A Genetic K-means Clustering Algorithm Applied to Gene Expression Data

One of the current main strategies to understand a biological process at genome level is to cluster genes by their expression data obtained from DNA microarray experiments. The classic K-means clustering algorithm is a deterministic search and may terminate in a locally optimal clustering. In this paper, a genetic K-means clustering algorithm, called GKMCA, for clustering in gene expression datasets is described. GKMCA is a hybridization of a genetic algorithm (GA) and the iterative optimal K-means algorithm (IOKMA). In GKMCA, each individual is encoded by a partition table which uniquely determines a clustering, and three genetic operators (selection, crossover, mutation) and an IOKM operator derived from IOKMA are employed. The superiority of the GKMCA over the IOKMA and over other GA-clustering algorithms without the IOKM operator is demonstrated for two real gene expression datasets.