DNA polymerase ε and δ variants drive mutagenesis in polypurine tracts in human tumors.

Alterations in the exonuclease domain of DNA polymerase ε cause ultramutated cancers. These cancers accumulate AGA>ATA transversions; however, their genomic features beyond the trinucleotide motifs are obscure. We analyze the extended DNA context of ultramutation using whole-exome sequencing data from 524 endometrial and 395 colorectal tumors. We find that G>T transversions in POLE-mutant tumors predominantly affect sequences containing at least six consecutive purines, with a striking preference for certain positions within polypurine tracts. Using this signature, we develop a machine-learning classifier to identify tumors with hitherto unknown POLE drivers and validate two drivers, POLE-E978G and POLE-S461L, by functional assays in yeast. Unlike other pathogenic variants, the E978G substitution affects the polymerase domain of Pol ε. We further show that tumors with POLD1 drivers share the extended signature of POLE ultramutation. These findings expand the understanding of ultramutation mechanisms and highlight peculiar mutagenic properties of polypurine tracts in the human genome.

Establishing the role of the FES tyrosine kinase in the pathogenesis, pathophysiology, and severity of sepsis and its outcomes.

Introduction: Sepsis is a result of initial over-activation of the immune system in response to an infection or trauma that results in reduced blood flow and life-threatening end-organ damage, followed by suppression of the immune system that prevents proper clearance of the infection or trauma. Because of this, therapies that not only limit the activation of the immune system early on, but also improve blood flow to crucial organs and reactivate the immune system in late-stage sepsis, may be effective treatments. The tyrosine kinase FES may fulfill this role. FES is present in immune cells and serves to limit immune system activation. We hypothesize that by enhancing FES in early sepsis and inhibiting its effects in late sepsis, the severity and outcome of septic illness can be improved. Methods and analysis: In vitro and in vivo modeling will be performed to determine the degree of inflammatory signaling, cytokine production, and neutrophil extracellular trap (NET) formation that occurs in wild-type (WT) and FES knockout (FES-/- ) mice. Clinically available treatments known to enhance or inhibit FES expression (lorlatinib and decitabine, respectively), will be used to explore the impact of early vs. late FES modulation on outcomes in WT mice. Bioinformatic analysis will be performed to examine FES expression levels in RNA transcriptomic data from sepsis patient cohorts, and correlate FES expression data with clinical outcomes (diagnosis of sepsis, illness severity, hospital length-of-stay). Ethics and dissemination: Ethics approval pending from the Queen's University Health Sciences & Affiliated Teaching Hospitals Research Ethics Board. Results will be disseminated through scientific publications and through lay summaries to patients and families.

Cancer cell genotype associated tumor immune microenvironment exhibits differential response to therapeutic STING pathway activation in high-grade serous ovarian cancer.

BACKGROUND: High-grade serous ovarian carcinoma (HGSC) is the most lethal gynecologic malignancy characterized by resistance to chemotherapy and high rates of recurrence. HGSC tumors display a high prevalence of tumor suppressor gene loss. Given the type 1 interferon regulatory function of BRCA1 and PTENgenes and their associated contrasting T-cell infiltrated and non-infiltrated tumor immune microenvironment (TIME) states, respectively, in this study we investigated the potential of stimulator of interferon genes (STING) pathway activation in improving overall survival via enhancing chemotherapy response, specifically in tumors with PTEN deficiency. METHODS: Expression of PTEN protein was evaluated in tissue microarrays generated using pretreatment tumors collected from a cohort of 110 patients with HGSC. Multiplex immunofluorescence staining was performed to determine spatial profiles and density of selected lymphoid and myeloid cells. In vivo studies using the syngeneic murine HGSC cell lines, ID8-Trp53 -/-; Pten -/- and ID8-Trp53 -/-; Brca1 -/-, were conducted to characterize the TIME and response to carboplatin chemotherapy in combination with exogenous STING activation therapy. RESULTS: Patient tumors with absence of PTEN protein exhibited a significantly decreased disease specific survival and intraepithelial CD68+ macrophage infiltration as compared with intact PTEN expression. In vivo studies demonstrated that Pten-deficient ovarian cancer cells establish an immunosuppressed TIME characterized by increased proportions of M2-like macrophages, GR1+MDSCs in the ascites, and reduced effector CD8+ cytotoxic T-cell function compared with Brca1-deficient cells; further, tumors from mice injected with Pten-deficient ID8 cells exhibited an aggressive behavior due to suppressive macrophage dominance in the malignant ascites. In combination with chemotherapy, exogenous STING activation resulted in longer overall survival in mice injected with Pten-deficient ID8 cells, reprogrammed intraperitoneal M2-like macrophages derived from Pten-deficient ascites to M1-like phenotype and rescued CD8+ cytotoxic T-cell activation. CONCLUSIONS: This study reveals the importance of considering the influence of cancer cell intrinsic genetic alterations on the TIME for therapeutic selection. We establish the rationale for the optimal incorporation of interferon activating therapies as a novel combination strategy in PTEN-deficient HGSC.

Expression of TCF3 target genes defines a subclass of diffuse large B-cell lymphoma characterized by up-regulation of MYC target genes and poor clinical outcome following R-CHOP therapy.

TCF3 is a lymphopoietic transcription factor that acquires somatic driver mutations in diffuse large B-cell lymphoma (DLBCL). Hypothesizing that expression patterns of TCF3-regulated genes can inform clinical management, we found that unsupervised clustering analysis with 15 TCF3-regulated genes and eight additional ones resolved local DLBCL cases into two main clusters, denoted Groups A and B, of which Group A manifested inferior overall survival (OS, p = 0.0005). We trained a machine learning model to classify samples into the Groups based on expression of the 23 transcripts in an independent validation cohort of 569 R-CHOP-treated DLBCL cases. Group A overlapped with the ABC cell-of-origin subgroup but its prognostic power was superior. GSEA analysis demonstrated asymmetric expression of 30 gene sets between the Groups, pointing to biological differences. We present, validate and make available a novel method to assign DLBCL cases into biologically-distinct groups with divergent OS following R-CHOP therapy.

A transcriptome analysis of basal and stimulated VWF release from endothelial cells derived from patients with type 1 VWD.

Type 1 von Willebrand disease (VWD) is associated with a reduction in qualitatively normal von Willebrand factor (VWF). Current diagnostic guidelines only take into consideration the contribution of basal VWF levels, despite a lack of correlation with bleeding severity. Defects in stimulated VWF release, which occurs after hemostatic challenge, may contribute to bleeding in type 1 VWD, but the pathogenic mechanisms are poorly defined. In this study, a layered multiomic approach including messenger RNA (mRNA) and microRNA (miRNA) sequencing was used to evaluate transcriptome-wide differences between type 1 VWD- and control-derived endothelial colony forming cells (ECFCs) during basal and stimulated VWF release. ECFCs from 8 patients with type 1 VWD and 4 other patients were included in this study as controls. VWF protein analysis revealed heterogenous responses to stimulation among type 1 VWD and control ECFCs. During basal VWF release, 64 mRNAs and 7 miRNAs were differentially regulated between type 1 VWD and control ECFCs, and 65 putatively pathogenic miRNA-mRNA interactions were identified. During stimulated VWF release, 190 mRNAs and 5 mRNAs were differentially regulated between type 1 VWD and control ECFCs, and 110 putatively pathogenic miRNA-mRNA interactions were identified. Five gene ontology terms including coagulation, regulation of cell shape, and regulation of cell signaling were also differentially regulated in type 1 VWD ECFCs during stimulated release. To our knowledge, we have shown for the first time that transcriptome-wide differences exist between type 1 VWD and control ECFCs. These differences may contribute to bleeding in type 1 VWD, and further investigation may reveal novel biomarkers and therapeutic targets.

A Computational Approach to Identification of Candidate Biomarkers in High-Dimensional Molecular Data.

Complex high-dimensional datasets that are challenging to analyze are frequently produced through '-omics' profiling. Typically, these datasets contain more genomic features than samples, limiting the use of multivariable statistical and machine learning-based approaches to analysis. Therefore, effective alternative approaches are urgently needed to identify features-of-interest in '-omics' data. In this study, we present the molecular feature selection tool, a novel, ensemble-based, feature selection application for identifying candidate biomarkers in '-omics' data. As proof-of-principle, we applied the molecular feature selection tool to identify a small set of immune-related genes as potential biomarkers of three prostate adenocarcinoma subtypes. Furthermore, we tested the selected genes in a model to classify the three subtypes and compared the results to models built using all genes and all differentially expressed genes. Genes identified with the molecular feature selection tool performed better than the other models in this study in all comparison metrics: accuracy, precision, recall, and F1-score using a significantly smaller set of genes. In addition, we developed a simple graphical user interface for the molecular feature selection tool, which is available for free download. This user-friendly interface is a valuable tool for the identification of potential biomarkers in gene expression datasets and is an asset for biomarker discovery studies.

Sex differences in the aging murine urinary bladder and influence on the tumor immune microenvironment of a carcinogen-induced model of bladder cancer.

Sex and age associated differences in the tumor immune microenvironment of non-muscle invasive bladder (NMIBC) cancer and associated clinical outcomes are emerging indicators of treatment outcomes. The incidence of urothelial carcinoma of the bladder is four times higher in males than females; however, females tend to present with a more aggressive disease, a poorer response to immunotherapy and suffer worse clinical outcomes. Recent findings have demonstrated sex differences in the tumor immune microenvironment of non-muscle invasive and muscle invasive bladder cancer and associated clinical outcomes. However, a significant gap in knowledge remains with respect to the current pre-clinical modeling approaches to more precisely recapitulate these differences towards improved therapeutic design. Given the similarities in mucosal immune physiology between humans and mice, we evaluated the sex and age-related immune alterations in healthy murine bladders. Bulk-RNA sequencing and multiplex immunofluorescence-based spatial immune profiling of healthy murine bladders from male and female mice of age groups spanning young to old showed a highly altered immune landscape that exhibited sex and age associated differences, particularly in the context of B cell mediated responses. Spatial profiling of healthy bladders, using markers specific to macrophages, T cells, B cells, activated dendritic cells, high endothelial venules, myeloid cells and the PD-L1 immune checkpoint showed sex and age associated differences. Bladders from healthy older female mice also showed a higher presence of tertiary lymphoid structures (TLSs) compared to both young female and male equivalents. Spatial immune profiling of N-butyl-N-(4-hydroxybutyl) nitrosamine (BBN) carcinogen exposed male and female bladders from young and old mice revealed a similar frequency of TLS formation, sex differences in the bladder immune microenvironment and, age associated differences in latency of tumor induction. These findings support the incorporation of sex and age as factors in pre-clinical modeling of bladder cancer and will potentially advance the field of immunotherapeutic drug development to improve clinical outcomes.

High interobserver variability and frequent overdiagnosis of dysplasia in fundic gland polyps can be improved by detecting atypia on the surface epithelium and an abrupt transition to non-neoplastic cells.

AIMS: Fundic gland polyps (FGPs) arise sporadically and in combination with familial adenomatous polyposis (FAP). Criteria for distinguishing low-grade dysplasia (LGD) from regenerative atypia in FGPs are not well established. The aims of study were to determine: (i) interobserver variability in diagnosing LGD in FGPs; (ii) bias in diagnosing LGD in FAP patients; and (iii) stringent criteria for LGD in FGPs. METHODS AND RESULTS: Five senior pathologists who were blinded to the clinical history reviewed 72 FAP-associated FGPs and 34 sporadic FGPs. Cases were classified as negative (score = 0) or positive (score = 1) for LGD. Each case was assigned a 'combined dysplasia score' (CDS) ranging from 0 to 5 to reflect all five opinions. Fleiss' kappa showed only moderate interobserver agreement (κ = 0.46). Forty-one FGPs were classified as negative for dysplasia by consensus (CDS = 0-1), including 10 (24%) originally diagnosed as LGD. In contrast, all 37 cases classified as LGD by consensus (CDS = 4-5) were originally diagnosed as LGD, indicating that overdiagnosis of dysplasia is more common than underdiagnosis (P = 0.0012). Cytological atypia in the surface epithelium and an abrupt transition between atypical and normal-appearing epithelium were the most sensitive (97% and 100%, respectively) and specific (100% and 98%, respectively) features of dysplasia (P < 0.0001 for both comparisons). Very good agreement was achieved when a diagnosis of dysplasia was based on the presence of both features (κ = 0.85). CONCLUSIONS: There is high interobserver variability and a tendency to overdiagnose LGD in FGPs. Strict criteria requiring both surface atypia and abrupt transition for LGD in FGPs result in low interobserver variability.

Discriminating Neoplastic from Nonneoplastic Tissues Using an miRNA-Based Deep Cancer Classifier.

Next-generation sequencing has enabled the collection of large biological data sets, allowing novel molecular-based classification methods to be developed for increased understanding of disease. miRNAs are small regulatory RNA molecules that can be quantified using next-generation sequencing and are excellent classificatory markers. Herein, a deep cancer classifier (DCC) was adapted to differentiate neoplastic from nonneoplastic samples using comprehensive miRNA expression profiles from 1031 human breast and skin tissue samples. The classifier was fine-tuned and evaluated using 750 neoplastic and 281 nonneoplastic breast and skin tissue samples. Performance of the DCC was compared with two machine-learning classifiers: support vector machine and random forests. In addition, performance of feature extraction through the DCC was also compared with a developed feature selection algorithm, cancer specificity. The DCC had the highest performance of area under the receiver operating curve and high performance in both sensitivity and specificity, unlike machine-learning and feature selection models, which often performed well in one metric compared with the other. In particular, deep learning had noticeable advantages with highly heterogeneous data sets. In addition, our cancer specificity algorithm identified candidate biomarkers for differentiating neoplastic and nonneoplastic tissue samples (eg, miR-144 and miR-375 in breast cancer and miR-375 and miR-451 in skin cancer).

Lack of definitive presurgical pathological diagnosis is associated with inadequate surgical margins in breast-conserving surgery.

PURPOSE: To determine the impact of definitive presurgical diagnosis on surgical margins in breast-conserving surgery (BCS) for primary carcinomas; clinicopathological features were also analyzed. METHODS: This retrospective study included women who underwent BCS for primary carcinomas in 2016 and 2017. Definitive presurgical diagnosis was defined as having a presurgical core needle biopsy (CNB) and not being upstaged between biopsy and surgery. Biopsy data and imaging findings including breast density were retrieved. Inadequate surgical margins (IM) were defined per latest ASCO and ASTRO guidelines. Univariable and multivariable analyses were performed. RESULTS: 360 women (median age, 66) met inclusion criteria with 1 having 2 cancers. 82.5% (298/361) were invasive cancers while 17.5% (63/361) were ductal carcinoma in situ (DCIS). Most biopsies were US-guided (284/346, 82.0%), followed by mammographic (60/346, 17.3%), and MRI-guided (2/346, 0.6%). US and mammographic CNB yielded median samples of 2 and 4, respectively, with a 14G needle. 15 patients (4.2%) lacked presurgical CNB. The IM rate was 30.0%. In multivariable analysis, large invasive cancers (>20 mm), dense breasts, and DCIS were associated with IM (p = 0.029, p = 0.010, and p = 0.013, respectively). Most importantly, lack of definitive presurgical diagnosis was a risk factor for IM (OR, 2.35; 95% CI: 1.23-4.51, p = 0.010). In contrast, neither patient age (<50) nor aggressive features (e.g., LVI) were associated with IM. CONCLUSION: Lack of a definitive presurgical diagnosis was associated with a two-fold increase of IM in BCS; other risk factors were dense breasts, large invasive cancers, and DCIS.

DNA Methylation, Deamination, and Translesion Synthesis Combine to Generate Footprint Mutations in Cancer Driver Genes in B-Cell Derived Lymphomas and Other Cancers.

Cancer genomes harbor numerous genomic alterations and many cancers accumulate thousands of nucleotide sequence variations. A prominent fraction of these mutations arises as a consequence of the off-target activity of DNA/RNA editing cytosine deaminases followed by the replication/repair of edited sites by DNA polymerases (pol), as deduced from the analysis of the DNA sequence context of mutations in different tumor tissues. We have used the weight matrix (sequence profile) approach to analyze mutagenesis due to Activation Induced Deaminase (AID) and two error-prone DNA polymerases. Control experiments using shuffled weight matrices and somatic mutations in immunoglobulin genes confirmed the power of this method. Analysis of somatic mutations in various cancers suggested that AID and DNA polymerases η and θ contribute to mutagenesis in contexts that almost universally correlate with the context of mutations in A:T and G:C sites during the affinity maturation of immunoglobulin genes. Previously, we demonstrated that AID contributes to mutagenesis in (de)methylated genomic DNA in various cancers. Our current analysis of methylation data from malignant lymphomas suggests that driver genes are subject to different (de)methylation processes than non-driver genes and, in addition to AID, the activity of pols η and θ contributes to the establishment of methylation-dependent mutation profiles. This may reflect the functional importance of interplay between mutagenesis in cancer and (de)methylation processes in different groups of genes. The resulting changes in CpG methylation levels and chromatin modifications are likely to cause changes in the expression levels of driver genes that may affect cancer initiation and/or progression.

Blood extracellular vesicles from healthy individuals regulate hematopoietic stem cells as humans age.

Hematopoietic stem cells (HSCs) maintain balanced blood cell production in a process called hematopoiesis. As humans age, their HSCs acquire mutations that allow some HSCs to disproportionately contribute to normal blood production. This process, known as age-related clonal hematopoiesis, predisposes certain individuals to cancer, cardiovascular and pulmonary pathologies. There is a growing body of evidence suggesting that factors outside cells, such as extracellular vesicles (EVs), contribute to the disruption of stem cell homeostasis during aging. We have characterized blood EVs from humans and determined that they are remarkably consistent with respect to size, concentration, and total protein content, across healthy subjects aged 20-85 years. When analyzing EV protein composition from mass spectroscopy data, our machine-learning-based algorithms are able to distinguish EV proteins based on age and suggest that different cell types dominantly produce EVs released into the blood, which change over time. Importantly, our data show blood EVs from middle and older age groups (>40 years) significantly stimulate HSCs in contrast to untreated and EVs sourced from young subjects. Our study establishes for the first time that although EV particle size, concentration, and total protein content remain relatively consistent over an adult lifespan in humans, EV content evolves during aging and potentially influences HSC regulation.

Classifying Lung Neuroendocrine Neoplasms through MicroRNA Sequence Data Mining.

Lung neuroendocrine neoplasms (NENs) can be challenging to classify due to subtle histologic differences between pathological types. MicroRNAs (miRNAs) are small RNA molecules that are valuable markers in many neoplastic diseases. To evaluate miRNAs as classificatory markers for lung NENs, we generated comprehensive miRNA expression profiles from 14 typical carcinoid (TC), 15 atypical carcinoid (AC), 11 small cell lung carcinoma (SCLC), and 15 large cell neuroendocrine carcinoma (LCNEC) samples, through barcoded small RNA sequencing. Following sequence annotation and data preprocessing, we randomly assigned these profiles to discovery and validation sets. Through high expression analyses, we found that miR-21 and -375 are abundant in all lung NENs, and that miR-21/miR-375 expression ratios are significantly lower in carcinoids (TC and AC) than in neuroendocrine carcinomas (NECs; SCLC and LCNEC). Subsequently, we ranked and selected miRNAs for use in miRNA-based classification, to discriminate carcinoids from NECs. Using miR-18a and -155 expression, our classifier discriminated these groups in discovery and validation sets, with 93% and 100% accuracy. We also identified miR-17, -103, and -127, and miR-301a, -106b, and -25, as candidate markers for discriminating TC from AC, and SCLC from LCNEC, respectively. However, these promising findings require external validation due to sample size.

Characterizing and classifying neuroendocrine neoplasms through microRNA sequencing and data mining.

Neuroendocrine neoplasms (NENs) are clinically diverse and incompletely characterized cancers that are challenging to classify. MicroRNAs (miRNAs) are small regulatory RNAs that can be used to classify cancers. Recently, a morphology-based classification framework for evaluating NENs from different anatomical sites was proposed by experts, with the requirement of improved molecular data integration. Here, we compiled 378 miRNA expression profiles to examine NEN classification through comprehensive miRNA profiling and data mining. Following data preprocessing, our final study cohort included 221 NEN and 114 non-NEN samples, representing 15 NEN pathological types and 5 site-matched non-NEN control groups. Unsupervised hierarchical clustering of miRNA expression profiles clearly separated NENs from non-NENs. Comparative analyses showed that miR-375 and miR-7 expression is substantially higher in NEN cases than non-NEN controls. Correlation analyses showed that NENs from diverse anatomical sites have convergent miRNA expression programs, likely reflecting morphological and functional similarities. Using machine learning approaches, we identified 17 miRNAs to discriminate 15 NEN pathological types and subsequently constructed a multilayer classifier, correctly identifying 217 (98%) of 221 samples and overturning one histological diagnosis. Through our research, we have identified common and type-specific miRNA tissue markers and constructed an accurate miRNA-based classifier, advancing our understanding of NEN diversity.

Distinguishing Tumor and Stromal Sources of MicroRNAs Linked to Metastasis in Cutaneous Melanoma.

MicroRNA (miRNA) dysregulation in cancer causes changes in gene expression programs regulating tumor progression and metastasis. Candidate metastasis suppressor miRNA are often identified by differential expression in primary tumors compared to metastases. Here, we performed comprehensive analysis of miRNA expression in The Cancer Genome Atlas (TCGA) skin cutaneous melanoma (SKCM) tumors (97 primary, 350 metastatic), and identified candidate metastasis-suppressor miRNAs. Differential expression analysis revealed miRNA significantly downregulated in metastatic tumors, including miR-205, miR-203, miR-200a-c, and miR-141. Furthermore, sequential feature selection and classification analysis identified miR-205 and miR-203 as the miRNA best able to discriminate between primary and metastatic tumors. However, cell-type enrichment analysis revealed that gene expression signatures for epithelial cells, including keratinocytes and sebocytes, were present in primary tumors and significantly correlated with expression of the candidate metastasis-suppressor miRNA. Examination of miRNA expression in cell lines revealed that candidate metastasis-suppressor miRNA identified in the SKCM tumors, were largely absent in melanoma cells or melanocytes, and highly restricted to keratinocytes and other epithelial cell types. Indeed, the differences in stromal cell composition between primary and metastatic tumor tissues is the main basis for identification of differential miRNA that were previously classified as metastasis-suppressor miRNAs. We conclude that future studies must consider tumor-intrinsic and stromal sources of miRNA in their workflow to identify bone fide metastasis-suppressor miRNA in cutaneous melanoma and other cancers.

Objective quantification of BCL2 protein by multiplex immunofluorescence in routine biopsy samples of diffuse large B-cell lymphoma demonstrates associations with survival and BCL2 gene alterations.

Up-regulation of BCL2 in cases of diffuse large B-cell lymphoma (DLBCL) can confer treatment resistance. Quantitative immunofluorescence (QIF) histology allows objective quantification of protein-based biomarkers. We investigated the utility of QIF for evaluating BCL2 as a biomarker in DLBCL by quantifying BCL2 selectively in CD20-expressing lymphoma cells in biopsy samples from 116 cases of DLBCL in two cohorts one of which consisted of relapsed/refractory cases from a clinical trial. BCL2 protein by QIF correlated with BCL2 mRNA abundance and was associated with both translocation and copy number gain of the BCL2 gene. Elevated BCL2 protein expression by QIF, but not immunohistochemistry or mRNA quantification, was associated with inferior overall and relapse-free survival in the relapsed/refractory cohort. QIF is an effective means of quantifying BCL2 protein objectively in routine cancer biopsy specimens and shows promise for identifying relapsed/refractory DLBCL patients at risk of inferior outcomes after salvage therapy.

Neutrophil recruitment and function in endometriosis patients and a syngeneic murine model.

Endometriosis is a chronic inflammatory, gynecological disease characterized by the presence of endometrial-like tissue lesions outside of the uterus. Neutrophils are elevated in the systemic circulation and peritoneal fluid of endometriosis patients; however, whether and how neutrophils contribute to endometriosis pathophysiology remain poorly understood. With emerging roles for neutrophils in chronic and sterile inflammatory conditions, we sought to provide in-depth characterization of neutrophil involvement in endometriosis. We demonstrate that neutrophils reside within patient endometriotic lesions and that patient lesions possess a microenvironment that may influence neutrophil recruitment and function. We also provide the first evidence that systemic circulating neutrophils from endometriosis patients display distinct transcriptomic differences compared neutrophils from healthy control subjects. Time course characterization of our syngeneic, immunocompetent mouse model of endometriosis revealed that neutrophils are rapidly recruited to the peritoneal environment early after endometriotic lesion establishment and remain present in murine lesions long term. In vivo neutrophil depletion altered the systemic and peritoneal immune microenvironment of mice with endometriosis as demonstrated by changes in pro-inflammatory and angiogenic mediators. Taken together, these findings highlight a novel role for neutrophils in early events such as angiogenesis and modulation of the local inflammatory environment associated with endometriosis pathogenesis.

Exploring microRNA Regulation of Cancer with Context-Aware Deep Cancer Classifier.

BACKGROUND: MicroRNAs (miRNAs) are small, non-coding RNA that regulate gene expression through post-transcriptional silencing. Differential expression observed in miRNAs, combined with advancements in deep learning (DL), have the potential to improve cancer classification by modelling non-linear miRNA-phenotype associations. We propose a novel miRNA-based deep cancer classifier (DCC) incorporating genomic and hierarchical tissue annotation, capable of accurately predicting the presence of cancer in wide range of human tissues. METHODS: miRNA expression profiles were analyzed for 1746 neoplastic and 3871 normal samples, across 26 types of cancer involving six organ sub-structures and 68 cell types. miRNAs were ranked and filtered using a specificity score representing their information content in relation to neoplasticity, incorporating 3 levels of hierarchical biological annotation. A DL architecture composed of stacked autoencoders (AE) and a multi-layer perceptron (MLP) was trained to predict neoplasticity using 497 abundant and informative miRNAs. Additional DCCs were trained using expression of miRNA cistrons and sequence families, and combined as a diagnostic ensemble. Important miRNAs were identified using backpropagation, and analyzed in Cytoscape using iCTNet and BiNGO. RESULTS: Nested four-fold cross-validation was used to assess the performance of the DL model. The model achieved an accuracy, AUC/ROC, sensitivity, and specificity of 94.73%, 98.6%, 95.1%, and 94.3%, respectively. CONCLUSION: Deep autoencoder networks are a powerful tool for modelling complex miRNA-phenotype associations in cancer. The proposed DCC improves classification accuracy by learning from the biological context of both samples and miRNAs, using anatomical and genomic annotation. Analyzing the deep structure of DCCs with backpropagation can also facilitate biological discovery, by performing gene ontology searches on the most highly significant features.

Validation of diagnostic gene sets to identify critically ill patients with sepsis.

PURPOSE: Gene expression diagnostics have been proposed to identify critically ill patients with sepsis. Three expression-based scores have been developed, but have not been compared in a prospective validation. We sought to validate these scores using an independent dataset and analysis. METHODS: We generated gene expression profiles from 61 critically ill patients. We validated the performance of 3 expression-based sepsis scores including 1) the Sepsis MetaScore (SMS); 2) the SeptiCyte™ Lab; and 3) the FAIM3:PLAC8 ratio. Sepsis was identified as the presence of definite, probable, or possible infection in the setting of organ dysfunction (SOFA score ≥ 2). RESULTS: For all 3 models, scores were significantly different between patients with and without sepsis. Discrimination was highest for the SMS (area under the receiver operating characteristics curve [AUROC 0.80 [95% CI 0.67-0.92]), with greater confidence in the presence of infection resulting in better model performance (max AUROC 0.93 [0.87-1.0]). CONCLUSIONS: All three scores distinguished septic from non-septic ICU patients, with the SMS showing the best performance overall in our cohort. Our results suggest that models developed from the co-analysis of multiple cohorts are more generalizable. Further work is needed to identify expression-based biomarkers of response to specific therapies.