Large-scale genomic analysis of ovarian carcinomas

Abstract

Epithelial ovarian cancers are typified by frequent genomic aberrations that have been difficult to unravel. Recently, high-resolution array technologies have provided the first glimpse of the remarkable complexity of these aberrations with some ovarian cancers containing hundreds of copy number breakpoints, micro-deletions and amplifications. Many of these alterations contain cancer-related genes suggesting that the majority is disease-associated and not just the product of random genomic instability. Future developments such as next-generation sequencing and integrated analysis of data from multiple array platforms on large numbers of samples are poised to revolutionize our understanding of this complex disease.

Figure 1

Comparison of CGH data with 500K SNP array data. (A) Frequency data from 528 ovarian tumors showing most frequent gains (green) and losses (red) from Progenetix (Baudis and Cleary, 2001). (B) Frequency data from 31 ovarian tumors by 500K SNP array (Gorringe et al., 2007). The overall frequency of gains and losses is broadly similar between the two data sets. Note the additional ability to detect allelic imbalance gives much additional information over losses alone (e.g. chr13, chr22).

Figure 2

SNP array data enables high‐resolution analysis of copy number and LOH. (A) Chromosome 17 plot (p‐ter at left to q‐ter at right) for a mucinous ovarian tumor showing a small amplification at ERBB2. LOH accompanied by copy number loss can be distinguished from copy number neutral LOH and allelic imbalance caused by copy number gain. (B) Frequency plot of small copy number alterations from Gorringe et al. (2007). Assuming these alterations would show an approximately normal distribution, it is clear that the detection of alterations less than 50 kb is compromised by the resolution of the platform.