BioMoby is an open source, extensible framework that enables the representation, discovery, retrieval, and integration of biological data from distributed data repositories and analysis services. By registering their analysis and data access services with BioMoby, service providers agree to use and provide service specifications in a shared semantic space. The BioMoby Central registry now hosts more than a thousand services in the United States , Canada , and several other countries across the world.

BioMoby uses a datatype hierarchy to facilitate automated discovery of Web services capable of handling specific input datatypes. As a minimal Web based interface, the Gbrowse Moby service browser can be used by biologists to discover and invoke biological Web services from the Moby registry and seamlessly chain these services to compose multi-step analytical workflows. The process is data centric, relying on input and output datatype specifications of the services. The Seahawk client interface can infer the datatype of the input data files directly and immediately present the biologist with a list of Web services that can process the input file. Users of the Seahawk interface are relieved of the necessity to familiarize themselves with datatype hierarchies, and instead are free to concentrate on the analytical aspects of their work.

The BioMoby service encyclopedia provides a query interface to the repository of services. A form allows users to retrieve the services whose description matches a given keyword. The output of the query is a table listing alphabetically the services registered in BioMoby matching the query. A scientist interested in the post-translational modifications that affect (or can affect) a given protein sequence, may enter the keyword "sequence" in the form and obtain a table containing more than 200 services. No refinement of query is allowed, the user can only extract manually the needed information. MOBY-S Web Service Browser retrieves bioinformatics resources with respect to a data type. A browsing session starts with the selection of a data type registered in BioMoby. To identify the services relevant to post-translational modifications of a given protein, the user has to search for all namespaces registered in BioMoby that capture the scientific concept "protein", and submit as many requests to the browser. Additional interfaces to BioMoby services include registry browsers that provide access to the complete list of registered BioMoby services organized in a HTML page.

These interfaces are convenient when searching for services with respect to a specific data format (input), but they are not suitable when searching for services with respect to their scientific meaning rather than their format. Another critical limitation of the approaches occurs when no single service achieves the task. In order to allow the discovery of the services that can be used to express scientific protocols, combinations of services must be retrieved.

The conceptual map generated by our algorithm can be viewed at http://www.eas.asu.edu/~bioinfo/BioMobyMap/. We present only a partial view of the concepts and services as the sheer number of services present in BioMoby cannot be displayed easily. The complete list of concepts and services can be downloaded from the following link (add the link for downloading the excel sheet).The boxes represent the services and the oval are the concepts. In the map, the concepts are not necessarily the actual inputs and outputs of the service but rather the more abstract (super class) data type of the hierarchy

The material presented in this page is partially extracted from the encyclopedia entry Biological Metadata Management and Resource Discovery authored by Zoé Lacroix , Arizona State University, Cartik R. Kothari, University of British Columbia, Canada, Peter Mork, The MITRE Corporation, Rami Rifaieh, San Diego Supercomputer Center, Mark Wilkinson, University of British Columbia, Canada, Sarah Cohen-Boulakia, Université Paris-Sud, France, and Juliana Freire, University of Utah, and to be published by Springer-Verlag.