common network workspace, interact easily in real time and even examine images concurrently. The network and computing resources required are not excessive and are within the reach of most researchers.
