Discovery of cancer linked biomarker genes through common subcluster mining

Abstract

Gene expression data from microarray experiments offer enormous scope for exploring the genetic relationship of deadly diseases. The motivation is to explore possible molecular biomarkers of such diseases with a view to early and periodic detection. A study has been reported in this paper with a methodology for common subcluster mining using FCM clustering. Subcluster refers to the peak formed through superimposition of clusters obtained from expressional data, both from the normal and diseased samples separately. Experiments are carried out on datasets of lung cancer, Acute Myeloid Leukemia(AML) and breast cancer employing the algorithm for common subcluster mining. Results are found to match to a large extent with those obtained in previous studies. Few genes emerge as indicative molecular biomarkers of respective diseases.