Title: Combining multiple perspective as intelligent agents into robust approach for biomarker detection in gene expression data
Authors: Mohammed Alshalalfa, Ghada Naji, Ala Qabaja, Reda Alhajj, Jon Rokne
Addresses: Department of Computer Science, University of Calgary, Calgary, Alberta, Canada; and The Palestine Polytechnic University, Hebron, Palestinian Authority. ' Department of Biology, Lebanese University, Tripoli, Lebanon. ' Department of Computer Science, University of Calgary, Calgary, Alberta, Canada. ' Department of Computer Science, University of Calgary, Calgary, Alberta, Canada; Department of Computer Science, Global University, Beirut, Lebanon. ' Department of Computer Science, University of Calgary, Calgary, Alberta, Canada
Abstract: Locating exceptional, abnormal or unusual trends in gene expression data to identifying disease biomarkers is the vital problem tackled in this paper. We developed a comprehensive framework that incorporates different perspectives each realised by an agent. Each agent applies its method to analyse the gene expression data and to come up with some candidate genes as potential cancer biomarkers. Further, gene enrichment, protein interaction, and miRNA regulation are given weight; they are used to confirm the discoveries by the major agents. We conducted experiments on two data sets; the obtained results are very encouraging with a high classification rate.
Keywords: cancer biomarkers; gene expression data; clustering; multi-agent systems; MAS; classification; gene ontology; functional analysis; microRNA; protein interactions; metabolic pathways; intelligent agents; agent-based systems; biomarker detection; gene enrichment; protein interaction; miRNA regulation.
International Journal of Data Mining and Bioinformatics, 2011 Vol.5 No.3, pp.332 - 350
Published online: 25 May 2011 *Full-text access for editors Access for subscribers Purchase this article Comment on this article