Identification of recurrent focal copy number variations and their putative targeted driver genes in ovarian cancer.

Liangcai Zhang,Li Zhang,Karen H Lu,Ying Yuan

doi:10.1186/s12859-016-1085-7

Abstract

BackgroundGenomic regions with recurrent DNA copy number variations (CNVs) are generally believed to encode oncogenes and tumor suppressor genes (TSGs) that drive cancer growth. However, it remains a challenge to delineate the key cancer driver genes from the regions encoding a large number of genes.ResultsIn this study, we developed a new approach to CNV analysis based on spectral decomposition of CNV profiles into focal CNVs and broad CNVs. We performed an analysis of CNV data of 587 serous ovarian cancer samples on multiple platforms. We identified a number of novel focal regions, such as focal gain of ESR1, focal loss of LSAMP, prognostic site at 3q26.2 and losses of sub-telomere regions in multiple chromosomes. Furthermore, we performed network modularity analysis to examine the relationships among genes encoded in the focal CNV regions. Our results also showed that the recurrent focal gains were significantly associated with the known oncogenes and recurrent losses associated with TSGs and the CNVs had a greater effect on the mRNA expression of the driver genes than that of the non-driver genes.ConclusionsOur results demonstrate that spectral decomposition of CNV profiles offers a new way of understanding the role of CNVs in cancer.Electronic supplementary materialThe online version of this article (doi:10.1186/s12859-016-1085-7) contains supplementary material, which is available to authorized users.

Highlights

Genomic regions with recurrent DNA copy number variations (CNVs) are generally believed to encode oncogenes and tumor suppressor genes (TSGs) that drive cancer growth
In this study, we developed a new approach to CNV analysis based on spectral decomposition CNV profile that separates focal CNV from broad CNVs
Our results yielded a list of interesting findings, such as focal gains around ESR1, focal loss around LSAMP, prognostic site at 3q26.2 and sub-telomeric losses. 29 of the 42 focal regions from our analysis overlapped with the focal regions reported by previous pancancer analysis using genomic identification of significant targets in cancer (GISTIC), which suggests that our results are in general agreements with previous analyses and offered new focal regions of interest, which demand further investigations

Summary

Introduction

Genomic regions with recurrent DNA copy number variations (CNVs) are generally believed to encode oncogenes and tumor suppressor genes (TSGs) that drive cancer growth. It remains a challenge to delineate the key cancer driver genes from the regions encoding a large number of genes. Solimini et al proposed a ‘gene island’ theory [16]: genes that stimulate/inhibit tumor growth may distribute very unevenly across the genome. Such genes are not classical oncogenes or

Methods

Results

Conclusion