Microarray-based Gene Expression Data Research Articles

Metaheuristic integrated machine learning classification of colon cancer using STFT LASSO and EHO feature extraction from microarray gene expressions

The microarray gene expression data poses a tremendous challenge due to their curse of dimensionality problem. The sheer volume of features far surpasses available samples, leading to overfitting and reduced classification accuracy. Thus the dimensionality of microarray gene expression data must be reduced with efficient feature extraction methods to reduce the volume of data and extract meaningful information to enhance the classification accuracy and interpretability. In this research, we discover the uniqueness of applying STFT (Short Term Fourier Transform), LASSO (Least Absolute Shrinkage and Selection Operator), and EHO (Elephant Herding Optimisation) for extracting significant features from lung cancer and reducing the dimensionality of the microarray gene expression database. The classification of lung cancer is performed using the following classifiers: Gaussian Mixture Model (GMM), Particle Swarm Optimization (PSO) with GMM, Detrended Fluctuation Analysis (DFA), Naive Bayes classifier (NBC), Firefly with GMM, Support Vector Machine with Radial Basis Kernel (SVM-RBF) and Flower Pollination Optimization (FPO) with GMM. The EHO feature extraction with the FPO-GMM classifier attained the highest accuracy in the range of 96.77, with an F1 score of 97.5, MCC of 0.92 and Kappa of 0.92. The reported results underline the significance of utilizing STFT, LASSO, and EHO for feature extraction in reducing the dimensionality of microarray gene expression data. These methodologies also help in improved and early diagnosis of lung cancer with enhanced classification accuracy and interpretability.

Prognostic gene biomarkers for c-Src inhibitor Si162 sensitivity in melanoma cells.

Early detection and treatment are crucial in combating malignant melanoma. Src is an important therapeutic target in melanoma due to its association with cancer progression. However, developing effective Src-targeting drugs remains challenging and personalized medicine relies on biomarkers and targeted therapies for precise and effective treatment. This study focuses on Si162, a newly synthesized c-Src inhibitor, to identify reliable biomarkers for predicting Si162 sensitivity and explore associated biological characteristics and pathways in melanoma cells. Primary melanoma cells (M1, M21, M24, M84, M133, M307, and M2025) were obtained from patients diagnosed with melanoma. Si162 cytotoxicity tests were performed using luminescent adenosine triphosphate detection and the half-maximal inhibitory concentration (IC50) values were calculated. Gene expression profiles were analyzed using microarray-based gene expression data. Differentially expressed genes between the resistant and sensitive groups were identified using Pearson correlation analysis. Gene coexpression, interactions, and pathways were investigated through clustering, network, and pathway analyses. Biological functions were examined using the Database for Annotation, Visualization, and Integrated Discovery. Molecular pathways associated with different responses to Si162 were identified using gene set enrichment analysis. The gene expressions were validated using reverse transcription-quantitative polymerase chain reaction. The cells revealed significant differences in response to Si162 based on the IC50 values (p < 0.05). A total of 36 differentially expressed genes associated with Si162 susceptibility were identified. Distinct expression patterns between the sensitive and resistant groups were observed in 9 genes (LRBA, MGMT, CAND1, ADD1, SETD2, CNTN6, FGF18, C18orf25, and RPL13). Coexpression among the differentially expressed genes was highlighted, and 9 genes associated with molecular pathways, including EMT, transforming growth factor-beta (TGF-β) signaling, and ribosomal protein synthesis, between groups. Genes involved in dysregulated immune response were observed in the resistant group. The involvement of 5 genes (ADD1, CNTN6, FGF18, C18orf25, and RPL13) in Si162 resistance was confirmed through qRT-PCR validation. These findings contribute to our understanding of the underlying biological differences among melanoma cells and suggest potential biomarkers and pathways associated with Si162 response and resistance.

Spatial clustering based gene selection for gene expression analysis in microarray data classification

A typical application of categorization in data mining is to uncover interesting distributions and significant patterns in the information that underlies it using density-based spatial clustering for workloads with noise. In these conditions, it is anticipated that the classification of the microarray gene expression database will have the necessary clustering property that may be utilized to emphasize the effects of the alterations. The proposed method typically guarantees that the subsequent identification of gene clusters’ best global arrangement of genes. It provides an iterative method for figuring out the precise number of clusters needed for each data collection. The technique is based on practices frequently used in statistical tests. The key idea is to coordinate gene redistribution optimization across clusters with the search for the optimal number of groups. An experiment that finds the most effective number of genes over time was used to evaluate the effectiveness of the suggested strategy. It used this stringent statistical test to show that our technique accurately clusters more than 95% of the genes. Finally, since the basic principles of gene development and gene cluster assignment have been well characterized by earlier studies and the technique was verified using real gene expression information.

Patients with schizophrenia and bipolar disorder display a similar global gene expression signature in whole blood that reflects elevated proportion of immature neutrophil cells with association to lipid changes

Schizophrenia (SCZ) and bipolar disorder (BD) share clinical characteristics, genetic susceptibility, and immune alterations. We aimed to identify differential transcriptional patterns in peripheral blood cells of patients with SCZ or BD versus healthy controls (HC). We analyzed microarray-based global gene expression data in whole blood from a cohort of SCZ (N = 329), BD (N = 203) and HC (N = 189). In total, 65 genes were significantly differentially expressed in SCZ and 125 in BD, as compared to HC, with similar ratio of up- and downregulated genes in both disorders. Among the top differentially expressed genes, we found an innate immunity signature that was shared between SCZ and BD, consisting of a cluster of upregulated genes (e.g., OLFM4, ELANE, BPI and MPO) that indicate an increased fraction of immature neutrophils. Several of these genes displayed sex differences in the expression pattern, and post-hoc analysis demonstrated a positive correlation with triglyceride and a negative correlation with HDL cholesterol. We found that many of the downregulated genes in SCZ and BD were associated with smoking. These findings of neutrophil granulocyte-associated transcriptome signatures in both SCZ and BD point at altered innate immunity pathways with association to lipid changes and potential for clinical translation.

Benchmarking causal reasoning algorithms for gene expression-based compound mechanism of action analysis

BackgroundElucidating compound mechanism of action (MoA) is beneficial to drug discovery, but in practice often represents a significant challenge. Causal Reasoning approaches aim to address this situation by inferring dysregulated signalling proteins using transcriptomics data and biological networks; however, a comprehensive benchmarking of such approaches has not yet been reported. Here we benchmarked four causal reasoning algorithms (SigNet, CausalR, CausalR ScanR and CARNIVAL) with four networks (the smaller Omnipath network vs. 3 larger MetaBase™ networks), using LINCS L1000 and CMap microarray data, and assessed to what extent each factor dictated the successful recovery of direct targets and compound-associated signalling pathways in a benchmark dataset comprising 269 compounds. We additionally examined impact on performance in terms of the functions and roles of protein targets and their connectivity bias in the prior knowledge networks.ResultsAccording to statistical analysis (negative binomial model), the combination of algorithm and network most significantly dictated the performance of causal reasoning algorithms, with the SigNet recovering the greatest number of direct targets. With respect to the recovery of signalling pathways, CARNIVAL with the Omnipath network was able to recover the most informative pathways containing compound targets, based on the Reactome pathway hierarchy. Additionally, CARNIVAL, SigNet and CausalR ScanR all outperformed baseline gene expression pathway enrichment results. We found no significant difference in performance between L1000 data or microarray data, even when limited to just 978 ‘landmark’ genes. Notably, all causal reasoning algorithms also outperformed pathway recovery based on input DEGs, despite these often being used for pathway enrichment. Causal reasoning methods performance was somewhat correlated with connectivity and biological role of the targets.ConclusionsOverall, we conclude that causal reasoning performs well at recovering signalling proteins related to compound MoA upstream from gene expression changes by leveraging prior knowledge networks, and that the choice of network and algorithm has a profound impact on the performance of causal reasoning algorithms. Based on the analyses presented here this is true for both microarray-based gene expression data as well as those based on the L1000 platform.

Association Between MRI-Assessed Patterns of Connectome Gradient and Gene-Expression Profiles in Two Independent Patient Cohorts With Hepatitis B Virus-Related Cirrhosis.

Patients with hepatitis B virus-related cirrhosis (HBV-RC) exhibit progressive neurologic dysfunction from primary sensorimotor to high-order cognition, as their disease advances. However, the exact neurobiologic mechanisms and the potential association with gene-expression profiles are not fully understood. To explore the hierarchical disorganization in the large-scale functional connectomes in HBV-RC patients and to investigate its potential underlying molecular basis. Prospective. Fifty HBV-RC patients and 40 controls (Cohort 1) and 30 HBV-RC patients and 38 controls (Cohort 2). Gradient-echo echo-planar and fast field echo sequences at 3.0 T (Cohort 1) and 1.5 T (Cohort 2). Data were processed with Dpabi and the BrainSpace package. Gradient scores were evaluated from global to voxel level. Cognitive measurement and patients grouping were based on psychometric hepatic encephalopathy scores. The whole-brain microarray-based gene-expression data were obtained from the AIBS website. One-way ANOVA, chi-square test, two-sample t-test, Kruskal-Wallis test, Spearman's correlation coefficient (r), the gaussian random field correction, false discovery rate (FDR) correction and the Bonferroni correction. Significance level: P < 0.05. HBV-RC patients exhibited a robust and replicable connectome gradient dysfunction, which was significantly associated with the gene-expression profiles in both cohorts (r = 0.52 and r = 0.56, respectively). The most correlated genes were enriched in γ-aminobutyric acid (GABA) and GABA-related receptor genes (FDR q value <0.05). Moreover, the connectome gradient dysfunction at network level observed in HBV-RC patients correlated with their poor cognitive performance (Cohort 2: visual network, r = -0.56; subcortical network, r = 0.66; frontoparietal network, r = 0.51). HBV-RC patients had hierarchical disorganization in the large-scale functional connectomes, which may underly their cognitive impairment. In addition, we showed the potential molecular mechanism of the connectome gradient dysfunction, which suggested the importance of GABA and GABA-related receptor genes. 2 TECHNICAL EFFICACY: Stage 2.

Late acceptance hill climbing aided chaotic harmony search for feature selection: An empirical analysis on medical data

In today’s era of data-driven digital society, there is a huge demand for optimized solutions that essentially reduce the cost of operation, thereby aiming to increase productivity. Processing a huge amount of data, like the Microarray based gene expression data, using machine learning and data mining algorithms has certain limitations in terms of memory and time requirements. This would be more concerning, when a dataset comes with redundant and non-important information. For example, many report-based medical datasets have several non-informative attributes which mislead the classification algorithms. To this end, researchers have been developing several feature selection algorithms that try to discard the redundant information from the raw datasets before feeding them to machine learning algorithms. Metaheuristic based optimization algorithms provide an excellent option to solve feature selection problems. In this paper, we propose a music-inspired harmony search (HS) algorithm based wrapper feature selection method. At the beginning, we use a chaotic mapping to initialize the population of the HS algorithm in order to better coverage of the search space. Further to complement the inferior exploitation of the HS algorithm, we integrate it with the Late Acceptance Hill Climbing (LAHC) method. Thus the combination of these two algorithms provides a good balance between the exploration and exploitation of the HS algorithm. We evaluate the proposed feature selection method on 15 UCI datasets and the obtained results are found to be better than many state-of-the-art methods both in terms of the classification accuracy and the number of features selected. To evaluate the effectiveness of our algorithm, we utilize a combination of precision, recall, F1 score, fitness value, and execution time as performance indicators. These metrics enable us to obtain a comprehensive assessment of the algorithm’s abilities and limitations. We also apply our method on 3 microarray based gene expression datasets used for prediction of cancer to ensure the scalability and robustness as a feature selection method in real-life scenarios. In addition to this, we test our approach using the COVID-19 dataset, and it performs better than several metaheuristic based optimization techniques.

A Machine Learning Method to Trace Cancer Primary Lesion Using Microarray-Based Gene Expression Data.

Cancer of unknown primary site (CUP) is a heterogeneous group of cancers whose tissue of origin remains unknown after detailed investigation by conventional clinical methods. The number of CUP accounts for roughly 3%–5% of all human malignancies. CUP patients are usually treated with broad-spectrum chemotherapy, which often leads to a poor prognosis. Recent studies suggest that the treatment targeting the primary lesion of CUP will significantly improve the prognosis of the patient. Therefore, it is urgent to develop an efficient method to accurately detect tissue of origin of CUP in clinical cancer research. In this work, we developed a novel framework that uses Extreme Gradient Boosting (XGBoost) to trace the primary site of CUP based on microarray-based gene expression data. First, we downloaded the microarray-based gene expression profiles of 59,385 genes for 57,08 samples from The Cancer Genome Atlas (TCGA) and 6,364 genes for 3,101 samples from the Gene Expression Omnibus (GEO). Both data were divided into training and independent testing data with a ratio of 4:1. Then, we obtained in the training data 200 and 290 genes from TCGA and the GEO datasets, respectively, to train XGBoost models for the identification of the primary site of CUP. The overall 5-fold cross-validation accuracies of our methods were 96.9% and 95.3% on TCGA and GEO training datasets, respectively. Meanwhile, the macro-precision for the independent dataset reached 96.75% and 98.8% on, respectively, TCGA and GEO. Experimental results demonstrated that the XGBoost framework not only can reduce the cost of clinical cancer traceability but also has high efficiency, which might be useful in clinical usage.

Hypoxia-inducible miR-196a modulates glioblastoma cell proliferation and migration through complex regulation of NRAS.

Glioblastoma (GBM) is the most common and aggressive malignant brain tumor in humans. Hypoxia has been correlated with the aggressive form of glial tumors, poor prognosis, recurrence and resistance to various therapies. MicroRNAs (miRNAs) have emerged as critical mediators of hypoxic responses and have shown great potential for cancer diagnostics and therapeutics. Here, we focus on the regulatory and functional characterization of miR-196a, a hypoxia-inducible miRNA, in GBM. Hypoxia/HIF regulation of miR-196a was assessed by RT-qPCR, promoter-luciferase and ChIP assays in GBM cell lines. miR-196a levels were analyzed inThe Cancer Genome Atlas (TCGA)-GBM, Chinese Glioma Genome Atlas (CGGA) and Indian GBM patient cohorts. miR-target interactions were studied using RNA/protein quantification and 3'UTR luciferase assays. The effect of miR-196a overexpression/inhibition wasassessed on cellular viability, migration and apoptosis under hypoxia and normoxia. Microarray-based gene expression profiling studies wereperformrd to study the effect of miR-196a ontheGBM cellular transcriptome under hypoxia. Weidentified miR-196a as a hypoxia-inducible and hypoxia-inducible factor (HIF)-regulated miRNA that plays an oncogenic role in GBM. miR-196a was found to be significantly up-regulated in TCGA-GBM, CGGA glioma as well as Indian GBM patient cohorts. miR-196a overexpression was found to induce cellular proliferation, migration, spheroid formation and colony formation and to inhibit apoptosis, while miR-196a inhibition using anti-miR-196ayielded opposite results, suggesting an oncogenic role of miR-196a in GBM. We further unveiled NRAS, AJAP1, TAOK1 and COL24A1 as direct targets of miR-196a. We also report a complex competitive regulation of oncogenic NRAS by miR-196a, miR-146a and let-7 in GBM. Analysis of microarray-based gene expression data obtained by miR-196a inhibition under hypoxia revealed a role of miR-196a inHIF, calcium adhesion, Wnt and cell adhesion pathways. Interestingly, miR-196a was found to positively regulate the expression of various genes involved in the induction or stabilization of HIFs andin maintenance of hypoxic conditions, thereby suggesting the existence of an indirect miR-196a/HIF positive feedback loop under hypoxia. Overall, our work identifies a novel association between hypoxia/HIF signalling and miR-196a in GBM and suggests its therapeutic significance.

The spatial organization of the chronnectome associates with cortical hierarchy and transcriptional profiles in the human brain

The chronnectome of the human brain represents dynamic connectivity patterns of brain networks among interacting regions, but its organization principle and related transcriptional signatures remain unclear. Using task-free fMRI data from the Human Connectome Project (681 participants) and microarray-based gene expression data from the Allen Institute for Brain Science (1791 brain tissue samples from six donors), we conduct a transcriptome-chronnectome association study to investigate the spatial configurations of dynamic brain networks and their linkages with transcriptional profiles. We first classify the dynamic brain networks into four categories of nodes according to their time-varying characteristics in global connectivity and modular switching: the primary sensorimotor regions with large global variations, the paralimbic/limbic regions with frequent modular switching, the frontoparietal cortex with both high global and modular dynamics, and the sensorimotor association cortex with limited dynamics. Such a spatial layout reflects the cortical functional hierarchy, microarchitecture, and primary connectivity gradient spanning from primary to transmodal areas, and the cognitive spectrum from perception to abstract processing. Importantly, the partial least squares regression analysis reveals that the transcriptional profiles could explain 28% of the variation in this spatial layout of network dynamics. The top-related genes in the transcriptional profiles are enriched for potassium ion channel complex and activity and mitochondrial part of the cellular component. These findings highlight the hierarchically spatial arrangement of dynamic brain networks and their coupling with the variation in transcriptional signatures, which provides indispensable implications for the organizational principle and cellular and molecular functions of spontaneous network dynamics.

Associations between dietary patterns and gene expression pattern in peripheral blood mononuclear cells: A cross-sectional study

Background and aimsDiet may alter gene expression in immune cells involved in atherosclerotic cardiovascular disease susceptibility. However, we still lack a robust understanding of the association between diet and immune cell-related gene expression in humans. Therefore, we examined associations between dietary patterns (DPs) and gene expression profiles in peripheral blood mononuclear cells (PBMCs) in a population of healthy, Norwegian adults (n = 130 women and 105 men). Methods and resultsWe used factor analysis to define a posteriori DPs from food frequency questionnaire-based dietary assessment data. In addition, we derived interpretable features from microarray-based gene expression data (13 967 transcripts) using two algorithms: CIBERSORT for estimation of cell subtype proportions, and weighted gene co-expression network analysis (WGCNA) for cluster discovery. Finally, we associated DPs with either CIBERSORT-predicted PBMC leukocyte distribution or WGCNA gene clusters using linear regression models. We detected three DPs that broadly reflected Western, Vegetarian, and Low carbohydrate diets. CIBERSORT-predicted percentage of monocytes associated negatively with the Vegetarian DP. For women, the Vegetarian DP associated with a large gene cluster consisting of 600 genes mainly involved in regulation of DNA transcription, whereas for men, the Western DP inversely associated with a smaller cluster of 36 genes mainly involved in regulation of metabolic and inflammatory processes. A subsequent protein–protein interaction network analysis suggested that genes within these clusters might physically interact in biological networks. ConclusionsAlthough the present findings are exploratory, our analysis pipeline serves as a useful framework for studying the association between diet and gene expression.

P0971IDENTIFICATION OF HUB GENES ASSOCIATED WITH THE DEPOSITION OF EXTRACELLULAR MATRIX AND SPECIFIC FOR DIABETIC NEPHROPATHY BY WEIGHTED GENE CO-EXPRESSION NETWORK ANALYSIS

Abstract Background and Aims Diabetic nephropathy (DN) and its most severe manifestation, end-stage renal disease (ESRD), remains one of the leading causes of reduced lifespan in people with diabetes. Identifying novel molecules that are involved in the pathogenesis of DN has both diagnostic and therapeutic implications. The gene co-expression network analysis (WGCNA) algorithm represents a novel systems biology approach that provide the approach of association between gene modules and clinical traits to find the module involvement into the certain phenotypic trait. The goal of this study was to identify hub genes and their roles in DN from the aspect of whole gene transcripts analysis. Method Various types of chronic kidney diseases (CKD), including DN, microarray-based mRNA gene expression data, listed in the Gene Expression Omnibus (GEO) database, were analyzed. Next, we constructed a weighted gene co-expression network and identified modules distinguishing DN from normal or other types of CKD by WGCNA. Functional annotations of the genes in modules specialized for DN were analyzed by Gene Ontology (GO) enrichment analysis. Through protein-protein interaction (PPI) analysis and hub gene screening, the hub genes specific for DN were obtained. Furthermore, we drew ROC curves to determine the diagnosis and differential diagnosis value to DN of hub genes. Finally, another study of microarray in the GEO database was selected to verify the expression level of hub genes and in the “Nephroseq” database, the correlation between the gene expression level and eGFR was analyzed. Results “GSE99339”, glomerular tissue microarray in 187 patients with a total of 10947 genes, was selected for analysis. After excluding the inappropriate cases, a total of 179 specimens were analyzed, including 14 cases of DN, 22 cases of focal segmental glomerulosclerosis (FSGS), 15 cases of hypertensive nephropathy (HT), 26 cases of IgA nephropathy (IgAN), 13 cases of minimal change disease (MCD), 21 cases of membranous nephropathy (MGN), 23 cases of rapidly progressive glomerulonephritis (RPGN), 30 cases of lupus nephritis (LN) and 14 cases of kidney tissue adjacent to tumor. Co-expression network analysis by WGCNA identified 23 distinct gene modules of the total 10947 genes and revealed “MEsaddlegreen” module was strongly correlated with DN (r=0,54), but not with other groups. GO functional annotation showed that these 64 genes in the “MEsaddlegreen” module mainly enriched in the deposition of extracellular matrix, which represents the specific and diagnostic pathophysiological process of DN. Further PPI and hub gene screening analysis revealed that LUM, ELN, FBLN1, MMP2, FBLN5 and FMOD can be served as hub genes, which had been proved to play an important role in the deposition of extracellular matrix. Furthermore, we found that the expression of hub genes was the highest in DN group and for the diagnosis value of DN by each gene, the area under the ROC curve is about 0.75∼0.95. The external verification of another study showed that compared with the normal control group, the expression of these hub genes was the highest in the DN group, and their expression level was negatively correlated with eGFR. Conclusion Using WGCNA and further bioinformatics approach, we identified six hub genes that appear to be identical to DN development. As such, they may represent potential diagnostic biomarkers as well as therapeutic targets with clinical utility.

Generating automated kidney transplant biopsy reports combining molecular measurements with ensembles of machine learning classifiers.

We previously reported a system for assessing rejection in kidney transplant biopsies using microarray-based gene expression data, the Molecular Microscope® Diagnostic System (MMDx). The present study was designed to optimize the accuracy and stability of MMDx diagnoses by replacing single machine learning classifiers with ensembles of diverse classifier methods. We also examined the use of automated report sign-outs and the agreement between multiple human interpreters of the molecular results. Ensembles generated diagnoses that were both more accurate than the best individual classifiers, and nearly as stable as the best, consistent with expectations from the machine learning literature. Human experts had ≈93% agreement (balanced accuracy) signing out the reports, and random forest-based automated sign-outs showed similar levels of agreement with the human experts (92% and 94% for predicting the expert MMDx sign-outs for T cell-mediated (TCMR) and antibody-mediated rejection (ABMR), respectively). In most cases disagreements, whether between experts or between experts and automated sign-outs, were in biopsies near diagnostic thresholds. Considerable disagreement with histology persisted. The balanced accuracies of MMDx sign-outs for histology diagnoses of TCMR and ABMR were 73% and 78%, respectively. Disagreement with histology is largely due to the known noise in histology assessments (ClinicalTrials.gov NCT01299168).

La-Related Protein 4 as a Suppressor for Motility of Ovarian Cancer Cells.

The La-related proteins (LARPs) are a family of RNA binding proteins that control the degradation and stabilization of RNAs. As emerging research reveals the biology of each LARP, it is evident that LARPs are dysregulated in some types of cancer. Upregulation of cell motility potentiates the metastatic potential of ovarian cancer cells; however, the roles of LARPs in cell motility remain unknown. In the present study, we investigated the roles of LARPs in the progression of ovarian cancer using SKOV3 human ovarian cancer cells and a public database that integrates microarray-based gene expression data and clinical data. To explore the involvement of LARPs in the cell motility, we performed RNA interference screening for LARP mRNAs in SKOV3 cells. The screening identified LARP4 as a potential suppressor of the formation of lamellipodia. Conversely, enforced expression of LARP4 suppressed the formation of lamellipodia. Moreover, cell migration was significantly increased in LARP4-depleted SKOV3 cells. Mechanistically, LARP4 depletion was associated with the decrease in RhoA protein expression. These results suggest that LARP4 may limit RhoA-dependent cell motility. In a mouse xenograft model with SKOV3 cells, LARP4 depletion potentiated peritoneal metastasis. Upon analysis of a public database of patients with ovarian cancer, the LARP4 mRNA-high expression group (n = 166) showed longer overall survival compared with the LARP4 mRNA-low expression group (n = 489), implying a positive correlation of LARP4 mRNA levels in ovarian cancer tissues with patient prognosis. Taken together, we propose that LARP4 could suppress motility and metastatic potential of ovarian cancer cells.

Assessment of data transformations for model-based clustering of RNA-Seq data.

Quality control, global biases, normalization, and analysis methods for RNA-Seq data are quite different than those for microarray-based studies. The assumption of normality is reasonable for microarray based gene expression data; however, RNA-Seq data tend to follow an over-dispersed Poisson or negative binomial distribution. Little research has been done to assess how data transformations impact Gaussian model-based clustering with respect to clustering performance and accuracy in estimating the correct number of clusters in RNA-Seq data. In this article, we investigate Gaussian model-based clustering performance and accuracy in estimating the correct number of clusters by applying four data transformations (i.e., naïve, logarithmic, Blom, and variance stabilizing transformation) to simulated RNA-Seq data. To do so, an extensive simulation study was carried out in which the scenarios varied in terms of: how genes were selected to be included in the clustering analyses, size of the clusters, and number of clusters. Following the application of the different transformations to the simulated data, Gaussian model-based clustering was carried out. To assess clustering performance for each of the data transformations, the adjusted rand index, clustering error rate, and concordance index were utilized. As expected, our results showed that clustering performance was gained in scenarios where data transformations were applied to make the data appear “more” Gaussian in distribution.

Abstract P5-03-05: Sphingosine-1-phosphate produced by sphingosine kinase 1 and exported via ABCC1 shortens survival of mice and humans with breast cancer

Abstract Background: Sphingosine-1-phosphate (S1P), a bioactive sphingolipid mediator that is generated by sphingosine kinase 1 (SphK1) when it is phosphorylated (pSphK1) inside cells, has been implicated in regulation of many process important for breast cancer progression. Previously we have shown that S1P is exported out of human breast cancer cells by ATP-binding cassette (ABC) transporter ABCC1, but not by ABCB1, both known multidrug resistance proteins that efflux chemotherapeutic agents. However, the pathological consequences of these events to breast cancer progression and metastasis have not been elucidated. Here, we report that high expression of ABCC1, but not ABCB1, is associated with poor prognosis in breast cancer patients via exporting S1P. Materials and methods: Microarray based gene expression data of 2509 patients associated with their survival were obtained from METABRIC database. Single gene survival analysis based on expressin of SphK1, and dual ABCC1 or ABCB1 and SphK1 survival analyses were perfomerd. For protein analyses, tissues were obrained from 275 patients with stage 1-3 breast cancers treated in Yokohama City University Medical Center in Japan between 2006 and 2008. The expression of pSphK1 was analyzed by immunohistochemistry and investigate the relationship with clinicopathological findings. For in vitro and in vivo experiments, breast cancer cell lines were transfected by ABCB1, ABCC1 or vector transiently or stably. BALB/c nu/nu mice and BALB/c mice were used for in vivo experiments. S1P was measured by LC-ESI-MS/MS. Results: SphK1 expression significantly associate with worse overall survival (median survival of 124 months with high SphK1 expression compared to 163 months for patients with low SphK1 expression, p=0.0014). Although patients with high ABCC1 expression had only a slightly worse overall survival of 150 months, those with high levels of both SphK1 and ABCC1 had much worse prognosis with median overall survival of 114 months (p &amp;lt; 0.0068). Such association was not observed with ABCB1 expression. The frequency of strong pSphK1 protein expression was higher in HER2 enrhiched or TNBC than in Luminal. pSphK1 was more prevalent and increased in a larger tumors and in tumors from patients with lymph node metastases. Patients with breast cancers that express both pSphK1 and ABCC1 proteins have significantly shorter disease free survival. Overexpression of ABCC1, but not ABCB1, in human MCF7 and murine 4T1 cells enhanced S1P secretion, proliferation and migration of breast cancer cells. Implantation of breast cancer cells overexpressing ABCC1, but not ABCB1, into the mammary pad markedly enhanced tumor growth, angiogenesis and lymphangiogenesis with concomitant increases in lymph node and lung metastases as well as shorter survival of mice. Interestingly, S1P exported via ABCC1 from breast cancer cells upregulated transcription of SphK1 and its own formation. Conclusions: Our findings suggest that production and export of S1P via ABCC1, but not ABCB1, is associated with worse overall and disease free survival of breast cancer patients and that S1P axis play a role in aggressive biology of breast cancer progression and metastasis. Citation Format: Yamada A, Nagahashi M, Aoyagi T, Huang W-C, Lima S, Miyazaki H, Narui K, Ishikawa T, Endo I, Waters MR, Milstien S, Spiegel S, Takabe K. Sphingosine-1-phosphate produced by sphingosine kinase 1 and exported via ABCC1 shortens survival of mice and humans with breast cancer [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P5-03-05.

Bayesian Network to Infer Drug-Induced Apoptosis Circuits from Connectivity Map Data.

The Connectivity Map (CMAP) project profiled human cancer cell lines exposed to a library of anticancer compounds with the goal of connecting cancer with underlying genes and potential treatments. As most targeted anticancer therapeutics aim to induce tumor-selective apoptosis, it is critical to understand the specific cell death pathways triggered by drugs. This can help to better understand the mechanism of how cancer cells respond to chemical stimulations and improve the treatment of human tumors. In this study, using Connectivity MAP microarray-based gene expression data, we applied a Bayesian network modeling approach and identified apoptosis as a major drug-induced cellular pathway. We focused on 13 apoptotic genes that showed significant differential expression across all drug-perturbed samples to reconstruct the apoptosis network. In our predicted subnetwork, 9 out of 15 high-confidence interactions were validated in literature, and our inferred network captured two major cell death pathways by identifying BCL2L11 and PMAIP1 as key interacting players for the intrinsic apoptosis pathway, and TAXBP1 and TNFAIP3 for the extrinsic apoptosis pathway. Our inferred apoptosis network also suggested the role of BCL2L11 and TNFAIP3 as "gateway" genes in the drug-induced intrinsic and extrinsic apoptosis pathways.

A Comparative Transcriptomics Workflow for Analyzing Microarray Data From CHO Cell Cultures.

Microarray-based comparative transcriptomics analysis is a powerful tool to understand therapeutic protein producing mammalian cell lines at the gene expression level. However, an integrated analysis workflow specifically designed for end-to-end analysis of microarray data for CHO cells, the most prevalent host for commercial recombinant protein production, is lacking. To address this gap, an automated data analysis workflow in R that leverages public domain analysis modules is developed to analyze microarray based gene expression data. In addition to testing the global transcriptome differences of CHO cells at different conditions, the workflow identifies differentially expressed genes and pathways with intuitive visualizations as the outputs. The utility of this automated workflow is demonstrated by comparing the transcriptomic profiles of recombinant protein expressing CHO cells with and without a temperature shift. Statistically significant differential expression at the gene, pathway, and global transcriptome levels are identified and visualized. An automated workflow like the one developed in this study will enable rapid translation of CHO culture microarray data into biologically relevant information for mechanism-driven cell line optimization and bioprocess development.

Expression of ATF6 as a marker of pre-cancerous atypical change in ulcerative colitis-associated colorectal cancer: a potential role in the management of dysplasia

BackgroundDiagnosis of low-grade dysplasia (LGD) is important in the management of ulcerative colitis (UC), but it is often difficult to distinguish LGD from inflammatory regenerative epithelium. The unfolded protein response (UPR) is activated in inflammatory bowel disease and malignancies. We aimed to identify a UPR-related gene that is involved in the development of non-UC and UC-associated colorectal cancer (CRC), and to investigate whether the target gene is useful for the diagnosis of LGD.MethodsUsing our microarray gene expression database of 152 CRCs, we identified activating transcription factor 6 (ATF6) as a target gene. Immunohistochemistry (IHC) of ATF6 were analyzed in 137 surgically resected CRCs, 95 endoscopically resected adenomas and pTis cancers, and 136 samples from 51 UC patients (93 colitis without neoplasia, 31 dysplasia, and 12 UC-associated CRC). The diagnostic accuracy of ATF6 and p53 as markers of LGD was assessed.ResultsATF6 expression was detectable in all CRCs but not in normal colonic mucosa, was elevated with increase in cellular atypia (adenoma with moderate atypia < severe atypia < pTis CRC, p < 0.001), and higher in dysplasia and CRC than in non-neoplastic colitis (p < 0.001). Notably, the difference between colitis and LGD was significant. Compared to p53-IHC, ATF6-IHC had better diagnostic accuracy for distinguishing LGD from background inflammatory mucosa (sensitivity 70.8 vs. 16.7%, specificity 78.5 vs.71.0%, respectively).ConclusionsATF6 was expressed in lesions undergoing pre-cancerous atypical change in both non-UC and UC-associated CRC and may be used to distinguish LGD from inflammatory regenerative epithelium in UC patients.

Subtype-specific prognostic impact of different immune signatures in node-negative breast cancer.

The role of different subtypes of immune cells is still a matter of debate. We compared the prognostic relevance for metastasis-free survival (MFS) of a B-cell signature (BS), a T-cell signature (TS), and an immune checkpoint signature (CPS) in node-negative breast cancer (BC) using mRNA expression. Microarray-based gene-expression data were analyzed in six previously published cohorts of node-negative breast cancer patients not treated with adjuvant therapy (n=824). The prognostic relevance of the individual immune markers was assessed using univariate analysis. The amount of independent prognostic information provided by each immune signature was then compared using a likelihood ratio statistic in the whole cohort as well as in different molecular subtypes. Univariate Cox regression in the whole cohort revealed prognostic significance of CD4 (HR 0.66, CI 0.50-0.87, p=0.004), CXCL13 (HR 0.86, CI 0.81-0.92, p<0.001), CD20 (HR 0.76, CI 0.64-0.89, p=0.001), IgκC (HR 0.81, CI 0.75-0.88, p<0.001), and CTLA-4 (HR 0.67, CI 0.46-0.97, p=0.032). Multivariate analyses of the immune signatures showed that both TS (p<0.001) and BS (p<0.001) showed a significant prognostic information in the whole cohort. After accounting for clinical-pathological variables, TS (p<0.001), BS (p<0.05), and CPS (p<0.05) had an independent effect for MFS. In subgroup analyses, the prognostic effect of immune cells was most pronounced in HER2+BC: BS as well as TS showed a strong association with MFS when included first in the model (p<0.001). Immune signatures provide subtype-specific additional prognostic information over clinical-pathological variables in node-negative breast cancer.