A Hybridized Multiobjective Evolutionary Approach for Microarray Biclustering

GALLO CRISTIAN; CARBALLIDO JESSICA ANDREA; PONZONI IGNACIO

Santa Fe (Argentina)

Conferencia; CLEI 2008 (Conferencia Latinoamericana de Informática); 2008

In this paper a new memetic algorithm that integrates a multiobjectiveevolutionary algorithm (the SPEA2) with a local search technique, for microarraydata analysis, is presented. The algorithm explores the gene expression data matrixin order to find biclusters that fulfill several objectives. The performance of the newstrategy was analyzed by means of a fair comparison with a well known methodpresented in the literature. The cases of study were two datasets corresponding toSaccharomyces cerevisiae and human B-cells Lymphoma. Our algorithm outperformedthe previous method in terms of Set Coverage and Spacing, yielding to undeniablypromising outcomes. Nonetheless, more experiments with data from otherorganisms are necessary, thus leading to more concluding results.