Abstract Various types of genomic aberration are observed in malignant tumours, including single nucleotide mutation and copy number variations (CNV). While the former are focal changes, the effects of which can be localised to individual genetic elements, the latter often contain megabases of genomic regions, covering hundreds of genetic elements. CNV has been successfully used on a coarse scale (cytoband level) as molecular features to define subtypes in many cancer entities. However, a systematic understanding of the CNV patterns in cancers is not established. Here, I use intervals as CNV aggregation units and perform enrichment analysis by gene. With ovarian cancer (601), lung cancer(554) and breast cancer (1105) data sets from TCGA, this method is able to enrich for frequently deleted and duplicated genes, previously described as cancer drivers. Also, I will show recent results in the attempt to use STRING protein network to further interpret the implications of copy number changes among interaction partners of drivers in cancer patient stratification.