A systematic comparison of whole genome sequencing and targeted panel sequencing for precision oncology.

person Ji-Hyung Park Graduate School of Medical Science and Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea info_outline Ji-Hyung Park, Erin Connolly-Strong, Young Tae Kim, Young Seok Ju

Authors person Ji-Hyung Park Graduate School of Medical Science and Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea info_outline Ji-Hyung Park, Erin Connolly-Strong, Young Tae Kim, Young Seok Ju Organizations Graduate School of Medical Science and Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea, Genome Insight, San Diego, CA, Seoul National University Cancer Research Institute, Department of Thoracic and Cardiovascular Surgery, Seoul National University Hospital, Seoul, South Korea, Genome Insight Inc., San Diego, CA Abstract Disclosures Research Funding No funding received None. Background: The analysis of the cancer genome provides clinically meaningful information such as actionable cancer driver mutations. Whole genome sequencing (WGS) provides a complete mutational landscape at an affordable cost thereby having increase clinical utility compared to the usual approach of targeted panel sequencing (TPS). To understand the capabilities of WGS in oncology, we conducted a study in lung adenocarcinomas. Methods: We produced WGS (surgically resected tumor tissues and matched blood; mean coverage 42x) and TPS (tumor only; SNUH panel encompassing 75 cancer genes; mean coverage 904x) through the Illumina sequencing platform and conducted comprehensive bioinformatic analyses. Oncogenic driver events, including single base substitution (SBS), insertion or deletion (INDEL), complex deletion-insertion, and fusion oncogene, were investigated in both datasets. Lastly, we compared the pattern-based mutational features, such as tumor mutational burden (TMB), mutational signatures, and copy number changes which is associated with genomic instability. Results: In 60 lung adenocarcinomas, the TPS method identified 120 driver mutations from the 75 targeted genes. Of these, 114 mutations were also detected by the WGS method, showing 95.0% sensitivity of WGS in detection of core driver mutations. In uncovered genomic regions by TPS, WGS captured 9 additional driver mutations (such as loss-of-function mutations in PIK3R2, ARID2, and APC) and 37 copy number variations (such as amplification of LMO and SMO, deletion of PIK3R1 and APC) in the cancer genes. TMB calculation is straightforward in WGS however, sophisticated considerations were necessary in TPS for TMB, to remove false positives and germline mutations. TPS showed a good capability for identifying hypermutator tumor samples, but a quantitative estimation was not always feasible due to the substantial background fluctuation, particularly for INDELs. Mutational signatures are unique combinations of mutation types in a given sample, which can give important clues to the mechanism underlying mutations. TPS showed adequate mutational signatures (cosine similarity > 0.7), but was not constantly satisfactory in tumors with high TMB ( > 20,000 genome wide SBS and INDELs). In TPS, approximately 30% of high TMB tumors represented marginal cosine similarity (0.5-0.7), and tobacco smoking related signature (SBS 4) found in WGS was captured only in 50% of the associated cases (9 out of 18). For DNA copy number changes and loss-of-heterozygosity analyses, WGS showed a much higher resolution. Conclusions: Our study demonstrates that WGS has comparable capabilities for the detection of driver mutations in core oncogenes and tumor suppressor genes. WGS was superior for pattern-based features, such as TMB, mutational signatures, and copy number changes. WGS is a medically necessary more comprehensive approach to precision oncology.