High-throughput Expression Data Research Articles

Identification of Breast Cancer Subtype Specific MicroRNAs Using Survival Analysis to Find Their Role in Transcriptomic Regulation.

The microRNA (miRNA) biomolecules have a significant role in the development of breast cancer, and their expression profiles are different in each subtype of breast cancer. Thus, our goal is to use the Next Generation Sequencing provided high-throughput miRNA expression and clinical data in an integrated fashion to perform survival analysis in order to identify breast cancer subtype specific miRNAs, and analyze associated genes and transcription factors. We select top 100 miRNAs for each of the four subtypes, based on the value of hazard ratio and p-value, thereafter, identify 44 miRNAs that are related to all four subtypes, which we call as four-star miRNAs. Moreover, 12, 14, 9, and 15 subtype specific, viz. one-star miRNAs, are also identified. The resulting miRNAs are validated by using machine learning methods to differentiate tumor cases from controls (for four-star miRNAs), and subtypes (for one-star miRNAs). The four-star miRNAs provide 95% average accuracy, while in case of one-star miRNAs 81% accuracy is achieved for HER2-Enriched. Differences in expression of miRNAs between cancer stages is also analyzed, and a subset of eight miRNAs is found, for which expression is increased in stage II relative to stage I, including hsa-miR-10b-5p, which contributes to breast cancer metastasis. Subsequently we prepare regulatory networks in order to identify the interactions among miRNAs, their targeted genes and transcription factors (TFs), that are targeting those miRNAs. In this way, key regulatory circuits are identified, where genes such as TP53, ESR1, BRCA1, MYC, and others, that are known to be important genetic factors for the cause of breast cancer, produce transcription factors that target the same genes as well as interact with the selected miRNAs. To provide further biological validation the Protein-Protein Interaction (PPI) networks are prepared and Kyoto Encyclopedia of Genes and Genomes pathway and gene ontology (GO) enrichment analysis are performed. Among the enriched pathways many are breast cancer-related, such as PI3K-Akt or p53 signaling pathways, and contain proteins such as TP53, also present in the regulatory networks. Moreover, we find that the genes are enriched in GO terms associated with breast cancer. Our results provide detailed analysis of selected miRNAs and their regulatory networks.

Genome-wide analysis of multi-view data of miRNA-seq to identify miRNA biomarkers for stomach cancer.

Stomach cancer is one of the leading causes of cancer-related deaths worldwide. More than 80% diagnosis of this cancer occur at later stages leading to low 5-year survival rate. This emphasizes the need to have better prognostic techniques for stomach cancer. In this regard, the Next-Generation Sequencing of whole genome and multi-view approach to omics may reveal the underlying molecular complexity of stomach cancer using high throughput expression data of miRNA. Generally, miRNAs are small, non-coding RNAs, which cause downregulation of target mRNAs. They also show differential expression for a specific biological condition like stage or histological type of stomach cancer, highlighting their importance as potential biomarkers. Analyzing miRNA expression data is a challenging task due to the existence of large number of miRNAs and less sample size. A small set of miRNAs will be helpful in designing efficient diagnostic and prognostic tool. In this regard, here a computational framework is proposed that selects different sets of miRNAs for five different categories of clinical outcomes viz. condition, clinical stage, age, histological type, and survival status. First, the miRNAs are ranked using four feature ranking methods. These ranks are used to find an ensemble rank based on adaptive weight. Second, the top 100 miRNAs from each category are used to find the miRNAs that are common to all categories as well as miRNAs that belong to only one category. Finally, the results have been validated quantitatively and through biological significance analysis.

Network meta-analysis correlates with analysis of merged independent transcriptome expression data

BackgroundUsing meta-analysis, high-dimensional transcriptome expression data from public repositories can be merged to make group comparisons that have not been considered in the original studies. Merging of high-dimensional expression data can, however, implicate batch effects that are sometimes difficult to be removed. Removing batch effects becomes even more difficult when expression data was taken using different technologies in the individual studies (e.g. merging of microarray and RNA-seq data). Network meta-analysis has so far not been considered to make indirect comparisons in transcriptome expression data, when data merging appears to yield biased results.ResultsWe demonstrate in a simulation study that the results from analyzing merged data sets and the results from network meta-analysis are highly correlated in simple study networks. In the case that an edge in the network is supported by multiple independent studies, network meta-analysis produces fold changes that are closer to the simulated ones than those obtained from analyzing merged data sets. Finally, we also demonstrate the practicability of network meta-analysis on a real-world data example from neuroinfection research.ConclusionsNetwork meta-analysis is a useful means to make new inferences when combining multiple independent studies of molecular, high-throughput expression data. This method is especially advantageous when batch effects between studies are hard to get removed.

Integrative workflows for network analysis.

Due to genetic heterogeneity across patients, the identification of effective disease signatures and therapeutic targets is challenging. Addressing this challenge, we have previously developed a network-based approach, which integrates heterogeneous sources of biological information to identify disease specific core-regulatory networks. In particular, our workflow uses a multi-objective optimization function to calculate a ranking score for network components (e.g. feedback/feedforward loops) based on network properties, biomedical and high-throughput expression data. High ranked network components are merged to identify the core-regulatory network(s) that is then subjected to dynamical analysis using stimulus-response and in silico perturbation experiments for the identification of disease gene signatures and therapeutic targets. In a case study, we implemented our workflow to identify bladder and breast cancer specific core-regulatory networks underlying epithelial-mesenchymal transition from the E2F1 molecular interaction map.In this study, we review our workflow and described how it has developed over time to understand the mechanisms underlying disease progression and prediction of signatures for clinical decision making.

Link prediction potentials for biological networks

Improvement of biological networks reconstructed from high-throughput expression data is an important challenge in systems biology. Link prediction is a problem of interest in many application domains that can be used for this purpose. In this paper after a short review of several biological networks, we present the latest definition of the link prediction problem and review it from several viewpoints. With a comprehensive search in the literature using PubMed, Science Direct and Google Scholar databases, and carefully reviewing the related papers having the 'link prediction' plus at least one of the biological network terms in their title, abstract or keywords, we classify the results based on the graph type and major link prediction outlooks. Finally, we analyse the preformed researches to find new insights about potential uses in addition to understanding the current state, and propose several hints and directions for future works.

Regulating the transcriptomes that mediate the conversion of fibroblasts to various nervous system neural cell types.

Our understanding of the mechanism of cell fate transition during the direct reprogramming of fibroblasts into various central nervous system (CNS) neural cell types has been limited by the lack of a comprehensive analysis on generated cells, independently and in comparison with other CNS neural cell types. Here, we applied an integrative approach on 18 independent high throughput expression data sets to gain insight into the regulation of the transcriptome during the conversion of fibroblasts into induced neural stem cells, induced neurons (iNs), induced astrocytes, and induced oligodendrocyte progenitor cells (iOPCs). We found common down-regulated genes to be mostly related to fibroblast-specific functions, and suggest their potential as markers for screening of the silencing of the fibroblast-specific program. For example, Tagln was significantly down-regulated across all considered data sets. In addition, we identified specific profiles of up-regulated genes for each CNS neural cell types, which could be potential markers for maturation and efficiency screenings. Furthermore, we identified the main TFs involved in the regulation of the gene expression program during direct reprogramming. For example, in the generation of iNs from fibroblasts, the Rest TF was the main regulator of this reprogramming. In summary, our computational approach for meta-analyzing independent expression data sets provides significant details regarding the molecular mechanisms underlying the regulation of the gene expression program, and also suggests potentially useful candidate genes for screening down-regulation of fibroblast gene expression profile, maturation, and efficiency, as well as candidate TFs for increasing the efficiency of the reprogramming process.

Comprehensive transcriptome mining of the direct conversion of mesodermal cells

The direct reprogramming of somatic cells is a promising approach for regenerative medicine, especially in the production of mesoderm layer-derived cells. Meta-analysis studies provide precise insight into the undergoing processes and help increase the efficiency of reprogramming. Here, using 27 high-throughput expression data sets, we analyzed the direct reprogramming of mesodermal cells in humans and mice. Fibroblast-derived cells showed a common expression pattern of up- and down-regulated genes that were mainly involved in the suppression of the fibroblast-specific gene expression program, and may be used as markers of the initiation of reprogramming. Furthermore, we found a specific gene expression profile for each fibroblast-derived cell studied, and each gene set appeared to play specific functional roles in its cell type, suggesting their use as markers for their mature state. Furthermore, using data from protein-DNA interactions, we identified the main transcription factors (TFs) involved in the conversion process and ranked them based on their importance in their gene regulatory networks. In summary, our meta-analysis approach provides new insights on the direct conversion of mesodermal somatic cells, introduces a list of genes as markers for initiation and maturation, and identifies TFs for which manipulating their expression may increase the efficiency of direct conversion.

Comprehensive Dissection of Transcriptome Data and Regulatory Factors in Pancreatic Cancer Cells.

Features of pancreatic cancers include high mortality rates caused by rapid tumor progression and a lack of effective therapy. Underpinning the molecular mechanisms involved in the alteration of the gene expression program in the pancreatic cancer remains to be understood. In the current study, we performed a comprehensive analysis using 282 pancreatic tumor and normal samples from seven independent expression data sets to provide a better view on the interactions between different transcription factors (TFs) and the most affected biological pathways in pancreatic cancer. We highlighted common differentially expressed genes (DEGs) and common affected processes within pancreatic cancer samples. We revealed 16 main DE-TFs that regulated gene expression alterations as well as the most significant processes in pancreatic cancer compared to normal cells. For example, we found the upregulated FOXM1 to be a top regulator of pancreatic cellular transformation based on results from different analyses, including from its regulation of gene regulatory networks, its presence in protein complex, its significant regulation of genes related to cancer pathways, and its regulation of most of the identified DE-TFs. Furthermore, we provided a model and assessed the role of different DE-TFs in the regulation of the most affected pancreatic- and cancer-specific processes. In conclusion, our bioinformatics meta-analysis of high throughput expression data sets, besides clarifying common affected genes and pathways, also showed the mechanisms involved in regulating these common profiles. Our results, especially for DE-TFs, could potentially be useful for screening for pancreatic cancer, and for confirming or determining novel pharmacological targets. J. Cell. Biochem. 118: 3976-3985, 2017. © 2017 Wiley Periodicals, Inc.

Model-Based Characterization of Inflammatory Gene Expression Patterns of Activated Macrophages.

Macrophages are cells with remarkable plasticity. They integrate signals from their microenvironment leading to context-dependent polarization into classically (M1) or alternatively (M2) activated macrophages, representing two extremes of a broad spectrum of divergent phenotypes. Thereby, macrophages deliver protective and pro-regenerative signals towards injured tissue but, depending on the eliciting damage, may also be responsible for the generation and aggravation of tissue injury. Although incompletely understood, there is emerging evidence that macrophage polarization is critical for these antagonistic roles. To identify activation-specific expression patterns of chemokines and cytokines that may confer these distinct effects a systems biology approach was applied. A comprehensive literature-based Boolean model was developed to describe the M1 (LPS-activated) and M2 (IL-4/13-activated) polarization types. The model was validated using high-throughput transcript expression data from murine bone marrow derived macrophages. By dynamic modeling of gene expression, the chronology of pathway activation and autocrine signaling was estimated. Our results provide a deepened understanding of the physiological balance leading to M1/M2 activation, indicating the relevance of co-regulatory signals at the level of Akt1 or Akt2 that may be important for directing macrophage polarization.

Nonlinear Network Reconstruction from Gene Expression Data Using Marginal Dependencies Measured by DCOL.

Reconstruction of networks from high-throughput expression data is an important tool to identify new regulatory relations. Given that nonlinear and complex relations exist between biological units, methods that can utilize nonlinear dependencies may yield insights that are not provided by methods using linear associations alone. We have previously developed a distance to measure predictive nonlinear relations, the Distance based on Conditional Ordered List (DCOL), which is sensitive and computationally efficient on large matrices. In this study, we explore the utility of DCOL in the reconstruction of networks, by combining it with local false discovery rate (lfdr)–based inference. We demonstrate in simulations that the new method named nlnet is effective in recovering hidden nonlinear modules. We also demonstrate its utility using a single cell RNA seq dataset. The method is available as an R package at https://cran.r-project.org/web/packages/nlnet.

Sparse factor model for co-expression networks with an application using prior biological knowledge.

Inference on gene regulatory networks from high-throughput expression data turns out to be one of the main current challenges in systems biology. Such networks can be very insightful for the deep understanding of interactions between genes. Because genes-gene interactions is often viewed as joint contributions to known biological mechanisms, inference on the dependence among gene expressions is expected to be consistent to some extent with the functional characterization of genes which can be derived from ontologies (GO, KEGG, …). The present paper introduces a sparse factor model as a general framework either to account for a prior knowledge on joint contributions of modules of genes to latent biological processes or to infer on the corresponding co-expression network. We propose an ℓ1 - regularized EM algorithm to fit a sparse factor model for correlation. We demonstrate how it helps extracting modules of genes and more generally improves the gene clustering performance. The method is compared to alternative estimation procedures for sparse factor models of relevance networks in a simulation study. The integration of a biological knowledge based on the gene ontology (GO) is also illustrated on a liver expression data generated to understand adiposity variability in chicken.

Revealing complex function, process and pathway interactions with high-throughput expression and biological annotation data.

The biological relationships both between and within the functions, processes and pathways that operate within complex biological systems are only poorly characterized, making the interpretation of large scale gene expression datasets extremely challenging. Here, we present an approach that integrates gene expression and biological annotation data to identify and describe the interactions between biological functions, processes and pathways that govern a phenotype of interest. The product is a global, interconnected network, not of genes but of functions, processes and pathways, that represents the biological relationships within the system. We validated our approach on two high-throughput expression datasets describing organismal and organ development. Our findings are well supported by the available literature, confirming that developmental processes and apoptosis play key roles in cell differentiation. Furthermore, our results suggest that processes related to pluripotency and lineage commitment, which are known to be critical for development, interact mainly indirectly, through genes implicated in more general biological processes. Moreover, we provide evidence that supports the relevance of cell spatial organization in the developing liver for proper liver function. Our strategy can be viewed as an abstraction that is useful to interpret high-throughput data and devise further experiments.

High-order dynamic Bayesian Network learning with hidden common causes for causal gene regulatory network.

BackgroundInferring gene regulatory network (GRN) has been an important topic in Bioinformatics. Many computational methods infer the GRN from high-throughput expression data. Due to the presence of time delays in the regulatory relationships, High-Order Dynamic Bayesian Network (HO-DBN) is a good model of GRN. However, previous GRN inference methods assume causal sufficiency, i.e. no unobserved common cause. This assumption is convenient but unrealistic, because it is possible that relevant factors have not even been conceived of and therefore un-measured. Therefore an inference method that also handles hidden common cause(s) is highly desirable. Also, previous methods for discovering hidden common causes either do not handle multi-step time delays or restrict that the parents of hidden common causes are not observed genes.ResultsWe have developed a discrete HO-DBN learning algorithm that can infer also hidden common cause(s) from discrete time series expression data, with some assumptions on the conditional distribution, but is less restrictive than previous methods. We assume that each hidden variable has only observed variables as children and parents, with at least two children and possibly no parents. We also make the simplifying assumption that children of hidden variable(s) are not linked to each other. Moreover, our proposed algorithm can also utilize multiple short time series (not necessarily of the same length), as long time series are difficult to obtain.ConclusionsWe have performed extensive experiments using synthetic data on GRNs of size up to 100, with up to 10 hidden nodes. Experiment results show that our proposed algorithm can recover the causal GRNs adequately given the incomplete data. Using the limited real expression data and small subnetworks of the YEASTRACT network, we have also demonstrated the potential of our algorithm on real data, though more time series expression data is needed.

BubbleGUM: automatic extraction of phenotype molecular signatures and comprehensive visualization of multiple Gene Set Enrichment Analyses.

BackgroundRecent advances in the analysis of high-throughput expression data have led to the development of tools that scaled-up their focus from single-gene to gene set level. For example, the popular Gene Set Enrichment Analysis (GSEA) algorithm can detect moderate but coordinated expression changes of groups of presumably related genes between pairs of experimental conditions. This considerably improves extraction of information from high-throughput gene expression data. However, although many gene sets covering a large panel of biological fields are available in public databases, the ability to generate home-made gene sets relevant to one’s biological question is crucial but remains a substantial challenge to most biologists lacking statistic or bioinformatic expertise. This is all the more the case when attempting to define a gene set specific of one condition compared to many other ones. Thus, there is a crucial need for an easy-to-use software for generation of relevant home-made gene sets from complex datasets, their use in GSEA, and the correction of the results when applied to multiple comparisons of many experimental conditions.ResultWe developed BubbleGUM (GSEA Unlimited Map), a tool that allows to automatically extract molecular signatures from transcriptomic data and perform exhaustive GSEA with multiple testing correction. One original feature of BubbleGUM notably resides in its capacity to integrate and compare numerous GSEA results into an easy-to-grasp graphical representation. We applied our method to generate transcriptomic fingerprints for murine cell types and to assess their enrichments in human cell types. This analysis allowed us to confirm homologies between mouse and human immunocytes.ConclusionsBubbleGUM is an open-source software that allows to automatically generate molecular signatures out of complex expression datasets and to assess directly their enrichment by GSEA on independent datasets. Enrichments are displayed in a graphical output that helps interpreting the results. This innovative methodology has recently been used to answer important questions in functional genomics, such as the degree of similarities between microarray datasets from different laboratories or with different experimental models or clinical cohorts. BubbleGUM is executable through an intuitive interface so that both bioinformaticians and biologists can use it. It is available at http://www.ciml.univ-mrs.fr/applications/BubbleGUM/index.html.Electronic supplementary materialThe online version of this article (doi:10.1186/s12864-015-2012-4) contains supplementary material, which is available to authorized users.

The assembly of miRNA-mRNA-protein regulatory networks using high-throughput expression data.

Inference of gene regulatory networks from high throughput measurement of gene and protein expression is particularly attractive because it allows the simultaneous discovery of interactive molecular signals for numerous genes and proteins at a relatively low cost. We developed two score-based local causal learning algorithms that utilized the Markov blanket search to identify direct regulators of target mRNAs and proteins. These two algorithms were specifically designed for integrated high throughput RNA and protein data. Simulation study showed that these algorithms outperformed other state-of-the-art gene regulatory network learning algorithms. We also generated integrated miRNA, mRNA, and protein expression data based on high throughput analysis of primary trophoblasts, derived from term human placenta and cultured under standard or hypoxic conditions. We applied the new algorithms to these data and identified gene regulatory networks for a set of trophoblastic proteins found to be differentially expressed under the specified culture conditions.

IAOseq: inferring abundance of overlapping genes using RNA-seq data.

BackgroundOverlapping transcription constitutes a common mechanism for regulating gene expression. A major limitation of the overlapping transcription assays is the lack of high throughput expression data.ResultsWe developed a new tool (IAOseq) that is based on reads distributions along the transcribed regions to identify the expression levels of overlapping genes from standard RNA-seq data. Compared with five commonly used quantification methods, IAOseq showed better performance in the estimation accuracy of overlapping transcription levels. For the same strand overlapping transcription, currently existing high-throughput methods are rarely available to distinguish which strand was present in the original mRNA template. The IAOseq results showed that the commonly used methods gave an average of 1.6 fold overestimation of the expression levels of same strand overlapping genes.ConclusionsThis work provides a useful tool for mining overlapping transcription levels from standard RNA-seq libraries. IAOseq could be used to help us understand the complex regulatory mechanism mediated by overlapping transcripts. IAOseq is freely available at http://lifecenter.sgst.cn/main/en/IAO_seq.jsp.

Integrated analysis of miRNA and mRNA expression profiles in response to Cd exposure in rice seedlings.

BackgroundIndependent transcriptome profile analyses of miRNAs or mRNAs under conditions of cadmium (Cd) stress have been widely reported in plants. However, a combined analysis of sRNA sequencing expression data with miRNA target expression data to infer the relative activities of miRNAs that regulate gene expression changes resulting from Cd stress has not been reported in rice. To elucidate the roles played by miRNAs in the regulation of changes in gene expression in response to Cd stress in rice (Oryza sativa L.), we simultaneously characterized changes in the miRNA and mRNA profiles following treatment with Cd.ResultsA total of 163 miRNAs and 2,574 mRNAs were identified to be differentially expressed under Cd stress, and the changes in the gene expression profile in the shoot were distinct from those in the root. At the miRNA level, 141 known miRNAs belonging to 48 families, and 39 known miRNAs in 23 families were identified to be differentially expressed in the root and shoot, respectively. In addition, we identified eight new miRNA candidates from the root and five from the shoot that were differentially expressed in response to Cd treatment. For the mRNAs, we identified 1,044 genes in the root and 448 genes in the shoot that were up-regulated, while 572 and 645 genes were down-regulated in the root and shoot, respectively. GO and KEGG enrichment analyses showed that genes encoding secondary, metabolite synthases, signaling molecules, and ABC transporters were significantly enriched in the root, while only ribosomal protein and carotenoid biosynthesis genes were significantly enriched in the shoot. Then 10 known miRNA-mRNA interaction pairs and six new candidate ones, that showed the opposite expression patterns, were identified by aligning our two datasets against online databases and by using the UEA sRNA toolkit respectively.ConclusionsThis study is the first to use high throughput DNA sequencing to simultaneously detect changes in miRNA and mRNA expression patterns in the root and shoot in response to Cd treatment. These integrated high-throughput expression data provide a valuable resource to examine global genome expression changes in response to Cd treatment and how these are regulated by miRNAs.Electronic supplementary materialThe online version of this article (doi:10.1186/1471-2164-15-835) contains supplementary material, which is available to authorized users.

Correcting transcription factor gene sets for copy number and promoter methylation variations.

Gene set analysis provides a method to generate statistical inferences across sets of linked genes, primarily using high-throughput expression data. Common gene sets include biological pathways, operons, and targets of transcriptional regulators. In higher eukaryotes, especially when dealing with diseases with strong genetic and epigenetic components such as cancer, copy number loss and gene silencing through promoter methylation can eliminate the possibility that a gene is transcribed. This, in turn, can adversely affect the estimation of transcription factor or pathway activity from a set of target genes, as some of the targets may not be responsive to transcriptional regulation. Here we introduce a simple filtering approach that removes genes from consideration if they show copy number loss or promoter methylation, and demonstrate the improvement in inference of transcription factor activity in a simulated dataset based on the background expression observed in normal head and neck tissue.

Network-based approach reveals Y chromosome influences prostate cancer susceptibility

The human Y chromosome contains a small number of genes that play a critical role in the determination of male-specific organs. Today’s advances have provided valuable resources for defining the functions of this chromosome in both normal and cancerous prostates. Despite the fact that generation of high-throughput expression data is becoming usual; the systematic methods of data analysis in a biological context are still an impediment.Here we have shown that constructing co-expression networks using Y-chromosome genes provides an alternative strategy for the detection of new candidate genes involved in prostate cancer. In our approach, independent co-expression networks from normal and cancerous stages are reconstructed using a reverse engineering approach. We then highlight crucial pathways, biological processes, and genes involved in the prostate cancer by analyzing each network individually and in concert. Thus, we have identified 18 critical pathways and processes related to prostate cancer, many of which have previously been shown to be involved in cancer. In particular, we identify 22 Y-chromosome genes putatively linked to prostate cancer, 13 of which have been already verified experimentally.Our novel network-based approach is useful for accurate inference of processes and essential regulators that mediate molecular changes during cancer progression.

A method for quantitative analysis of standard and high-throughput qPCR expression data based on input sample quantity.

Over the past decade rapid advances have occurred in the understanding of RNA expression and its regulation. Quantitative polymerase chain reactions (qPCR) have become the gold standard for quantifying gene expression. Microfluidic next generation, high throughput qPCR now permits the detection of transcript copy number in thousands of reactions simultaneously, dramatically increasing the sensitivity over standard qPCR. Here we present a gene expression analysis method applicable to both standard polymerase chain reactions (qPCR) and high throughput qPCR. This technique is adjusted to the input sample quantity (e.g., the number of cells) and is independent of control gene expression. It is efficiency-corrected and with the use of a universal reference sample (commercial complementary DNA (cDNA)) permits the normalization of results between different batches and between different instruments – regardless of potential differences in transcript amplification efficiency. Modifications of the input quantity method include (1) the achievement of absolute quantification and (2) a non-efficiency corrected analysis. When compared to other commonly used algorithms the input quantity method proved to be valid. This method is of particular value for clinical studies of whole blood and circulating leukocytes where cell counts are readily available.