Visual Mining for Microarray Knowledge Discovery

We present Genome3DExploerer, an immersive visual mining tool to explore microarray data with other sources. The paper addresses the problem of exploring huge genome expression data, allowing biologists to group gene expression data in immersive environment. This could be a solution to explore augmented microarray data with other gene information sources. A microarray data set represents thousands of genes' expression levels in various experimental conditions. Genome3DExploerer handles co-expression patterns, namely clusters of genes with correlated expression profiles. Classical clustering methods offer biologists to group genes in distinct groups. Nevertheless, these methods can not deal with genes similar to several ones or those shared by distinct clusters. Moreover, the visualization techniques associated with these methods are not well adapted in order to explore huge amount of microarray data. We present in this work a new approach, based on dynamic graph immersive visualization, which offers a solution to process these microarray data characteristics.