Abstract 2487: CONSERTING: an accurate method for detecting focal and gross somatic copy number alterations in cancer genome by next generation sequencing

Abstract

Abstract The Identification of somatically acquired DNA copy number alterations (CNA) is extremely important for finding causal genes in cancer. Although next-generation sequencing (NGS) can assay a genome at base-pair resolution, accurate identification of CNA from NGS data has been hindered by sequencing artifacts, coverage bias and paralogous genomic duplication. As a result, existing algorithms tend to generate many CNA segments that are not corroborated by an alternative assay such as SNP array, suggesting high false discovery rate. To improve CNA analysis for NGS data, we developed Copy Number Segmentation by Regression Tree in Next Generation Sequencing (CONSERTING), a novel copy number detection method that integrates read depth analysis with structural variation (SV) detection and adjustment for sequencing artifacts. Analysis of simulated data shown that CONSERTING outperforms existing methods by achieving high sensitivity of focal copy number detection (CONSERTING: 0.80, CNV-Seq: 0.26, SegSeq: 0.19, FREEC: 0.02) at a low false discovery rate (CONSERTING: 0.09, CNV-Seq: 0.19, SegSeq: 0.46, FREEC: 0.07). When applied to matching tumor/normal samples of hematopoietic malignancies and solid tumors, CONSERTING showed a high concordance with SNP array, with a mean of 18 uncorroborated CNA segments compared with 104 by CNVSeq, 182 by SegSeq and 63 by FREEC. In addition, CONSERTING is able to discover bona fide focal CNAs in genes of biological importance that were not detectable by SNP array, such as a 36 kb homozygous deletion in SH2B3 and a 20 kb deletion in CDK6. Finally, CONSERTING is able to find CNAs and SVs present in tumor subclone that are missed by SNP array or SV analysis of NGS data. The novel CNAs and SVs reported here were experimentally validated, demonstrating that the high sensitivity and accuracy of CONSERTING enable discovery of key genetic lesions in cancer. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 2487. doi:1538-7445.AM2012-2487