γ-aminobutyric acid receptor B signaling drives glioblastoma in females in an immune-dependent manner.

Sex differences in immune responses impact cancer outcomes and treatment response, including in glioblastoma (GBM). However, host factors underlying sex specific immune-cancer interactions are poorly understood. Here, we identify the neurotransmitter γ-aminobutyric acid (GABA) as a driver of GBM-promoting immune response in females. We demonstrated that GABA receptor B (GABBR) signaling enhances L-Arginine metabolism and nitric oxide synthase 2 (NOS2) expression in female granulocytic myeloid-derived suppressor cells (gMDSCs). GABBR agonist and GABA analog promoted GBM growth in females in an immune-dependent manner, while GABBR inhibition reduces gMDSC NOS2 production and extends survival only in females. Furthermore, female GBM patients have enriched GABA transcriptional signatures compared to males, and the use of GABA analogs in GBM patients is associated with worse short-term outcomes only in females. Collectively, these results highlight that GABA modulates anti-tumor immune response in a sex-specific manner, supporting future assessment of GABA pathway inhibitors as part of immunotherapy approaches.

Identification and bioinformatic characterization of a serum miRNA signature for early detection of laryngeal squamous cell carcinoma.

BACKGROUND: The growing understanding of cancer biology and the establishment of new treatment modalities has not yielded the expected results in terms of survival for Laryngeal Squamous Cell Cancer (LSCC). Early diagnosis, as well as prompt identification of patients with high risk of relapse would ensure greater chance of therapeutic success. However, this goal remains a challenge due to the absence of specific biomarkers for this neoplasm. METHODS: Serum samples from 45 LSCC patients and 23 healthy donors were collected for miRNA expression profiling by TaqMan Array analysis. Additional 20 patients and 42 healthy volunteers were included for the validation set, reaching an equal number of clinical samples for each group. The potential diagnostic ability of the such identified three-miRNA signature was confirmed by ROC analysis. Moreover, each miRNA was analyzed for the possible correlation with HNSCC patients' survival and TNM status by online databases Kaplan-Meier (KM) plotter and OncomiR. In silico analysis of common candidate targets and their network relevance to predict shared biological functions was finally performed by PANTHER and GeneMANIA software. RESULTS: We characterized serum miRNA profile of LSCC patients identifying a novel molecular signature, including miR-223, miR-93 and miR-532, as circulating marker endowed with high selectivity and specificity. The oncogenic effect and the prognostic significance of each miRNA was investigated by bioinformatic analysis, denoting significant correlation with OS. To analyse the molecular basis underlying the pro-tumorigenic role of the signature, we focused on the simultaneously regulated gene targets-IL6ST, GTDC1, MAP1B, CPEB3, PRKACB, NFIB, PURB, ATP2B1, ZNF148, PSD3, TBC1D15, PURA, KLF12-found by prediction tools and deepened for their functional role by pathway enrichment analysis. The results showed the involvement of 7 different biological processes, among which inflammation, proliferation, migration, apoptosis and angiogenesis. CONCLUSIONS: In conclusion, we have identified a possible miRNA signature for early LSCC diagnosis and we assumed that miR-93, miR-223 and miR-532 could orchestrate the regulation of multiple cancer-related processes. These findings encourage the possibility to deepen the molecular mechanisms underlying their oncogenic role, for the desirable development of novel therapeutic opportunities based on the use of short single-stranded oligonucleotides acting as non-coding RNA antagonists in cancer.

Reinventing gene expression connectivity through regulatory and spatial structural empowerment via principal node aggregation graph neural network.

The intricacies of the human genome, manifested as a complex network of genes, transcend conventional representations in text or numerical matrices. The intricate gene-to-gene relationships inherent in this complexity find a more suitable depiction in graph structures. In the pursuit of predicting gene expression, an endeavor shared by predecessors like the L1000 and Enformer methods, we introduce a novel spatial graph-neural network (GNN) approach. This innovative strategy incorporates graph features, encompassing both regulatory and structural elements. The regulatory elements include pair-wise gene correlation, biological pathways, protein-protein interaction networks, and transcription factor regulation. The spatial structural elements include chromosomal distance, histone modification and Hi-C inferred 3D genomic features. Principal Node Aggregation models, validated independently, emerge as frontrunners, demonstrating superior performance compared to traditional regression and other deep learning models. By embracing the spatial GNN paradigm, our method significantly advances the description of the intricate network of gene interactions, surpassing the performance, predictable scope, and initial requirements set by previous methods.

Quantifying intra-tumoral genetic heterogeneity of glioblastoma toward precision medicine using MRI and a data-inclusive machine learning algorithm.

BACKGROUND AND OBJECTIVE: Glioblastoma (GBM) is one of the most aggressive and lethal human cancers. Intra-tumoral genetic heterogeneity poses a significant challenge for treatment. Biopsy is invasive, which motivates the development of non-invasive, MRI-based machine learning (ML) models to quantify intra-tumoral genetic heterogeneity for each patient. This capability holds great promise for enabling better therapeutic selection to improve patient outcome. METHODS: We proposed a novel Weakly Supervised Ordinal Support Vector Machine (WSO-SVM) to predict regional genetic alteration status within each GBM tumor using MRI. WSO-SVM was applied to a unique dataset of 318 image-localized biopsies with spatially matched multiparametric MRI from 74 GBM patients. The model was trained to predict the regional genetic alteration of three GBM driver genes (EGFR, PDGFRA and PTEN) based on features extracted from the corresponding region of five MRI contrast images. For comparison, a variety of existing ML algorithms were also applied. Classification accuracy of each gene were compared between the different algorithms. The SHapley Additive exPlanations (SHAP) method was further applied to compute contribution scores of different contrast images. Finally, the trained WSO-SVM was used to generate prediction maps within the tumoral area of each patient to help visualize the intra-tumoral genetic heterogeneity. RESULTS: WSO-SVM achieved 0.80 accuracy, 0.79 sensitivity, and 0.81 specificity for classifying EGFR; 0.71 accuracy, 0.70 sensitivity, and 0.72 specificity for classifying PDGFRA; 0.80 accuracy, 0.78 sensitivity, and 0.83 specificity for classifying PTEN; these results significantly outperformed the existing ML algorithms. Using SHAP, we found that the relative contributions of the five contrast images differ between genes, which are consistent with findings in the literature. The prediction maps revealed extensive intra-tumoral region-to-region heterogeneity within each individual tumor in terms of the alteration status of the three genes. CONCLUSIONS: This study demonstrated the feasibility of using MRI and WSO-SVM to enable non-invasive prediction of intra-tumoral regional genetic alteration for each GBM patient, which can inform future adaptive therapies for individualized oncology.

Mutational landscape of inflammatory breast cancer.

BACKGROUND: Inflammatory breast cancer (IBC) is the most pro-metastatic form of BC. Better understanding of its enigmatic pathophysiology is crucial. We report here the largest whole-exome sequencing (WES) study of clinical IBC samples. METHODS: We retrospectively applied WES to 54 untreated IBC primary tumor samples and matched normal DNA. The comparator samples were 102 stage-matched non-IBC samples from TCGA. We compared the somatic mutational profiles, spectra and signatures, copy number alterations (CNAs), HRD and heterogeneity scores, and frequencies of actionable genomic alterations (AGAs) between IBCs and non-IBCs. The comparisons were adjusted for the molecular subtypes. RESULTS: The number of somatic mutations, TMB, and mutational spectra were not different between IBCs and non-IBCs, and no gene was differentially mutated or showed differential frequency of CNAs. Among the COSMIC signatures, only the age-related signature was more frequent in non-IBCs than in IBCs. We also identified in IBCs two new mutational signatures not associated with any environmental exposure, one of them having been previously related to HIF pathway activation. Overall, the HRD score was not different between both groups, but was higher in TN IBCs than TN non-IBCs. IBCs were less frequently classified as heterogeneous according to heterogeneity H-index than non-IBCs (21% vs 33%), and clonal mutations were more frequent and subclonal mutations less frequent in IBCs. More than 50% of patients with IBC harbored at least one high-level of evidence (LOE) AGA (OncoKB LOE 1-2, ESCAT LOE I-II), similarly to patients with non-IBC. CONCLUSIONS: We provide the largest mutational landscape of IBC. Only a few subtle differences were identified with non-IBCs. The most clinically relevant one was the higher HRD score in TN IBCs than in TN non-IBCs, whereas the most intriguing one was the smaller intratumor heterogeneity of IBCs.

Epigenetic remodeling to improve the efficacy of immunotherapy in human glioblastoma: pre-clinical evidence for development of new immunotherapy approaches.

BACKGROUND: Glioblastoma multiforme (GBM) is a highly aggressive primary brain tumor, that is refractory to standard treatment and to immunotherapy with immune-checkpoint inhibitors (ICI). Noteworthy, melanoma brain metastases (MM-BM), that share the same niche as GBM, frequently respond to current ICI therapies. Epigenetic modifications regulate GBM cellular proliferation, invasion, and prognosis and may negatively regulate the cross-talk between malignant cells and immune cells in the tumor milieu, likely contributing to limit the efficacy of ICI therapy of GBM. Thus, manipulating the tumor epigenome can be considered a therapeutic opportunity in GBM. METHODS: Microarray transcriptional and methylation profiles, followed by gene set enrichment and IPA analyses, were performed to study the differences in the constitutive expression profiles of GBM vs MM-BM cells, compared to the extracranial MM cells and to investigate the modulatory effects of the DNA hypomethylating agent (DHA) guadecitabine among the different tumor cells. The prognostic relevance of DHA-modulated genes was tested by Cox analysis in a TCGA GBM patients' cohort. RESULTS: The most striking differences between GBM and MM-BM cells were found to be the enrichment of biological processes associated with tumor growth, invasion, and extravasation with the inhibition of MHC class II antigen processing/presentation in GBM cells. Treatment with guadecitabine reduced these biological differences, shaping GBM cells towards a more immunogenic phenotype. Indeed, in GBM cells, promoter hypomethylation by guadecitabine led to the up-regulation of genes mainly associated with activation, proliferation, and migration of T and B cells and with MHC class II antigen processing/presentation. Among DHA-modulated genes in GBM, 7.6% showed a significant prognostic relevance. Moreover, a large set of immune-related upstream-regulators (URs) were commonly modulated by DHA in GBM, MM-BM, and MM cells: DHA-activated URs enriched for biological processes mainly involved in the regulation of cytokines and chemokines production, inflammatory response, and in Type I/II/III IFN-mediated signaling; conversely, DHA-inhibited URs were involved in metabolic and proliferative pathways. CONCLUSIONS: Epigenetic remodeling by guadecitabine represents a promising strategy to increase the efficacy of cancer immunotherapy of GBM, supporting the rationale to develop new epigenetic-based immunotherapeutic approaches for the treatment of this still highly deadly disease.

Identification of therapeutic targets in osteoarthritis by combining heterogeneous transcriptional datasets, drug-induced expression profiles, and known drug-target interactions.

BACKGROUND: Osteoarthritis (OA) is a multifactorial, hypertrophic, and degenerative condition involving the whole joint and affecting a high percentage of middle-aged people. It is due to a combination of factors, although the pivotal mechanisms underlying the disease are still obscure. Moreover, current treatments are still poorly effective, and patients experience a painful and degenerative disease course. METHODS: We used an integrative approach that led us to extract a consensus signature from a meta-analysis of three different OA cohorts. We performed a network-based drug prioritization to detect the most relevant drugs targeting these genes and validated in vitro the most promising candidates. We also proposed a risk score based on a minimal set of genes to predict the OA clinical stage from RNA-Seq data. RESULTS: We derived a consensus signature of 44 genes that we validated on an independent dataset. Using network analysis, we identified Resveratrol, Tenoxicam, Benzbromarone, Pirinixic Acid, and Mesalazine as putative drugs of interest for therapeutics in OA for anti-inflammatory properties. We also derived a list of seven gene-targets validated with functional RT-qPCR assays, confirming the in silico predictions. Finally, we identified a predictive subset of genes composed of DNER, TNFSF11, THBS3, LOXL3, TSPAN2, DYSF, ASPN and HTRA1 to compute the patient's risk score. We validated this risk score on an independent dataset with a high AUC (0.875) and compared it with the same approach computed using the entire consensus signature (AUC 0.922). CONCLUSIONS: The consensus signature highlights crucial mechanisms for disease progression. Moreover, these genes were associated with several candidate drugs that could represent potential innovative therapeutics. Furthermore, the patient's risk scores can be used in clinical settings.

MiR-449a antagonizes EMT through IL-6-mediated trans-signaling in laryngeal squamous cancer.

MicroRNAs (miRNAs) are involved in post-transcriptional gene expression regulation and in mechanisms of cancer growth and metastases. In this light, miRNAs could be promising therapeutic targets and biomarkers in clinical practice. Therefore, we investigated if specific miRNAs and their target genes contribute to laryngeal squamous cell carcinoma (LSCC) development. We found a significant decrease of miR-449a in LSCC patients with nodal metastases (63.3%) compared with patients without nodal involvement (44%). The AmpliSeq Transcriptome of HNO-210 miR-449a-transfected cell lines allowed the identification of IL6-R as a potential target. Moreover, the downregulation of IL6-R and the phosphorylation reduction of the downstream signaling effectors, suggested the inhibition of the IL-6 trans-signaling pathway. These biochemical effects were paralleled by a significant inhibition of invasion and migration in vitro and in vivo, supporting an involvement of epithelial-mesenchymal transition. These findings indicate that miR-449a contributes to suppress the metastasization of LSCC by the IL-6 trans-signaling block and affects sensitivity to external stimuli that mimic pro-inflammatory conditions.

A community challenge to predict clinical outcomes after immune checkpoint blockade in non-small cell lung cancer.

BACKGROUND: Predictive biomarkers of immune checkpoint inhibitor (ICI) efficacy are currently lacking for non-small cell lung cancer (NSCLC). Here, we describe the results from the Anti-PD-1 Response Prediction DREAM Challenge, a crowdsourced initiative that enabled the assessment of predictive models by using data from two randomized controlled clinical trials (RCTs) of ICIs in first-line metastatic NSCLC. METHODS: Participants developed and trained models using public resources. These were evaluated with data from the CheckMate 026 trial (NCT02041533), according to the model-to-data paradigm to maintain patient confidentiality. The generalizability of the models with the best predictive performance was assessed using data from the CheckMate 227 trial (NCT02477826). Both trials were phase III RCTs with a chemotherapy control arm, which supported the differentiation between predictive and prognostic models. Isolated model containers were evaluated using a bespoke strategy that considered the challenges of handling transcriptome data from clinical trials. RESULTS: A total of 59 teams participated, with 417 models submitted. Multiple predictive models, as opposed to a prognostic model, were generated for predicting overall survival, progression-free survival, and progressive disease status with ICIs. Variables within the models submitted by participants included tumor mutational burden (TMB), programmed death ligand 1 (PD-L1) expression, and gene-expression-based signatures. The best-performing models showed improved predictive power over reference variables, including TMB or PD-L1. CONCLUSIONS: This DREAM Challenge is the first successful attempt to use protected phase III clinical data for a crowdsourced effort towards generating predictive models for ICI clinical outcomes and could serve as a blueprint for similar efforts in other tumor types and disease states, setting a benchmark for future studies aiming to identify biomarkers predictive of ICI efficacy. TRIAL REGISTRATION: CheckMate 026; NCT02041533, registered January 22, 2014. CheckMate 227; NCT02477826, registered June 23, 2015.

Decomprolute is a benchmarking platform designed for multiomics-based tumor deconvolution.

Tumor deconvolution enables the identification of diverse cell types that comprise solid tumors. To date, however, both the algorithms developed to deconvolve tumor samples, and the gold-standard datasets used to assess the algorithms are geared toward the analysis of gene expression (e.g., RNA sequencing) rather than protein levels. Despite the popularity of gene expression datasets, protein levels often provide a more accurate view of rare cell types. To facilitate the use, development, and reproducibility of multiomic deconvolution algorithms, we introduce Decomprolute, a Common Workflow Language framework that leverages containerization to compare tumor deconvolution algorithms across multiomic datasets. Decomprolute incorporates the large-scale multiomic datasets produced by the Clinical Proteomic Tumor Analysis Consortium (CPTAC), which include matched mRNA expression and proteomic data from thousands of tumors across multiple cancer types to build a fully open-source, containerized proteogenomic tumor deconvolution benchmarking platform. http://pnnl-compbio.github.io/decomprolute.

The epitranscriptome of high-grade gliomas: a promising therapeutic target with implications from the tumor microenvironment to endogenous retroviruses.

Glioblastoma (GBM) comprises 45.6% of all primary malignant brain cancers and is one of the most common and aggressive intracranial tumors in adults. Intratumoral heterogeneity with a wide range of proteomic, genetic, and epigenetic dysregulation contributes to treatment resistance and poor prognosis, thus demanding novel therapeutic approaches. To date, numerous clinical trials have been developed to target the proteome and epigenome of high-grade gliomas with promising results. However, studying RNA modifications, or RNA epitranscriptomics, is a new frontier within neuro-oncology. RNA epitranscriptomics was discovered in the 1970s, but in the last decade, the extent of modification of mRNA and various non-coding RNAs has emerged and been implicated in transposable element activation and many other oncogenic processes within the tumor microenvironment. This review provides background information and discusses the therapeutic potential of agents modulating epitranscriptomics in high-grade gliomas. A particular emphasis will be placed on how combination therapies that include immune agents targeting hERV-mediated viral mimicry could improve the treatment of GBM.

Guadecitabine plus ipilimumab in unresectable melanoma: five-year follow-up and integrated multi-omic analysis in the phase 1b NIBIT-M4 trial.

Association with hypomethylating agents is a promising strategy to improve the efficacy of immune checkpoint inhibitors-based therapy. The NIBIT-M4 was a phase Ib, dose-escalation trial in patients with advanced melanoma of the hypomethylating agent guadecitabine combined with the anti-CTLA-4 antibody ipilimumab that followed a traditional 3 + 3 design (NCT02608437). Patients received guadecitabine 30, 45 or 60 mg/m2/day subcutaneously on days 1 to 5 every 3 weeks starting on week 0 for a total of four cycles, and ipilimumab 3 mg/kg intravenously starting on day 1 of week 1 every 3 weeks for a total of four cycles. Primary outcomes of safety, tolerability, and maximum tolerated dose of treatment were previously reported. Here we report the 5-year clinical outcome for the secondary endpoints of overall survival, progression free survival, and duration of response, and an exploratory integrated multi-omics analysis on pre- and on-treatment tumor biopsies. With a minimum follow-up of 45 months, the 5-year overall survival rate was 28.9% and the median duration of response was 20.6 months. Re-expression of immuno-modulatory endogenous retroviruses and of other repetitive elements, and a mechanistic signature of guadecitabine are associated with response. Integration of a genetic immunoediting index with an adaptive immunity signature stratifies patients/lesions into four distinct subsets and discriminates 5-year overall survival and progression free survival. These results suggest that coupling genetic immunoediting with activation of adaptive immunity is a relevant requisite for achieving long term clinical benefit by epigenetic immunomodulation in advanced melanoma patients.

Immunologic constant of rejection as a predictive biomarker of immune checkpoint inhibitors efficacy in non-small cell lung cancer.

BACKGROUND: Anti-PD1/PDL1 immune checkpoint inhibitors (ICI) transformed the prognosis of patients with advanced non-small cell lung cancer (NSCLC). However, the response rate remains disappointing and toxicity may be life-threatening, making urgent identification of biomarkers predictive for efficacy. Immunologic Constant of Rejection signature (ICR) is a 20-gene expression signature of cytotoxic immune response with prognostic value in some solid cancers. Our objective was to assess its predictive value for benefit from anti-PD1/PDL1 in patients with advanced NSCLC. METHODS: We retrospectively profiled 44 primary tumors derived from NSCLC patients treated with ICI as single-agent in at least the second-line metastatic setting. Transcriptomic analysis was performed using the nCounter® analysis system and the PanCancer Immune Profiling Panel. We then pooled our data with clinico-biological data from four public gene expression data sets, leading to a total of 162 NSCLC patients treated with single-agent anti-PD1/PDL1. ICR was applied to all samples and correlation was searched between ICR classes and the Durable Clinical Benefit (DCB), defined as stable disease or objective response according to RECIST 1.1 for a minimum of 6 months after the start of ICI. RESULTS: The DCB rate was 29%; 22% of samples were classified as ICR1, 30% ICR2, 22% ICR3, and 26% ICR4. These classes were not associated with the clinico-pathological variables, but showed enrichment from ICR1 to ICR4 in quantitative/qualitative markers of immune response. ICR2-4 class was associated with a 5.65-fold DCB rate when compared with ICR1 class. In multivariate analysis, ICR classification remained associated with DCB, independently from PDL1 expression and other predictive immune signatures. By contrast, it was not associated with disease-free survival in 556 NSCLC TCGA patients untreated with ICI. CONCLUSION: The 20-gene ICR signature was independently associated with benefit from anti-PD1/PDL1 ICI in patients with advanced NSCLC. Validation in larger retrospective and prospective series is warranted.

Pancancer network analysis reveals key master regulators for cancer invasiveness.

BACKGROUND: Tumor invasiveness reflects numerous biological changes, including tumorigenesis, progression, and metastasis. To decipher the role of transcriptional regulators (TR) involved in tumor invasiveness, we performed a systematic network-based pan-cancer assessment of master regulators of cancer invasiveness. MATERIALS AND METHODS: We stratified patients in The Cancer Genome Atlas (TCGA) into invasiveness high (INV-H) and low (INV-L) groups using consensus clustering based on an established robust 24-gene signature to determine the prognostic association of invasiveness with overall survival (OS) across 32 different cancers. We devise a network-based protocol to identify TRs as master regulators (MRs) unique to INV-H and INV-L phenotypes. We validated the activity of MRs coherently associated with INV-H phenotype and worse OS across cancers in TCGA on a series of additional datasets in the Prediction of Clinical Outcomes from the Genomic Profiles (PRECOG) repository. RESULTS: Based on the 24-gene signature, we defined the invasiveness score for each patient sample and stratified patients into INV-H and INV-L clusters. We observed that invasiveness was associated with worse survival outcomes in almost all cancers and had a significant association with OS in ten out of 32 cancers. Our network-based framework identified common invasiveness-associated MRs specific to INV-H and INV-L groups across the ten prognostic cancers, including COL1A1, which is also part of the 24-gene signature, thus acting as a positive control. Downstream pathway analysis of MRs specific to INV-H phenotype resulted in the identification of several enriched pathways, including Epithelial into Mesenchymal Transition, TGF-ß signaling pathway, regulation of Toll-like receptors, cytokines, and inflammatory response, and selective expression of chemokine receptors during T-cell polarization. Most of these pathways have connotations of inflammatory immune response and feasibility for metastasis. CONCLUSION: Our pan-cancer study provides a comprehensive master regulator analysis of tumor invasiveness and can suggest more precise therapeutic strategies by targeting the identified MRs and downstream enriched pathways for patients across multiple cancers.

MOViDA: multiomics visible drug activity prediction with a biologically informed neural network model.

MOTIVATION: The process of drug development is inherently complex, marked by extended intervals from the inception of a pharmaceutical agent to its eventual launch in the market. Additionally, each phase in this process is associated with a significant failure rate, amplifying the inherent challenges of this task. Computational virtual screening powered by machine learning algorithms has emerged as a promising approach for predicting therapeutic efficacy. However, the complex relationships between the features learned by these algorithms can be challenging to decipher. RESULTS: We have engineered an artificial neural network model designed specifically for predicting drug sensitivity. This model utilizes a biologically informed visible neural network, thereby enhancing its interpretability. The trained model allows for an in-depth exploration of the biological pathways integral to prediction and the chemical attributes of drugs that impact sensitivity. Our model harnesses multiomics data derived from a different tumor tissue sources, as well as molecular descriptors that encapsulate the properties of drugs. We extended the model to predict drug synergy, resulting in favorable outcomes while retaining interpretability. Given the imbalanced nature of publicly available drug screening datasets, our model demonstrated superior performance to state-of-the-art visible machine learning algorithms. AVAILABILITY AND IMPLEMENTATION: MOViDA is implemented in Python using PyTorch library and freely available for download at https://github.com/Luigi-Ferraro/MOViDA. Training data, RIS score and drug features are archived on Zenodo https://doi.org/10.5281/zenodo.8180380.

An integrated tumor, immune and microbiome atlas of colon cancer.

The lack of multi-omics cancer datasets with extensive follow-up information hinders the identification of accurate biomarkers of clinical outcome. In this cohort study, we performed comprehensive genomic analyses on fresh-frozen samples from 348 patients affected by primary colon cancer, encompassing RNA, whole-exome, deep T cell receptor and 16S bacterial rRNA gene sequencing on tumor and matched healthy colon tissue, complemented with tumor whole-genome sequencing for further microbiome characterization. A type 1 helper T cell, cytotoxic, gene expression signature, called Immunologic Constant of Rejection, captured the presence of clonally expanded, tumor-enriched T cell clones and outperformed conventional prognostic molecular biomarkers, such as the consensus molecular subtype and the microsatellite instability classifications. Quantification of genetic immunoediting, defined as a lower number of neoantigens than expected, further refined its prognostic value. We identified a microbiome signature, driven by Ruminococcus bromii, associated with a favorable outcome. By combining microbiome signature and Immunologic Constant of Rejection, we developed and validated a composite score (mICRoScore), which identifies a group of patients with excellent survival probability. The publicly available multi-omics dataset provides a resource for better understanding colon cancer biology that could facilitate the discovery of personalized therapeutic approaches.

A variational algorithm to detect the clonal copy number substructure of tumors from scRNA-seq data.

Single-cell RNA sequencing is the reference technology to characterize the composition of the tumor microenvironment and to study tumor heterogeneity at high resolution. Here we report Single CEll Variational ANeuploidy analysis (SCEVAN), a fast variational algorithm for the deconvolution of the clonal substructure of tumors from single-cell RNA-seq data. It uses a multichannel segmentation algorithm exploiting the assumption that all the cells in a given copy number clone share the same breakpoints. Thus, the smoothed expression profile of every individual cell constitutes part of the evidence of the copy number profile in each subclone. SCEVAN can automatically and accurately discriminate between malignant and non-malignant cells, resulting in a practical framework to analyze tumors and their microenvironment. We apply SCEVAN to datasets encompassing 106 samples and 93,322 cells from different tumor types and technologies. We demonstrate its application to characterize the intratumor heterogeneity and geographic evolution of malignant brain tumors.

Integrative multi-omics networks identify PKCδ and DNA-PK as master kinases of glioblastoma subtypes and guide targeted cancer therapy.

Despite producing a panoply of potential cancer-specific targets, the proteogenomic characterization of human tumors has yet to demonstrate value for precision cancer medicine. Integrative multi-omics using a machine-learning network identified master kinases responsible for effecting phenotypic hallmarks of functional glioblastoma subtypes. In subtype-matched patient-derived models, we validated PKCδ and DNA-PK as master kinases of glycolytic/plurimetabolic and proliferative/progenitor subtypes, respectively, and qualified the kinases as potent and actionable glioblastoma subtype-specific therapeutic targets. Glioblastoma subtypes were associated with clinical and radiomics features, orthogonally validated by proteomics, phospho-proteomics, metabolomics, lipidomics and acetylomics analyses, and recapitulated in pediatric glioma, breast and lung squamous cell carcinoma, including subtype specificity of PKCδ and DNA-PK activity. We developed a probabilistic classification tool that performs optimally with RNA from frozen and paraffin-embedded tissues, which can be used to evaluate the association of therapeutic response with glioblastoma subtypes and to inform patient selection in prospective clinical trials.

Fast, accurate, and racially unbiased pan-cancer tumor-only variant calling with tabular machine learning.

Accurately identifying somatic mutations is essential for precision oncology and crucial for calculating tumor-mutational burden (TMB), an important predictor of response to immunotherapy. For tumor-only variant calling (i.e., when the cancer biopsy but not the patient's normal tissue sample is sequenced), accurately distinguishing somatic mutations from germline variants is a challenging problem that, when unaddressed, results in unreliable, biased, and inflated TMB estimates. Here, we apply machine learning to the task of somatic vs germline classification in tumor-only solid tumor samples using TabNet, XGBoost, and LightGBM, three machine-learning models for tabular data. We constructed a training set for supervised classification using features derived exclusively from tumor-only variant calling and drawing somatic and germline truth labels from an independent pipeline using the patient-matched normal samples. All three trained models achieved state-of-the-art performance on two holdout test datasets: a TCGA dataset including sarcoma, breast adenocarcinoma, and endometrial carcinoma samples (AUC > 94%), and a metastatic melanoma dataset (AUC > 85%). Concordance between matched-normal and tumor-only TMB improves from R2 = 0.006 to 0.71-0.76 with the addition of a machine-learning classifier, with LightGBM performing best. Notably, these machine-learning models generalize across cancer subtypes and capture kits with a call rate of 100%. We reproduce the recent finding that tumor-only TMB estimates for Black patients are extremely inflated relative to that of white patients due to the racial biases of germline databases. We show that our approach with XGBoost and LightGBM eliminates this significant racial bias in tumor-only variant calling.

Pan-cancer onco-signatures reveal a novel mitochondrial subtype of luminal breast cancer with specific regulators.

BACKGROUND: Somatic alterations in cancer cause dysregulation of signaling pathways that control cell-cycle progression, apoptosis, and cell growth. The effect of individual alterations in these pathways differs between individual tumors and tumor types. Recognizing driver events is a complex task requiring integrating multiple molecular data, including genomics, epigenomics, and functional genomics. A common hypothesis is that these driver events share similar effects on the hallmarks of cancer. The availability of large-scale multi-omics studies allows for inferring these common effects from data. Once these effects are known, one can then deconvolve in every individual patient whether a given genomics alteration is a driver event. METHODS: Here, we develop a novel data-driven approach to identify shared oncogenic expression signatures among tumors. We aim to identify gene onco-signature for classifying tumor patients in homogeneous subclasses with distinct prognoses and specific genomic alterations. We derive expression pan-cancer onco-signatures from TCGA gene expression data using a discovery set of 9107 primary pan-tumor samples together with respective matched mutational data and a list of known cancer-related genes from COSMIC database. RESULTS: We use the derived ono-signatures to state their prognostic significance and apply them to the TCGA breast cancer dataset as proof of principle of our approach. We uncover a "mitochondrial" sub-group of Luminal patients characterized by its biological features and regulated by specific genetic modulators. Collectively, our results demonstrate the effectiveness of onco-signatures-based methodologies, and they also contribute to a comprehensive understanding of the metabolic heterogeneity of Luminal tumors. CONCLUSIONS: These findings provide novel genomics evidence for developing personalized breast cancer patient treatments. The onco-signature approach, demonstrated here on breast cancer, is general and can be applied to other cancer types.