Set Of Gene Expression Profiles Research Articles

Transcriptomic analyses and machine-learning methods reveal dysregulated key genes and potential pathogenesis in human osteoarthritic cartilage.

This study aimed to explore the biological and clinical importance of dysregulated key genes in osteoarthritis (OA) patients at the cartilage level to find potential biomarkers and targets for diagnosing and treating OA. Six sets of gene expression profiles were obtained from the Gene Expression Omnibus database. Differential expression analysis, weighted gene coexpression network analysis (WGCNA), and multiple machine-learning algorithms were used to screen crucial genes in osteoarthritic cartilage, and genome enrichment and functional annotation analyses were used to decipher the related categories of gene function. Single-sample gene set enrichment analysis was performed to analyze immune cell infiltration. Correlation analysis was used to explore the relationship among the hub genes and immune cells, as well as markers related to articular cartilage degradation and bone mineralization. A total of 46 genes were obtained from the intersection of significantly upregulated genes in osteoarthritic cartilage and the key module genes screened by WGCNA. Functional annotation analysis revealed that these genes were closely related to pathological responses associated with OA, such as inflammation and immunity. Four key dysregulated genes (cartilage acidic protein 1 (CRTAC1), iodothyronine deiodinase 2 (DIO2), angiopoietin-related protein 2 (ANGPTL2), and MAGE family member D1 (MAGED1)) were identified after using machine-learning algorithms. These genes had high diagnostic value in both the training cohort and external validation cohort (receiver operating characteristic > 0.8). The upregulated expression of these hub genes in osteoarthritic cartilage signified higher levels of immune infiltration as well as the expression of metalloproteinases and mineralization markers, suggesting harmful biological alterations and indicating that these hub genes play an important role in the pathogenesis of OA. A competing endogenous RNA network was constructed to reveal the underlying post-transcriptional regulatory mechanisms. The current study explores and validates a dysregulated key gene set in osteoarthritic cartilage that is capable of accurately diagnosing OA and characterizing the biological alterations in osteoarthritic cartilage; this may become a promising indicator in clinical decision-making. This study indicates that dysregulated key genes play an important role in the development and progression of OA, and may be potential therapeutic targets.

Identification of immune-related biomarkers co-occurring in acute ischemic stroke and acute myocardial infarction.

Acute ischemic stroke (AIS) and acute myocardial infarction (AMI) share several features on multiple levels. These two events may occur in conjunction or in rapid succession, and the occurrence of one event may increase the risk of the other. Owing to their similar pathophysiologies, we aimed to identify immune-related biomarkers common to AIS and AMI as potential therapeutic targets. We identified differentially expressed genes (DEGs) between the AIS and control groups, as well as AMI and control groups using microarray data (GSE16561 and GSE123342). A weighted gene co-expression network analysis (WGCNA) approach was used to identify hub genes associated with AIS and/or AMI progression. The intersection of the four gene sets identified key genes, which were subjected to functional enrichment and protein-protein interaction (PPI) network analyses. We confirmed the expression levels of hub genes using two sets of gene expression profiles (GSE58294 and GSE66360), and the ability of the genes to distinguish patients with AIS and/or AMI from control patients was assessed by calculating the receiver operating characteristic values. Finally, the investigation of transcription factor (TF)-, miRNA-, and drug-gene interactions led to the discovery of therapeutic candidates. We identified 477 and 440 DEGs between the AIS and control groups and between the AMI and control groups, respectively. Using WGCNA, 2,776 and 2,811 genes in the key modules were identified for AIS and AMI, respectively. Sixty key genes were obtained from the intersection of the four gene sets, which were used to identify the 10 hub genes with the highest connection scores through PPI network analysis. Functional enrichment analysis revealed that the key genes were primarily involved in immunity-related processes. Finally, the upregulation of five hub genes was confirmed using two other datasets, and immune infiltration analysis revealed their correlation with certain immune cells. Regulatory network analyses indicated that GATA2 and hsa-mir-27a-3p might be important regulators of these genes. Using comprehensive bioinformatics analyses, we identified five immune-related biomarkers that significantly contributed to the pathophysiological mechanisms of both AIS and AMI. These biomarkers can be used to monitor and prevent AIS after AMI, or vice versa.

A Simple, Test-Based Method to Control the Overestimation Bias in the Analysis of Potential Prognostic Tumour Markers

The early evaluation of prognostic tumour markers is commonly performed by comparing the survival of two groups of patients identified on the basis of a cut-off value. The corresponding hazard ratio (HR) is usually estimated, representing a measure of the relative risk between patients with marker values above and below the cut-off. A posteriori methods identifying an optimal cut-off are appropriate when the functional form of the relation between the marker distribution and patient survival is unknown, but they are prone to an overestimation bias. In the presence of a small sample size, which is typical of rare diseases, the external validation sets are hardly available and internal cross-validation could be unfeasible. We describe a new method to obtain an unbiased estimate of the HR at an optimal cut-off, exploiting the simple relation between the HR and the associated p-value estimated by a random permutation analysis. We validate the method on both simulated data and set of gene expression profiles from two large, publicly available data sets. Furthermore, a reanalysis of a previously published study, which included 134 Stage 4S neuroblastoma patients, allowed for the identification of E2F1 as a new gene with potential oncogenic activity. This finding was confirmed by an immunofluorescence analysis on an independent cohort.

Alpha-2-Heremans-Schmid-glycoprotein (AHSG) a potential biomarker associated with prognosis of chromophobe renal cell carcinoma: The PROPOLIS study.

Background and AimsChromophobe renal cell carcinoma (chRCC) is the third common pathological subtype in renal cancers. However, the underlying mechanisms of specific genetic characteristics of chRCC are currently unclear. In this study, protein expression profiles, gene ontology (GO), and survival plots were provided by integrated bioinformatics analysis to investigate key genes associated with the mechanism of tumorigenesis and prognosis of chRCC.MethodsThe chRCC data set of gene expression profiles and clinical data were obtained from the gdc‐client (https://portal.gdc.cancer.gov) deposited on The Cancer Genome Atlas (TCGA) data portal. Differentially expressed genes (DEGs) in chRCC, compared with normal samples, were analyzed by R packages “DESeq2,” “edgeR,” and “limma.” Heat maps, volcano plots, and principal component analysis (PCA) were performed for integrated analyses. GUniGO, mutant analysis, and survival plots were performed by R packages. A protein–protein interaction (PPI) network was generated and analyzed by R packages, online String software, and Cytoscape software. Survival analysis and gene expressing comparison in tumor and normal samples were used to detect the core genes of chRCC. Furthermore, the top interacting proteins were reanalyzed.ResultsA total of 306 upregulated genes and 678 downregulated genes were identified by a Venn diagram. Ten hub genes were extracted from PPI network. Furthermore, Alpha‐2‐Heremans‐Schmid‐glycoprotein (AHSG), one of 10 hub genes, was found to be associated with chRCC, and had a big difference in expression between survival and dead events. AHSG could predict potential prognostic and may be a diagnostic biomarker in chRCC.ConclusionThis study illustrated that AHSG may be a potential therapeutic target and prognostic genetic marker for chRCC.

Identification of crucial genes correlated with esophageal cancer by integrated high-throughput data analysis.

Esophageal cancer (ESCA) is one of the most deadly malignancies in the world. Although the management and treatment of patients with ESCA have improved, the overall 5-year survival rate is still very poor. The study aimed to identify potential key genes associated with the pathogenesis and prognosis of ESCA. In the study, integrated bioinformatics methods were used to screen differentially expressed genes (DEGs) between ESCA and normal tissue in the data set of gene expression profiles. The hub gene in DEGs was further analyzed by protein-protein interaction (PPI) network and survival analysis to explore its relationship with the pathogenesis and poor prognosis of ESCA. 134 up-regulated genes and 183 down-regulated genes were obtained in ESCA compared with normal tissues. Moreover, the PPI network was established with 176 nodes and 800 interactions. Ten hub genes (AURKA, CDC20, BUB1, TOP2A, ASPM, DLGAP5, TPX2, CENPF, UBE2C, and NEK2) were filtered out based on the degree value. Functional enrichment analysis indicated that a variety of extracellular related items and ECM-receptor interaction pathway were all correlated with the ESCA. The results of this study would provide some guidance for further study of diagnostic and prognostic biomarkers to promote ESCA treatment.

Analysis of the autophagy gene expression profile of pancreatic cancer based on autophagy-related protein microtubule-associated protein 1A/1B-light chain 3.

BACKGROUNDPancreatic cancer is a highly invasive malignant tumor. Expression levels of the autophagy-related protein microtubule-associated protein 1A/1B-light chain 3 (LC3) and perineural invasion (PNI) are closely related to its occurrence and development. Our previous results showed that the high expression of LC3 was positively correlated with PNI in the patients with pancreatic cancer. In this study, we further searched for differential genes involved in autophagy of pancreatic cancer by gene expression profiling and analyzed their biological functions in pancreatic cancer, which provides a theoretical basis for elucidating the pathophysiological mechanism of autophagy in pancreatic cancer and PNI.AIMTo identify differentially expressed genes involved in pancreatic cancer autophagy and explore the pathogenesis at the molecular level.METHODSTwo sets of gene expression profiles of pancreatic cancer/normal tissue (GSE16515 and GSE15471) were collected from the Gene Expression Omnibus. Significance analysis of microarrays algorithm was used to screen differentially expressed genes related to pancreatic cancer. Gene Ontology (GO) analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis were used to analyze the functional enrichment of the differentially expressed genes. Protein interaction data containing only differentially expressed genes was downloaded from String database and screened. Module mining was carried out by Cytoscape software and ClusterOne plug-in. The interaction relationship between the modules was analyzed and the pivot nodes between the functional modules were determined according to the information of the functional modules and the data of reliable protein interaction network.RESULTSBased on the above two data sets of pancreatic tissue total gene expression, 6098 and 12928 differentially expressed genes were obtained by analysis of genes with higher phenotypic correlation. After extracting the intersection of the two differential gene sets, 4870 genes were determined. GO analysis showed that 14 significant functional items including negative regulation of protein ubiquitination were closely related to autophagy. A total of 986 differentially expressed genes were enriched in these functional items. After eliminating the autophagy related genes of human cancer cells which had been defined, 347 differentially expressed genes were obtained. KEGG pathway analysis showed that the pathways hsa04144 and hsa04020 were related to autophagy. In addition, 65 clustering modules were screened after the protein interaction network was constructed based on String database, and module 32 contains the LC3 gene, which interacts with multiple autophagy-related genes. Moreover, ubiquitin C acts as a pivot node in functional modules to connect multiple modules related to pancreatic cancer and autophagy.CONCLUSIONThree hundred and forty-seven genes associated with autophagy in human pancreatic cancer were concentrated, and a key gene ubiquitin C which is closely related to the occurrence of PNI was determined, suggesting that LC3 may influence the PNI and prognosis of pancreatic cancer through ubiquitin C.

Regularization and grouping -omics data by GCA method: A transcriptomic case.

The paper presents the application of Grade Correspondence Analysis (GCA) and Grade Correspondence Cluster Analysis (GCCA) for ordering and grouping -omics datasets, using transcriptomic data as an example. Based on gene expression data describing 256 patients with Multiple Myeloma it was shown that the GCA method could be used to find regularities in the analyzed collections and to create characteristic gene expression profiles for individual groups of patients. GCA iteratively permutes rows and columns to maximize the tau-Kendall or rho-Spearman coefficients, which makes it possible to arrange rows and columns in such a way that the most similar ones remain in each other’s neighbourhood. In this way, the GCA algorithm highlights regularities in the data matrix. The ranked data can then be grouped using the GCCA method, and after that aggregated in clusters, providing a representation that is easier to analyze–especially in the case of large sets of gene expression profiles. Regularization of transcriptomic data, which is presented in this manuscript, has enabled division of the data set into column clusters (representing genes) and row clusters (representing patients). Subsequently, rows were aggregated (based on medians) to visualise the gene expression profiles for patients with Multiple Myeloma in each collection. The presented analysis became the starting point for characterisation of differentiated genes and biochemical processes in which they are involved. GCA analysis may provide an alternative analytical method to support differentiation and analysis of gene expression profiles characterising individual groups of patients.

Human umbilical cord matrix mesenchymal stem cells suppress the growth of breast cancer by expression of tumor suppressor genes.

Human and rat umbilical cord matrix mesenchymal stem cells (UCMSC) possess the ability to control the growth of breast carcinoma cells. Comparative analyses of two types of UCMSC suggest that rat UCMSC-dependent growth regulation is significantly stronger than that of human UCMSC. Their different tumoricidal abilities were clarified by analyzing gene expression profiles in the two types of UCMSC. Microarray analysis revealed differential gene expression between untreated naïve UCMSC and those co-cultured with species-matched breast carcinoma cells. The analyses screened 17 differentially expressed genes that are commonly detected in both human and rat UCMSC. The comparison between the two sets of gene expression profiles identified two tumor suppressor genes, adipose-differentiation related protein (ADRP) and follistatin (FST), that were specifically up-regulated in rat UCMSC, but down-regulated in human UCMSC when they were co-cultured with the corresponding species’ breast carcinoma cells. Over-expression of FST, but not ADRP, in human UCMSC enhanced their ability to suppress the growth of MDA-231 cells. The growth of MDA-231 cells was also significantly lower when they were cultured in medium conditioned with FST, but not ADRP over-expressing human UCMSC. In the breast carcinoma lung metastasis model generated with MDA-231 cells, systemic treatment with FST-over-expressing human UCMSC significantly attenuated the tumor burden. These results suggest that FST may play an important role in exhibiting stronger tumoricidal ability in rat UCMSC than human UCMSC and also implies that human UCMSC can be transformed into stronger tumoricidal cells by enhancing tumor suppressor gene expression.

Plant Omics Data Center: an integrated web repository for interspecies gene expression networks with NLP-based curation.

Comprehensive integration of large-scale omics resources such as genomes, transcriptomes and metabolomes will provide deeper insights into broader aspects of molecular biology. For better understanding of plant biology, we aim to construct a next-generation sequencing (NGS)-derived gene expression network (GEN) repository for a broad range of plant species. So far we have incorporated information about 745 high-quality mRNA sequencing (mRNA-Seq) samples from eight plant species (Arabidopsis thaliana, Oryza sativa, Solanum lycopersicum, Sorghum bicolor, Vitis vinifera, Solanum tuberosum, Medicago truncatula and Glycine max) from the public short read archive, digitally profiled the entire set of gene expression profiles, and drawn GENs by using correspondence analysis (CA) to take advantage of gene expression similarities. In order to understand the evolutionary significance of the GENs from multiple species, they were linked according to the orthology of each node (gene) among species. In addition to other gene expression information, functional annotation of the genes will facilitate biological comprehension. Currently we are improving the given gene annotations with natural language processing (NLP) techniques and manual curation. Here we introduce the current status of our analyses and the web database, PODC (Plant Omics Data Center; http://bioinf.mind.meiji.ac.jp/podc/), now open to the public, providing GENs, functional annotations and additional comprehensive omics resources.

Molecular Predictors of Locoregional Recurrence

Molecular Predictors of Locoregional Recurrence

Biomedical Informatics and Computational Biology for High-Throughput Data Analysis

The last few years have witnessed the emergence of several high-throughput (HTP) platforms that are based on the evolution of omic science through transcriptomics, proteomics, and metabolomics. With the huge data accumulation, the informatics and computational methods have become essential to understand the complexity of biomedical data. We lunched the special issue to address the demand for statistical models, biological omic data analyses, and meta-analysis of biomarker for complex diseases. Statistical models are always important for the understanding of complex data; in this issue, Y. Liang et al. proposed a new adaptive L1/2 shooting regularization method for variable selection based on the Cox's proportional hazards mode. Simulation and the real gene expression dataset analysis showed that the method is more accurate for variable selection than Lasso and adaptive Lasso methods. L. Tian et al. employed a nonlinear model to analyze time course gene expression data. They firstly developed a method for estimating the parameters in the nonlinear model and then utilized the model to perform the significance analysis of individually differentially expressed genes and clustering analysis of a set of gene expression profiles. The simulation and real-life biological data analyses showed that their methods outperform some existing methods. For the application of bioinformatics tools, Z. E et al. analyzed the gene expression in rice leaf blades at different temperatures. They analyzed the next-generation sequencing (NGS) datasets and characterized the transcription profiles of the samples in rice seedling leaf blades at 25°C and 30°C. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases were applied to study the effect of the temperature. Finally, they observed that temperature markedly regulates several superfamilies of transcription factors, including bZIP, MYB, and WRKY. Chronobiology is a field of biology that studies biological rhythms. R. Lopes et al. reviewed the bioinformatics or computational systems biology methods, tools, and databases which are very helpful to the understanding of the patterns and biological rhythms found in living organisms and the review will benefit the community for the application of bioinformatics in their future researches. For the understanding of protein structure and structure-function relationship, two papers are selected in this issue. M. Shambhu et al. studied the surface accessibility of hydrophobic residues by considering their conservation score and knowledge of flanking regions. The accuracy of prediction is therefore improved. In the other work, D. Lu et al. applied molecular dynamics to investigate the effects of X-linked agammaglobulinaemia associated with amino acid mutations in pleckstrin homology (PH) domain on the binding of the domain with inositol 1,3,4,5-tetrakisphosphate (Ins(1,3,4,5)P4). The mutations were then classified as “functional mutations”, and “folding mutations.” The result provided new insights into the biological function of the Btk-PH domain and related mutation-causing diseases. Finally, two works in this special issue are medical studies. The first is the work by B. Lv et al. They analyzed the 347 studies for the diagnostic value of tumor necrosis factor-α (TNF-α) and found that the northern hemisphere group in the TNF-α test has higher sensitivity and specificity than those of southern hemisphere group. The second medical study is an association study. Z. Liu et al. reported the significantly association of NCK2 with opiates addiction in African-origin men by a comprehensive analysis of a dataset from the Study of Addiction: Genetics and Environment (SAGE). They employed both SNP and gene based methods of analysis and identified a strong and significant association between a SNP in the NCK2 gene on chromosome 2 with opiates addiction in African-origin men. Bairong Shen Jian Ma Jiajun Wang Junbai Wang

Nonlinear-Model-Based Analysis Methods for Time-Course Gene Expression Data

Microarray technology has produced a huge body of time-course gene expression data and will continue to produce more. Such gene expression data has been proved useful in genomic disease diagnosis and drug design. The challenge is how to uncover useful information from such data by proper analysis methods such as significance analysis and clustering analysis. Many statistic-based significance analysis methods and distance/correlation-based clustering analysis methods have been applied to time-course expression data. However, these techniques are unable to account for the dynamics of such data. It is the dynamics that characterizes such data and that should be considered in analysis of such data. In this paper, we employ a nonlinear model to analyse time-course gene expression data. We firstly develop an efficient method for estimating the parameters in the nonlinear model. Then we utilize this model to perform the significance analysis of individually differentially expressed genes and clustering analysis of a set of gene expression profiles. The verification with two synthetic datasets shows that our developed significance analysis method and cluster analysis method outperform some existing methods. The application to one real-life biological dataset illustrates that the analysis results of our developed methods are in agreement with the existing results.

Identification of target genes of PAX3-FOXO1 in alveolar rhabdomyosarcoma

Rhabdomyosarcoma (RMS) is a soft tissue sarcoma categorized into two major subtypes: alveolar RMS (ARMS) and embryonal RMS (ERMS). Most ARMS express the PAX3-FOXO1 (P3F) fusion oncoprotein generated by the 2;13 chromosomal translocation. In the present study, the downstream target genes of P3F were identified by analyzing two independent sets of gene expression profiles: primary RMS tumors and RD ERMS cells transduced with inducible P3F constructs. We found 34 potential target genes (27 upregulated and 7 downregulated) that were significantly and differentially expressed between P3F-positive and P3F-negative categories, both in primary RMS tumors and in the inducible P3F cell culture system. Gene ontology analysis of microarray data of the inducible P3F cell culture system employed indicated apoptosis, cell death, development, and signal transduction as overrepresented significant functional categories found in both upregulated and downregulated genes. Therefore, among the 34 potential target genes, the expression of cell death-related [Gremlin1, cysteine knot superfamily 1, BMP antagonist 1 (GREM1) and death-associated protein kinase 1 (DAPK1)] and development-related [myogenic differentiation 1 (MYOD1) and hairy/enhancer-of-split related with YRPW motif 1 (HEY1)] genes were further investigated. The differential expression of GREM1, DAPK1, MYOD1 and HEY1 was confirmed in independent tumors and inducible cell culture systems. The expression of GREM1, DAPK1 and MYOD1 were significantly upregulated; HEY1 was significantly downregulated in independent P3F-positive ARMS tumors and transcriptionally active P3F cells, compared to those in ERMS tumors and transcriptionally inactive P3F cells. This study identified target genes of P3F and suggested that four downstream targets (GREM1, DAPK1, MYOD1 and HEY1) can contribute to the biological activities of P3F involved in growth suppression or cell death and myogenic differentiation.

Abstract 220: Human umbilical cord matrix-derived stem cells control tumor growth by tumor suppressor gene expression.

Abstract Umbilical cord matrix derived stem cells (UCMSC) have the potential to treat various diseases including cancer. We have shown that naïve human and rat UCMSC significantly attenuate proliferation of multiple cancer cells (Ganta et al., Cancer Res., 2009. Ayuzawa et al., Cancer Lett., 2009). However, the growth attenuation ability of rat UCMSC is stronger than that of human UCMSC. To clarify their different tumoricidal abilities, differential gene expression profiles of these two cells were studied by micro-array analysis using Illumina HumanRef-8 V2 for human and RatRef-12 BeadChip for rat UCMSC. The gene expression profiles were compared between UCMSC untreated and co-cultured with either human (MDA-231) or rat breast carcinoma cells (Mat B III). Selection criteria used for the screening of candidate genes associated with UCMSC-dependent tumoricidal ability are as follows; 1) gene expression difference should be at least a 1.5 fold between naive UCMSC and those co-cultured with mammary tumor cells; 2) they must encode secretory proteins; and 3) they must encode cell growth regulation-related proteins. These analyses screened 16 common genes from both human and rat UCMSC. The comparison between two sets of gene expression profiles identified that two genes, adipose-differentiation related protein (ADRP) and follistatin (FST), were specifically up-regulated in rat UCMSC, whereas they were down-regulated in human UCMSC when they were co-cultured with the corresponding species’ breast cancer cells. The suppression of either protein by adding specific neutralizing antibody in culture of rat UCMSC significantly abrogated their ability to attenuate cell growth. Over-expression of both genes by adenovirus vector in human UCMSC enhanced their ability to suppress the growth of MDA-231 cells. FST over-expression in MDA 231 cells also decreased their growth significantly. In the experimental lung metastasis model with MDA 231 cells in immunodeficient mice, the treatment with FST-over-expressing human UCMSC was administered three times, at one week intervals, via tail vein injection, starting 6 days after tumor cell inoculation. Examination 4 weeks after tumor inoculation showed a significant decrease in the number of tumor nodules in the lungs. Both ADRP and FST are considered to be tumor suppressor genes. These results suggest that both ADRP and FST may play important roles in exhibiting a stronger tumoricidal ability in rat UCMSC than human UCMSC. These results imply that human UCMSC can be transformed into stronger tumoricidal cells by enhancing tumor suppressor gene expression. This work was supported by the Kansas State University (KSU) Terry C. Johnson Center for Basic Cancer Research, KSU College of Veterinary Medicine Dean's Fund, NIH RR017686, RR15563, CA135599, Kansas Bioscience Authority research grant. Citation Format: Naomi Ohta, Susumu Ishiguro, Atsushi Kawabata, Deepthi Uppalapati, Marla Pyle, Deryl Troyer, Supriyo De, Yongqing Zhang, Kevin G. Becker, Masaaki Tamura. Human umbilical cord matrix-derived stem cells control tumor growth by tumor suppressor gene expression. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 220. doi:10.1158/1538-7445.AM2013-220

SERPINA6, BEX1, AGTR1, SLC26A3, and LAPTM4B are markers of resistance to neoadjuvant chemotherapy in HER2-negative breast cancer

Response rates to chemotherapy remain highly variable in breast cancer patients. We set out to identify genes associated with chemotherapy resistance. We analyzed what is currently the largest single-institute set of gene expression profiles derived from breast cancers prior to a single neoadjuvant chemotherapy regimen (dose-dense doxorubicin and cyclophosphamide). We collected, gene expression-profiled, and analyzed 178 HER2-negative breast tumor biopsies ("NKI dataset"). We employed a recently developed approach for detecting imbalanced differential signal (DIDS) to identify markers of resistance to treatment. In contrast to traditional methods, DIDS is able to identify markers that show aberrant expression in only a small subgroup of the non-responder samples. We found a number of markers of resistance to anthracycline-based chemotherapy. We validated our findings in three external datasets, totaling 456 HER2-negative samples. Since these external sets included patients who received differing treatment regimens, the validated markers represent markers of general chemotherapy resistance. There was a highly significant overlap in the markers identified in the NKI dataset and the other three datasets. Five resistance markers, SERPINA6, BEX1, AGTR1, SLC26A3, and LAPTM4B, were identified in three of the four datasets (p value overlap < 1 × 10(-6)). These five genes identified resistant tumors that could not have been identified by merely taking ER status or proliferation into account. The identification of these genes might lead to a better understanding of the mechanisms involved in (clinically) observed chemotherapy resistance and could possibly assist in the recognition of breast cancers in which chemotherapy does not contribute to response or survival.

Identification of a Gene Expression Signature Common to Distinct Cancer Pathways

Mutations in cancer-causing genes induce changes in gene expression programs critical for malignant cell transformation. Publicly available gene expression profiles produced by modulating the expression of distinct cancer genes may therefore represent a rich resource for the identification of gene signatures common to seemingly unrelated cancer genes. We combined automatic retrieval with manual validation to obtain a data set of high-quality gene microarray profiles. This data set was used to create logical models of the signaling events underlying the observed expression changes produced by various cancer genes and allowed to uncover unknown and verifiable interactions. Data clustering revealed novel sets of gene expression profiles commonly regulated by distinct cancer genes. Our method allows retrieval of significant new information and testable hypotheses from a pool of deposited cancer gene expression experiments that are otherwise not apparent or appear insignificant from single measurements. The complete results are available through a web-application at http://biodata.ethz.ch/cgi-bin/geologic.

Integrative inference of gene-regulatory networks in Escherichia coli using information theoretic concepts and sequence analysis

BackgroundAlthough Escherichia coli is one of the best studied model organisms, a comprehensive understanding of its gene regulation is not yet achieved. There exist many approaches to reconstruct regulatory interaction networks from gene expression experiments. Mutual information based approaches are most useful for large-scale network inference.ResultsWe used a three-step approach in which we combined gene regulatory network inference based on directed information (DTI) and sequence analysis. DTI values were calculated on a set of gene expression profiles from 19 time course experiments extracted from the Many Microbes Microarray Database. Focusing on influences between pairs of genes in which one partner encodes a transcription factor (TF) we derived a network which contains 878 TF - gene interactions of which 166 are known according to RegulonDB. Afterward, we selected a subset of 109 interactions that could be confirmed by the presence of a phylogenetically conserved binding site of the respective regulator. By this second step, the fraction of known interactions increased from 19% to 60%. In the last step, we checked the 44 of the 109 interactions not yet included in RegulonDB for functional relationships between the regulator and the target and, thus, obtained ten TF - target gene interactions. Five of them concern the regulator LexA and have already been reported in the literature. The remaining five influences describe regulations by Fis (with two novel targets), PhdR, PhoP, and KdgR. For the validation of our approach, one of them, the regulation of lipoate synthase (LipA) by the pyruvate-sensing pyruvate dehydrogenate repressor (PdhR), was experimentally checked and confirmed.ConclusionsWe predicted a set of five novel TF - target gene interactions in E. coli. One of them, the regulation of lipA by the transcriptional regulator PdhR was validated experimentally. Furthermore, we developed DTInfer, a new R-package for the inference of gene-regulatory networks from microarrays using directed information.

Heuristic Approach to Sparse Approximation of Gene Regulatory Networks

Determining the structure of the gene regulatory network using the information in genomewide profiles of mRNA abundance, such as microarray data, poses several challenges. Typically, "static" rather than dynamical profile measurements, such as those taken from steady state tissues in various conditions, are the starting point. This makes the inference of causal relationships between genes difficult. Moreover, the paucity of samples relative to the gene number leads to problems such as overfitting and underconstrained regression analysis. Here we present a novel method for the sparse approximation of gene regulatory networks that addresses these issues. It is formulated as a sparse combinatorial optimization problem which has a globally optimal solution in terms of l(0) norm error. In order to seek an approximate solution of the l(0) optimization problem, we consider a heuristic approach based on iterative greedy algorithms. We apply our method to a set of gene expression profiles comprising of 24,102 genes measured over 79 human tissues. The inferred network is a signed directed graph, hence predicts causal relationships. It exhibits typical characteristics of regulatory networks organism with partially known network topology, such as the average number of inputs per gene as well as the in-degree and out-degree distribution.

Site-Specific Molecular Signatures Predict Aggressive Disease in HNSCC

It is known that head and neck squamous cell carcinomas (HNSCC) originating from different anatomic locations can exhibit varying behavior that is not predictable by histopathology of the primary tumor. Using a microarray containing 27,323 cDNA clones, we generated sets of gene expression profiles for 36 HNSCC primary tumors (12 oral cavity, 12 oropharynx, and 12 larynx/hypopharynx). From these datasets, we ranked genes according to their ability to differentiate between patients whose disease progressed within a 24 month period (aggressive phenotype) and those that did not (non-aggressive phenotype) based on levels of gene expression. A merging of datasets from the three sites revealed that only a fraction of identified genes were shared between any two sites. This contrasted greatly with the significant overlap (approximately 50%) in down-regulated genes identified in tumor/normal comparisons using cases both from oropharynx and larynx/hypopharynx. From these data, we conclude that HNSCC tumors originating from different anatomic sites share consistent changes in gene expression when comparing primary tumors to normal adjacent mucosa; these common changes most likely reflect alterations required for tumor development. In contrast, once a tumor has developed, tumor-host interactions at the different anatomic sites are likely responsible for the site-specific signatures associated with aggressive versus non-aggressive disease. Predictions of outcome based on gene expression profiling are therefore heavily influenced by the anatomic site of the primary tumor.

Measles virus induces cell-type specific changes in gene expression

Measles virus (MV) causes various responses including the induction of immune responses, transient immunosuppression and establishment of long-lasting immunity. To obtain a comprehensive view of the effects of MV infection on target cells, DNA microarray analyses of two different cell-types were performed. An epithelial (293SLAM; a 293 cell line stably expressing SLAM) and lymphoid (COBL-a) cell line were inoculated with purified wild-type MV. Microarray analyses revealed significant differences in the regulation of cellular gene expression between these two different cells. In 293SLAM cells, upregulation of genes involved in the antiviral response was rapidly induced; in the later stages of infection, this was followed by regulation of many genes across a broad range of functional categories. On the other hand, in COBL-a cells, only a limited set of gene expression profiles was modulated after MV infection. Since it was reported that V protein of MV inhibited the IFN signaling pathway, we performed a microarray analysis using V knockout MV to evaluate V protein's effect on cellular gene expression. The V knockout MV displayed a similar profile to that of parental MV. In particular, in COBL-a cells infected with the virus, no alteration of cellular gene expression, including IFN signaling, was observed. Furthermore, IFN signaling analyzed in vitro was completely suppressed by MV infection in the COBL-a cells. These results reveal that MV induces different cellular responses in a cell-type specific manner. Microarray analyses will provide us useful information about potential mechanisms of MV pathogenesis.