Desenvolvimento baseado em modelos de serviços adaptadores para ferramentas de bioinformática

Abstract

An in silico experiment demands the execution of a number of activities supported by one or more dedicated command-line tools. A analysis web service may be created to adapt an existing analysis tool, exposing a suitable web interface to its user while still executing the original tool over user-submitted data in the background. However, the development of an adapter web service requires technical knowledge unusually detained by bioinformaticians and biologists. Additionally, the lack of a well-known reference model to guide the development of mature analysis services leads to the proliferation of services that have distinct characteristics, hindering the reuse of these services in different contexts. Thus, the availability of a development process that facilitates the development of mature and consistent adapter services is highly desirable. This research aimed at studying the use of model-based approaches to support the development and use of analysis web services in bioinformatics. First, this work proposes a reference model for bioinformatics analysis services to guide the development of mature RESTful analysis services. Then, the work proposes a model-based development process for adapter services, their clients and their descriptions, and provides a support infrastructure for the development process. Finally, the work details the application of the proposed development process and support infrastructure in the reengineering of an existing service repository in the functional genomics domain. We believe our work promotes the development and use of analysis services in the bioinformatics domain by allowing the bioinformatician to quickly obtain a mature analysis service to adapt an existing analysis tool, as well as its clients and associate service description.