Comparative genomic hybridization by representational oligonucleotide microarray analysis.

The central cause to any cancer ultimately lies in the genome and the initial alterations that result in changes in gene expression that are reflected in the phenotype of the cancer cell. The gene expression data are rich in information but the primary lesions responsible for carcinogenesis are obscured due to the complex cascade of expression changes that can occur. The primary lesions can be characterized by the smallest of point mutations to small insertions and deletions (in/dels) to much larger deletions and amplifications (for simplicity all copy number gains will be referred to as amplifications) as well as balanced or unbalanced translocations. In addition to these mutations there are a myriad of epigenetic alterations that affect the cells phenotype. Any gene if important to tumor growth will be altered by mutation or by deletion/amplification eventually, and if a large number of tumor samples is analyzed the majority of these genes will be detected. This chapter describes a variation of comparative genomic hybridization, called Representational oligonucleotide microarray analysis (ROMA), that surveys reduced-complexity representations of tumor genomic DNA to discover deletions and amplifications (and the underlying cancer genes).