Non-coding RNAs (ncRNAs) are an important subset of the transcripts produced in the cells of organisms, since they aect many cellular processes. Although there are ecient and fast computational methods to identify proteins, annotation of ncRNAs is now focus of intensive research once their characteristics and signals are not yet entirely known. In this context, in this thesis, we present an architecture for ncRNAs annotation based on the multi-agent system paradigm. The implementation of a system, called ncRNA-Agents, uses collaborative agents, where each agent has knowledge and reasonig (simulating biologists) about a specic aspect of RNA, which contributes to a curated ncRNA annotation, with associated quality and explanations based on the results of the tools used by the system to recommend the annotation. In addition, we performed three case studies with three fungi, Saccharomyces cerevisiae, Schizosaccharomyces pombe and Paracoccidioides brasiliensis, to evaluate the performance of the system and its ability to annotate known ncRNAs and predict new ncRNAs. This tool is publicly available at http://www.biomol.unb.br/ncrna-agents.

Noncoding RNAs (ncRNAs) have been focus of intense research over the last few years. Since characteristics and signals of ncRNAs are not entirely known, researchers use different computational tools together with their biological knowledge to predict putative ncRNAs. In this context, this work presents ncRNA-Agents, a multi-agent system to annotate ncRNAs based on the output of different tools, using inference rules to simulate biologists' reasoning. Experiments with data from the fungus Saccharomyces cerevisiae allowed to measure the performance of ncRNA-Agents, with better sensibility, when compared to Infernal, a widely used tool for annotating ncRNA. Besides, data of the Schizosaccharomyces pombe and Paracoccidioides brasiliensis fungi identified novel putative ncRNAs, which demonstrated the usefulness of our approach. NcRNA-Agents can be be found at: http://www.biomol.unb.br/ncrna-agents.

. In recent years, non-coding RNAs (ncRNAs) have been focus of intensive research. Since the characteristics and signals of ncRNAs are not entirely known, researchers use different computational tools together with their biological knowledge to predict potential ncRNAs. In this context, this work presents a multiagent system to annotate ncRNAs based on the output of different tools, using inference rules to simulate biologists' reasoning. Experiments with real data of fungi allowed to identify novel putative ncRNAs, which shows the usefulness of our approach.

No abstract