Biclustering is a very useful data mining technique which identifies coherent patterns from microarray gene expression data. A bicluster of a gene expression dataset is a subset of genes which exhibit similar expression patterns along a subset of conditions. Biclustering is a powerful analytical tool for the biologist and has generated considerable interest over the past few decades. Many biclustering algorithms optimize a mean squared residue to discover biclusters from a gene expression dataset. In this paper a Two-Phase method of finding a bicluster is developed. In the first phase, a modified version of k-means algorithm is applied to the gene expression data to generate k clusters. In the second phase, an iterative search is performed to check the possibility of removing more genes and conditions within the given threshold value of mean squared residue score. Experimental results on yeast dataset show that our approach can effectively find high quality biclusters