Somatic Copy Number Variations Research Articles

Mucinous adenocarcinoma is a pharmacogenomically distinct subtype of colorectal cancer.

Mucinous colorectal cancer is a unique histological subtype that is known to respond poorly to cytotoxic chemotherapy and radiotherapy. There are a number of genes known to be associated with resistance to 5-fluorouracil (5-FU), oxaliplatin, and irinotecan. The aim of this study was to compare the somatic mutation frequency and copy number variation (CNV) in these genes between mucinous and non-mucinous colorectal cancer. A systematic search of PubMed was performed to identify papers investigating drug resistance in colorectal cancer. From this review, a list of 26 drug-resistance-associated genes was compiled. Using patient data from The Cancer Genome Atlas (TCGA), the somatic mutation rate and CNV was compared between patients with mucinous and non-mucinous colorectal cancer. Statistical analysis was carried out using GraphPad PRISM® version 5.00. Data were available on 531 patients (464 non-mucinous, 67 mucinous). A statistically significant difference in the somatic mutation rate between the two cohorts was identified in the TYMP (p = 0.0179), ATP7B (p = 0.0465), SRPK1 (p = 0.0135), ABCB1 (p = 0.0423), and ABCG2 (p = 0.0102) genes. A statistically significant difference in CNV was identified between the two cohorts in the GSTP1 (p = 0.0405), CCS (p = 0.0063), and TOP1 (p = 0.0048) genes. Differences in somatic mutation rate and CNV in genes associated with resistance to 5-FU, oxaliplatin, and irinotecan may partly account for the pattern of resistance observed in mucinous colorectal cancers. These genetic alterations may prove useful when deciding on a personalized approach to chemotherapy and may also represent potential therapeutic targets going forward.

Therapeutic Implication of Genomic Landscape of Adult Metastatic Sarcoma.

This study investigated therapeutic potential of integrated genome and transcriptome profiling of metastatic sarcoma, a rare but extremely heterogeneous group of aggressive mesenchymal malignancies with few systemic therapeutic options. Forty-three adult patients with advanced or metastatic non-GI stromal tumor sarcomas of various histology subtypes who were enrolled in the Personalized OncoGenomics program at BC Cancer were included in this study. Fresh tumor tissues along with blood samples underwent whole-genome and transcriptome sequencing. The most frequent genomic alterations in this cohort are large-scale structural variation and somatic copy number variation. Outlier RNA expression as well as somatic copy number variations, structural variations, and small mutations together suggest the presence of one or more potential therapeutic targets in the majority of patients in our cohort. Point mutations or deletions in known targetable cancer genes are rare; for example, tuberous sclerosis complex 2 provides a rationale for targeting the mammalian target of rapamycin pathway, resulting in a few patients with exceptional clinical benefit from everolimus. In addition, we observed recurrent 17p11-12 amplifications, which seem to be a sarcoma-specific event. This may suggest that this region harbors an oncogene(s) that is significant for sarcoma tumorigenesis. Furthermore, some sarcoma tumors carrying a distinct mutational signature suggestive of homologous recombination deficiency seem to demonstrate sensitivity to double-strand DNA-damaging agents. Integrated large-scale genomic analysis may provide insights into potential therapeutic targets as well as novel biologic features of metastatic sarcomas that could fuel future experimental and clinical research and help design biomarker-driven basket clinical trials for novel therapeutic strategies.

Investigation of somatic CNVs in brains of synucleinopathy cases using targeted SNCA analysis and single cell sequencing

Synucleinopathies are mostly sporadic neurodegenerative disorders of partly unexplained aetiology, and include Parkinson’s disease (PD) and multiple system atrophy (MSA). We have further investigated our recent finding of somatic SNCA (α-synuclein) copy number variants (CNVs, specifically gains) in synucleinopathies, using Fluorescent in-situ Hybridisation for SNCA, and single-cell whole genome sequencing for the first time in a synucleinopathy. In the cingulate cortex, mosaicism levels for SNCA gains were higher in MSA and PD than controls in neurons (> 2% in both diseases), and for MSA also in non-neurons. In MSA substantia nigra (SN), we noted SNCA gains in > 3% of dopaminergic (DA) neurons (identified by neuromelanin) and neuromelanin-negative cells, including olig2-positive oligodendroglia. Cells with CNVs were more likely to have α-synuclein inclusions, in a pattern corresponding to cell categories mostly relevant to the disease: DA neurons in Lewy-body cases, and other cells in the striatonigral degeneration-dominant MSA variant (MSA-SND). Higher mosaicism levels in SN neuromelanin-negative cells may correlate with younger onset in typical MSA-SND, and in cingulate neurons with younger death in PD. Larger sample sizes will, however, be required to confirm these putative findings. We obtained genome-wide somatic CNV profiles from 169 cells from the substantia nigra of two MSA cases, and pons and putamen of one. These showed somatic CNVs in ~ 30% of cells, with clonality and origins in segmental duplications for some. CNVs had distinct profiles based on cell type, with neurons having a mix of gains and losses, and other cells having almost exclusively gains, although control data sets will be required to determine possible disease relevance. We propose that somatic SNCA CNVs may contribute to the aetiology and pathogenesis of synucleinopathies, and that genome-wide somatic CNVs in MSA brain merit further study.

Identification of Circulating Genomic and Metabolic Biomarkers in Intrahepatic Cholangiocarcinoma.

Intrahepatic cholangiocarcinoma (ICC) is an aggressive cancer arising from the bile ducts with a need for earlier diagnosis and a greater range of treatment options. KRAS/NRAS mutations are common in ICC tumours and 6–32% of patients also have isocitrate dehydrogenase 1 and 2 (IDH1 and IDH2) gene mutations associated with metabolic changes. This feasibility study investigated sequencing circulating tumour DNA (ctDNA) combined with metabolite profiling of plasma as a method for biomarker discovery in ICC patients. Plasma was collected from four ICC patients receiving radio-embolisation and healthy controls at multiple time points. ctDNA was sequenced using Ampliseq cancer hotspot panel-v2 on Ion Torrent PGM for single nucleotide variants (SNV) detection and with Illumina whole genome sequencing for copy number variants (CNV) and further targeted examination for SNVs. Untargeted analysis of metabolites from patient and control plasma was performed using liquid chromatography coupled with high-resolution tandem mass spectrometry (LC-MS/MS). Metabolite identification was performed using multi-parameter comparisons with analysis of authentic standards, and univariate statistical analysis was performed to identify differences in metabolite abundance between patient and control samples. Recurrent somatic SNVs and CNVs were identified in ctDNA from three out of four patients that included both NRAS and IDH1 mutations linked to ICC. Plasma metabolite analysis revealed biomarker metabolites associated with ICC and in particular 2-hydroxyglutarate (2-HG) levels were elevated in both samples from the only patient showing a variant allele in IDH1. A reduction in the number of CNVs was observed with treatment. This study demonstrates that ctDNA and metabolite levels can be identified and correlated in ICC patient blood samples and differentiated from healthy controls. We conclude that combining genomic and metabolic analysis of plasma offers an effective approach to biomarker identification with potential for disease stratification and early detection studies.Intrahepatic cholangiocarcinoma (ICC) is an aggressive cancer arising from the bile ducts with a need for earlier diagnosis and a greater range of treatment options. KRAS/NRAS mutations are common in ICC tumours and 6–32% of patients also have isocitrate dehydrogenase 1 and 2 (IDH1 and IDH2) gene mutations associated with metabolic changes. This feasibility study investigated sequencing circulating tumour DNA (ctDNA) combined with metabolite profiling of plasma as a method for biomarker discovery in ICC patients. Plasma was collected from four ICC patients receiving radio-embolisation and healthy controls at multiple time points. ctDNA was sequenced using Ampliseq cancer hotspot panel-v2 on Ion Torrent PGM for single nucleotide variants (SNV) detection and with Illumina whole genome sequencing for copy number variants (CNV) and further targeted examination for SNVs. Untargeted analysis of metabolites from patient and control plasma was performed using liquid chromatography coupled with high-resolution tandem mass spectrometry (LC-MS/MS). Metabolite identification was performed using multi-parameter comparisons with analysis of authentic standards, and univariate statistical analysis was performed to identify differences in metabolite abundance between patient and control samples. Recurrent somatic SNVs and CNVs were identified in ctDNA from three out of four patients that included both NRAS and IDH1 mutations linked to ICC. Plasma metabolite analysis revealed biomarker metabolites associated with ICC and in particular 2-hydroxyglutarate (2-HG) levels were elevated in both samples from the only patient showing a variant allele in IDH1. A reduction in the number of CNVs was observed with treatment. This study demonstrates that ctDNA and metabolite levels can be identified and correlated in ICC patient blood samples and differentiated from healthy controls. We conclude that combining genomic and metabolic analysis of plasma offers an effective approach to biomarker identification with potential for disease stratification and early detection studies.

Clinical Implications of Copy Number Variant Detection from Panel-Based Next-Generation Sequencing Data in Myelodysplastic Syndrome

Some of the recurrently occurring somatic mutations are known to be diagnostic or prognostic in myelodysplastic syndrome (MDS). Targeted gene capture and next-generation sequencing (NGS) has rapidly become routine clinical tools to detect the somatic mutations in patients with MDS. Copy number variants (CNVs) may have additional clinical significance in MDS. Chromosomal microarray analysis is a standard technique for genome-wide CNV detection, but multiple testing strategies require high costs and time. Recent advancements in NGS technologies have developed more cost-effective and rapid methods to allow simultaneous identification of targeted CNVs as well as somatic mutations using the same panel-based NGS data. In this study, we investigated whether the detection of CNVs using the targeted NGS data provided an additional value other than the clinical implications of somatic mutations. We performed targeted deep sequencing analysis on bone marrow samples obtained from 266 patients with MDS using an MDS panel targeting 28 well-known MDS-related genes (NRAS, DNMT3A, SF3B1, IDH1, TET2, NPM1, LAMB4, EZH2, JAK2, CBL, ETV6, KRAS, FLT3, IDH2, PRPF8, TP53, NF1, SRSF2, SETBP1, DNMT1, ASXL1, RUNX1, U2AF1, ZRSR2, ATRX, STAG2, MMP8, and ARID2). Sequencing libraries were generated using the AmpliSeq Library Kit 2.0 (Life Technologies, Carlsbad, CA) and the MDS panel was then sequenced using the Ion Torrent Proton system (Life Technologies) according to the manufacturer's instructions. The multiscale reference module and Rank Segmentation statistical algorithm in NEXUS software v9.0 (Biodiscovery) were used to define the CNVs for each sample. Overall survival (OS) and acute myeloid leukemia (AML)-free survival (AFS) were estimated from the date of MDS diagnosis to death or AML progression using the Kaplan-Meier method, and the differences in survival were compared using the log-rank test (for univariate analysis) and the Cox proportional hazards model (for multivariate analysis). Overall, 215 patients (80.8%) carried at least one somatic mutations, and 67 (25.2%) had one or more CNVs. The number of mutated genes per patient ranged from 0 to 6, and the number of genes with CNVs per patient ranged from 0 to 10. Of 51 patients who did not have somatic mutations, 12 (23.5%) had the targeted CNVs. The mutated genes in more than 10% of patients were 8: U2AF1 (21.8%), TET2 (17.7%), ASXL1 (13.5%), TP53 (13.2%), SETBP1 (12.8%), NF1 (10.9%), SF3B1 (10.5%), and RUNX1 (10.5%). The genes with CNVs detected in 10 or more patients were 5: EZH2 (loss in 7q, 6.8%), KRAS (gain and loss in 12p, 5.3%), ASXL1 (gain and loss in 20q, 4.5%), LAMB4 (loss in 7q, 3.8%), and RUNX1 (gain and loss in 21q, 3.8%). Interestingly, all five patients with TP53 deletion exhibited TP53 mutation as well, suggesting a bi-allelic alteration (mutation + copy loss). The higher number of genes with CNVs per patient were significantly associated with inferior OS (P <0.001) and AFS (P=0.007) (Figure). Considering clinical variables and somatic mutations only, multivariate analysis demonstrated that sex, IPSS-R, NRAS mutation, and TP53 mutation were independent prognostic factors for OS, and sex, IPSS-R, NRAS mutation, DNMT3A mutation, and complex karyotype/TP53 mutation were independent prognostic factors for AFS. When we consider clinical variables and both somatic mutations and CNVs, genetic alterations in TET2, LAMB4, U2AF1, and CBL showed additional significant impact on OS or AFS (Table). Our study suggests that simultaneous detection of targeted CNVs as well as somatic mutations using the same panel-based NGS data add clinically useful information on the prognosis of MDS patients. Disclosures No relevant conflicts of interest to declare.

Tumor Mutational Burden and Genomic Alterations in Chinese Small Cell Lung Cancer Measured by Whole-Exome Sequencing.

Purpose Studies on genetic alterations of the heterogenous small cell lung cancer (SCLC) are rare. We carried out the present study to clarify the genomic alterations and TMB levels of Chinese SCLC patients by whole-exome sequencing. Materials and Methods Whole-exome sequencing by next-generation sequencing technique was implemented on twenty SCLC samples. Significant somatic mutations and copy number variations were screened, followed by comparison with the data extracted from COSMIC. Besides, altered signaling pathways were examined in order to figure out actionable targets. Results A total of 8,062 nonsynonymous mutations were defined. The number of mutations for each case ranged from 98 to 864. As for base substitutions, a total of 15,817 substitutions were detected with C > A conversion which was correlated to smoking occupying 25.57%. The TMB values ranged from 2.51/Mb to 22.1/Mb with a median value of 9.95/Mb. RB1 was the most frequently mutated gene altered in 18 (90%) cases, followed by TP53 altered in 17 (85%) cases. Other commonly changed genes were PTEN, and RBL1, with frequencies of 55% and 50%, respectively. SOX2 significantly amplified in 6 (30%) cases and MYCN amplified in 1 (5%) patient. Notch signaling pathway and PI3K/AKT/mTOR signaling pathway were universally and significantly changed. Major genomic alterations were in consistency with data from COSMIC, but frequencies of less common mutations were different. Conclusion TP53 and RB1 inactivations were universally detected in SCLC. The Notch and PI3K/AKT/mTOR signaling pathways were both significantly altered, implying potential actionable targets.

Copy number variations in critical cell-signalling genes with potential targeted therapeutic application

Abstract Background Brain tumours are among “hard-to treat” world malignancy with the median survival time of 15 months for the most aggressive form - Glioblastoma multiforme (GBM). It is extremely complex and heterogeneous brain cancer, characterized by rapid progression, infiltration and therapeutic resistance. Currently, many trials for target therapy are focused on understanding the molecular heterogeneity of GBM and designing molecules that inhibit most common genetic alterations. Methods In the present study, somatic copy number variations (CNVs) in critical signalling genes suitable for future target therapy are examined among Bulgarian patients. Fresh brain biopsies were collected from patients with primary and secondary GBM. Genomic DNA was isolated and CNVs were analysed by Multiplex ligation – dependent probe amplification (MLPA). Results CDKN2A deletion is the most frequent genetic alteration (78%) among Bulgarian patients with GBM. The loss of this gene is a progression-associated genetic marker. Nearly 60% of GBM patients showed EGFR amplification. It is a prognostic and predictive marker and some patients with this aberration are suggested to respond better to EGFR inhibitors in combination therapy. PTEN deletion is another common event with the incidence rate of 33% and its loss corelates with the choice of mTOR and MEK inhibitors, as well as, the resistance to some EGFR inhibitors. TP53 deletion was found in only 11% of the analysed patients which is a positive marker and the presence of wild-type TP53 is targeted by a MDM2 inhibitors. In addition, MDM2 amplification was observed in one patient and this event also serve as a potential target by MDM2 inhibitors.Amplification in PDGFRA is observed in one patient. Depending on the status of other receptor tyrosine kinases, it could be targeted by PDGFR inhibitor and is also suggested to be a poor prognostic marker. Conclusions Personalized treatment, customized for an individual patient genetics, has the potential to improve the therapy response. CNVs data could be useful for designing more effective personalized therapy. This is a small study on CNVs with potentially targeted therapeutic application in GBM for Bulgaria and recruitment of more patients is in progress. Legal entity responsible for the study The authors. Funding Medical University Sofia, Grant number D-127/2019. Disclosure All authors have declared no conflicts of interest.

Identification and meta-analysis of copy number variation-driven circadian clock genes for colorectal cancer.

Both copy number variation (CNV) and circadian clock genes play a critical role in the etiology and pathogenesis of colorectal cancer (CRC); however, a comprehensive analysis of CNV-driven circadian clock genes is urgently required. The present study aimed to investigate the systematic associations between somatic cell CNVs and circadian clock gene expression in patients with CRC. Using somatic CNV, legacy clinical information and gene expression data from The Cancer Genome Atlas, 295 genes that were significantly differentially expressed and with significantly different CNV were obtained, and the expression of the genes, among which 15 were circadian clock genes, was significantly associated with CNV. Further analysis revealed that aryl hydrocarbon receptor nuclear translocator-like 2 (ARNTL2) expression and CNV in these circadian clock genes were significantly associated with survival time in patients with CRC, and the expression of ARNTL2 was also significantly associated with the pathological stage of CRC. Gene set enrichment analysis found that ARNTL2 is enriched for gene sets associated with CRC pathogenesis such as the p53 signaling pathway. These results suggest that ARNTL2 may be a promising prognostic biomarker for patients with CRC, and that circadian clock genes play an important role in CRC through CNV.

Targeted sequencing of cancer-related genes in nasopharyngeal carcinoma identifies mutations in the TGF-β pathway.

Approximately, 25% of nasopharyngeal carcinoma (NPC) patients develop recurrent disease. NPC may involve relatively few genomic alterations compared to other cancers due to its association with Epstein‐Barr virus (EBV). We envisioned that in‐depth sequencing of tumor tissues might provide new insights into the genetic alterations of this cancer. Thirty‐three NPC paired tumor/adjacent normal or peripheral blood mononuclear cell samples were deep‐sequenced (>1000×) with respect to a panel of 409 cancer‐related genes. Newly identified mutations and its correlation with clinical outcomes were evaluated. Profiling of somatic mutations and copy number variations (CNV) in NPC tumors identified alterations in RTK/RAS/PI3K, NOTCH, DNA repair, chromatin remodeling, cell cycle, NF‐κB, and TGF‐β pathways. In addition, patients harbored CNV among 409 cancer‐related genes and missense mutations in TGF‐β/SMAD signaling were associated with poor overall survival and poor recurrence‐free survival, respectively. The CNV events were correlated with plasma EBV copies, while mutations in TGFBR2 and SMAD4 abrogate SMAD‐dependent TGF‐β signaling. Functional analysis revealed that the new TGFBR2 kinase domain mutants were incapable of transducing the signal, leading to failure of phosphorylation of SMAD2/3 and activation of downstream TGF‐β‐mediated cell growth arrest. This study provides evidence supporting CNV and dysregulated TGF‐β signaling contributes to exacerbating the NPC pathogenesis.

Pan-Cancer and Single-Cell Modeling of Genomic Alterations Through Gene Expression.

Cancer is a disease often characterized by the presence of multiple genomic alterations, which trigger altered transcriptional patterns and gene expression, which in turn sustain the processes of tumorigenesis, tumor progression, and tumor maintenance. The links between genomic alterations and gene expression profiles can be utilized as the basis to build specific molecular tumorigenic relationships. In this study, we perform pan-cancer predictions of the presence of single somatic mutations and copy number variations using machine learning approaches on gene expression profiles. We show that gene expression can be used to predict genomic alterations in every tumor type, where some alterations are more predictable than others. We propose gene aggregation as a tool to improve the accuracy of alteration prediction models from gene expression profiles. Ultimately, we show how this principle can be beneficial in intrinsically noisy datasets, such as those based on single-cell sequencing.

Abstract 1417: Multiple gastric cancer and predisposing genes

Abstract BACKGROUND: Multiple gastric cancer (MGC) is one special type of gastric cancer with more than two different tumors at various locations of stomach. However, the clonal relationship and carcinogenesis of MGC are still remain unclear. Therefore, we investigated the clonal relationship and role of germline mutations playing in the carcinogenesis of MGC. METHODS: We collected 16 multiple gastric cancer patients who underwent subtotal and total gastrectomy at Peking Cancer Hospital form January 2016 to December 2017. Thirty-three tumor samples and sixteen normal gastric tissue or blood samples were obtained for experiment. We also conducted analysis for 208 gastric cancer and 49 esophagogastric junction cancer (GC-EGJ) tumors from TCGA. DNA extraction from our samples was conducted for whole exome sequencing. We analyzed nonsynonymous mutations and somatic copy number variations (CNVs) based on exome data. Driver mutations, germline and cancer predisposing genes, significantly mutated genes (SMGs) were also analyzed. RESULTS: Tumor mutation burden (TMB) was not statically significant within database and our data in GCs-EGJ groups (P=0.0591). And the TMB level was also similar between two databases in GCs groups (P=0.3113). The mutation spectrum and mutation signatures also showed uniform distributions in GCs and GCs-EGJ groups within our data and TCGA database. Among sixteen patients, four were identified as monoclonal, in which 11, 10, 26 and 6 somatic mutations were shared within different tumors of P7, P8, P9 and P16 respectively. However, no common mutation between different tumors of same patient was found from the other 12 patients in spite of thousands of somatic mutations in P2. After identifying predisposing genes, we found that germline MSH2 and NCOR2 mutations were significantly dominant in 8/12 and 10/12 of genetic MGCs patients. And all patients were identified MSH2 mutations in cancer samples of those same patients. Interestingly, the NCOR2 mutations were not detected significantly with non-frameshift INDELs in only three patients. Taking genetic MGCs as a whole, we identified that TP53 were significantly mutated in 14 of 25 tumor samples. And MSH2 mutations were found in all patients. The correlation analysis between clinical information and mutation signatures showed that the signature B was associated with the level of CA72.4. CONCLUSIONS: Whole exome sequencing analyses are suggestive of monoclonal and polyclonal origin of MGC, which may promote the classification MGC into genetic and metastatic MGC. For genetic MGC patients, germline MSH2 mutations may contribute to the carcinogenesis of them, may thus giving rise to the consideration of more radical surgery and PD-1/PD-L1 therapy. Citation Format: Anqiang Wang, Zhongwu Li, Xin Ji, Tao Fu, Xiaojiang Wu, Ji Zhang, Zhaode Bu, Jiafu Ji. Multiple gastric cancer and predisposing genes [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 1417.

Abstract 2471: Pan-cancer catalog of Differentially Methylated Regions by Rocker-meth, a new computational method

Abstract DNA methylation (DNAm) is an essential player of gene regulation and therefore one of the most studied epigenetic mechanisms. Alterations of DNAm have been associated with a wide range of diseases, including cancer. Most studies on DNAm focused on the characterization of methylation sites, despite the recognition that DNAm changes spanning entire genomic regions are common in cancer tissues with respect to benign cells, referred to as Differentially Methylated Regions (DMRs). Here we report on a new computational method to identify DMRs both from array and Bisulfite Sequencing data. Rocker-meth consists of four main modules: 1) computation of Area Under the Curve (AUC) values from Receiver operating characteristic (ROC) Curve analysis of methylation levels (i.e., beta values) in tumor versus normal samples; 2) segmentation of AUC values by a tailored Heterogeneous Shifting Level Model (HSLM); 3) estimation of the statistical significance of AUC segments by Wilcoxon-Mann-Whitney test on beta values of tumor versus normal samples; 4) identification of sample specific DMRs by Z-score statistics. Tests on synthetic DNAm profiles demonstrated that Rocker-meth outperforms current state-of-the-art tools. Specifically, it showed markedly higher precision and recall in the presence of the lowest signal-to-noise ratios that mimic more than 40% of real human samples data. By exploiting well known features of DNAm in human cancers, we also demonstrated that Rocker-meth outperforms the most used array-based DMR detection tools in real data. We then drafted the first pan-cancer catalog of DMRs based on the characterization of more than 6,000 human samples DNAm data of 14 tumor types from TCGA and other recent studies and studied the characteristics of hyper and hypo DMRs in terms of structure, genomic locations and functional role. Pan-cancer integrative analysis of DMRs and gene expression recapitulated well-known associations, validating the accuracy of our method. Finally, we demonstrated the reliability of Rocker-meth to perform sample-wise analysis. This was possible thanks to the intrinsic characteristic of DMRs to be extremely recurrent within a given tumor type, supporting our previous observation for methylation sites and in contrast to genomic alterations such as Single Nucleotide Variants (SNVs) and Somatic Copy Number Variants (SCNAs). Strikingly, DMRs detected by Rocker-meth were informative in segregating prostate and breast adenocarcinoma based on their molecular subtypes, confirming the results from DNAm site analyses. The pan-cancer catalog of DMRs with sample level characterization could be used to investigate across cancer mechanisms of tumorigenesis or to query synthetic lethal events. Citation Format: Matteo Benelli, Alberto Magi, Dario Romagnoli, Gian Marco Franceschini, Chiara Biagioni, Ilenia Migliaccio, Luca Malorni, Angelo Di Leo, Francesca Demichelis. Pan-cancer catalog of Differentially Methylated Regions by Rocker-meth, a new computational method [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 2471.

Abstract 2533: The genomic and transcriptomic landscape of left versus right sided breast cancer in 410 cases

Abstract Background: Studies have shown that unilateral breast cancer is more common on the left (L) than right (R) and that the L breast is often larger than the R. No clear explanation for this finding has been described in the literature. Hypotheses range from complex genomic effects to environmental, hormonal, and anatomical causes. Methods: We performed a retrospective analysis of data from the NantHealth database of 410 commercial cases where WGS (35x germline, 75x somatic) or WES (∼150x). Primary and metastatic breast tumors were included in the analysis. Of these 410 cases, 39.3% were R, 38.5% were L, and 22% were unspecified. Bases adjacent to variant sites were used to determine genomic context (i.e. mutation signatures) as previously reported. Additionally, paired deep whole transcriptomic sequencing (RNA-Seq; ∼200x106 reads per tumor) was performed on 376 of the 410 cases. Nant Breast Cancer Intrinsic Subtypes sorting based on RNAseq assay classified breast tumors into 4 subtypes (LumA/B, Basal-like, and HER2-enriched). Results: Luminal A subtype was significantly overrepresented in R v L (OR=0.53, p=0.015). Similarly, R-side cases were significantly enriched for estrogen response gene expression by GSEA. However, only slight differential expression of ESR1 and MKI67 was observed specifically in individuals between 40 and 50 years old. Germline variants in the following genes were associated with sidedness: MEN1, KIT, JAK2, TET2, EP300, MET, FLCN, and KMT2D, and somatic variants within the following genes were associated with sidedness: CACNA1E, AGRN, RIMS2, TTYH3, KRTAP10-6, and DDX60L. However, significance of the variant frequency differences was lost after multiple-hypothesis correction. No significant differences between L and R were noted in tumor stage, HER2, ER, PR, somatic CNV, and mutation signatures, including signature 3 (BRCAness). Conclusions: Despite performing WGS/WES a paucity of differentially mutated genes were seen between left and right sided breast tumors. RNAseq analysis reveals that right sided breast cancers may be more likely to involve ESR1-signaling and be Luminal A subtype. Right-sided ESR1-signaling may be especially relevant in perimenopausal cases, however the current cohort is underpowered to achieve significance in this analysis. Citation Format: Yulia Newton, Christopher Szeto, Charles Vaske, Leia Reddy, Sandeep Reddy. The genomic and transcriptomic landscape of left versus right sided breast cancer in 410 cases [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 2533.

Abstract 4233: Improving the diagnostic yield of a 124 gene pediatric solid tumor panel through somatic copy number variation analysis

Abstract In cancer genomes, both large-scale CNVs (>100 kb) spanning chromosome arms, and smaller CNVs limited to a few genes are prevalent. While large CNVs are readily detected with existing methodologies (e.g. SNP array), gene-level CNVs require a higher resolution not routinely available in current tools, including next generation sequencing (NGS)-based clinical cancer mutation panels. We sought to leverage NGS data generated by one of these panels: the TCH Pediatric Solid Tumor panel (124 cancer genes), and built a clinical-grade analytic pipeline for detection of somatic CNVs. After reviewing literature, CNVkit was selected for its ability to perform unmatched (requiring no matched normal specimen) CNV analysis, customize segmentation algorithms, and provide superior visualization. CNVkit performs circular binary segmentation on the log2 difference of binned read depths (on- and off-target) from tumor and pooled-normal blood samples to identify CNVs. Although the panel captures only 124 genes (~1Mbp) at > 300X coverage, the low background coverage of 0.05X (off-target) due to imperfect hybridization capture allows us to detect chromosome-arm level changes. As a pilot study, we optimized CNVkit to detect gene-level and chromosome arm-level CNVs in a reference aneuploid colon cancer cell line (HT-29) characterized by aCGH and observed high correlation (r = 0.89) between the aCGH and CNVkit fold change. Through in-silico and bench experiments we performed limit of detection analysis, by diluting different proportions of CNV positive tumor samples with normal sample, to set thresholds for calling amplification and losses. We show the ability of the pipeline to detect shallow and deep deletions if they are present in at least 40% and 80% of the sample sequenced, respectively. During validation, we used 24 pediatric cancer samples with diverse histology to confirm and detect numerous CNVs of clinical significance, including deletions of SMARCB1, PTEN, BRCA2, RB1, and ARID1B, and amplifications of MYCN, MYC, CCND1 and KRAS. As the panel densely tiles tumor suppressors, we are also able to infer apparent intragenic deletions in BRCA1, BRCA2, ATRX, RB1 and PTEN, highlighting the theoretical resolution of this tool for detecting intragenic events over other methods. We have also developed a Python based interactive CNV Viewer for assessing the copy number analysis results from NGS data. The browser like interface allows the user to zoom, link out to UCSC Genome Browser, hover for information, take high-quality screenshots, etc. Hence, the CNV pipeline we have developed will allow more comprehensive evaluation of the existing integrated DNA and RNA analysis pipeline, increasing the diagnostic yield of mutation panel testing for childhood cancer patients. Citation Format: Raghu Chandramohan, Jacquelyn Reuther, Ilavarasi Gandhi, Horatiu Voicu, Karla R. Alvarez, Kevin E. Fisher, Dolores H. Lopez-Terrada, Donald W. Parsons, Angshumoy Roy. Improving the diagnostic yield of a 124 gene pediatric solid tumor panel through somatic copy number variation analysis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 4233.

Abstract 3705: Identification of frequent HER2 activating mutations in canine primary pulmonary adenocarcinoma

Abstract Spontaneous primary canine lung cancers are aggressive malignancies that are increasingly common in dogs. They share clinicopathologic features with human lung cancers in never-smokers, but their genetic underpinnings are unknown. As never-smoker lung cancer incidence increases in humans, a need exists for improved biologic understanding and development of new models to fuel translational research. Here, we describe for the first time the detailed clinical and genetic features of primary canine lung cancers through case review and multi-platform sequencing of 90 primary canine lung tumors and cell lines. We performed multi-platform genomic profiling including whole exome sequencing (WES), amplicon panel-based sequencing of 250 regions of 51 cancer genes, and single nucleotide polymorphism (SNP) array-based analysis of tumors and matched constitutional DNA. We profiled 76 cPAC, 11 canine pulmonary adenosquamous carcinomas (cPASC) and three canine pulmonary squamous cell carcinomas (cPSCC). The median age at the time of diagnosis was 11 year old. The most common affected pure breed was the Labrador retriever (19%). Five cPACs assessed by WES displayed low mutation burden with a median of 64 somatic single nucleotide variants (SNVs), 19 somatic copy number variants (CNVs), and 1 structural variant (SV), and frequent C>T substitutions (80%) at NpCpG trinucleotides, a common mutation signature seen across human and canine cancers. Based on WES and amplicon sequencing, we discovered somatic, coding HER2 (ERRB2) point mutations in 37% of cPAC, but none in cPASC or cPSCC. Additional recurrently mutated genes in cPAC included CDKN2A/B (~40%), TP53 (~12%), K/HRAS (~7%), SMAD4 (~5%), and PTEN (~5%). In cPASC, KRAS and TP53 were the most frequently mutated genes (18% each). In cPSCC, no recurrently mutated genes were identified, but individual somatic coding mutations were found in BRAF and PTPN11. The majority (93%) of HER2 mutations were hotspot V659E, located on the transmembrane domain (TMD), and comparable to activating mutations at this same site in human cancer. We were able to detect this mutation not only in biopsied tumors, but also in the plasma of 33% (2/6) of dogs with localized V659E-positive tumors. Other somatic HER2 mutations identified were similarly located in the HER2 juxtamembrane domain and TMD including A664T and K676E. HER2 point mutations occurred in the absence of significant focal amplification (assessed by WES and SNP array) or overexpression (assessed by qRT-PCR and IHC). We also showed that HER2 mutation correlated with constitutive phosphorylation of Akt in cPAC cell lines. Furthermore, HER2 V659E-mutant lines displayed hypersensitivity to the HER2 kinase inhibitors neratinib (IC50, 23nM) and lapatinib (IC50, 168nM) relative to HER2 wild-type cell lines (IC50, >2500nM). These data bear implications for comparative understanding of HER2-mutant lung cancer across species. Citation Format: Gwendolen Lorch, Karthigayini Sivaprakasam, Victoria Zissman, Nieves Perdigones, Tania Contente-Cuomo, Alexandra Nazareno, Salvatore Facista, ShukMei Wong, Winnie Liang, Joseph M. Amann, Sara L. Sinicropi-Yao, Michael J. Koenig, Krista La Perle, Timothy G. Whitsett, Muhammed Murtaza, Jeffrey Trent, David P. Carbone, William P. Hendricks. Identification of frequent HER2 activating mutations in canine primary pulmonary adenocarcinoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 3705.

Abstract 3521: Critical Assessment of CNV Calling Using Next Generation Sequencing

Abstract Tumor-derived Copy Number Variations (CNVs) are among the most important genomic aberrations in cancers. Since a CNV amplification is often associated with oncogene activation, and a CNV deletion is frequently attributed to the gene inactivation, an accurate detection of CNVs plays a crucial role in cancer diagnosis and treatment. In this study, using the next generation sequencing, we systematically interrogated somatic CNVs in paired tumor-normal cell lines to identify factors affecting detection reproducibility and accuracy. Whole-genome and whole-exome sequencing were carried out at six sequencing centers followed by processing with six CNV callers to evaluate their reproducibility. Different types of samples with varying input amount and tumor purity were processed using multiple library construction protocols and read depths. We compared the consistency of reference replicates, the effect of experimental, technical, and biological confounding factors, and benchmarked our callers using cytogenetic array. Our evaluation was based on two approaches; bin-based and gene-based approach. The bin-based comparison used a sliding window of 10 Kb, while gene-based comparison used segments at the gene level. Our evaluations indicate variations among CNV calls are mainly driven by callers and less by the sequencing centers. In addition to caller effect, PCA analyses also show site-specific effect, largely pairwise distances on the PC plot for pairs between different centers than for those within the same centers. Among confounding factors, namely FFPE, library concentration, and tumor purity with several sequencing depths, we report that tumor purity has a dominant and major effect in caller’s performance. PCA analysis shows low tumor purity (<= 50%) has a large effect on CNV calls which are evidenced by low purity samples being away from reference samples on PC plots for all callers. Tumor purity <= 50% leads to severe missing calls and lower copy numbers in amplified regions while sequencing depth seems to affect the outcome. We illustrated that most of the other confounding factors such as WES, FFPE, and LBP are caller-specific. Comparing to other factors, the effect of low sequence depth seems to be minimal. FFPE also affects CNV calls by increased variability in several callers. Consensus call and slightly stringent cutoffs for calling amplification/deletion can produce high confidence calls. Comparing to cytogenetic array, we found high correlations between the array and WGS segments in some callers while other callers show large variability. Citation Format: Yun-Ching Chen, Fayaz Seifuddin, Cu Nguyen, Chunhua Yan, Qingrong Chen, Wenming Xiao, Mehdi Pirooznia, Daoud Meerzaman. Critical Assessment of CNV Calling Using Next Generation Sequencing [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 3521.

A novel computational tool for copy number variation detection in targeted circulating tumor DNA.

e13051 Background: Targeted sequencing of circulating tumor DNA (ctDNA) has been used in early tumor detection and guidance of clinical cancer treatment. Accurate somatic copy number variation (SCNV) estimation is useful for better decisions in clinics, but remain challenging due to low percentage of tumor-cell-released DNA in circulating blood, in addition to the low signal-noise ratio and usual lack of normal control. Methods: To overcome these challenges and call SCNV at gene level, we develop a novel bioinformatics tool ctCNV. After correction for GC content, target region length and read counts, genome local scores at the gene level (GCS) were calculated, using 30 normal blood samples as control. Statistically significant cutoffs were determined using control. Results: The new method has been benchmarked with other SCNV calling tools such as CNVkit, which have been developed mainly for whole exome sequencing instead of targeted DNAseq. The validation experiment involves 6 mixed samples consisting of different proportions of known cell lines and real blood samples. ROC curve obtained through comparing the SCNV calling results with Droplet Digital PCR results show improvement of our method comparing to existing methods. Conclusions: As a summary, our study has three main results/contributions. First, we developed a novel computational tool for SCNV calling in targeted sequencing of ctDNA, with good sensitivity and specificity. Second, we thoroughly evaluated the performance of current available CNV calling tools in targeted sequenced ctDNA. Third, the raw data generated from our six real samples of gradient mixture of different cell lines and human’s samples can serve as an evaluation standard to other further computational tools for CNV calling in targeted sequenced ctDNA.

Germline Predisposition and Copy Number Alteration in Pre-stage Lung Adenocarcinomas Presenting as Ground-Glass Nodules.

Objective: Synchronous multiple ground-glass nodules (SM-GGNs) are a distinct entity of lung cancer which has been emerging increasingly in recent years in China. The oncogenesis molecular mechanisms of SM-GGNs remain elusive.Methods: We investigated single nucleotide variations (SNV), insertions and deletions (INDEL), somatic copy number variations (CNV), and germline mutations of 69 SM-GGN samples collected from 31 patients, using target sequencing (TRS) and whole exome sequencing (WES).Results: In the entire cohort, many known driver mutations were found, including EGFR (21.7%), BRAF (14.5%), and KRAS (6%). However, only one out of the 31 patients had the same somatic missense or truncated events within SM-GGNs, indicating the independent origins for almost all of these SM-GGNs. Many germline mutations with a low frequency in the Chinese population, and genes harboring both germline and somatic variations, were discovered in these pre-stage GGNs. These GGNs also bore large segments of copy number gains and/or losses. The CNV segment number tended to be positively correlated with the germline mutations (r = 0.57). The CNV sizes were correlated with the somatic mutations (r = 0.55). A moderate correlation (r = 0.54) was also shown between the somatic and germline mutations.Conclusion: Our data suggests that the precancerous unstable CNVs with potentially predisposing genetic backgrounds may foster the onset of driver mutations and the development of independent SM-GGNs during the local stimulation of mutagens.

Identification of oncogenic long noncoding RNAs CASC9 and LINC00152 in oral carcinoma through genome-wide comprehensive analysis.

Oral carcinoma (OC) is the major cancer type in the head and neck region; however, the molecular mechanisms of its pathogenesis and progression remain poorly understood. In recent years, the long noncoding RNAs (lncRNAs) have been uncovered as critical regulators in the development and progression of multiple human cancers, but most of the lncRNAs expression patterns, clinical relevance, and biological functions in OC are still unclear. To better understand the significance of lncRNAs in OC carcinogenesis, we analyzed the expression levels of lncRNAs between OC and healthy oral mucosa using The Cancer Genome Atlas Cancer Genome RNA sequencing data and another three independent microarray datasets from Gene Expression Omnibus. As a result, we found that thousands of lncRNAs expression are dysregulated in OC, and further somatic copy number variation analyses showed that some of these lncRNAs alterations are associated with copy number amplification or loss in OC. Moreover, lots of lncRNAs expression levels are associated with OC patients' overall survival and recurrence-free survival; for example, higher CASC9, LINC01232, and MIR4435-1HG expression levels are related to shorter overall survival and recurrence-free survival in OC patients. Finally, the potential function of two lncRNAs (CASC9 and LINC00152) that were upregulated in OC tissues and associated with patients' survival time was verified by loss-of-function assays in OC cells. Our findings indicate that these altered lncRNAs might play important roles in the development of OC, and our data can provide a valuable list of lncRNAs candidates for further investigation of their function and mechanisms in OC.

The genomic mutation spectrums of breast fibroadenomas in Chinese population by whole exome sequencing analysis.

Fibroadenomas (FAs) are the most common fibroepithelial lesions and the most common benign tumors of the breast in women of reproductive age. Although MED12 mutations, an overwhelming majority of all mutations, and some other gene mutations have been found in FAs, the genomic landscapes of FAs are still not completely clear and the genomic mutation spectrums of FAs in Chinese population remains unknown. Here, by performing whole exome sequencing of 12 FAs and the corresponding normal breast tissues in Chinese Han population, we observed the somatic and germline landscapes of genetic alterations. We identified 16 recurrently mutated genes with 37 nonsynonymous or frameshift somatic mutations and 27 recurrent somatic copy number variants (CNVs). In these mutated genes, MED12 was the most common in FAs, harboring 6 nonsynonymous/frameshift somatic mutations and 1 CNV. In addition, 6 germline mutations of tumor susceptibility genes in 5 FAs were identified and the tumor mutational burden of the 5 FAs was significantly higher than the other 7 FAs without germline mutations. This study provides genomic mutation spectrums of FAs in Chinese population and expand the genetic spectrum of FAs.