Gene expression data analysis

Microarrays are one of the latest breakthroughs in experimental molecular biology, which allow monitoring of gene expression for tens of thousands of genes in parallel and are already producing huge amounts of valuable data. Analysis and handling of such data is becoming one of the major bottlenecks in the utilization of the technology. The raw microarray data are images, which have to be transformed into gene expression matrices, tables where rows represent genes, columns represent various samples such as tissues or experimental conditions, and numbers in each cell characterize the expression level of the particular gene in the particular sample. These matrices have to be analyzed further if any knowledge about the underlying biological processes is to be extracted. In this paper we concentrate on discussing bioinformatics methods used for such analysis. We briefly discuss supervised and unsupervised data analysis and its applications, such as predicting gene function classes and cancer classification as well as some possible future directions.     

Gene expression analysis in sections and tissue microarrays of archival tissues by mRNA in situ hybridization

Altered expression of genes in diseased tissues can prognosticate a distinct natural progression of the disease as well as predict sensitivity or resistance to particular therapies. Archival tissues from patients with a known medical history and treatments are an invaluable resource to validate the utility of candidate genes for prognosis and prediction of therapy outcomes. However, stored tissues with associated long-term follow-up information typically are formalin-fixed, paraffin-embedded specimen and this can severely restrict the methods applicable for gene expression analysis. We report here on the utility of tissue microarrays (TMAs) that use valuable tissues sparingly and provide a platform for simultaneous analysis of gene expression in several hundred samples. In particular, we describe a stable method applicable to mRNA expression screening in such archival tissues. TMAs are constructed from sections of small drill cores, taken from tissue blocks of archival tissues and multiple samples can thus be arranged on a single microscope slide. We used mRNA in situ hybridization (ISH) on >500 full sections and >100 TMAs for >10 different cDNAs that yielded >10,000 data points. We provide detailed experimental protocols that can be implemented without major hurdles in a molecular pathology laboratory and discuss quantitative analysis and the advantages and limitations of ISH. We conclude that gene expression analysis in archival tissues by ISH is reliable and particularly useful when no protein detection methods are available for a candidate gene.     

Bone to pick: the importance of evaluating reference genes for RT-qPCR quantification of gene expression in craniosynostosis and bone-related tissues and cells

ABSTRACT: BACKGROUND: RT-qPCR is a common tool for quantification of gene expression, but its accuracy is dependent on the choice and stability (steady state expression levels) of the reference gene/s used for normalization. To date, in the bone field, there have been few studies to determine the most stable reference genes and, usually, RT-qPCR data is normalised to non-validated reference genes, most commonly GAPDH, ACTB and 18 S rRNA. Here we draw attention to the potential deleterious impact of using classical reference genes to normalise expression data for bone studies without prior validation of their stability. RESULTS: Using the geNorm and Normfinder programs, panels of mouse and human genes were assessed for their stability under three different experimental conditions: 1) disease progression of Crouzon syndrome (craniosynostosis) in a mouse model, 2) proliferative culture of cranial suture cells isolated from craniosynostosis patients and 3) osteogenesis of a mouse bone marrow stromal cell line. We demonstrate that classical reference genes are not always the most 'stable' genes and that gene 'stability' is highly dependent on experimental conditions. Selected stable genes, individually or in combination, were then used to normalise osteocalcin and alkaline phosphatase gene expression data during cranial suture fusion in the craniosynostosis mouse model and strategies compared. Strikingly, the expression trends of alkaline phosphatase and osteocalcin varied significantly when normalised to the least stable, the most stable or the three most stable genes. CONCLUSION: To minimise errors in evaluating gene expression levels, analysis of a reference panel and subsequent normalization to several stable genes is strongly recommended over normalization to a single gene. In particular, we conclude that use of single, non-validated "housekeeping" genes such as GAPDH, ACTB and 18 S rRNA, currently a widespread practice by researchers in the bone field, is likely to produce data of questionable reliability when changes are 2 fold or less, and such data should be interpreted with due caution.     

Phylogenetic diversification of the globin gene superfamily in chordates

Phylogenetic reconstructions provide a means of inferring the branching relationships among members of multigene families that have diversified via successive rounds of gene duplication and divergence. Such reconstructions can illuminate the pathways by which particular expression patterns and protein functions evolved. For example, phylogenetic analyses can reveal cases in which similar expression patterns or functional properties evolved independently in different lineages, either through convergence, parallelism, or evolutionary reversals. The purpose of this article is to provide a robust phylogenetic framework for interpreting experimental data and for generating hypotheses about the functional evolution of globin proteins in chordate animals. To do this, we present a consensus phylogeny of the chordate globin gene superfamily. We document the relative roles of gene duplication and whole-genome duplication in fueling the functional diversification of vertebrate globins, and we unravel patterns of shared ancestry among globin genes from representatives of the three chordate subphyla (Craniata, Urochordata, and Cephalochordata). Our results demonstrate the value of integrating phylogenetic analyses with genomic analyses of conserved synteny to infer the duplicative origins and evolutionary histories of globin genes. We also discuss a number of case studies that illustrate the importance of phylogenetic information when making inferences about the evolution of globin gene expression and protein function. Finally, we discuss why the globin gene superfamily presents special challenges for phylogenetic analysis, and we describe methodological approaches that can be used to meet those challenges.     

遗传与基因表达数据的整合—— eQTL的方法及应用

高通量的基因型分析和芯片技术的发展使人们能够进一步研究哪些遗传差异最终影响基因的表达。通过表达数量性状座位(eQTL)作图方法可对基因表达水平的遗传基础进行解析。与传统的QTL分析方法一样, eQTL的主要目标是鉴别表达性状座位所在的染色体区域。但由于表达谱数据成千上万, 而传统的QTL分析方法最多分析几十个性状, 因此需要考虑这类实验设计的特点以及统计分析方法。本文详细介绍了eQTL定位过程及其研究方法, 重点从个体选择、基因芯片实验设计、基因表达数据的获得与标准化、作图方法及结果分析等方面进行了综述, 指出了当前eQTL研究存在的问题和局限性。最后介绍了eQTL研究在估计基因表达遗传率、挖掘候选基因、构建基因调控网络、理解基因间及基因与环境的互作的应用进展。     

Parametric approach to genomic imprinting analysis with applications to Angelman's syndrome

Genomic imprinting is a mechanism by which only one copy of a gene pair is expressed, and this expression is determined by the parental origin of the copy. The deregulation of imprinted genes has been implicated in a number of human diseases. The Imprinted Gene Catalogue now has more than 200 genes listed, and estimates based on mouse models suggest many more may exist in humans. Therefore, the development of methods to identify such genes is important. In this communication, we present a parametric model-based approach to analyzing arbitrary-sized pedigree data for genomic imprinting. We have modified widely used LINKAGE program to incorporate our proposed approach. In addition, our approach allows for the use of sex-specific recombinations in the analysis, which is of particular importance in a genome-wide analysis for imprinted genes. We compared our imprinting analysis approach to that implemented in the GENEHUNTER-IMPRINT program using simulation studies as well as by analyzing causal genes in Angelman's syndrome families, which are known to be imprinted. These analyses showed that the proposed approach is very powerful for detecting imprinted genes in large pedigrees.     

Gene expression alterations in connexin null mice extend beyond the gap junction

Connexin43 (Cx43) is the principal gap junction protein between astrocytes in the neonatal brain and also interconnects neural precursor cells during CNS development. In an attempt to understand global effects of expression of the Cx43 gap junction gene on development and function of the nervous system, we have compared gene expression patterns in cultured astrocytes and brains from wildtype mice with those in which Cx43 is deleted as well as in spinal cords of experimental autoimmune encepahlomyelitis (EAE) mice. One surprising result obtained from high densitity mouse cDNA studies was the large number of genes that were statistically altered in mice with decreased expression of Cx43. These altered genes encode proteins with a wide range of functions within cells, and thus deletion of normal gap junction expression appears to result in globally altered glial functions in addition to disruption of intercellular communication. Here we discuss those results in the context of the strategies and data analysis paradigms that we are using in such studies.     

Mining co-regulated gene profiles for the detection of functional associations in gene expression data

MOTIVATION: Association pattern discovery (APD) methods have been successfully applied to gene expression data. They find groups of co-regulated genes in which the genes are either up- or down-regulated throughout the identified conditions. These methods, however, fail to identify similarly expressed genes whose expressions change between up- and down-regulation from one condition to another. In order to discover these hidden patterns, we propose the concept of mining co-regulated gene profiles. Co-regulated gene profiles contain two gene sets such that genes within the same set behave identically (up or down) while genes from different sets display contrary behavior. To reduce and group the large number of similar resulting patterns, we propose a new similarity measure that can be applied together with hierarchical clustering methods. RESULTS: We tested our proposed method on two well-known yeast microarray data sets. Our implementation mined the data effectively and discovered patterns of co-regulated genes that are hidden to traditional APD methods. The high content of biologically relevant information in these patterns is demonstrated by the significant enrichment of co-regulated genes with similar functions. Our experimental results show that the Mining Attribute Profile (MAP) method is an efficient tool for the analysis of gene expression data and competitive with bi-clustering techniques.     

Gene microarray analysis using angular distribution decomposition.

Clustering techniques have been widely used in the analysis of microarray data to group genes with similar expression profiles. The similarity of expression profiles and hence the results of clustering greatly depend on how the data has been transformed. We present a method that uses the relative expression changes between pairs of conditions and an angular transformation to define the similarity of gene expression patterns. The pairwise comparisons of experimental conditions can be chosen to reflect the purpose of clustering allowing control the definition of similarity between genes. A variational Bayes mixture modeling approach is then used to find clusters within the transformed data. The purpose of microarray data analysis is often to locate groups genes showing particular patterns of expression change and within these groups to locate specific target genes that may warrant further experimental investigation. We show that the angular transformation maps data to a representation from which information, in terms of relative regulation changes, can be automatically mined. This information can be then be used to understand the "features" of expression change important to different clusters allowing potentially interesting clusters to be easily located. Finally, we show how the genes within a cluster can be visualized in terms of their expression pattern and intensity change, allowing potential target genes to be highlighted within the clusters of interest.     

Statistical analysis of oligonucleotide microarray data

Microchip arrays have become one of the most rapidly growing techniques for monitoring gene expression at the genomic level and thereby gaining valuable insight about various important biological mechanisms. Examples of such mechanisms are: identifying disease-causing genes, genes involved in the regulation of some aspect of the cell cycle, etc. In this article, we discuss the problem of estimating gene expression based on a proper statistical model. More precisely, we show how the model introduced by Li and Wong can be used in its full bivariate generality to provide a new measure of gene expression from high-density oligonucleotide arrays. We also present a second gene expression index based on a new way of reducing the model into a simpler univariate model. In both cases, the gene expression indices are shown to be unbiased and to have lower variance than the established ones. Moreover, we present a bootstrap method aiming at providing non-parametric confidence intervals for the expression index.     

Biasogram: Visualization of Confounding Technical Bias in Gene Expression Data

Gene expression profiles of clinical cohorts can be used to identify genes that are correlated with a clinical variable of interest such as patient outcome or response to a particular drug. However, expression measurements are susceptible to technical bias caused by variation in extraneous factors such as RNA quality and array hybridization conditions. If such technical bias is correlated with the clinical variable of interest, the likelihood of identifying false positive genes is increased. Here we describe a method to visualize an expression matrix as a projection of all genes onto a plane defined by a clinical variable and a technical nuisance variable. The resulting plot indicates the extent to which each gene is correlated with the clinical variable or the technical variable. We demonstrate this method by applying it to three clinical trial microarray data sets, one of which identified genes that may have been driven by a confounding technical variable. This approach can be used as a quality control step to identify data sets that are likely to yield false positive results.     

Relationships between gene expression and brain wiring in the adult rodent brain

We studied the global relationship between gene expression and neuroanatomical connectivity in the adult rodent brain. We utilized a large data set of the rat brain "connectome" from the Brain Architecture Management System (942 brain regions and over 5000 connections) and used statistical approaches to relate the data to the gene expression signatures of 17,530 genes in 142 anatomical regions from the Allen Brain Atlas. Our analysis shows that adult gene expression signatures have a statistically significant relationship to connectivity. In particular, brain regions that have similar expression profiles tend to have similar connectivity profiles, and this effect is not entirely attributable to spatial correlations. In addition, brain regions which are connected have more similar expression patterns. Using a simple optimization approach, we identified a set of genes most correlated with neuroanatomical connectivity, and find that this set is enriched for genes involved in neuronal development and axon guidance. A number of the genes have been implicated in neurodevelopmental disorders such as autistic spectrum disorder. Our results have the potential to shed light on the role of gene expression patterns in influencing neuronal activity and connectivity, with potential applications to our understanding of brain disorders. Supplementary data are available at http://www.chibi.ubc.ca/ABAMS.     

Phytochrome regulation of nuclear gene expression in plants

The photoregulation of gene expression in higher plants was extensively studied during the 1980s, in particular the light-responsive cis -acting elements and trans -acting factors of the Lhcb and rbcS genes. However, little has been discovered about: (1) which plant genes are regulated by light, and (2) which photoreceptors control the expression of these genes. In the 1990s, the functional analysis of the various photoreceptors has progressed rapidly using photoreceptor-deficient mutants, including those of the phytochrome gene family. More recently however, advanced techniques for gene expression analysis, such as fluorescent differential display and DNA microarray technology, have become available enabling the global identification of genes that are regulated by particular photoreceptors. In this paper we describe distinct and overlapping effects of individual phytochromes on gene expression in Arabidopsis thaliana.     

Spearman Correlation Identifies Statistically Significant Gene Expression Clusters in Spinal Cord Development and Injury

An important problem in the analysis of large-scale gene expression data is the validation of gene expression clusters. By examining the temporal expression patterns of 74 genes expressed in rat spinal cord under three different experimental conditions, we have found evidence that some genes cluster together under multiple conditions. Using RT-PCR data from spinal cord development and two sets of microarray data from spinal injury, we applied Spearman correlation to identify clusters and to assign P values to pairs of genes with highly similar temporal expression patterns. We found that 15% of genes occurred in statistically significant pairs in all three experimental conditions, providing both statistical and experimental support for the idea that genes that cluster together are co-regulated. In addition, we demonstrated that DNA microarray and RT-PCR data are comparable, and can be combined to confirm gene expression relationships.