Informative Gene Selection for Leukemia Cancer Using Weighted K-Means Clustering

Gene expression data analysis is playing a vital role in diagnosing the diseases and drug designing. Many researchers realized that most of the cancers could be diagnosed based on the gene expression data. This paper focusses on identifying the prominent genes, which are mainly causing the Leukemia cancer using computational methods. Clustering methods are used to identify the components of a data set without the prior knowledge of the data set. The Weighted K-Means clustering method is proposed in a novel manner to analyze the Leukemia Cancer data set. However, the resultant clusters are again clustered sample wise using the KMeans clustering approach to understand more meaningful biological inferences. The proposed method selects the most significant genes and produces high accuracy in cancer classification.