MASSA: Multi-agent system to support functional annotation

Abstract

Functional annotation aims to predict the biological function of DNA sequences. This complex and time-consuming task has to process huge amounts of data and get high quality results. In order to guarantee the quality of the outcome, the annotation should be carried out by human experts, but the great volume of biological data produced lately demands a high degree of automation. The features of this problem (i.e., knowledge-based, distributed resources, and an evolving environment) make it suitable for an agent approach. This paper presents MASSA, a Multi-Agent System to support functional annotation. MASSA combines the potentialities of the agent approach with a Rule-Based Expert System to reproduce the annotation steps, including the human reasoning, at the inference stage. The expert system integrates knowledge on Biology and tools. A case study on the annotation of sequences of four phylogenetically distinct species illustrates the results and use of MASSA. © 2014 Springer International Publishing Switzerland.