Abstract. The availability of massive amounts of biological data, distributed in various data sources, has prompted the development of a wide range of data analysis tools. However, due to the heterogeneity and complexity of the domain, performing a specific analysis by combining tools is complicated. In this paper, we argue that the software agent metaphor constitutes a flexible approach towards integration of Bioinformatics tools and dynamic composition-execution of data analysis workflows. To substantiate this assertion, a scalable engineering methodology is presented, aiming to address simple as well as complex cases, in which generic Multiagent Systems are described to address different functional scenarios, incorporating agent coordination strategies and semantic domain descriptions. Based on this methodology, test cases are discussed on gene prediction tools to illustrate the applicability of the proposed systems.