Clustering techniques from significance analysis of microarrays

Abstract

Microarray technology is a prominent tool that analyzes many thousands of gene expressions in a single experiment as well as to realize the primary genetic causes of various human diseases. There are abundant applications of this technology and its dataset is of high dimension and it is difficult to analyze the whole gene sets. In this paper, the SAM technique is used in a Golub microarray dataset which helps in identifying significant genes. Then the identified genes are clustered using three clustering techniques, namely, Hierarchical, k-means and Fuzzy C-means clustering algorithms. It helps in forming groups or clusters that share similar characteristics, which are useful when unknown dataset is used for analysis. From the results, it is shown that the hierarchical clustering performs well in exactly forming 27 samples in first cluster (ALL) and 11 samples in the second cluster (AML). They will provide an idea regarding the characteristics of the dataset.