A robust method for finding somatic mutations to form clusters

Abstract

Major goal of malignant (cancer) genomics is to pinpoint which physically changed qualities are engaged with tumor commencement and movement. The goal of this paper is to speak to all focuses in a high dimensional source space by focuses in a low dimensional target space by natural neural systems and to discover subspace clustering adequately and proficiently. Here a mechanism is applied to identify the somatic mutational genes in the form of mutational patterns to categorize clusters. To achieve this target a model based clustering method SOM2C is applied for effective clustering of high dimensional data. This proposed novel approach begins by taking 584 patients’ data from COSMIC, and processes the data and forms the somatic mutational genes in one cluster and non –cancerous cells in another cluster. The experimental results show breast cancerous related cancerous(somatic mutational) and non-cancerous clusters with classification accuracy.