Biclustering of Gene Expression Data with a New Hybrid Multi-Objective Evolutionary Algorithm of NSGA-II and EDA

Gene expression data produced by DNA microarray experiments advance the study of functions of genes. Clustering is able to find gene groups with identical biological function based on the principle that co-expression means so-regulation. Different from traditional clustering, biclustering is simultaneous clustering of both genes and conditions and searches maximal submatrices with maximal subgroups of genes and conditions where the genes exhibit highly correlated activities over a range of conditions. Maximizing the submatrices as much as possible and obtaining enough highly coherency among genes are usually conflicted. Therefore, multi-objective evolutionary algorithm is suitable for biclustering. We combine NSGA-II and EDA to generate a new multi-objective evolutionary algorithm for biclustering with advantages of the both methods. Finally the improved algorithm is applied to the dataset and gets better result..