A semantic framework for web service annotation, matching and classification in bioinformatics

The success of Web service technology has brought a lot of interest from a large number of research communities such as Software Engineering, Artificial Intelligence, Semantic Web, Semantic Grid, etc. Despite several efforts towards automating service discovery and composition, users still search for services via online repositories and compose them manually. In our opinion, this is due to the lack of semantic annotations (metadata) to describe service semantics and support an effective and efficient discovery of services. Semantic annotation is commonly recognized as one of the cornerstones of the semantic Web and also, an expensive, time consuming and error prone process. Thus, approaches to automatically derive annotations that would describe rapidly changing Web services repositories are extremely required. In this paper, we propose a semantic framework for bioinfor-matics Web service annotation, matching and classification. We propose a semi-automatic extraction approach of lightweight semantic annotations from textual description of Web services. We investigate the use of NLP techniques to derive service properties given a corpus of textual description of bioinformatics services. We evaluate the performance of the annotation extraction method and the importance of lightweight annotations to match, reason and classify bioinformatics Web services in order to bootstrap the service discovery process. Based on extracted annotations, we propose an inference and block clustering approaches, the two approaches are complementary. The former relies on semantic annotations and explicit background knowledge to match a discovery query and a set of Web services. The latter approach aims to deduce implicit associations between services and annotations highly correlated by applying an accelerated version of the Croki2 algorithm.