The analysis of microarray data is a computationalchallenge due to the characteristics of these data.Clustering techniques are widely applied to create groups ofgenes that exhibit a similar behavior under the conditionstested. Biclustering emerges as an improvement of classicalclustering since it relaxes the constraints for grouping allowinggenes to be evaluated only under a subset of the conditionsand not under all of them. However, this technique is notappropriate for the analysis of temporal microarray data inwhich the genes are evaluated under certain conditions atseveral time points. In this paper, we propose the TriGenalgorithm, which finds triclusters that take into account theexperimental conditions and the time points, using evolutionarycomputation, in particular genetic algorithms, enabling theevaluation of the gene’s behavior under subsets of conditionsand of time points.
展开▼