Linear Models For Microarray Data Research Articles

The Effect of Statins on Blood Gene Expression in COPD.

BackgroundCOPD is currently the fourth leading cause of death worldwide. Statins are lipid lowering agents with documented cardiovascular benefits. Observational studies have shown that statins may have a beneficial role in COPD. The impact of statins on blood gene expression from COPD patients is largely unknown.ObjectiveIdentify blood gene signature associated with statin use in COPD patients, and the pathways underpinning this signature that could explain any potential benefits in COPD.MethodsWhole blood gene expression was measured on 168 statin users and 451 non-users from the ECLIPSE study using the Affymetrix Human Gene 1.1 ST microarray chips. Factor Analysis for Robust Microarray Summarization (FARMS) was used to process the expression data. Differential gene expression analysis was undertaken using the Linear Models for Microarray data (Limma) package adjusting for propensity score and surrogate variables. Similarity of the expression signal with published gene expression profiles was performed in ProfileChaser.Results25 genes were differentially expressed between statin users and non-users at an FDR of 10%, including LDLR, CXCR2, SC4MOL, FAM108A1, IFI35, FRYL, ABCG1, MYLIP, and DHCR24. The 25 genes were significantly enriched in cholesterol homeostasis and metabolism pathways. The resulting gene signature showed correlation with Huntington’s disease, Parkinson’s disease and acute myeloid leukemia gene signatures.ConclusionThe blood gene signature of statins’ use in COPD patients was enriched in cholesterol homeostasis pathways. Further studies are needed to delineate the role of these pathways in lung biology.

In search of key genes associated with endometriosis using bioinformatics approach

ObjectiveThe aim of this study was to identify key genes associated with endometriosis. Study designMicroarray data GSE7846 was downloaded from Gene Expression Omnibus. The differentially expressed genes (DEGs) in human endometrial endothelial cells derived from eutopic endometrium of patients with endometriosis compared with controls without endometriosis were identified using Linear Models for Microarray Data (LIMMA) package in R. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway and gene ontology (GO) enrichment analyses were performed using DAVID. Moreover, the protein–protein interaction (PPI) network was constructed by STRING and subsequently significantly enriched modules were mined by ClusterONE. Finally, protein domains and KEGG pathway enrichment analyses were performed for PPI modules. ResultsA total of 687 DEGs were identified, including 584 up- and 103 down-regulated genes. The up-regulated DEGs, such as epidermal growth factor (EGF) and interleukin 1 beta (IL-1β) were significantly enriched in KEGG pathways of focal adhesion, ECM–receptor interaction and calcium signaling pathway. Similarly, only one module was obtained form PPI network, and the genes, like angiotensin II receptor, Type 1 (AGTR1) in the module were mainly enriched in protein domain of rhodopsin-like G protein-coupled receptors and most altered pathways of neuroactive ligand–receptor interaction, calcium signaling pathway and vascular smooth muscle contraction. ConclusionsOur findings indicate that EGF, IL-1β and AGTR1 may play important roles in the pathogenesis of endometriosis.

Abstract 2885: Differential expression of long non-coding RNAs in ovarian cancer-associated fibroblasts versus normal ovarian fibroblasts

Abstract Ovarian cancer is the most lethal gynecological malignancy in women. Expression profiling of the high-grade serous subtype has revealed that patients exhibiting a “stromal expression signature” demonstrate the poorest prognosis1. However, genetic aberrations in ovarian cancer-associated fibroblasts (CAFs) are extremely rare2, raising the possibility that alternative mechanisms that regulate gene expression, such as long non-coding RNAs (lncRNAs), are present in CAFs. LncRNAs are polyadenylated RNA transcripts that do not encode for protein, but have been shown to correlate with cancer progression and outcome. Therefore, our aim was to investigate the presence of differentially expressed lncRNAs in ovarian CAFs versus normal ovarian fibroblasts. CAFs were laser-capture microdissected from 51 advanced-stage high-grade serous ovarian cancers and 10 normal ovaries removed from women for non-cancer reasons. RNA was extracted from the microdissected samples and expression analyzed using Affymetrix U133 Plus 2.0 Arrays. Probes previously identified as lncRNAs were used in this analysis3. Samples were normalized and background corrected using the robust multi-arrray average (RMA) method and expression values were log2 transformed. Differentially expressed probes between cancer versus normal samples were calculated using linear models for microarray data (limma) package from Bioconductor and a moderated t-statistic used to assess significance. Based on a significance cutoff of fold-change > 2 and a p-value < 0.05, 54 lncRNAs were identified as differentially expressed in cancer versus normal fibroblasts. These included lncRNAs previously implicated in ovarian cancer such as XIST and H19, the p53-regulated lncRNA TUG1, as well as several lncRNAs that are known to play a role in other cancer types such as NEAT1, GAS5 and CASC2. In addition, 16 lncRNAs were differentially expressed between patients with short (<15 months) versus patients with long (>15 months) overall survival. These included MALAT1, MEG3, TUG1, XIST and CRNDE. In summary, we have identified lncRNAs that are differentially expressed in ovarian CAFs compared to normal ovarian fibroblasts, as well as lncRNAs that are differentially expressed based on survival. Given that CAFs are known to promote tumor progression and metastasis, these lncRNAs present suitable candidates for their role in the tumor-promoting function of CAFs.

Differential Expression of microRNA in Cerebrospinal Fluid as a Potential Novel Biomarker for Alzheimer’s Disease

The need to find a better reflection of Alzheimer's disease (AD) pathophysiology led us to investigate differential expression of microRNA (miRNA) in cerebrospinal fluid (CSF) of AD patients compared to matched controls, using a genome-wide data-driven approach. From the Amsterdam Dementia Cohort, we selected 19 AD patients with CSF indicative of AD pathophysiology and 19 age and gender-matched controls without CSF evidence of AD (67 ± 6 years old, 20 [53%] female). We measured 754 miRNA in CSF using qRT-PCR (Taqman Array MicroRNA cards A and B, v3.0) according to the Megaplex Taqman protocol. Hierarchical cluster analysis was performed and groups were compared using Linear Models for Microarray Data, a modified t-test. We performed validation analysis using qRT-PCR single assays. 144 ± 66 miRNA could be detected using Megaplex array analysis (19% ). Mean Ct (average 32.4 ± 0.5) was correlated to age (r = 0.52, p = 0.001). Five miRNA were differentially expressed in CSF of AD patients. None of these could be replicated. After stratification by age, seven miRNA showed differential expression in late-onset AD, of which lower abundance of let-7a was replicated (log10RQ -1.46, p < 0.05). In early-onset AD, twelve miRNA were differentially expressed of which lower abundance of miRNA-532-3p remained borderline significant (log10RQ -1.27, p = 0.05). Although we could not consistently separate AD patients and controls in the whole group, we have found indications miRNA in CSF are able to reflect aging and perhaps also heterogeneity in AD. Further investigation requires optimizing RNA input, while maintaining strict age matching.

Comparison of High-Level Microarray Analysis Methods in the Context of Result Consistency.

MotivationWhen we were asked for help with high-level microarray data analysis (on Affymetrix HGU-133A microarray), we faced the problem of selecting an appropriate method. We wanted to select a method that would yield "the best result" (detected as many "really" differentially expressed genes (DEGs) as possible, without false positives and false negatives). However, life scientists could not help us – they use their "favorite" method without special argumentation. We also did not find any norm or recommendation. Therefore, we decided to examine it for our own purpose. We considered whether the results obtained using different methods of high-level microarray data analyses – Significant Analysis of Microarrays, Rank Products, Bland-Altman, Mann-Whitney test, T test and the Linear Models for Microarray Data – would be in agreement. Initially, we conducted a comparative analysis of the results on eight real data sets from microarray experiments (from the Array Express database). The results were surprising. On the same array set, the set of DEGs by different methods were significantly different. We also applied the methods to artificial data sets and determined some measures that allow the preparation of the overall scoring of tested methods for future recommendation.ResultsWe found a very low level concordance of results from tested methods on real array sets. The number of common DEGs (detected by all six methods on fixed array sets, checked on eight array sets) ranged from 6 to 433 (22,283 total array readings). Results on artificial data sets were better than those on the real data. However, they were not fully satisfying. We scored tested methods on accuracy, recall, precision, f-measure and Matthews correlation coefficient. Based on the overall scoring, the best methods were SAM and LIMMA. We also found TT to be acceptable. The worst scoring was MW. Based on our study, we recommend: 1. Carefully taking into account the need for study when choosing a method, 2. Making high-level analysis with more than one method and then only taking the genes that are common to all methods (which seems to be reasonable) and 3. Being very careful (while summarizing facts) about sets of differentially expressed genes: different methods discover different sets of DEGs.

Gene expression profiling of Gram-negative bacteria-induced inflammation in human whole blood: The role of complement and CD14-mediated innate immune response.

Non-sterile pathogen-induced sepsis and sterile inflammation like in trauma or ischemia–reperfusion injury may both coincide with the life threatening systemic inflammatory response syndrome and multi-organ failure. Consequently, there is an urgent need for specific biomarkers in order to distinguish sepsis from sterile conditions. The overall aim of this study was to uncover putative sepsis biomarkers and biomarker pathways, as well as to test the efficacy of combined inhibition of innate immunity key players complement and Toll-like receptor co-receptor CD14 as a possible therapeutic regimen for sepsis. We performed whole blood gene expression analyses using microarray in order to profile Gram-negative bacteria-induced inflammatory responses in an ex vivo human whole blood model. The experiments were performed in the presence or absence of inhibitors of complement proteins (C3 and CD88 (C5a receptor 1)) and CD14, alone or in combination. In addition, we used blood from a C5-deficient donor. Anti-coagulated whole blood was challenged with heat-inactivated Escherichia coli for 2 h, total RNA was isolated and microarray analyses were performed on the Affymetrix GeneChip Gene 1.0 ST Array platform. The initial experiments were performed in duplicates using blood from two healthy donors. C5-deficiency is very rare, and only one donor could be recruited. In order to increase statistical power, a technical replicate of the C5-deficient samples was run. Subsequently, log2-transformed intensities were processed by robust multichip analysis and filtered using a threshold of four. In total, 73 microarray chips were run and analyzed. The normalized and filtered raw data have been deposited in NCBI's Gene Expression Omnibus (GEO) and are accessible with GEO Series accession number GSE55537. Linear models for microarray data were applied to estimate fold changes between data sets and the respective multiple testing adjusted p-values (FDR q-values). The interpretation of the data has been published by Lau et al. in an open access article entitled “CD14 and Complement Crosstalk and Largely Mediate the Transcriptional Response to Escherichia coli in Human Whole Blood as revealed by DNA Microarray” (Lau et al., 2015).

MP68-07 THE MEDIATOR COMPLEX SUBUNIT MED12 IS IMPLICATED IN THE PROGRESSION OF UROTHELIAL CELL CARCINOMA OF THE BLADDER

INTRODUCTION AND OBJECTIVES: Epigenetic changes, including promoter CpG island hypermethylation, occur frequently in bladder cancer (BC) and may be exploited as a means for BC detection and distinction between high-grade (HG) and low-grade (LG) disease. METHODS: To determine epigenetic differences between LG and HG BC we performed genome-wide methylation analysis with Agilent Human CpG Island Microarrays on a panel of fresh frozen BC tissue samples. Using Linear Models for Microarray Data (LIMMA) and local-pooled-error (LPE) approaches and unsupervised hierarchical clustering we identified 990 probes comprising a 32-gene panel that completely distinguished LG from HG BC based on methylation status. We selected five representative differentially methylated genes (DMGs) from our 32-gene panel for methylation validation by real-time PCRbased MethyLight in a series of 40 LG cases ageand sex-matched to 40 HG cases. RESULTS: Among the DMGs tested, including EOMES, GP5, PAX6, TCF4, and ZSCAN12, we identified that EOMES, GP5, and ZSCAN12 methylation significantly differs between normal, LG, and HG disease. GP5 and ZSCAN12, two novel methylated genes in BC, are significantly hypermethylated in HG versus LG BC (p1⁄40.006 and 0.028, respectively). Pathway enrichment analysis and functional annotation determined the most frequently methylated pathways in HG BC were enriched for anterior/posterior pattern specification, embryonic skeletal system development, and neuron fate commitment. The molecular functions of the most enriched genes were involved in DNA binding and transcription factor activity. CONCLUSIONS: These results indicate the ability to distinguish normal tissue from cancer, as well as LG from HG tumours, and reveal important pathways dysregulated in HG BC. Ultimately, the creation of a methylation panel able to distinguish between disease phenotypes will improve disease management and patient outcomes.

MP68-06 TISSUE-BASED DNA METHYLATION PROFILING ESTABLISHES A NOVEL PANEL OF BIOMARKERS FOR DISCRIMINATION OF HIGH-GRADE VERSUS LOW-GRADE BLADDER CANCER

You have accessJournal of UrologyBladder Cancer: Basic Research IV1 Apr 2015MP68-06 TISSUE-BASED DNA METHYLATION PROFILING ESTABLISHES A NOVEL PANEL OF BIOMARKERS FOR DISCRIMINATION OF HIGH-GRADE VERSUS LOW-GRADE BLADDER CANCER Andrea Savio, Ekaterina Olkhov-Mitsel, Ken Kron, Thomas Hermanns, Bas van Rhijn, Alexandre Zlotta, Theodorus van der Kwast, and Bharati Bapat Andrea SavioAndrea Savio More articles by this author , Ekaterina Olkhov-MitselEkaterina Olkhov-Mitsel More articles by this author , Ken KronKen Kron More articles by this author , Thomas HermannsThomas Hermanns More articles by this author , Bas van RhijnBas van Rhijn More articles by this author , Alexandre ZlottaAlexandre Zlotta More articles by this author , Theodorus van der KwastTheodorus van der Kwast More articles by this author , and Bharati BapatBharati Bapat More articles by this author View All Author Informationhttps://doi.org/10.1016/j.juro.2015.02.2468AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookTwitterLinked InEmail INTRODUCTION AND OBJECTIVES Epigenetic changes, including promoter CpG island hypermethylation, occur frequently in bladder cancer (BC) and may be exploited as a means for BC detection and distinction between high-grade (HG) and low-grade (LG) disease. METHODS To determine epigenetic differences between LG and HG BC we performed genome-wide methylation analysis with Agilent Human CpG Island Microarrays on a panel of fresh frozen BC tissue samples. Using Linear Models for Microarray Data (LIMMA) and local-pooled-error (LPE) approaches and unsupervised hierarchical clustering we identified 990 probes comprising a 32-gene panel that completely distinguished LG from HG BC based on methylation status. We selected five representative differentially methylated genes (DMGs) from our 32-gene panel for methylation validation by real-time PCR-based MethyLight in a series of 40 LG cases age- and sex-matched to 40 HG cases. RESULTS Among the DMGs tested, including EOMES, GP5, PAX6, TCF4, and ZSCAN12, we identified that EOMES, GP5, and ZSCAN12 methylation significantly differs between normal, LG, and HG disease. GP5 and ZSCAN12, two novel methylated genes in BC, are significantly hypermethylated in HG versus LG BC (p=0.006 and 0.028, respectively). Pathway enrichment analysis and functional annotation determined the most frequently methylated pathways in HG BC were enriched for anterior/posterior pattern specification, embryonic skeletal system development, and neuron fate commitment. The molecular functions of the most enriched genes were involved in DNA binding and transcription factor activity. CONCLUSIONS These results indicate the ability to distinguish normal tissue from cancer, as well as LG from HG tumours, and reveal important pathways dysregulated in HG BC. Ultimately, the creation of a methylation panel able to distinguish between disease phenotypes will improve disease management and patient outcomes. © 2015 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 193Issue 4SApril 2015Page: e860 Advertisement Copyright & Permissions© 2015 by American Urological Association Education and Research, Inc.MetricsAuthor Information Andrea Savio More articles by this author Ekaterina Olkhov-Mitsel More articles by this author Ken Kron More articles by this author Thomas Hermanns More articles by this author Bas van Rhijn More articles by this author Alexandre Zlotta More articles by this author Theodorus van der Kwast More articles by this author Bharati Bapat More articles by this author Expand All Advertisement Advertisement PDF downloadLoading ...

Predicting tumour stages of lung cancer adenocarcinoma tumours from pooled microarray data using machine learning methods

This paper involved a novel method combination of predicting lung cancer adenocarcinoma stages using differential expression analysis for gene selection (linear modelling) and machine learning methods (support vector machines (SVMs) and random forest) on a pooled dataset from multiple publicly available microarray experiments. The raw data of 123 tumour microarray samples were initially preprocessed and analysed using robust multi-array average (RMA) and linear models for microarray data (LIMMA) to screen a list of significantly differential expressed genes, where two gene lists were identified according to different experimental settings. These two gene lists were then placed into the SVM model and random forest (RF) model for further investigation to build the prediction models. As result, both the SVM and RF models provided a lung cancer stage prediction model with the accuracy ranging from 67% to 71%.

T-Test at the Probe Level: An Alternative Method to Identify Statistically Significant Genes for Microarray Data.

Microarray data analysis typically consists in identifying a list of differentially expressed genes (DEG), i.e., the genes that are differentially expressed between two experimental conditions. Variance shrinkage methods have been considered a better choice than the standard t-test for selecting the DEG because they correct the dependence of the error with the expression level. This dependence is mainly caused by errors in background correction, which more severely affects genes with low expression values. Here, we propose a new method for identifying the DEG that overcomes this issue and does not require background correction or variance shrinkage. Unlike current methods, our methodology is easy to understand and implement. It consists of applying the standard t-test directly on the normalized intensity data, which is possible because the probe intensity is proportional to the gene expression level and because the t-test is scale- and location-invariant. This methodology considerably improves the sensitivity and robustness of the list of DEG when compared with the t-test applied to preprocessed data and to the most widely used shrinkage methods, Significance Analysis of Microarrays (SAM) and Linear Models for Microarray Data (LIMMA). Our approach is useful especially when the genes of interest have small differences in expression and therefore get ignored by standard variance shrinkage methods.

Identification of an Epithelial-to-Mesenchymal Transition (EMT)-like Programme in t(4;14)-Positive Multiple Myeloma Reveals Novel Targets for Therapeutic Intervention

Introduction: Prognosis for multiple myeloma (MM) patients has improved dramatically since the introduction of novel therapies. However, prognosis in the t(4;14) subgroup, characterized by expression of the histone methyltransferase MMSET, remains relatively poor due to the acquisition of a highly aggressive, motile and invasive phenotype and a generally modest response to therapy. We have previously demonstrated that MMSET promotes an epithelial-to-mesenchymal transition (EMT) in prostate cancer1. While the term EMT is not commonly used to describe MM, we hypothesise that an EMT-like process plays a critical role in t(4;14)­-positive MM disease pathogenesis. In this study, we conducted a comprehensive evaluation of the association between t(4;14) and an EMT-related gene expression signature in MM and identify N-cadherin as a therapeutically-targetable EMT-related gene in t(4;14)-positive MM.Methods and results: Expression of a core EMT-related signature, comprising 169 mesenchymal genes and 49 epithelial genes, was assessed in CD138-selected MM plasma cells from newly-diagnosed MM patients in four independent microarray datasets (E-GEOD-19784 [n = 327 MM patients], E-GEOD-26863 [n = 304], E-MTAB-317 [n = 226] and E-MTAB-363 [n = 155]), accessed through ArrayExpress (EMBL-EBI). In each dataset, gene expression was compared in t(4;14)-positive and t(4;14)-negative patients using linear models for microarray data (LIMMA). Twenty six mesenchymal genes were found to be upregulated in t(4;14)-positive patients across the 4 datasets (p < 0.05, Fisher’s method). A general loss of epithelial genes was not observed, likely due to the haematopoietic origins of MM. Upregulated genes included key EMT drivers (TWIST1, SOX9, TCF4) and genes associated with the cytoskeleton (VIM), adhesion and migration (CDH2, ITGB1, NCAM1) and signalling pathways involved in EMT (BMPR1A, IL6R, TGFB2). The t(4;14)-mediated regulation of EMT-related genes including TWIST1, CDH2, BMPR1A and ITGB1 was confirmed by assessing the effects of MMSET knockdown, knockout and add-back in the t(4;14)-positive human myeloma cell line KMS-11.We have previously demonstrated that N-cadherin (CDH2) expression is elevated in plasma cells from approximately 50% of newly diagnosed multiple myeloma (MM) patients and that elevated serum N-cadherin is associated with poor prognosis2. In order to identify whether N-cadherin is a potential therapeutic target in t(4;14)-positive myeloma, the effects of an N-cadherin peptide inhibitor ADH-1 (Exherin™) or shRNA-mediated N-cadherin knock-down were assessed on MM PC adhesion and proliferation in vivo and in vitro. The role of N-cadherin in MM tumour establishment and intramedullary growth was investigated using the C57BL/KalwRijHsd mouse model of MM. In this model, intravenously injected luciferase-expressing mouse MM PC cells (5TGM1-SFG) home to the bone marrow and initiate systemic MM disease. C57BL/KalwRijHsd mice bearing 5TGM1-SFG N-cadherin knock-down cells had significantly reduced tumour burden, as assessed by bioluminescent imaging and serum paraprotein levels, after 4 weeks compared with mice bearing control 5TGM1-SFG cells. Furthermore, daily intraperitoneal ADH-1 administration (100 mg/kg/day) to 5TGM1-SFG cell-bearing mice significantly significantly decreased tumour burden when administered at the time of tumour inoculation, but had no effect on established tumours. These results suggest that N-cadherin plays a role in extravasation, homing and/or tumour establishment in vivo.Conclusion: This study identified an extensive EMT-like gene expression signature driven by MMSET in t(4;14)-positive MM patients. This provides insight as to how the t(4;14) translocation leads poor prognostic outcomes in up to 20% of MM patients. Furthermore, these studies demonstrate a potential role for N-cadherin in MM tumour dissemination in t(4;14)-positive MM and suggest that N-cadherin represents a novel candidate for therapeutic targeting in these patients.1Ezponda et al. Oncogene, 20132Vandyke et al. Br J Haematol, 2013 DisclosuresNo relevant conflicts of interest to declare.

Molecular mechanisms of osteoarthritis using gene microarrays

This study aimed to investigate the molecular mechanisms of osteoarthritis (OA) by microarray analysis. Three gene expression datasets GSE1919, 19664 and 55235 were downloaded from the Gene Expression Omnibus, and data of OA samples and healthy controls were used. After data preprocessing, differential expression analysis between the OA group and controls was performed using LIMMA (Linear Models for Microarray Data) package and genes with |log2FC (fold change)|>1 and P<0.05 were screened as DEGs (differentially expressed genes). The screened DEGs were then subject to functional annotation and pathway enrichment analysis using DAVID (Database for Annotation Visualization and Integrated Discovery). Next, gene-set enrichment analysis was performed using Enrichment map Cytoscape plug-in, followed by detecting sub-networks using clusterONE. Finally, risk subpathways were screened using iSubpathwayMiner package. A total of 141 DEGs were screened, including 52 up-regulated ones and 89 down-regulated ones. These DEGs were enriched in 48 GO terms that were mainly related to locomotory behavior, taxis, adhesion, and 11 pathways that were related to cytokine-cytokine receptor interaction, ECM-receptor interaction, focal adhesion, as well as several signaling pathways. The enrichment map enriched gene-sets mainly related to cell death and apoptosis, and extracellular components. The risk pathways up-regulated DEGs were exclusively related to arachidonic acid metabolism and glycosphingolipid biosynthesis, and the top two risk pathways were tyrosine metabolism for the down-regulated ones. From this study we conclude that genes involved in cell death and apoptosis, as well as cell–extracellular matrix interaction, may be essential for OA pathogenesis by altering multiple signaling pathways.

A three-gene panel that distinguishes benign from malignant thyroid nodules.

Reliable preoperative diagnosis of malignant thyroid tumors remains challenging because of the inconclusive cytological examination of fine-needle aspiration biopsies. Although numerous studies have successfully demonstrated the use of high-throughput molecular diagnostics in cancer prediction, the application of microarrays in routine clinical use remains limited. Our aim was, therefore, to identify a small subset of genes to develop a practical and inexpensive diagnostic tool for clinical use. We developed a two-step feature selection method composed of a linear models for microarray data (LIMMA) linear model and an iterative Bayesian model averaging model to identify a suitable gene set signature. Using one public dataset for training, we discovered a three-gene signature dipeptidyl-peptidase 4 (DPP4), secretogranin V (SCG5) and carbonic anhydrase XII (CA12). We then evaluated the robustness of our gene set using three other independent public datasets. The gene signature accuracy was 85.7, 78.8 and 85.7%, respectively. For experimental validation, we collected 70 thyroid samples from surgery and our three-gene signature method achieved an accuracy of 94.3% by quantitative polymerase chain reaction (QPCR) experiment. Furthermore, immunohistochemistry in 29 samples showed proteins expressed by these three genes are also differentially expressed in thyroid samples. Our protocol discovered a robust three-gene signature that can distinguish benign from malignant thyroid tumors, which will have daily clinical application.

Key genes associated with osteoporosis revealed by genome wide gene expression analysis.

Gene expression profiles of circulating monocytes were analyzed to identify key genes associated with osteoporosis. Raw microarray data were downloaded from gene expression omnibus under accession number GSE7158, including 8 microarray dataset for patients with high peak bone mass (PBM) and 8 for low PBM. Package linear models for microarray data of R was adopted to screen out differentially expressed genes (DEGs). Gene ontology enrichment analysis and Kyoto encyclopedia of genes and genomes pathway analysis were performed with plug-ins of cytoscape. Protein-protein interaction network was constructed using FunCoup. A total of 283 DEGs were identified in low-PBM group, including 135 up- and 148 down-regulated genes. A considerable part of DEGs were localized in plasma membrane. Several ion transport-related pathways were revealed, such as mineral absorption and carbohydrate digestion and absorption. A range of DEGs were identified and some of them were related to calcium transport as well as osteoporosis. These findings are helpful in disclosing the pathogenetic mechanisms of osteoporosis.

Poster #M104 SETTING PSYCHOPATHOLOGY IN MOTION: A NETWORK PERSPECTIVE

In the current study, we planned to predict the optimal gene functions for osteosarcoma (OS) by integrating network-based method with guilt by association (GBA) principle (called as network-based gene function inference approach) based on gene ontology (GO) data and gene expression profile. To begin with, differentially expressed genes (DEGs) were extracted using linear models for microarray data (LIMMA) package. Then, construction of differential co-expression network (DCN) relying on DEGs was implemented, and sub-DCN was identified using Spearman correlation coefficient (SCC). Subsequently, GO annotations for OS were collected according to known confirmed database and DEGs. Ultimately, gene functions were predicted by means of GBA principle based on the area under the curve (AUC) for GO terms, and we determined GO terms with AUC >0.7 as the optimal gene functions for OS. Totally, 123 DEGs and 137 GO terms were obtained for further analysis. A DCN was constructed, which included 123 DEGs and 7503 interactions. A total of 105 GO terms were identified when the threshold was set as AUC >0.5, which had a good classification performance. Among these 105 GO terms, 2 functions had the AUC >0.7 and were determined as the optimal gene functions including angiogenesis (AUC =0.767) and regulation of immune system process (AUC =0.710). These gene functions appear to have potential for early detection and clinical treatment of OS in the future.

The characteristic direction: a geometrical approach to identify differentially expressed genes

BackgroundIdentifying differentially expressed genes (DEG) is a fundamental step in studies that perform genome wide expression profiling. Typically, DEG are identified by univariate approaches such as Significance Analysis of Microarrays (SAM) or Linear Models for Microarray Data (LIMMA) for processing cDNA microarrays, and differential gene expression analysis based on the negative binomial distribution (DESeq) or Empirical analysis of Digital Gene Expression data in R (edgeR) for RNA-seq profiling.ResultsHere we present a new geometrical multivariate approach to identify DEG called the Characteristic Direction. We demonstrate that the Characteristic Direction method is significantly more sensitive than existing methods for identifying DEG in the context of transcription factor (TF) and drug perturbation responses over a large number of microarray experiments. We also benchmarked the Characteristic Direction method using synthetic data, as well as RNA-Seq data. A large collection of microarray expression data from TF perturbations (73 experiments) and drug perturbations (130 experiments) extracted from the Gene Expression Omnibus (GEO), as well as an RNA-Seq study that profiled genome-wide gene expression and STAT3 DNA binding in two subtypes of diffuse large B-cell Lymphoma, were used for benchmarking the method using real data. ChIP-Seq data identifying DNA binding sites of the perturbed TFs, as well as known drug targets of the perturbing drugs, were used as prior knowledge silver-standard for validation. In all cases the Characteristic Direction DEG calling method outperformed other methods. We find that when drugs are applied to cells in various contexts, the proteins that interact with the drug-targets are differentially expressed and more of the corresponding genes are discovered by the Characteristic Direction method. In addition, we show that the Characteristic Direction conceptualization can be used to perform improved gene set enrichment analyses when compared with the gene-set enrichment analysis (GSEA) and the hypergeometric test.ConclusionsThe application of the Characteristic Direction method may shed new light on relevant biological mechanisms that would have remained undiscovered by the current state-of-the-art DEG methods. The method is freely accessible via various open source code implementations using four popular programming languages: R, Python, MATLAB and Mathematica, all available at: http://www.maayanlab.net/CD.

Analysis of long non-coding RNA expression profiles in gastric cancer

To investigate the expression patterns of long non-coding RNAs (lncRNAs) in gastric cancer. Two publicly available human exon arrays for gastric cancer and data for the corresponding normal tissue were downloaded from the Gene Expression Omnibus (GEO). We re-annotated the probes of the human exon arrays and retained the probes uniquely mapping to lncRNAs at the gene level. LncRNA expression profiles were generated by using robust multi-array average method in affymetrix power tools. The normalized data were then analyzed with a Bioconductor package linear models for microarray data and genes with adjusted P-values below 0.01 were considered differentially expressed. An independent data set was used to validate the results. With the computational pipeline established to re-annotate over 6.5 million probes of the Affymetrix Human Exon 1.0 ST array, we identified 136053 probes uniquely mapping to lncRNAs at the gene level. These probes correspond to 9294 lncRNAs, covering nearly 76% of the GENCODE lncRNA data set. By analyzing GSE27342 consisting of 80 paired gastric cancer and normal adjacent tissue samples, we identified 88 lncRNAs that were differentially expressed in gastric cancer, some of which have been reported to play a role in cancer, such as LINC00152, taurine upregulated 1, urothelial cancer associated 1, Pvt1 oncogene, small nucleolar RNA host gene 1 and LINC00261. In the validation data set GSE33335, 59% of these differentially expressed lncRNAs showed significant expression changes (adjusted P-value < 0.01) with the same direction. We identified a set of lncRNAs differentially expressed in gastric cancer, providing useful information for discovery of new biomarkers and therapeutic targets in gastric cancer.

CorrelaGenes: a new tool for the interpretation of the human transcriptome

CorrelaGenes: a new tool for the interpretation of the human transcriptome

Determinants of Excess Genetic Risk of Acute Myocardial Infarction - A Matched Case-control Study

Background: Myocardial infarction and stroke represent a major public health problem in most developing countries. This study explores genetic predisposition of acute myocardial infarction in the Czech population. Methods and Results: Genome-wide expression study used matched case-control design. Peripheral blood samples of the controls were matched to those of cases based on gender, age, status of diabetes mellitus and smoking status. Six months cardiovascular survival status of the cases was used to identify two distinct subgroups among the cases. Linear models for microarray data were employed to identify differential gene expression. Shrunken centroids technique helped in identifying the subsets of differentially expressed genes with predictive properties in independent samples. Predictive properties were evaluated using bootstrap sampling. Sixty transcripts were found to be both clinically and statistically differentially expressed among the cases not surviving the six months follow-up period relative to controls, while no such transcripts were observed among other surviving cases. The two subgroups of cases exhibited fourteen differentially expressed transcripts. Predictive modeling indicated sixteen out of sixty transcripts to best discriminate between the controls and cases that died during the follow-up period from cardiovascular causes, while for the surviving cases the already non-significant set of transcripts could not be further reduced. Eleven out of fourteen transcripts were found to best discriminate between the two groups of cases using shrunken centroids. Conclusions: The study identified genes associated with excess genetic risk of acute myocardial infarction, including those associated with the six months fatality of the cases.

BISPHENOL A-ASSOCIATED EPIGENOMIC CHANGES IN PREPUBESCENT GIRLS FROM GHARBIAH, EGYPT

Background and Aims: There is now compelling evidence that epigenetic modifications link adult disease susceptibility to environmental exposures during specific life stages, including pre-pubertal development. Animal studies indicate that bisphenol A (BPA) may impact health through epigenetic mechanisms. Epidemiological data associate BPA levels with metabolic disorders and reproductive effects. Here, we conducted an environmental epidemiology study of BPA exposure and epigenetics in pre-adolescent girls in Gharbiah, Egypt, hypothesizing that CpG methylation profiles will exhibit exposure-dependent trends. Methods: Total urinary BPA was measured by the CDC for 46 girls aged 9 to 12. We concurrently evaluated genome-wide CpG methylation using the Infinium HumanMethylation27 array in saliva DNA samples. The Linear Models for Microarray Data (Limma) library in R program was used to model age, BMI, specific gravity, creatinine, and log BPA to predict % change in methylation. Top hits below a p-value of 0.05 were subjected to enrichment analysis including pathway and gene ontology via Molecular Concepts Map (MCM). Results: Overall, urinary BPA levels were lower in Egyptian girls compared to similarly aged U.S. females as reported in NHANES. CpG methylation varied widely between girls, and BPA levels were associated with less genomic methylation. Pathway analyses indicated that genes with decrease in methylation % associated with BPA were involved in immune function, transport activity, metabolism, and caspase activity. In particular, hypomethylation of CpG targets on chromosome X was associated with BPA. Using the Comparative Toxicogenomics Database, we identified candidate genes with upregulated expression associated with BPA and found that many of the targets had BPA-associated hypomethylation in our sample, and visa versa. Conclusions: Our data indicate that BPA may affect human health through specific epigenomic modification of genes in relevant pathways. Epigenetic epidemiology holds promise for the identification biomarkers of previous exposure and the development of epigenetic-based diagnostic and therapeutic strategies.