Paired Sequencing Data Research Articles

Abstract GS03-11: Germline-mediated immunoediting sculpts breast cancer subtypes and metastatic proclivity

Abstract Background: Somatic genomic aberrations are acquired within the context of germline genomes which differ across individuals at millions of polymorphic sites. However, the role of germline variation in somatic evolution remains poorly understood. The most compelling example is is that deleterious germline variants in BRCA1 and, to a lesser extent, BRCA2 are preferentially associated with the development of triple-negative breast cancer (BC). The variable frequency of BC subtypes across ancestral populations further suggests a role for germline contributions. On the other hand, various lines of evidence indicate that avoidance of the adaptive immune system is a strong determinant of which somatic mutations persist within a tumor. Whether and how germline differences influence immunoediting has not been studied. Building on these observations, we sought to investigate whether germline variation mediates somatic evolution through immunoediting. Specifically, we hypothesize inherited variation in oncogenes would be subject to varied immunoediting pressures during malignant transformation and progression. A high burden of germline-derived epitopes in recurrently amplified oncogenes is predicted to select against amplification of the cognate gene during malignant transformation because this would increase epitope availability, the likelihood of epitope presentation, and immune-mediated cell death. Instead, immune pressures may select for amplification of an alternate driver gene with a lower germline-mediated epitope burden. We evaluated this hypothesis in a collection of 3,855 BC, spanning ductal carcinoma in situ (DCIS) to invasive BC, and metastatic lesions. Methods: We analyzed paired tumor and normal sequencing data from 1,087 primary and 702 metastatic breast cancer patients as well as somatic genomic profiles from 341 patients with DCIS using a novel algorithm to estimate germline-derived epitope burden (GEB) based on an individual’s genotype and class 1 HLA alleles. The relationship between GEB and subtype commitment, defined by the acquisition of focal oncogenic amplifications in five prognostic subgroups of BC: HER2+ disease and four high-risk of relapse ER+/HER2 integrative subgroups/clusters (ICs) which we previously described (IC1: 17q23, IC2: 11q13, IC6: 8p12, and IC9: 8q24) was evaluated. Specifically, we evaluated the association between the GEB per gene and whether an individual developed the corresponding subtype via logistic regression, correcting for the first six genetic principal components and somatic mutation burden. Outcome associations were evaluated via Cox Proportional Hazards Models. Results: Interrogating 3,855 breast cancer lesions, we demonstrate that germline-derived epitopes in recurrently amplified genes influence somatic evolution by mediating immunoediting. Individuals with a high GEB in ERBB2/HER2 are significantly less likely to develop HER2-positive breast cancer compared to other subtypes. The same holds true for recurrent amplicons that define four aggressive, high-risk of relapse, ER-positive integrative subgroups. Thus, GEB selects against cognate oncogene amplification. Tumors that overcome such immune-mediated negative selection are more aggressive and exhibited microenvironments depleted of lymphocytes, consistent with “immune cold” tumors. Conclusions: We demonstrate that inherited variation sculpts breast cancer subtypes, aggressivity, and immune landscapes by mediating anti-tumor immune responses. The implications of these findings are severalfold. First, GEB is prognostic, complementing other molecular measurements. Second, immunoediting pressures differ across the disease course, with implications for the timing of therapeutic interventions. Third, these data illuminate a broad source of currently under-appreciated immunogenic antigens. Citation Format: Kathleen Houlahan, Aziz Khan, Noah Greenwald, Robert West, Michael Angelo, Christina Curtis. Germline-mediated immunoediting sculpts breast cancer subtypes and metastatic proclivity [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr GS03-11.

Abstract NG01: Germline-mediated immunoediting sculpts breast cancer subtypes and metastatic proclivity

Abstract NG01: Germline-mediated immunoediting sculpts breast cancer subtypes and metastatic proclivity

Multidimensional Transcriptomics Unveils RNF34 as a Prognostic Biomarker and Potential Indicator of Chemotherapy Sensitivity in Wilms' Tumour.

Nephroblastoma, colloquially known as Wilms' tumour (WT), is the predominant malignant renal neoplasm arising in the paediatric population. Modern therapeutic approaches for WT incorporate a synergistic combination of surgical intervention, radiotherapy, and chemotherapy, which substantially ameliorate the overall patient survival rate. Despite this, the optimal sequence of chemotherapy and surgical intervention remains a matter of contention, with each strategy presenting its own strengths and weaknesses that could influence clinical decision-making. To make some headway on this clinical dilemma, we deployed a multidimensional transcriptomics integration approach by analysing bulk RNA sequencing data with 136 samples, as well as single-nucleus RNA sequencing (snRNA-seq) and paired spatial transcriptome sequencing (stRNA) data from 32 WT specimens. Our findings identified a distinct elevation of RNF34 expression within WT samples, which correlated with unfavourable prognostic outcomes. Leveraging the Genomics of Drug Sensitivity in Cancer (GDSC), we simultaneously revealed that patients with high expression of RNF34 have higher sensitivity to commonly used chemotherapy drugs for WT. Furthermore, our analysis of snRNA and stRNA data unveiled a reduced proportion of RNF34 expression in neoplastic cells after chemotherapy. Moreover, stRNA data delineated a significant association between a higher proportion of RNF34 expression in cancer cells and adverse features such as anaplastic histology and tumour recurrence. Intriguingly, we also observed a close association between elevated RNF34 expression and a characteristic exhausted tumour immune microenvironment. Collectively, our findings underscore the pivotal role of RNF34 in the prognostic prediction potential and treatment sensitivity of WT. This comprehensive analysis can potentially inform and refine clinical decision-making for WT patients and guide future studies towards the development of optimized, rational therapeutic strategies.

Single-cell RNA-sequencing to interrogate TCR-Repertoire and antigen-specificity in human atherosclerosis

Abstract Background Atherosclerosis is a chronic inflammatory disease characterized by the accumulation of distinct immune cells and lipids in the arterial wall. Despite recent advances in immunophenotyping, the molecular characteristics of T cells and their T cell receptor (TCR) repertoire in patients with atherosclerosis remain elusive. Studies using bulkRNA-sequencing of the TCR in both mice and humans have suggested a significant proportion of T cells in atherosclerotic plaques arises from a small number of antigen-specific T cells that express a limited TCR repertoire. However, it is unknown whether this restriction of TCR usage is caused by CD8+ or CD4+ T cells. Methods Here, we performed paired Single-cell RNA-sequencing (scRNAseq) and TCR-sequencing on atherosclerotic plaques (n=8) and peripheral blood mononuclear cells (PBMCs) from patients with and without coronary artery disease (CAD: n=31, noCAD: n=15). To analyze the paired scRNAseq and TCR sequencing data, we developed a bioinformatic tool that enables the identification of clonally expanded T cells in scRNAseq datasets. This tool allowed to investigate the T cell response in human atherosclerosis at a single-cell level and link clonal expansion with specific gene expression. Results Our analysis revealed that CD8+, rather than CD4+ T cells, exhibit an oligoclonal phenotype in the atherosclerotic plaques. By applying bioinformatic pseudotime analysis, we were able to recapitulate the developmental relationships within the T cell compartment. Our results demonstrate that clonally expanded CD8+ T cells show a TEMRA phenotype (Effector Memory–Expressing CD45RA) and originate from naive T cells within the plaque microenvironment. Pathway analyses of enriched CD8+ Plaque TEMRAs revealed increased signalling for cellular activation and migration, response to stress, and inflammation. In addition, clonally expanded CD8+ cells were also detected in PBMCs from patients with CAD and appear to be overrepresented in comparison to patients without CAD. Conclusion Our study provides new insights into the T cell response in human atherosclerosis. We demonstrate that CD8+ T cells show an oligoclonal phenotype in the atherosclerotic plaque and are also present in PBMCs from patients suffering from CAD. Furthermore, these cells appear to be dysregulated, suggesting a possible mechanism by which they contribute to disease progression. These findings highlight the power of paired scRNAseq and TCR sequencing which identify dysregulated genes and pathways in clonally expanded T cells, providing new insights into their potential role in atherosclerosis.

MSI-XGNN: an explainable GNN computational framework integrating transcription- and methylation-level biomarkers for microsatellite instability detection.

Microsatellite instability (MSI) is a hypermutator phenotype caused by DNA mismatch repair deficiency. MSI has been reported in various human cancers, particularly colorectal, gastric and endometrial cancers. MSI is a promising biomarker for cancer prognosis and immune checkpoint blockade immunotherapy. Several computational methods have been developed for MSI detection using DNA- or RNA-based approaches based on next-generation sequencing. Epigenetic mechanisms, such as DNA methylation, regulate gene expression and play critical roles in the development and progression of cancer. We here developed MSI-XGNN, a new computational framework for predicting MSI status using bulk RNA-sequencing and DNA methylation data. MSI-XGNN is an explainable deep learning model that combines a graph neural network (GNN) model to extract features from the gene-methylation probe network with a CatBoost model to classify MSI status. MSI-XGNN, which requires tumor-only samples, exhibited comparable performance with two well-known methods that require tumor-normal paired sequencing data, MSIsensor and MANTIS and better performance than several other tools. MSI-XGNN also showed good generalizability on independent validation datasets. MSI-XGNN identified six MSI markers consisting of four methylation probes (EPM2AIP1|MLH1:cg14598950, EPM2AIP1|MLH1:cg27331401, LNP1:cg05428436 and TSC22D2:cg15048832) and two genes (RPL22L1 and MSH4) constituting the optimal feature subset. All six markers were significantly associated with beneficial tumor microenvironment characteristics for immunotherapy, such as tumor mutation burden, neoantigens and immune checkpoint molecules such as programmed cell death-1 and cytotoxic T-lymphocyte antigen-4. Overall, our study provides a powerful and explainable deep learning model for predicting MSI status and identifying MSI markers that can potentially be used for clinical MSI evaluation.

Natural restoration exhibits better soil bacterial network complexity and stability than artificial restoration on the Loess Plateau, China

Natural restoration (NR, e.g., secondary succession) and artificial restoration (AR, e.g., afforestation) are key approaches for rehabilitating degraded land; however, a comparative assessment of microbial network between these approaches is lacking. We compared bacterial networks under NR and AR in two different watersheds on the Loess Plateau. Our findings revealed significantly heightened network complexity under NR compared to AR, including metrics such as node, edge, modularity, degree, centrality, and keystone nodes. NR's network robustness exceeded AR by 19.45–35.9% and 7.79–17.74% in the two watersheds, aligning with the ecological principle that complexity begets stability. The significantly higher negative/positive cohesion and natural connectivity under NR also support its better network stability than AR. Integrated analysis of paired sequencing data from five Loess Plateau studies conducted on the Loess Plateau further confirmed the higher complexity and stability of bacterial networks under NR. Further analysis unveiled “biological interactions” as primary drivers of bacterial co-occurrence (on average 84.21% of links), surpassing the influence of environmental filtering (5.17%) or dispersal limitation (4.2%). Importantly, networked communities under NR exhibited generally stronger linkages with various ecosystem function than AR. Overall, our study provides insights into vegetation restoration strategies from the perspective of microbial network, underscoring natural regeneration's potential as a superior remedy for degraded-land restoration.

Associations Between Cancer Predisposition Mutations and Clonal Hematopoiesis in Patients With Solid Tumors.

Clonal hematopoiesis (CH), the expansion of clones in the hematopoietic system, has been linked to different internal and external features such as aging, genetic ancestry, smoking, and oncologic treatment. However, the interplay between mutations in known cancer predisposition genes and CH has not been thoroughly examined in patients with solid tumors. We used prospective tumor-blood paired sequencing data from 46,906 patients who underwent Memorial Sloan Kettering-Integrated Mutation Profiling of Actionable Cancer Targets (MSK-IMPACT) testing to interrogate the associations between CH and rare pathogenic or likely pathogenic (P/LP) germline variants. We observed an enrichment of CH-positive patients among those carrying P/LP germline mutations and identified a significant association between P/LP germline variants in ATM and CH. Germline and CH comutation patterns in ATM, TP53, and CHEK2 suggested biallelic inactivation as a potential mediator of clonal expansion. Moreover, we observed that CH-PPM1D mutations, similar to somatic tumor-associated PPM1D mutations, were depleted in patients with P/LP germline mutations in the DNA damage response (DDR) genes ATM, CHEK2, and TP53. Patients with solid tumors and harboring P/LP germline mutations, CH mutations, and mosaicism chromosomal alterations might be at an increased risk of developing secondary leukemia while germline variants in TP53 were identified as an independent risk factor (hazard ratio, 36; P < .001) for secondary leukemias. Our results suggest a close relationship between inherited variants and CH mutations within the DDR genes in patients with solid tumors. Associations identified in this study might translate into enhanced clinical surveillance for CH and associated comorbidities in patients with cancer harboring these germline mutations.

Multi-omics analysis of multiple myeloma patients with differential response to first-line treatment.

The genome backgrounds of multiple myeloma (MM) would affect the efficacy of specific treatment. However, the mutational and transcriptional landscapes in MM patients with differential response to first-line treatment remains unclear. We collected paired whole-exome sequencing (WES) and transcriptomic data of over 200 MM cases from MMRF-COMPASS project. R package, maftools was applied to analyze the somatic mutations and mutational signatures across MM samples. Differential expressed genes (DEG) was calculated using R package, DESeq2. The feature selection of the predictive model was determined by LASSO regression. In silico analysis revealed newly discovered recurrent mutated genes such as TTN, MUC16. TP53 mutation was observed more frequent in nonCR (complete remission) group with poor prognosis. DNA repair-associated mutational signatures were enriched in CR patients. Transcriptomic profiling showed that the activity of NF-kappa B and TGF-β pathways was suppressed in CR patients. A transcriptome-based response predictive model was constructed and showed promising predictive accuracy in MM patients receiving first-line treatment. Our study delineated distinctive mutational and transcriptional landscapes in MM patients with differential response to first-line treatment. Furthermore, we constructed a 20-gene predictive model which showed promising accuracy in predicting treatment response in newly diagnosed MM patients.

Precise characterization of somatic complex structural variations from tumor/control paired long-read sequencing data with nanomonsv.

We present our novel software, nanomonsv, for detecting somatic structural variations (SVs) using tumor and matched control long-read sequencing data with a single-base resolution. The current version of nanomonsv includes two detection modules, Canonical SV module, and Single breakend SV module. Using tumor/control paired long-read sequencing data from three cancer and their matched lymphoblastoid lines, we demonstrate that Canonical SV module can identify somatic SVs that can be captured by short-read technologies with higher precision and recall than existing methods. In addition, we have developed a workflow to classify mobile element insertions while elucidating their in-depth properties, such as 5' truncations, internal inversions, as well as source sites for 3' transductions. Furthermore, Single breakend SV module enables the detection of complex SVs that can only be identified by long-reads, such as SVs involving highly-repetitive centromeric sequences, and LINE1- and virus-mediated rearrangements. In summary, our approaches applied to cancer long-read sequencing data can reveal various features of somatic SVs and will lead to a better understanding of mutational processes and functional consequences of somatic SVs.

Single-cell RNA and T cell receptor (TCR) sequencing in patients with HPV-positive head and neck squamous cell carcinoma (HNSCC) after induction CTLA4 and PD-1 immune checkpoint blockade (ICB).

6056 Background: Combination CTLA-4 and PD-1 ICB results in durable responses in patients with recurrent and metastatic HNSCC and therefore justifies further investigation in the induction or neoadjuvant setting. Methods: We investigated features of the tumor immune microenvironment (TIME) associated with MHR after induction CTLA-4 and PD-1 ICB in patients with newly diagnosed HPV-positive HNSCC. Tumor samples were collected at baseline and on treatment from 31 patients enrolled in a phase 2 trial of a 6-week cycle of CTLA-4 and PD-1 ICB induction followed by dose/volume-adapted IMRT (50-66Gy) concurrent with cycle 2. Tumors were evaluated for histological response and single-cell RNA sequencing (scRNA-seq) and scTCR-seq were performed on the 10X genomics platform. After stringent clustering and annotation of scRNA-seq data, 76,319 CD8+ and 78,622 CD4+ T cells with paired TCR sequence data were evaluable for transcriptomic, TCR repertoire and clonotype differentiation analysis. Results: Percent reduction in tumor viability on histology correlated significantly with reduction in tumor cell proportion in scRNA-seq data (R2= 0.7). Thirteen (48%) of 27 patients with paired samples (missing data due to COVID) had MHR, defined as ≤10% residual tumor viability on ICB treatment. Through TCR tracing, over 100 T cell clonotypes that expanded significantly (p &lt; 0.05, Fisher’s exact) post-ICB were identified. Highly expanded clonotypes resided mainly in exhausted CD8+ T cell clusters and exhibited a tissue resident memory (TRM) phenotype. Examination of their functional states revealed a variety of T cell modulatory trajectories, including reverse transition from exhausted to less exhausted states and from memory to functional states. When compared to patients without MHR, those with MHR had significantly higher ICB-induced TCR clonotype expansion (p = 0.025, Wilcoxon), reductions of activated Tregs (p = 0.003, Wilcoxon) and terminally exhausted CD8+ T cells (p = 0.003, Wilcoxon), and increase of polyfunctional CD8+ T cells (p = 0.014, Wilcoxon). A lasso regression model associated T cell expansion, basal 7-gene TRM and 5-gene effector scores, and on-treatment proliferation score and effector cell proportion with percent reduction in tumor viability (R2 = 0.66). Conclusions: In HPV-HNSCC, induction CTLA4 and PD1 ICB lead to extensive reinvigoration of CD8+ T cell clonotypes with both an exhausted and TRM phenotype, revealing a high degree of plasticity of distinct T cell clones. Paired RNA and TCR profiling facilitated identification of potentially tumor reactive and ICB-responsive T cells and clonotypes and identified several TIME features associated with tumor cell death. Clinical trial information: NCT03799445 .

Genomic–transcriptomic evolution in lung cancer and metastasis

Intratumour heterogeneity (ITH) fuels lung cancer evolution, which leads to immune evasion and resistance to therapy1. Here, using paired whole-exome and RNA sequencing data, we investigate intratumour transcriptomic diversity in 354 non-small cell lung cancer tumours from 347 out of the first 421 patients prospectively recruited into the TRACERx study2,3. Analyses of 947 tumour regions, representing both primary and metastatic disease, alongside 96 tumour-adjacent normal tissue samples implicate the transcriptome as a major source of phenotypic variation. Gene expression levels and ITH relate to patterns of positive and negative selection during tumour evolution. We observe frequent copy number-independent allele-specific expression that is linked to epigenomic dysfunction. Allele-specific expression can also result in genomic–transcriptomic parallel evolution, which converges on cancer gene disruption. We extract signatures of RNA single-base substitutions and link their aetiology to the activity of the RNA-editing enzymes ADAR and APOBEC3A, thereby revealing otherwise undetected ongoing APOBEC activity in tumours. Characterizing the transcriptomes of primary–metastatic tumour pairs, we combine multiple machine-learning approaches that leverage genomic and transcriptomic variables to link metastasis-seeding potential to the evolutionary context of mutations and increased proliferation within primary tumour regions. These results highlight the interplay between the genome and transcriptome in influencing ITH, lung cancer evolution and metastasis.

Abstract 6463: Persistence of smoking mutational signatures in the non-small cell lung cancer genome

Abstract Mutational signatures offer insights into cancer etiology. Smoking tobacco is associated with three signatures involving single base substitutions (SBS4), double base substitutions (DBS2) and insertions/deletions (ID3). Combining mutational signatures with clinical data allows examination of temporal and quantitative aspects of carcinogenic exposures. We analyzed 132 non-small cell lung carcinoma (NSCLC) whole genomes, sequenced through the 100,000 Genomes Project (Genomics England), with paired clinical data available for 130/132 (98.5%). Clinical data included smoking status, duration, pack-years (py), and duration of smoking cessation. The majority were “ever-smokers” (ex-smokers: 90/130, 69.2%; current smokers: 32/130, 24.6%) and 120/122 (98.4%) had at least one smoking signature. DBS2 was the most sensitive signature for ever-smokers (118/122, 96.7%), although SBS4 was the optimal signature for diagnosing smoking-related NSCLCs (SBS4 positive likelihood ratio 7.15). An absence of ID3 best identified true never smoker NSCLCs (ID3 negative likelihood ratio 0.08). Ever-smokers had a higher tumor mutational burden (TMB) than never smokers (median 8.25 mutations/Mb vs 1.18 mutations/Mb, p = &amp;lt;0.0001), while TMB positively correlated with smoking pack-years (p = 0.044). Following smoking cessation, the median length of discontinuation was 11 years before a smoking-related NSCLC was diagnosed and, in some cases, smoking signatures persisted for up to 50 years prior to NSCLC diagnosis. To complement existing epidemiological data, 40 pack years was estimated as a critical cut-off for minimizing ongoing NSCLC risk. Prolonged smoking cessation did not reduce TMB (p = 0.055) or SBS4 mutations (p = 0.11), although this did produce lower DBS2 counts (p = 0.0067). Further signatures observed included SBS92 (previously seen in the bladder urothelium of ever-smokers) and a novel signature hypothesized to be linked to DNA mismatch repair deficiencies. NSCLCs from two patients with prior platinum chemo-radiotherapy harbored signatures of both chemotherapy- and radiotherapy-induced damage in their recurrent biopsy sample. These results highlight the importance of paired sequencing and clinical data, revealing that smoking-induced DNA damage can have long-lived effects, and strengthening health guidance of abstinence from smoking at the earliest possibility. Citation Format: Colin R. Lindsay, Pantelis Nicola, Andreas Gruber, Kate Brown, Mathew Carter, Helen Adderley, Shereen Rafee, Anna Moss, Andrew Wallace, George Burghel, Helene Schlecht, Katie Baker, Sharzad Moghadam, Jane Rogan, Jamie Weaver, Angeliki Malliri, William Newman, Fiona Blackhall, David Wedge, Genomics England. Persistence of smoking mutational signatures in the non-small cell lung cancer genome. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 6463.

Abstract 5214: Expanding the spectrum of germline-driven cancers by leveraging population-scale targeted tumor and normal sequencing

Abstract Large case-control and familial studies have established clear cancer-specific risk profiles for several key cancer predisposition genes (CPGs). For example, germline pathogenic variants (PVs) in BRCA1/2 (gBRCA) are associated with increased risk for developing breast, ovarian, pancreatic, and prostate cancers. However, the extent to which gBRCA mutations are involved in mediating the tumorigenesis of other cancer types remains challenging to characterize. We hypothesized that integrating orthogonal features such as selection for biallelic inactivation of the PVs and depletion of canonical somatic drivers among the carriers can enrich the signal for identifying novel gene and cancer type associations. We then extend this framework to identify novel CPGs as well as to understand how tumors arise in patients with PVs in oncogenes. To study this, we leveraged the prospective MSK-IMPACT matched tumor-normal sequencing cohort of 49,291 patients across 77 major cancer types. We study 90 well-known CPGs as well as &amp;gt;300 cancer genes not previously associated with cancer predisposition. Overall, 8% (N=3,964) of patients harbored a PV in high or moderate penetrance CPGs. We identified 90 gene and cancer type associations with enrichment for biallelic inactivation (q&amp;lt;0.05), including 19 novel findings among clinically actionable genes. For example, we find enrichment for biallelic inactivation of BRCA1/2 in unexpected lineages such as hepatobiliary, endometrioid, and ampullary cancers. These tumors were also significantly depleted for somatic gain-of-function driver alterations. Hepatobiliary cancers with gBRCA1 mutations were also enriched for somatic loss of NF1. Among carriers of PVs in oncogenes, we observe two possible mechanisms of first somatic hit towards malignant transformation. We find enrichment for copy number gain or copy neutral loss of heterozygosity of the germline PV in thyroid cancers with a PV in RET. We also find that lung cancers with a germline PV in EGFR frequently developed additional somatic point mutations located in cis with the PV. Investigating genes with no prior association with germline predisposition to cancer, we find evidence for KEAP1 and CIC as likely novel CPGs. Lung (n=8) and thyroid (n=4) cancers with deleterious germline variants in KEAP1 were characterized by loss of the wild-type allele, co-occurring somatic STK11 mutations, and depletion of canonical drivers such as EGFR. We also found biallelic loss of CIC in two patients with Neuroblastoma, each carrying a different germline loss-of-function mutation in CIC. Both tumors were also negative for MYCN and ALK defects. Collectively, our findings expand our understanding of cancer predisposition in cancer, shed new insights into how tumors arise in germline carriers, and provide a framework for identifying new CPGs using population scale tumor-normal paired clinical sequencing data. Citation Format: Miika Mehine, Rebecca Caeser, Yelena Kemel, Daniel Muldoon, Sebastià Franch-Expósito, A. Rose Brannon, Aijazuddin Syed, Ozge Ceyhan-Birsoy, Maksym Misyura, Panieh Terraf, David B. Solit, Marc Ladanyi, Kenneth Offit, Zsofia K. Stadler, Diana L. Mandelker, Yonina R. Murciano-Goroff, Charles M. Rudin, Michael F. Berger, Chaitanya Bandlamudi. Expanding the spectrum of germline-driven cancers by leveraging population-scale targeted tumor and normal sequencing. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5214.

Clinical severity of Omicron subvariants BA.1, BA.2, and BA.5 in a population-based cohort study in British Columbia, Canada.

The SARS-CoV-2 variant Omicron emerged in late 2021. In British Columbia (BC), Canada, and globally, three genetically distinct subvariants of Omicron, BA.1, BA.2, and BA.5, emerged and became dominant successively within an 8-month period. SARS-CoV-2 subvariants continue to circulate in the population, acquiring new mutations that have the potential to alter infectivity, immunity, and disease severity. Here, we report a propensity-matched severity analysis from residents of BC over the course of the Omicron wave, including 39,237 individuals infected with BA.1, BA.2, or BA.5 based on paired high-quality sequence data and linked to comprehensive clinical outcomes data between December 23, 2021 and August 31, 2022. Relative to BA.1, BA.2 cases were associated with a 15% and 28% lower risk of hospitalization and intensive care unit (ICU) admission (aHRhospital = 1.17; 95% confidence interval [CI] = 1.096-1.252; aHRICU = 1.368; 95% CI = 1.152-1.624), whereas BA.5 infections were associated with an 18% higher risk of hospitalization (aHRhospital = 1.18; 95% CI = 1.133-1.224) after accounting for age, sex, comorbidities, vaccination status, geography, and social determinants of health. Phylogenetic analysis revealed no specific subclades associated with more severe clinical outcomes for any Omicron subvariant. In summary, BA.1, BA.2, and BA.5 subvariants were associated with differences in clinical severity, emphasizing how variant-specific monitoring programs remain critical components of patient and population-level public health responses as the pandemic continues.

EPCO-25. MULTI-OMIC ANALYSIS OF THE GLIOBLASTOMA EPIGENOME AND TRANSCRIPTOME INFORMS OF MIGRATORY INTERNEURON-LIKE DEVELOPMENTAL REGULATORS

Abstract Recent studies have demonstrated that, despite their nomenclature, gliomas recapitulate an interneuron progenitor-like state that drives tumor progression. During human neurodevelopment, interneurons arise from the subcortical ganglionic eminences and migrate tangentially into the neocortex, settling in the cortical plate where they integrate local neurocircuitry. Analogously, malignant glioblastoma (GBM) cells migrate from the tumor core into the surrounding healthy tissue. This innate infiltrative property renders these malignant cells elusive to surgical resection, leading to tumor recurrence. To understand the regulatory networks that drive tumor infiltration from a neurodevelopmental perspective, we generated a single-nucleus Assay for Transposase-Accessible Chromatin sequencing (snATAC-seq) dataset of 41,000 nuclei from the core and infiltrative edge of surgically resected GBM specimens (n = 4). Concurrently, we sequenced 46,000 nuclei from non-pathological, postmortem samples of second- and third-trimester neocortices (n = 17). We integrated these datasets with paired single-nucleus RNA sequencing (snRNA-seq) data and identified candidate regulatory TFs that exhibit high correlation between motif enrichment and TF expression. Using single-trajectory inference and pseudo-time analyses, we identified TCF12 as a potential driver of interneuron lineage fate in developing cortical progenitors. Given its implication in projection neuron migration, we were intrigued to find that TCF12 activity is highest in GBM cells with a migratory interneuron signature, hinting at its putative role in tumor infiltration. To understand the significance of these findings, we will interrogate other genes in the TCF12 regulatory network with the ultimate goal of identifying therapeutic targets that inhibit GBM infiltration.

DeepPerVar: a multi-modal deep learning framework for functional interpretation of genetic variants in personal genome.

Understanding the functional consequence of genetic variants, especially the non-coding ones, is important but particularly challenging. Genome-wide association studies (GWAS) or quantitative trait locus analyses may be subject to limited statistical power and linkage disequilibrium, and thus are less optimal to pinpoint the causal variants. Moreover, most existing machine-learning approaches, which exploit the functional annotations to interpret and prioritize putative causal variants, cannot accommodate the heterogeneity of personal genetic variations and traits in a population study, targeting a specific disease. By leveraging paired whole-genome sequencing data and epigenetic functional assays in a population study, we propose a multi-modal deep learning framework to predict genome-wide quantitative epigenetic signals by considering both personal genetic variations and traits. The proposed approach can further evaluate the functional consequence of non-coding variants on an individual level by quantifying the allelic difference of predicted epigenetic signals. By applying the approach to the ROSMAP cohort studying Alzheimer's disease (AD), we demonstrate that the proposed approach can accurately predict quantitative genome-wide epigenetic signals and in key genomic regions of AD causal genes, learn canonical motifs reported to regulate gene expression of AD causal genes, improve the partitioning heritability analysis and prioritize putative causal variants in a GWAS risk locus. Finally, we release the proposed deep learning model as a stand-alone Python toolkit and a web server. https://github.com/lichen-lab/DeepPerVar. Supplementary data are available at Bioinformatics online.

The relationship of mRNA with protein expression in CD8+ T cells associates with gene class and gene characteristics.

T cells are key players in our defence against infections and malignancies. When T cells differentiate or become activated, they undergo substantial alterations in gene expression. Even though RNA expression levels are now well documented throughout different stages of T cells, it is not well understood how mRNA expression translates into the protein landscape. By combining paired RNA sequencing and mass spectrometry data of primary human CD8+ T cells, we report that mRNA expression is a poor proxy for the overall protein output, irrespective of the differentiation or activation status. Yet, gene class stratification revealed a function-specific correlation of mRNA with protein expression. This gene class-specific expression pattern associated with differences in gene characteristics such as sequence conservation and untranslated region (UTR) lengths. In addition, the presence of AU-rich elements in the 3'UTR associated with alterations in mRNA and protein abundance T cell activation dependent, gene class-specific manner. In conclusion, our study highlights the role of gene characteristics as a determinant for gene expression in T cells.

Pangenomic analysis of Chinese gastric cancer

Pangenomic study might improve the completeness of human reference genome (GRCh38) and promote precision medicine. Here, we use an automated pipeline of human pangenomic analysis to build gastric cancer pan-genome for 185 paired deep sequencing data (370 samples), and characterize the gene presence-absence variations (PAVs) at whole genome level. Genes ACOT1, GSTM1, SIGLEC14 and UGT2B17 are identified as highly absent genes in gastric cancer population. A set of genes from unaligned sequences with GRCh38 are predicted. We successfully locate one of predicted genes GC0643 on chromosome 9q34.2. Overexpression of GC0643 significantly inhibits cell growth, cell migration and invasion, cell cycle progression, and induces cell apoptosis in cancer cells. The tumor suppressor functions can be reversed by shGC0643 knockdown. The GC0643 is approved by NCBI database (GenBank: MW194843.1). Collectively, the robust pan-genome strategy provides a deeper understanding of the gene PAVs in the human cancer genome.

One-stop molecular classification of endometrial carcinoma using comprehensive next-generation sequencing.

This study aims to investigate the feasibility of molecular classification using only comprehensive next-generation sequencing-based techniques and its relationship with survival outcomes in patients with endometrial cancer. Paired tumor-normal sequencing data of 1021 cancer-related genes using tumor tissues or peripheral blood samples and clinical data were retrospectively collected from a cohort of endometrial cancers. The microsatellite instability status was inferred using the MSIsensor (v0.5) with a cut-off of 8%. Sixty patients were classified into four groups: POLEMUT group (13.3%), MSI-H group (20%), TP53WT group (45%) and TP53MUT group (21.7%). Patients within TP53MUT group were more common in serous carcinoma compared to endometrioid carcinoma (P=.0098). TP53WT was significantly correlated with early stage and low grade. TP53MUT group was associated with significantly worse DFS compared to MSI-H group and TP53WT group (P=.014 and .004, respectively). Comprehensive next-generation sequencing is a reliable and simple method to stratify the prognosis of endometrial carcinoma. It can be potentially used to guide treatment of patients with endometrial cancer in routine practice.

Abstract 5042: Reconstructing the spatial evolution of cancer through subclone detection on copy number profiles in tumor sequencing data

Abstract Cancer results from somatic mutations such as copy number alterations (CNAs) that continue to accumulate during disease progression. These mutations can lead to functional heterogeneity within tumors and can influence the efficacy of cancer therapy. Therefore, studying the functional characteristics and spatial distribution of genetically distinct subclones is crucial to the understanding of tumor evolution and the design of cancer treatment. Here, we present Spatioscope, a method for subclone detection using copy number profiles that can be applied to spatial transcriptomics (ST) data and data from single-cell sequencing platforms such as scRNA-seq, scATAC-seq and scDNA-seq. Spatioscope implements a nested Chinese restaurant process, which mimics the tumor evolutionary process, to identify de novo subclones within one or multiple samples from the same patient. Spatioscope incorporates prior information from paired whole-genome or whole-exome sequencing (WGS/WES) data to achieve more reliable subclone detection and malignant cell labeling. We first applied Spatioscope on ST data from breast cancer, colorectal cancer and squamous cell carcinoma, as well as on scRNA-seq from three primary and metastatic gastrointestinal tumor samples. Spatioscope successfully distinguished malignant cells from stromal cells, and identified genetically distinct subclones which were validated using matched WGS/WES data, pathology annotations, or scDNA-seq data. On ST data, we show that Spatioscope accurately delineates the tumor’s invasive front, allowing for detailed characterization of interactions between stromal cells and malignant cells of different subclonal origins. In previous work, we showed the pervasive occurrence of highly complex subclonal allele-specific copy number alterations, and thus, we extended Spatioscope to identify subclones with different allele-specific copy number profiles. On three gastrointestinal tumor samples with scDNA-seq and two additional scATAC-seq datasets from a basal cell carcinoma and a gastric cancer cell line, Spatioscope successfully reconstructed complex recurrently mutated subclonal copy number regions. This is especially useful for data with sparse signals such as scATAC-seq when matched scDNA-seq data are unavailable. Using Spatioscope, we detected subclones based on copy number profiles in spatial and single cell tumor sequencing, enabling the investigation of the interplay between genome, transcriptome, and spatial environment during tumor evolution. Citation Format: Chi-Yun Wu, Paul R. Hess, Anuja Sathe, Jiazhen Rong, Billy T. Lau, Susan M. Grimes, Hanlee P. Ji, Nancy R. Zhang. Reconstructing the spatial evolution of cancer through subclone detection on copy number profiles in tumor sequencing data [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 5042.