| Popis: |
Here, we describe analyses towards the first tumor/normal benchmark from the Genome in a Bottle (GIAB) consortium using new cell lines from a broadly-consented pancreatic ductal adenocarcinoma patient. This effort builds on the widely-used germline variant benchmarks for seven normal DNA reference materials. GIAB benchmarks are designed to reliably identify false positive and false negative variant calls, and are formed by combining multiple sequencing technologies and variant callers. To combine technologies and variant callers, we first understand their strengths and weaknesses for different variant types and repetitive regions. In this initial work on the new tumor/normal cell line, we have analyzed sequencing from HiC for copy number and structural variant analysis and 100to150x PCR-free Illumina WGS from normal and tumor cells for small variants. Preliminary HiC results identified substantial aneuploidy common in pancreatic tumors, with ~15 large inversions and translocations and 16 chromosomes with extensive loss of heterozygosity due to missing >30% of one copy. The tumor contains the common G12V mutation in KRAS, and interestingly the ~2Mbp region containing this mutation is likely triplicated. About 300 (76% of) SNVs and 10,000 (69% of) small indels agree between mutect2 and strelka2 with default filtering. After excluding difficult regions, 95% of SNVs and 94% of small indels agree, with remaining disagreements caused by systematic sequencing errors, alignment errors, low frequency, and proximity to germline variants. Examining discordant SNVs indifficult regions, most are erroneous somatic variants inregions with loss of heterozygosity that also have mapping errors due to segmental duplications, sequences missing from GRCh38, and/or tandem repeats. After excluding difficult regions, we also found that somatic SNVs fell in 4 primary classes: ~5100 in most cells in diploid regions, ~2100 in most cells in haploid regions, ~760 in only some cells in diploid regions, and ~450 in only some cells in haploid regions. GIAB plans to perform PacBioHiFi and ultra long ONT sequencing to generate phased assemblies for the normal and tumor, and is forming an open working group to develop the first authoritative benchmarks from publicly available, broadly-consented tumor-normal cell lines. Citation Format: Gail Rosen, Justin Wagner, Andrew Liss, Justin Zook. Sequencing a new broadly-consented tumor/normal cell line for a Genome in a Bottle Benchmark [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 876. |
