Microarray,SAGE and their applications to cardiovascular diseases

The wealth of DNA data generated by the human genome project coupling with recently invented high-throughput gene expression profiling techniques has dramatically sped up the process for biomedical researchers on elucidating the role of genes in human diseases. One powerful method to reveal insight into gene functions is the systematic analysis of gene expression. Two popular high-throughput gene expression technologies, microarray and Serial Analysis of Gene Expression (SAGE) are capable of producing large amounts of gene expression data with the potential of providing novel insights into fundamental disease processes, especially complex syndromes such as cardiovascular disease, whose etiologies are due to multiple genetic factors and their interplay with the environment. Microarray and SAGE have already been used to examine gene expression patterns of cell-culture, animal and human tissues models of cardiovascular diseases. In this review, we will first give a brief introduction of microarray and SAGE     

Analysis of allelic differential expression in the human genome using allele-specific serial analysis of gene expression tags

Recent reports have demonstrated that a significant proportion of human genes display allelic differential expression (ADE). ADE is associated with phenotypic variability and may contribute to complex genetic diseases. Here, we present a computational analysis of ADE using allele-specific serial analysis of gene expression (SAGE) tags representing 1295 human genes. We identified 472 genes for which unequal representation (>3-fold) of allele-specific SAGE tags was observed in at least one SAGE library, suggesting the occurrence of ADE. For 235 out of these 472 genes, the difference in the expression level between both allele-specific SAGE tags was statistically significant (p < 0.05). Eleven candidate genes were then subjected to experimental validation and ADE was confirmed for 8 out of these 11 genes. Our results suggest that at least 25% of the human genes display ADE and that allele-specific SAGE tags can be efficiently used for the identification of such genes.     

Comparison of microarray and sage techniques in gene expression analysis of human glioblastoma

To enhance glioblastoma (GB) marker discovery we compared gene expression in GB with human normal brain (NB) by accessing SAGE Genie web site and compared obtained results with published data. Nine GB and five NB SAGE-libraries were analyzed using the Digital Gene Expression Displayer (DGED), the results of DGED were tested by Northern blot analysis and RT-PCR of arbitrary selected genes. Review of available data from the articles on gene expression profiling by microarray-based hybridization showed as few as 35 overlapped genes with increased expression in GB. Some of them were identified in four articles, but most genes in three or even in two investigations. There was found also some differences between SAGE results of GB analysis. Digital Gene Expression Displayer approach revealed 676 genes differentially expressed in GB vs. NB with cut-off ratio: twofold change and P < or = 0.05. Differential expression of selectedgenes obtained by DGED was confirmed by Northern analysis and RT-PCR. Altogether, only 105 of 955 genes presented in published investigations were among the genes obtained by DGED. Comparison of the results obtained by microarrays and SAGE is very complicated because authors present only the most prominent differentially expressed genes. However, even available data give quite poor overlapping of genes revealed by microarrays. Some differences between results obtained by SAGE in different investigations can be explained by high dependence on the statistical methods used. As for now, the best solution to search for molecular tumor markers is to compare all available results and to select only those genes, which significant expression in tumor combined with very low expression in normal tissues was reproduced in several articles. 105 differentially expressed genes, common to both methods, can be included in the list of candidates for the molecular typing of GBs. Some genes, encoded cell surface or extra-cellular proteins may be useful for targeting gliomas with antibody-based therapy.     

Comparison of microarray and SAGE techniques in gene expression analysis of human glioblastoma

To enhance glioblastoma (GB) marker discovery, we compared gene expression in GB with human normal brain (NB) by accessing the SAGE Genie web site and compared the results with published data. Nine GB and five NB SAGE libraries were analyzed using the Digital Gene Expression Displayer (DGED); the results of DGED were tested by Northern blot analysis and RT-PCR of arbitrarily selected genes. Review of available data from the articles on gene expression profiling by microarray-based hybridization showed as few as 35 overlapped genes with increased expression in GB. Some of them were identified in four articles, but most genes were identified in three or even in two investigations. Some differences were also found between SAGE results of GB analysis. The Digital Gene Expression Displayer approach revealed 676 genes differentially expressed in GB vs. NB with cutoff ratio: twofold change and P ≤ 05. Differential expression of selected genes obtained by DGED was confirmed by Northern analysis and RT-PCR. Altogether, only 105 of 955 genes presented in published investigations were among the genes obtained by DGED. Comparison of the results obtained by microarrays and SAGE is very complicated because the authors present only the most prominent differentially expressed genes. However, even available data give quite poor overlapping of genes revealed by microarrays. Some differences between results obtained by SAGE in different investigations can be explained by high dependence on the statistical methods used. As for now, the best solution to search for molecular tumor markers is to compare all available results and to select only those genes where significant expression in tumors combined with very low expression in normal tissues was reproduced in several articles. One hundred five differentially expressed genes, common to both methods, can be included in the list of candidates for the molecular typing of GBs. Some genes, encoded cell surface or extracellular proteins may be useful for targeting gliomas with antibody-based therapy. The text was submitted by the authors in English.     

Visualizing metabolic activity on a genome-wide scale

MOTIVATION: To enhance the exploration of gene expression data in a metabolic context, one requires an application that allows the integration of this data and which represents this data in a (genome-wide) metabolic map. The layout of this metabolic map must be highly flexible to enable discoveries of biological phenomena. Moreover, it must allow the simultaneous representation of additional information about genes and enzymes. Since the layout and properties of existing maps did not fulfill our requirements, we developed a new way of representing gene expression data in metabolic charts. RESULTS: ViMAc generates user-specified (genome-wide) metabolic maps to explore gene expression data. To enhance the interpretation of these maps information such as sub-cellular localization is included. ViMAc can be used to analyse human or yeast expression data obtained with DNA microarrays or SAGE. We introduce our metabolic map method and demonstrate how it can be applied to explore DNA microarray data for yeast. Availability: ViMAc is freely available for academic institutions on request from the authors.     

Interactive gene expression pattern in prostate cancer cells exposed to phenolic antioxidants

Dietary phenolic compounds are known to elicite vital cellular responses such as cell cycle arrest, apoptosis and differentiation by activating a cascade of molecular events. As there is an increasing interest to improve the efficacy of these compounds for use as potential chemopreventive agents, we wanted to understand the impact of phenolic compounds on target genes in prostate cancer. In this study we used human cDNA microarrays with 2400 clones consisting of 17 prosite motifs to characterize alterations in gene expression pattern in response to the phenolic antioxidants ellagic acid (EA) and resveratrol (RE). Over a 48-hr exposure of androgen - sensitive LNCaP cells to EA and RE, a total of 593 and 555 genes respectively, showed more than a two fold difference in expression. A distinct set of genes in both EA-and RE-treated cells may represent the signature profile of phenolic antioxidant-induced gene expression in LNCaP cells. Although extensive similarity was found between effects of EA - and RE - responsive genes in prostate cancer cells, out of 246 genes with overlapping responses, 25 genes showed an opposite effect. Quantitative RT-PCR was used to verify and validate the differential expression of selected genes identified from cDNA microarrays. In-depth analysis of the data from this study provided insight into the alterations in the p53 - responsive genes, p300, Apaf-1, NF-kBp50 and p65 and PPAR families of genes, suggesting the activation of multiple signaling pathways that leads to growth inhibition of LNCaP cells. This is a first study to look for changes in a large number of human genes in response to dietary compounds.     

Integration of text- and data-mining using ontologies successfully selects disease gene candidates

Genome-wide techniques such as microarray analysis, Serial Analysis of Gene Expression (SAGE), Massively Parallel Signature Sequencing (MPSS), linkage analysis and association studies are used extensively in the search for genes that cause diseases, and often identify many hundreds of candidate disease genes. Selection of the most probable of these candidate disease genes for further empirical analysis is a significant challenge. Additionally, identifying the genes that cause complex diseases is problematic due to low penetrance of multiple contributing genes. Here, we describe a novel bioinformatic approach that selects candidate disease genes according to their expression profiles. We use the eVOC anatomical ontology to integrate text-mining of biomedical literature and data-mining of available human gene expression data. To demonstrate that our method is successful and widely applicable, we apply it to a database of 417 candidate genes containing 17 known disease genes. We successfully select the known disease gene for 15 out of 17 diseases and reduce the candidate gene set to 63.3% (±18.8%) of its original size. This approach facilitates direct association between genomic data describing gene expression and information from biomedical texts describing disease phenotype, and successfully prioritizes candidate genes according to their expression in disease-affected tissues.     

Evolutionary conservation of expression profiles between human and mouse orthologous genes

Mouse models are often used to study human genes because it is believed that the expression and function are similar for the majority of orthologous genes between the two species. However, recent comparisons of microarray data from thousands of orthologous human and mouse genes suggested rapid evolution of gene expression profiles under minimal or no selective constraint. These findings appear to contradict non-array-based observations from many individual genes and imply the uselessness of mouse models for studying human genes. Because absolute levels of gene expression are not comparable between species when the data are generated by species-specific microarrays, use of relative mRNA abundance among tissues (RA) is preferred to that of absolute expression signals. We thus reanalyze human and mouse genome-wide gene expression data generated by oligonucleotide microarrays. We show that the mean correlation coefficient among expression profiles detected by different probe sets of the same gene is only 0.38 for humans and 0.28 for mice, indicating that current measures of expression divergence are flawed because the large estimation error (discrepancy in expression signal detected by different probe sets of the same gene) is mistakenly included in the between-species divergence. When this error is subtracted, 84% of human-mouse orthologous gene pairs show significantly lower expression divergence than that of random gene pairs. In contrast to a previous finding, but consistent with the common sense, expression profiles of orthologous tissues between species are more similar to each other than to those of nonorthologous tissues. Furthermore, the evolutionary rate of expression divergence and that of coding sequence divergence are found to be weakly, but significantly positively correlated, when RA and the Euclidean distance are used to measure expression-profile divergence. These results highlight the importance of proper consideration of various estimation errors in comparing the microarray data between species.