Microarray Biclustering: a novel Memetic Approach based on the PISA Plataform

GALLO CRISTIAN; CARBALLIDO JESSICA ANDREA; PONZONI IGNACIO

LECTURE NOTES IN COMPUTER SCIENCE

Springer-Verlag

Año: 2009 vol. 5483 p. 44 - 55

0302-9743

In this paper, a new memetic approach that integrates a multiobjectiveevolutionary algorithm (MOEA) with local search for microarraybiclustering is presented. The original features of this proposal are theconsideration of opposite regulation and incorporation of a mechanism fortuning the balance between the size and row variance of the biclusters. Theapproach was developed according to the PISA framework, thus achieving thepossibility of testing and comparing several different memetic MOEAs. Theperformance of the MOEA strategy that performed better was also comparedwith the biclusters obtained by a multi-objective approach recently published.The benchmarks were two datasets corresponding to Saccharomyces cerevisiaeand human B-cells Lymphoma. Our proposal achieves a better proportion ofcoverage of the gene expression data matrix, and it also obtains biclusters withnew features that the former existing evolutionary strategies can not detect.