For a growing number of biologists DNA or protein data are typically retrieved and managed on the Web, and not in the laboratory. A large number of bioinformatics datasets from primary and (thousands of) secondary databases are scattered on the Web in various formats. A biologist end-user might need to access and use tens of databases and tools every day. For this reason, the bioinformatics community is developing more and more service-oriented architectures (SOAs): software architecture of loosely coupled software services that can be accessed without knowledge of, or control over, their internal architecture. Data-processing and analysis tasks can be automated by having free access to bioinformatics Web services (WSs) that are the building blocks of the SOAs. In this paper we introduce a new bioinformatics Web server, mepsMAP (mapping epitopes on protein surface: Mining Annotated Proteins), developed to identify the recognition sites between antibodies and their cognate antigens. In some cases, the recognition site is represented by a continuous segment of the antigen sequence, but much more often the epitope is "conformational," i.e., the antibody recognizes the location and type of exposed antigen side chains that are not necessarily contiguous in the antigen's sequence, but brought together by its three-dimensional structure. A facility on the server allows the user to search putative conformational epitopes on protein surface, querying the system for proteins with a given annotation. The mepsMAP server has been implemented as a SOA composed by a database and a set of four WSs. We present here the software architecture of the system with a detailed description of the WS dataflow that has been optimized to provide the best computing performance while maintaining the easiest end-user access to the system via a Web interface.