Lynette M. Sholl, Allison D. Manning, Matthew D. Ducar, Neal I. Lindeman, Yonghui Jia, Elizabeth P. Garcia, Laura E. MacConaill, Ruchi Joshi, Priyanka Shivdasani, Lawrence P. Chung, Xin Gong, Bernard J. Fendler, and Phani K. Davineni
The recent understanding of cancer as a genomic disease, coupled with the advent of targeted therapeutic agents that more specifically and effectively treat a patient's cancer, has paved the way for genomic profiling of tumor samples for prognostic, diagnostic and predictive purposes. Next-generation sequencing (NGS) enables rapid, high-throughput simultaneous detection of multiple types of variations, in hundreds or thousands of genes. From a clinical perspective, accurate and high-quality variant identification is of great importance in identifying these genetic variants. Several factors challenge the reliability of NGS data, including GC content of target DNA, PCR amplification, and hybridization conditions. While bioinformatic solutions can overcome some of these challenges, improvements to the laboratory protocols that generate the sequence data are also important. KAPA HiFi polymerase is a new recombinant DNA polymerase reported to have better processivity, accuracy and improved coverage across GC- and AT-rich templates. In this study, we optimized KAPA library preparation protocol in OncoPanel, a targeted next generation sequencing Panel interrogating 300 cancer-related genes for single nucleotide variants (SNVs), small insertions/deletions (indels), copy number variants (CNV) and a limited number of structural variants. Briefly, indexed sequencing libraries were prepared from tumor specimens using either KAPA or TruSeq LT reagents, followed by pooling, solution-based hybrid capture, and massively parallel sequencing on Illumina HiSeq 2500. Sequencing QC metrics and variant calls were generated using an internally-developed analysis pipeline. We compared the sequencing metrics from an equal number of samples processed with KAPA or TruSeq protocol (n = 800). Use of KAPA protocol improved several quality metrics: (1) a significant increase in mean target coverage (from 133 to 209, increase of 36.1%; p value: 6.3E-61); (2)a significant reduction in the duplication rate (from 74% to 60%, reduction of 18.2%, p value: 5.2E-110) and (3) a significantly lower GC bias as measured by average slope of coverage vs GC content (from 3.0 to 2.6, 14.9% lower, p value: 0.007). Validation with orthogonally tested tumor specimen showed increased sensitivities for SNV/indels (from 95.2% to 100%), CNV (from 76.9% to 91.7%) and structural variants (from 85.7% to 87.5%). Overallsensitivity was 94.8% and 93.4% for KAPA and TruSeq LT, respectively, while 100% specificity was observed for both methods. In summary, the KAPA library protocol developed for OncoPanel significantly improved sequencing quality and sensitivities for calling of genomic alterations. Citation Format: Yonghui Jia, Allison D. Manning, Ruchi A. Joshi, Bernard J. Fendler, Priyanka Shivdasani, Lawrence P. Chung, Phani K. Davineni, Xin Gong, Matthew D. Ducar, Lynette M. Sholl, Neal I. Lindeman, Laura E. Macconaill, Elizabeth P. Garcia. Reducing GC bias and increasing complexity: Clinical implementation and validation of KAPA library preparation protocol for Oncopanel, a targeted next generation sequencing panel. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 4257. doi:10.1158/1538-7445.AM2015-4257