Genomic Data Commons Research Articles

The NCI Genomic Data Commons

The National Cancer Institute (NCI) Genomic Data Commons (GDC) contains more than 2.9 petabytes of genomic and associated clinical data from more than 60 NCI-funded and other contributed cancer genomics research projects. The GDC consists of five applications over a common data model and a common application programming interface.

Abstract OT-18-01: The metastatic breast cancer project: Generating the clinical and genomic landscape of metastatic breast cancer through patient-partnered research

Abstract The Metastatic Breast Cancer Project (MBCproject) is an ongoing research study that directly engages patients (pts) through social media and advocacy groups, and empowers them to share their samples, clinical information, and experiences. The goal is to create a publicly available dataset of linked genomic, clinical, and pt-reported data to enable research. In collaboration with pts, advocates, and advocacy groups, a website (MBCproject.org) was developed that allows pts with metastatic breast cancer (MBC) anywhere in the US or Canada to register. From 10/20/15-3/31/20, 5708 women and men with MBC registered for the MBCproject. Registered pts are sent an online consent form that asks for permission to obtain and analyze their medical records and samples. Consented pts are sent a saliva and/or blood kit and asked to mail back a saliva sample, which is used to extract germline DNA, and/or a blood sample, which is used to extract germline DNA and cell free DNA (cfDNA). We contact participants’ medical providers to obtain medical records and a portion of their stored tumor biopsies. 3245 pts receiving care at over 1700 different institutions have consented to share medical records and tumor/saliva/blood samples and to have genomic analysis performed. Whole exome sequencing (WES) is performed on tumor DNA, germline DNA, and cfDNA; transcriptome sequencing (RNA-seq) is performed on tumor RNA. Medical records and pt-reported data are abstracted to create a detailed clinical record for each pt. Table 1 highlights clinical data collection, biospecimen acquisition, and genomic data generation to date. Examples of clinicogenomic analyses are shown in Table 2. De-identified linked genomic, clinical, and pt-reported data is shared regularly via public and semi-public databases (mbcproject.org, cBioPortal, dbGaP, NCI Genomic Data Commons). To date, this data has been cited in over 20 published journal articles. Study updates are shared with participants regularly. The MBCproject continues to enroll new patients, generate additional data, and perform integrated clinical and genomic analyses with the goal of building a dataset that is representative of patients with MBC. We have partnered with over 30 non-profit breast cancer advocacy groups. We also have several community engagement efforts underway to more directly reach patients in underrepresented communities, including partnerships with faith-based organizations and colleges/universities, as well as targeted engagement with the African American community. In addition, in partnership with Latinx patients, advocates, and researchers, the project has been translated into Spanish and is expected to launch in late 2020. Partnering directly with pts rapidly enables thousands of pts to remotely share tumors, blood, saliva, and medical records to accelerate research. The resulting publicly shared clinically annotated database is a resource that allows researchers to identify patients with specific phenotypes, who have often been challenging to identify with traditional approaches. Clinical data collection, biospecimen acquisition, and genomic data generation:NumberConsent signed (US & CA)3245 ptsSurvey #1 submitted(demographics, diagnosis details, receptor status, clinical experiences)3245 ptsSurvey #2 submitted(pathology details, sites of metastasis, treatments with start and stop dates)1638 ptsMedical record received1352 ptsSaliva sample received2004 ptsBlood sample received1121 ptsTumor samples received585 tumor samples from 424 ptsDigital image of tumor slide H&E generated585 tumor samplesWES from germline complete458 germline samplesWES from tumor (primary and metastatic) samples complete343 tumor samplesRNA-seq from tumor (primary and metastatic) samples complete228 tumor samplesULP-WGS from cfDNA (taken in metastatic setting) complete993 blood samplesWES from circulating tumor DNA (taken in metastatic setting) complete143 blood samples CohortConsented (US & CA)Tumor WES completeTumor RNA-seq completePts diagnosed < 40 yrs of age107312071De novo MBC112712183Late recurrence (>5 years after dx)8307752Long term survivors (MBC > 10yrs)158115Resistance to CDK4/6 inhibitors70914839NED at time of f/u survey4238939Triple Negative Breast Cancer3107531Patients with 2 or more tumor biopsies / cfDNA samples collected by the MBCproject2876138 Citation Format: Nikhil Wagle, Corrie Painter, Elana Anastasio, Michael Dunphy, Mary McGillicuddy, Esha Jain, Brett Tomson, Tania G. Hernandez, Beena Thomas, Dewey Kim, Alyssa L. Damon, Shahrayz Shah, Rafael Ramos, Colleen Nguyen, Lee O'Neil, Sarah Winnicki, Sara Balch, Rachel Stoddard, Taylor Cusher, Parker Chastain, Jorge Gomez Tejeda Zanudo, Jorge Buendia-Buendia, Ofir Cohen, Netsanet Tsegai, Lauren Sterlin, Ulcha F. Ulysse, Imani Boykin, Kate Sine, Oyin Alao, Jacqueline Lucia, Eric S. Lander, Todd R. Golub. The metastatic breast cancer project: Generating the clinical and genomic landscape of metastatic breast cancer through patient-partnered research [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr OT-18-01.

TAp73β Can Promote Hepatocellular Carcinoma Dedifferentiation

Simple SummaryHepatocellular carcinoma (HCC) is a highly complex and heterogeneous type of cancer. Hepatocyte dedifferentiation is one of the important steps in the development of HCC. However, its molecular mechanisms are not well known. In this study, we report that transcriptionally active TAp73 isoforms are overexpressed in HCC. We also show that TAp73β suppresses the expression of the hepatocyte markers including CYP3A4, AFP, ALB, HNF4α, while increasing the expression of several cholangiocyte markers in HCC cell lines. In conclusion, this report reveals a pro-oncogenic role for TAp73β in liver cancer.Hepatocyte dedifferentiation is a major source of hepatocellular carcinoma (HCC), but its mechanisms are unknown. We explored the p73 expression in HCC tumors and studied the effects of transcriptionally active p73β (TAp73β) in HCC cells. Expression profiles of p73 and patient clinical data were collected from the Genomic Data Commons (GDC) data portal and the TSVdb database, respectively. Global gene expression profiles were determined by pan-genomic 54K microarrays. The Gene Set Enrichment Analysis method was used to identify TAp73β-regulated gene sets. The effects of TAp73 isoforms were analyzed in monolayer cell culture, 3D-cell culture and xenograft models in zebrafish using western blot, flow cytometry, fluorescence imaging, real-time polymerase chain reaction (RT-PCR), immunohistochemistry and morphological examination. TAp73 isoforms were significantly upregulated in HCC, and high p73 expression correlated with poor patient survival. The induced expression of TAp73β caused landscape expression changes in genes involved in growth signaling, cell cycle, stress response, immunity, metabolism and development. Hep3B cells overexpressing TAp73β had lost hepatocyte lineage biomarkers including ALB, CYP3A4, AFP, HNF4α. In contrast, TAp73β upregulated genes promoting cholangiocyte lineage such as YAP, JAG1 and ZO-1, accompanied with an increase in metastatic ability. Our findings suggest that TAp73β may promote malignant dedifferentiation of HCC cells.

Bringing Structural Implications and Deep Learning-Based Drug Identification for KRAS Mutants.

Colorectal cancer is considered one of the leading causes of death that is linked with the Kirsten Rat Sarcoma (KRAS) harboring codons 13 and 61 mutations. The objective for this study is to search for clinically important codon 61 mutations and analyze how they affect the protein structural dynamics. Additionally, a deep-learning approach is used to carry out a similarity search for potential compounds that might have a comparatively better affinity. Public databases like The Cancer Genome Atlas and Genomic Data Commons were accessed for obtaining the data regarding mutations that are associated with colon cancer. Multiple analysis such as genomic alteration landscape, survival analysis, and systems biology-based kinetic simulations were carried out to predict dynamic changes for the selected mutations. Additionally, a molecular dynamics simulation of 100 ns for all the seven shortlisted codon 61 mutations have been conducted, which revealed noticeable deviations. Finally, the deep learning-based predicted compounds were docked with the KRAS 3D conformer, showing better affinity and good docking scores as compared to the already existing drugs. Taking together the outcomes of systems biology and molecular dynamics, it is observed that the reported mutations in the SII region are highly detrimental as they have an immense impact on the protein sensitive sites' native conformation and overall stability. The drugs reported in this study show increased performance and are encouraged to be used for further evaluation regarding the situation that ascends as a result of KRAS mutations.

RNA splicing and aggregate gene expression differences in lung squamous cell carcinoma between patients of West African and European ancestry

ObjectivesDespite disparities in lung cancer incidence and mortality, the molecular landscape of lung cancer in patients of African ancestry remains underexplored, and race-related differences in RNA splicing remain unexplored. Materials and methodsWe identified differentially spliced genes (DSGs) and differentially expressed genes (DEGs) in biobanked lung squamous cell carcinoma (LUSC) between patients of West African and European ancestry, using ancestral genotyping and Affymetrix Clariom D array. DSGs and DEGs were validated independently using the National Cancer Institute Genomic Data Commons. Associated biological processes, overlapping canonical pathways, enriched gene sets, and cancer relevance were identified using Gene Ontology Consortium, Ingenuity Pathway Analysis, Gene Set Enrichment Analysis, and CancerMine, respectively. Association with LUSC survival was conducted using The Cancer Genome Atlas. Results4,829 DSGs and 267 DEGs were identified, including novel targets in NSCLC as well as genes identified previously to have relevance to NSCLC. RNA splicing events within 3 DSGs as well as 1 DEG were validated in the independent cohort. 853 DSGs and 29 DEGs have been implicated as potential drivers, oncogenes and/or tumor suppressor genes. Biological processes enriched among DSGs and DEGs included metabolic process, biological regulation, and multicellular organismal process and, among DSGs, ion transport. Overlapping canonical pathways among DSGs included neuronal signaling pathways and, among DEGs, cell metabolism involving biosynthesis. Gene sets enriched among DSGs included KRAS Signaling, UV Response, E2 F Targets, Glycolysis, and Coagulation. 355 RNA splicing events within DSGs and 18 DEGs show potential association with LUSC patient survival. ConclusionThese DSGs and DEGs, which show potential biological and clinical relevance, could have the ability to drive novel biomarker and therapeutic development to mitigate LUSC disparities.

The intratumor microbiome predicts prognosis across gender and subtypes in papillary thyroid carcinoma

While the intratumor microbiome has become increasingly implicated in cancer development, the microbial landscape of papillary thyroid carcinoma (PTC) is essentially uninvestigated. PTC is characterized by varied prognosis between gender and cancer subtype, but the cause for gender and subtype-based dissimilarities is unclear. Women are more frequently diagnosed with PTC, while men suffer more advanced-staged PTC. In addition, tall cell variants are more aggressive than classical and follicular variants of PTC. We hypothesized that intratumor microbiome composition distinctly alters the immune landscape and predicts clinical outcome between PTC subtypes and between patient genders. Raw whole-transcriptome RNA-sequencing, Level 3 normalized mRNA expression read counts, and DNA methylation 450 k sequencing data for untreated, nonirradiated tumor, and adjacent normal tissue were downloaded from the Genomic Data Commons (GDC) legacy archive for 563 thyroid carcinoma patients. Microbe counts were extracted using Pathoscope 2.0 software. We correlated microbe abundance to clinical variables and immune-associated gene expression. Gene-set enrichment, mutation, and methylation analyses were conducted to correlate microbe abundance to characterize microbes’ roles. Overall, PTC tumor tissue significantly lacked microbes that are populated in adjacent normal tissue, which suggests presence of microbes may be critical in controlling immune cell expression and regulating immune and cancer pathways to mitigate cancer growth. In contrast, we also found that microbes distinctly abundant in tall cell and male patient cohorts were also correlated with higher mutation expression and methylation of tumor suppressors. Microbe dysbiosis in specific PTC types may explain observable differences in PTC progression and pathogenesis. These microbesprovide a basis for developing specialized prebiotic and probiotic treatments for varied PTC tumors.

Metabolism-Associated Molecular Classification of Colorectal Cancer.

The high heterogeneity of colorectal cancer (CRC) is the main clinical challenge for individualized therapies. Molecular classification will contribute to drug discovery and personalized management optimizing. Here, we aimed to characterize the molecular features of CRC by a classification system based on metabolic gene expression profiles. 435 CRC samples from the Genomic Data Commons data portal were chosen as training set while 566 sample in GSE39582 were selected as testing set. Then, a non-negative matrix factorization clustering was performed, and three subclasses of CRC (C1, C2, and C3) were identified in both training set and testing set. Results showed that subclass C1 displayed high metabolic activity and good prognosis. Subclass C2 was associated with low metabolic activities and displayed high immune signatures as well as high expression of immune checkpoint genes. C2 had the worst prognosis among the three subtypes. Subclass C3 displayed intermediate metabolic activity, high gene mutation numbers and good prognosis. Finally, a 27-gene metabolism-related signature was identified for prognosis prediction. Our works deepened the understanding of metabolic hallmarks of CRC, and provided valuable information for “multi-molecular” based personalized therapies.

OpenGDC: Unifying, Modeling, Integrating Cancer Genomic Data and Clinical Metadata

Next Generation Sequencing technologies have produced a substantial increase of publicly available genomic data and related clinical/biospecimen information. New models and methods to easily access, integrate and search them effectively are needed. An effort was made by the Genomic Data Commons (GDC), which defined strict procedures for harmonizing genomic and clinical data of cancer, and created the GDC data portal with its application programming interface (API). In this work, we enhance GDC harmonization by applying a state of the art data model (called Genomic Data Model) made of two components: the genomic data, in Browser Extensible Data (BED) format, and the related metadata, in a tab-delimited key-value format. Furthermore, we extend the GDC genomic data with information extracted from other public genomic databases (e.g., GENCODE, HGNC and miRBase). For metadata, we implemented automatic procedures to extract and normalize them, recognizing and eliminating redundant ones, from both Clinical/Biospecimen Supplements and GDC Data Model, that are present on the two sources of GDC (i.e., data portal and API). We developed and released the OpenGDC software, which is able to extract, integrate, extend, and standardize genomic and clinical data of The Cancer Genome Atlas (TCGA) from the GDC. Additionally, we created a publicly accessible repository, containing such homogenized and enhanced TCGA data (resulting in about 1.3 TB). Our approach, implemented in the OpenGDC software, provides a step forward to the effective and efficient management of big genomic and clinical data of cancer. The strong usability of our data model and utility of our work is demonstrated through the application of the GenoMetric Query Language (GMQL) on the transformed TCGA data from the GDC, achieving promising results, facilitating information retrieval and knowledge discovery analyses.

Abstract IA18: How digital disintermediation (“apps”) and cloud computing change the precision prevention landscape

Abstract The emergence of cloud-based data commons is both a response to the call for fair scientific data stewardship and to the challenge of scaling consumer-facing precision medicine. The advancement towards cloud-hosted cancer data platforms mirrors a broader trend towards stateless API ecosystems for user-governed application development. This is particularly the case in cancer prevention, where the goal ultimately is to avoid the need for treatment. Accordingly, reference data will typically be exposed publicly, while its contextualization will rely on the individual’s own mobile computing resources through web applications (“apps”). This discussion will be illustrated with a number of practical applications, ranging from serverless traversal of the 20 million patient-encounter records of New York’s Statewide Planning and Research Cooperative System (SPARCS) to the contextualization of NCI’s Genomic Data Commons (GDC). Presentation material at bit.ly/appCommons; open source code at github.com/episphere. Citation Format: Jonas S. Almeida. How digital disintermediation (“apps”) and cloud computing change the precision prevention landscape [abstract]. In: Proceedings of the AACR Special Conference on Modernizing Population Sciences in the Digital Age; 2019 Feb 19-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2020;29(9 Suppl):Abstract nr IA18.

Developing and Using a Data Commons for Understanding the Molecular Characteristics of Germ Cell Tumors.

Germ cell tumors (GCTs) are a rare disease, but they account for 15% of all malignancies diagnosed during adolescence. The biological mechanisms underpinning their development are only starting to be explored. Current GCT treatment may be associated with significant toxicity. Therefore, there is an urgent need to understand the molecular basis of GCT and identify biomarkers to tailor the therapy for individual patients. However, this research is severely hamstrung by the rarity of GCTs in individual hospitals/institutes. A publicly available genomic data commons with GCT datasets compiled from different institutes/studies would be a valuable resource to facilitate such research. In this study, we first reviewed publicly available web portals containing GCT genomics data, focusing on comparing data availability, data access, and analysis tools, and the limitations of using these resources for GCT molecular studies. Next, we specifically designed a GCT data commons with a web portal, GCT Explorer, to assist the research community to store, manage, search, share, and analyze data. The goal of this work is to facilitate GCT molecular basis exploration and translational research.

Genomic and epigenetic aberrations of chromosome 1p36.13 have prognostic implications in malignancies.

Deletions of chromosome 1p36 are common in malignancies; however, there is limited information regarding the biological and prognostic implications of 1p36 in cancer. Steroid Receptor-Associated and Regulated Protein (SRARP) is a tumor suppressor on chromosome 1p36.13 that its inactivation predicts poor cancer outcome, indicating that the 1p36.13 segment requires further studies. Therefore, a comprehensive multi-omics analysis of The Cancer Genome Atlas (TCGA), the Pan-Cancer Analysis of Whole Genomes (PCAWD), the International Cancer Genome Consortium (ICGC), and the Genomic Data Commons (GDC) Pan-Cancer datasets was conducted to investigate the prognostic implications of 1p36.13 in malignancies. This study revealed that expression and DNA methylation of multiple genes on 1p36.13 are significantly associated with survival in primary tumors and normal adjacent tissues. In addition, copy-number loss in every gene on 1p36.13 predicts poor cancer outcome. Importantly, copy-number loss and somatic mutations of chromosome 1p36.13 segment are associated with worse survival in primary tumors, and DNA hypermethylation of 1p36.13 predicts poor outcome in normal adjacent tissues. Therefore, genomic and epigenetic aberrations of chromosome 1p36.13 have promising prognostic implications in cancer.

The Veterans Affairs Precision Oncology Data Repository, a Clinical, Genomic, and Imaging Research Database

SummaryThe Veterans Affairs Precision Oncology Data Repository (VA-PODR) is a large, nationwide repository of de-identified data on patients diagnosed with cancer at the Department of Veterans Affairs (VA). Data include longitudinal clinical data from the VA's nationwide electronic health record system and the VA Central Cancer Registry, targeted tumor sequencing data, and medical imaging data including computed tomography (CT) scans and pathology slides. A subset of the repository is available at the Genomic Data Commons (GDC) and The Cancer Imaging Archive (TCIA), and the full repository is available through the Veterans Precision Oncology Data Commons (VPODC). By releasing this de-identified dataset, we aim to advance Veterans' health care through enabling translational research on the Veteran population by a wide variety of researchers.

Sphingosine Kinase 1 is Associated With Immune Cell–Related Gene Expressions in Human Breast Cancer

BackgroundAlthough previous experiments have implicated sphingosine-1-phosphate (S1P) as a links between immune reactions and cancer progression, the exact mechanism of this interaction has not comprehensively studied in clinical human samples. This study sought to evaluate the S1P regulation by sphingosine kinase 1 (SPHK1), an S1P-producing enzyme, in the immunity/immuno-reactivity of clinical human breast cancer surgical specimens. MethodsS1P levels were examined in tumor, peritumoral, and normal human breast samples using mass spectrometry. Genomics Data Commons data portal of The Cancer Genome Atlas cohort was used to assess the expression of S1P-related and immune-related genes. ResultsS1P levels were significantly higher in tumor samples compared to peritumoral (P < 0.05) or normal human breast samples (P < 0.001). SPHK1 gene expression was elevated in tumoral samples compared to normal breast samples (P < 0.01). Furthermore, the elevated expression of SPHK1 in breast cancer tissue was associated with an increased expression of the different kinds of immune-related genes, such as CD68, CD163, CD4, and FOXP3 (forkhead box P3), in HER2-negative breast cancer. Network analysis showed the central role of SPHK1 in the interaction of S1P signaling and expression of immune cell–related proteins. ConclusionsWe demonstrated that S1P is mainly produced by tumor tissue, rather than peritumoral tissue, in breast cancer patients. Our data revealed the involvement of S1P signaling in the regulation of immune-related genes, suggesting the links between S1P and complicated immune–cancer interactions in breast cancer patients.

Abstract 5862: NCI Office of Cancer Genomics supports multidisciplinary genomics research initiatives to advance precision oncology

Abstract The National Cancer Institute's (NCI) Office of Cancer Genomics' (OCG) mission is to advance the molecular understanding of cancers to improve clinical outcomes. To accomplish this goal, OCG develops and collaboratively manages molecular characterization and translational genomics research initiatives. OCG currently supports four innovative and complementary initiatives. These programs generate comprehensive genomic datasets, develop bioinformatics tools to analyze these datasets, create biologically-relevant human tumor-derived Next-Generation Cancer Models (NGCMs) and valuable experimental reagents and resources. The Cancer Genome Characterization Initiative (CGCI) (https://ocg.cancer.gov/programs/cgci) and the Therapeutically Applicable Research to Generate Effective Treatments (TARGET) (https://ocg.cancer.gov/programs/target) are the molecular characterization programs. These programs aim to identify therapeutic targets and biomarkers in understudied, rare, and high-risk adult and pediatric cancers. Clinically annotated tumor and normal matched tissue samples are collected and analyzed by whole-genome, whole-exome, and transcriptome sequencing. The datasets are available to the research community through program-specific Data Matrices: CGCI Data Matrix and TARGET Data Matrix. CGCI and TARGET data are also available through NCI's Genomic Data Commons (GDC)(https://portal.gdc.cancer.gov/). The Pediatric Genomic Data Inventory (PGDI) is a global inventory of pediatric molecular and clinical datasets. The CTD2 Network (https://ocg.cancer.gov/programs/ctd2) aims to understand cancer metastasis, tumor heterogeneity and drug resistance, and develop optimal combinations of pharmacologic and immunological agents. The CTD2 Network develops bioinformatics tools, generates diverse raw/primary datasets that can be accessed through the CTD2 Data Portal and further validates subsets of these data listed at the CTD2 Dashboard. The Human Cancer Models Initiative (HCMI) (https://ocg.cancer.gov/programs/HCMI) is an international consortium that is generating human tumor-derived NGCMs from diverse tumor subtypes including rare and pediatric cancers. The models, together with related clinical and genomic data, are available as a community resource. The HCMI Searchable Catalog allows users to search the list of available models and associated clinical and molecular data. The clinical and molecular data are stored at the NCI's GDC. Data, analytical tools, and resources generated by OCG initiatives are made publicly available to further the collaborative goal of enabling precision oncology to improve patient care. The OCG website (https://ocg.cancer.gov/) also provides data usage policies, guides to access data, experimental methods, standard operating procedures, and educational and helpful links for the community. Citation Format: Cindy W. Kyi, Pamela C. Birriel, Tanja M. Davidsen, Martin L. Ferguson, Patee Gesuwan, Nicholas B. Griner, Yiwen He, Lauren M. Hurd, Subhashini Jagu, Eva Tonsing-Carter, Daniela S. Gerhard. NCI Office of Cancer Genomics supports multidisciplinary genomics research initiatives to advance precision oncology [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 5862.

Abstract 35: Construction of ce-RNA networks and comprehensive analysis with tumor infiltrating immune cells in hepatocellular carcinoma

Abstract Background: Hepatocellular carcinoma (HCC) is one of the most common cancers worldwide with a high recurrence rate. Although immune-checkpoint treatment provides survival benefits for patients with HCC, the mortality of HCC is still high. Therefore, it is necessary to understand the molecular mechanism, immune microenvironment and related prognostic factors of HCC. Methods: HCC patients with both RNAseq and miRNA expression data and their clinical information were downloaded from the NCI's Genomic Data Commons. Gene expression data were analyzed to identify differentially expressed genes (DEGs). The results were further used to construct the competing endogenous RNA (ceRNA) networks related to HCC. Meanwhile, the proportions of 22 immune cell types in HCC were estimated by using CIBERSORT (https://cibersort.stanford.edu). Subsequently, two nomograms based on the expression value of ceRNAs and the percentage of immune cells were constructed to predict the prognosis of HCC patients, respectively. The calibration and time series ROC curves were used to evaluate these models. Finally, the correlation between ceRNAs and immune cells were explored to reveal the underlying molecular mechanism. Results: Gene expression data and clinical information of 367 primary HCC and 50 adjacent normal samples were downloaded from the TCGA database. These data were used to identify DEGs. A total of 12 lncRNAs, 71 mRNAs and 39 miRNAs were identified to be in ceRNA networks according to starBase v2.0 database. The results showed that stages of cancer, the expression values of CBX2, MAFG, DNM3, BUB1, ARL5B and other 23 genes were important predictors. The AUCs of the constructed nomogram in predicting the 3- and 5-year OS were 0.79 and 0.81 respectively. Stages of cancer, proportions of resting memory CD4+ T cell, M0 macrophages and resting mast cells constructed another nomogram with AUC of 0.79 and 0.81 in predicting the 3- and 5-year OS. CIBERSORT results showed that the proportions of regulatory T cells, CD8+ T cells, follicular helper T cells and M0 macrophages were slightly higher while the proportions of M2 macrophages, monocytes and activated mast cells were slightly lower in HCC than in adjacent normal samples. Integration analysis of CIBERSORT results, ceRNAs and gene set enrichment analysis of immunologic signatures identified BUB1, E2F8, H2AFZ, JPT1, LPCAT1, MAFG and PSMC3IP to be positively related with regulatory T cells, follicular helper T cells, activated memory CD4+ T cells and M0 macrophages. This is consistent with current knowledge that immune activity is suppressed in certain types of cancers. However, the underlying mechanism warrant further exploration. Conclusions: In this study, two nomograms were constructed to predict the prognosis of HCC patients, which might be useful in predicting the prognosis of HCC patients. The results suggest that the immune suppress effect of HCC may be related to BUB1, E2F8, H2AFZ, JPT1, LPCAT1, MAFG and PSMC3IP genes. These genes may serve as potential therapeutic targets. Acknowledgments: This work was supported by the National Science and Technology Major Project of China (2018ZX10302205). Citation Format: Jianxiang Shi, Chi Cui, Yaru Duan, Hua Ye, Peng Wang, Jitian Li, Shuiling Jin, Liping Dai, Jianying Zhang. Construction of ce-RNA networks and comprehensive analysis with tumor infiltrating immune cells in hepatocellular carcinoma [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 35.

Abstract 6580: Proteogenomic characterization of endometrial carcinoma

Abstract Endometrial cancer (EC) is the sixth most common cancer in women globally and is one of the few cancers for which mortality is increasing; to address the need to develop new methods for the diagnosis and treatment of this disease, we performed a comprehensive proteogenomic characterization of 95 endometrial carcinomas. We prospectively collected 83 endometrioid and 12 serous tumors, along with 49 normal tissue samples, for multi-omic characterization. Each sample underwent whole exome sequencing, whole genome sequencing, total RNA and miRNA sequencing, and DNA methylation array analyses. In addition, we evaluated all samples for protein, phosphorylation, and acetylation levels using an integrated, isobaric tandem mass tag labeling–based quantitative proteomics workflow. This dataset, publicly available through the Genomic Data Commons, CPTAC data portal, and other sources, provides a valuable resource for researchers and clinicians. Our analyses of these tumors revealed, among other things, a potential role for circRNAs in the epithelial-mesenchymal transition and new information about proteomic markers of clinical and genomic tumor subgroups, including relationships to known druggable pathways. We found possible new consequences of perturbations to the Wnt/β-catenin pathway, particularly CTNNB1 hotspot mutations; moreover, from the genome-wide acetylation survey, we found potential regulatory mechanisms linking Wnt signaling and histone acetylation. Phosphorylation data analysis suggests an elevated DNA damage response in serous tumors, which we hypothesize may be targetable by certain FDA-approved drugs. We found that TP53 hotspot mutations play unique trans-acting roles as compared to the same mutations in colon and ovarian cancers. We also characterized aspects of the tumor immune landscape, including immunogenic alterations, neoantigens, common cancer/testis antigens, and the immune microenvironment, all of which can inform immunotherapy decisions; of particular translational interest, we identified a subset of patients who, despite having high amounts of microsatellite instability, exhibit biomarkers which suggest that immunotherapy would be an ineffective treatment for their disease. By integrating global measurements of protein abundance, phosphorylation, and acetylation with next-generation genomics and transcriptomics, we have provided novel insights into biological processes underlying the development of endometrial cancers and generated strong hypotheses for new approaches to personalized treatment of individuals with endometrial cancer. Citation Format: Emily A. Kawaler, Yongchao Dou, Daniel Cui Zhou, Marina A. Gritsenko, Karin D. Rodland, Li Ding, Bing Zhang, Tao Liu, David Fenyö, Clinical Proteomic Tumor Analysis Consortium. Proteogenomic characterization of endometrial carcinoma [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 6580.

Abstract LB-242: Proteomic Data Commons: A resource for proteogenomic analysis

Abstract The objective of the National Cancer Institutes' Proteomic Data Commons (PDC) is to make cancer-related proteomic datasets accessible to the public. The PDC provides the cancer research community with a unified data repository that enables data sharing across cancer proteomic studies and also enables multi-omic integration in support of precision medicine. As a domain-specific repository within the Cancer Research Data Commons (CRDC), the vision for the PDC is to provide researchers the ability to find and analyze proteomic data across a wide variety of tumor types. Currently, the PDC houses data, supported by a large collection of metadata attributes, for nearly 40 datasets from over 12 cancer types produced by several large-scale cancer research programs, each with cohort sizes greater than 100 patients. The PDC facilitates the analysis of proteomic, genomic, and imaging data derived from the same tumor. Most of the datasets in the PDC also have corresponding genomic and imaging data available in the Genomic Data Commons and The Cancer Imaging Archive respectively. Researchers can discover which genomic variants are detectable at the protein-level or better understand associations between gene expression, copy number variation, and protein abundance. The resource is currently available to the public in beta phase (https://pdc.esacinc.com) and will be officially launched on the cancer.gov domain in March 2020. The PDC data portal is supported by a robust and extensible data model and provides user-friendly exploration, visualization and data analysis. This allows researchers to search for and visualize expression of proteins (through their mapped genes) across all studies, analyze protein abundance for all cases in a study through heatmaps, build and explore pan-cancer cohorts using highly curated, clinical metadata, and comprehensively view a study without needing to download the data. The PDC provides quick access to mapping of peptide identities and quantities on the human genome as well as protein databases containing patient/tumor-specific variants and novel splicing events. It also enables fast, accurate, and convenient proteomic validation of novel genomic alterations through the PepQuery algorithm. Through a highly versatile application programming interface (API), PDC allows users to interact with data programmatically and facilitates integration with data from other resources in their scripts for multi-omic analysis. Big data interoperability is critical for progress in precision medicine. PDC is designed to interoperate with other resources including the CRDC nodes, allowing users to analyze PDC data with the tools and pipelines available on the NCI cloud resources. It further allows users to use their own tools to co-analyze genomic and proteomic data available from a common sample on Amazon Web Services (AWS) platform or on a local system. The presentation will provide an overview of the PDC and it's available datasets, as well as a discussion of how it facilitates multi-omic data analyses. Citation Format: Ratna Rajesh Thangudu, Paul A. Rudnick, Michael Holck, Deepak Singhal, Michael J. MacCoss, Nathan J. Edwards, Karen A. Ketchum, Christopher R. Kinsinger, Erika Kim, Anand Basu. Proteomic Data Commons: A resource for proteogenomic analysis [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr LB-242.

Linked Entity Attribute Pair (LEAP): A Harmonization Framework for Data Pooling.

PURPOSEAs data-sharing projects become increasingly frequent, so does the need to map data elements between multiple classification systems. A generic, robust, shareable architecture will result in increased efficiency and transparency of the mapping process, while upholding the integrity of the data.MATERIALS AND METHODSThe American Association for Cancer Research’s Genomics Evidence Neoplasia Information Exchange (GENIE) collects clinical and genomic data for precision cancer medicine. As part of its commitment to open science, GENIE has partnered with the National Cancer Institute’s Genomic Data Commons (GDC) as a secondary repository. After initial efforts to submit data from GENIE to GDC failed, we realized the need for a solution to allow for the iterative mapping of data elements between dynamic classification systems. We developed the Linked Entity Attribute Pair (LEAP) database framework to store and manage the term mappings used to submit data from GENIE to GDC.RESULTSAfter creating and populating the LEAP framework, we identified 195 mappings from GENIE to GDC requiring remediation and observed a 28% reduction in effort to resolve these issues, as well as a reduction in inadvertent errors. These results led to a decrease in the time to map between OncoTree, the cancer type ontology used by GENIE, and International Classification of Disease for Oncology, 3rd Edition, used by GDC, from several months to less than 1 week.CONCLUSIONThe LEAP framework provides a streamlined mapping process among various classification systems and allows for reusability so that efforts to create or adjust mappings are straightforward. The ability of the framework to track changes over time streamlines the process to map data elements across various dynamic classification systems.

Abstract A04: DNA damage repair gene mutations and associated tumor immune landscape as biomarkers of response to immunotherapy in muscle-invasive urothelial cancer

Abstract Molecular subtyping of urothelial cancer (UC) has significantly advanced the understanding of bladder tumor heterogeneity and development of prognostic and predictive biomarkers. Evolving evidence across cancers strongly suggests that tumor immunoediting has a profound impact on the behavior of cancer cells and their adaptation to the coevolving microenvironment and response to treatment. In alignment with this concept, recent immune checkpoint blockade (ICB) therapies in UC have demonstrated the predictive potential of mutations in the DNA damage repair (DDR). A comprehensive understanding of DDR mutation-associated expression of immune regulatory genes could thus aid in expansion of current immunotherapies and predictive biomarkers for the design of patient-tailored combination treatments. We thus investigated the pretreatment tumor transcriptomic profiles of the five recently described molecular subtypes of muscle-invasive urothelial cancer (MIUC; n=408) from the Genomic Data Commons, to determine subtype-specific immune cell abundance, expression of 67 immune regulatory genes, and association with DDR gene mutation profiles. Analysis using CIBERSORT immune cell abundance determination tool showed significant differences in immune cell profiles and abundance between MIUC subtypes. Expression patterns of a selected panel of 67 genes including both immune-stimulatory and -inhibitory genes showed significant associations with subtypes and DDR mutation status. Findings from our study provide compelling evidence for coexpression of multiple immune checkpoint genes including PD-1, PD-L1, IDO1, TIGIT, TIM-3, TGFB1, LAG3, and others, that potentially contribute to compensatory immune evasion in bladder tumors. Our findings also emphasize the urgent need for biomarker discovery approaches that combine molecular subtype, DDR gene mutation status, tumor immune landscape classification, and immune checkpoint gene expression to increase the number of patients responding to immunotherapies. Citation Format: Thiago Vidotto, Sarah Nersesian, Charles Graham, David Robert Siemens, Madhuri Koti. DNA damage repair gene mutations and associated tumor immune landscape as biomarkers of response to immunotherapy in muscle-invasive urothelial cancer [abstract]. In: Proceedings of the AACR Special Conference on Bladder Cancer: Transforming the Field; 2019 May 18-21; Denver, CO. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(15_Suppl):Abstract nr A04.

Oncogenic Features in Histologically Normal Mucosa: Novel Insights Into Field Effect From a Mega-Analysis of Colorectal Transcriptomes.

INTRODUCTION:Colorectal cancer is a common malignancy that can be cured when detected early, but recurrence among survivors is a persistent risk. A field effect of cancer in the colon has been reported and could have implications for surveillance, but studies to date have been limited. A joint analysis of pooled transcriptomic data from all available bulk RNA-sequencing data sets of healthy, histologically normal tumor-adjacent, and tumor tissues was performed to provide an unbiased assessment of field effect.METHODS:A novel bulk RNA-sequencing data set from biopsies of nondiseased colon from screening colonoscopy along with published data sets from the Genomic Data Commons and Sequence Read Archive were considered for inclusion. Analyses were limited to samples with a quantified read depth of at least 10 million reads. Transcript abundance was estimated with Salmon, and downstream analysis was performed in R.RESULTS:A total of 1,139 samples were analyzed in 3 cohorts. The primary cohort consisted of 834 independent samples from 8 independent data sets, including 462 healthy, 61 tumor-adjacent, and 311 tumor samples. Tumor-adjacent gene expression was found to represent an intermediate state between healthy and tumor expression. Among differentially expressed genes in tumor-adjacent samples, 1,143 were expressed in patterns similar to tumor samples, and these genes were enriched for cancer-associated pathways.DISCUSSION:Novel insights into the field effect in colorectal cancer were generated in this mega-analysis of the colorectal transcriptome. Oncogenic features that might help explain metachronous lesions in cancer survivors and could be used for surveillance and risk stratification were identified.