Integrative analysis of multiple gene expression profiles applied to liver cancer study

Regucalcin downregulation in human cancer

Regucalcin is a unique calcium-binding protein first discovered in rat liver in 1978. Regucalcin has multiple functions as an inhibitor of various cellular signaling pathways that regulate cell activity. The expression of the regucalcin gene can be altered by various physiological and pathological factors such as diet (nutrients), hormones, diabetes, alcohol and drugs. Several transcription factors have been identified on the regucalcin gene, including AP-1, NF1-A1, RGPR-p117, β-catenin, NF-κB, STAT3 and hypoxia-inducible factor-1α (HIF-1α). Notably, regucalcin plays an important role in the development of several cancers by controlling cell growth. Clinically, many studies have reported that the expression of the regucalcin gene is downregulated in various human cancers. In addition, higher expression of regucalcin in tumor tissue has been associated with longer patient survival, suggesting that regucalcin may act as a potential suppressor of various types of human cancer. Regucalcin may offer a novel therapeutic strategy and diagnostic tool for cancer treatment. However, the underlying mechanism by which regucalcin expression is reduced in human cancer is still unclear. A deeper understanding of regucalcin reduction and function in cancer is needed to discover potential resistance mechanisms and biomarkers, and to improve regucalcin-targeting agents. We review recent findings on regucalcin gene expression in cancer. We discuss the possible mechanisms by which regucalcin expression is downregulated in cancer cells to facilitate understanding of how regucalcin regulates cell growth function. This mini-review may lead to better therapeutic targets with regucalcin.

A tensor decomposition-based integrated analysis applicable to multiple gene expression profiles without sample matching

The integrated analysis of multiple gene expression profiles previously measured in distinct studies is problematic since missing both sample matches and common labels prevent their integration in fully data-driven, unsupervised training. In this study, we propose a strategy to enable the integration of multiple gene expression profiles among multiple independent studies with neither labeling nor sample matching using tensor decomposition unsupervised feature extraction. We apply this strategy to Alzheimer's disease (AD)-related gene expression profiles that lack precise correspondence among samples, including AD single-cell RNA sequence (scRNA-seq) data. We were able to select biologically reasonable genes using the integrated analysis. Overall, integrated gene expression profiles can function analogously to prior- and/or transfer-learning strategies in other machine-learning applications. For scRNA-seq, the proposed approach significantly reduces the required computational memory.

Controversial roles of hepatocyte nuclear receptor 4 α on tumorigenesis

Hepatocyte nuclear receptor 4 α (HNF4α) is known to be a master transcription regulator of gene expression in multiple biological processes, particularly in liver development and liver function. To date, the function of HNF4α in human cancers has been widely investigated; however, the critical roles of HNF4α in tumorigenesis remain unclear. Numerous controversies exist, even in studies from different research groups but on the same type of cancer. In the present review, the critical roles of HNF4α in tumorigenesis will be summarized and discussed. Furthermore, HNF4α expression profile and alterations will be examined by pan-cancer analysis through bioinformatics, in order to provide a better understanding of the functional roles of this gene in human cancers.

Hepatic Tumor Formation in Adult Mice Developmentally Exposed to Organotin

BACKGROUND

Tributyltin (TBT) is a persistent and bioaccumulative environmental toxicant. Developmental exposure to TBT has been shown to cause fatty liver disease (steatosis), as well as increased adiposity in many species, leading to its characterization as an obesogen.

OBJECTIVE

We aimed to determine the long-term effects of developmental TBT exposure on the liver.

METHODS

C57BL/6J mice were exposed to a dose of TBT (0.5 mg / kg body weight per day; 3.07 μ M ) below the current developmental no observed adverse effect level (NOAEL) via drinking water, or drinking water alone, provided to the dam from preconception through lactation. Sires were exposed during breeding and lactation. Pups from two parity cycles were included in this study. Animals were followed longitudinally, and livers of offspring were analyzed by pathological evaluation, immunohistochemistry, immunoblotting, and RNA sequencing.

RESULTS

Developmental exposure to TBT led to increased adiposity and hepatic steatosis at 14 and 20 weeks of age and increased liver adenomas at 45 weeks of age in male offspring. Female offspring displayed increased adiposity as compared with males, but TBT did not lead to an increase in fatty liver or tumor development in female offspring. Liver tumors in male mice were enriched in pathways and gene signatures associated with human and rodent nonalcoholic fatty liver disease (NAFLD) and hepatocellular carcinoma (HCC). This includes down-regulation of growth hormone receptor (GHR) and of STAT5 signaling, which occurred in response to TBT exposure and preceded liver tumor development.

CONCLUSIONS

These data reveal a previously unappreciated ability of TBT to increase risk for liver tumorigenesis in mice in a sex-specific manner. Taken together, these findings provide new insights into how early life environmental exposures contribute to liver disease in adulthood. https://doi.org/10.1289/EHP5414.

Construction of Traf3 knockout liver cancer cell line using CRISPR/Cas9 system

PURPOSE

The gene editing technology in CRISPR/Cas9 system was used to construct the Traf3 knockout HepG2 cell line to explore the role of Traf3 in the development of liver cancer.

METHODS

Five sgRNA sites were designed for the exons of Traf3. The recombinant plasmid of Lentiviral vector2-Traf3-sgRNA was constructed and transformed into Stbl3 competent cells. The recombinants were screened and sequenced, and the effectiveness of the designed gRNA was verified by sequencing. The constructed vector was transfected into HepG2 cells by lentiviral, and the monoclonal antibody was selected to detect the knockout effect of Traf3 gene in HepG2 cells by Western blot. PCR amplification and gene sequencing were performed to obtain the cell line, which the Traf3 gene was knocked out. MTT and Transwell assays were used to detect the effect of Traf3-knockout on HepG2 cell proliferation and cell invasion, respectively.

RESULTS

The Lentiviral vector2-sgRNA expression vector was successfully constructed. PCR amplification electrophoresis and gene sequencing showed that the Trep3-knockdown HepG2 cells were successfully constructed. Compared with the wild HepG2 cells group, the proliferation and invasion ability of HepG2 cells were enhanced in the Traf3 knockout group.

CONCLUSION

Knockout Traf3 gene by CRISPR/Cas9 system enhanced the proliferation, migration, and invasion of HepG2 cells, and provided an effective tool for studying the function and mechanism of Traf3.

Identification of key signaling pathways in cerebral small vessel disease using differential pathway network analysis

Cerebral small vessel disease (CSVD) primarily affects the perforating cerebral arterioles and capillaries, and results in injury to subcortical grey and white matter. Despite advances in determining the genetic basis of CSVD, the molecular mechanisms underlying the development and progression of CSVD remain unclear. The present study aimed to identify significant signaling pathways associated with CSVD based on differential pathway network analysis. Combining CSVD microarray data with human protein-protein interaction data and data from the Reactome pathway database, pathway interactions were constructed using the Spearman's correlation coefficient strategy. Pathway interactions with weight values of >0.95 were selected to construct the differential pathway network, which contained 715 differential pathway interactions covering 312 nodes and was visualized using Cytoscape software. A total of 15 hub pathways with a top 5% degree distribution in the differential pathway network were identified. The top 5 hub pathways were associated with the synthesis and metabolism of fatty acids. The results of the present study indicate that the synthesis and metabolism of fatty acids is associated with the occurrence and development of CSVD, and may thus provide insights to improve the early diagnosis and treatment of CSVD.

Merging Microarray Data, Robust Feature Selection, and Predicting Prognosis in Prostate Cancer

Motivation

Individual microarray studies searching for prognostic biomarkers often have few samples and low statistical power; however, publicly accessible data sets make it possible to combine data across studies.

Method

We present a novel approach for combining microarray data across institutions and platforms. We introduce a new algorithm, robust greedy feature selection (RGFS), to select predictive genes.

Results

We combined two prostate cancer microarray data sets, confirmed the appropriateness of the approach with the Kolmogorov-Smirnov goodness-of-fit test, and built several predictive models. The best logistic regression model with stepwise forward selection used 7 genes and had a misclassification rate of 31%. Models that combined LDA with different feature selection algorithms had misclassification rates between 19% and 33%, and the sets of genes in the models varied substantially during cross-validation. When we combined RGFS with LDA, the best model used two genes and had a misclassification rate of 15%.

Availability

Affymetrix U95Av2 array data are available at http://www.broad.mit.edu/cgi-bin/cancer/datasets.cgi . The cDNA microarray data are available through the Stanford Microarray Database ( http://cmgm.stanford.edu/pbrown/ ). GeneLink software is freely available at http://bioinformatics.mdanderson.org/GeneLink/ . DNA-Chip Analyzer software is publicly available at http://biosun1.harvard.edu/complab/dchip/ .

Integrative transcriptome analysis of liver cancer profiles identifies upstream regulators and clinical significance of ACSM3 gene expression

PURPOSE

Hepatocellular carcinoma (HCC) is one of the most common human malignancies. It has frequently been associated with metabolic perturbations and liver damages. Various members of the family of acyl-CoA synthetases are known to be involved in the production of bioactive fatty acids, and altered expression of its encoding genes has been found to be involved in metabolic perturbations. For the development of novel diagnostic and therapeutic HCC options, a fundamental understanding of the mechanisms associated with the deregulation of candidate genes involved in metabolic perturbation is required.

METHODS

A meta-analysis of multiple HCC mRNA profiles was performed to identify consistently deregulated genes. Expression of the acyl-CoA synthetase medium chain family member 3 (ACSM3) gene was subsequently assessed in different HCC tumor stages and correlated with various clinicopathological features. Transcription regulation, survival and pathway-associated features of the ACSM3 gene were investigated using integrative functional genomic and molecular cell biological methods.

RESULTS

We found that expression of the ACSM3 gene was significantly reduced in HCC tissues and was frequently downregulated in patients exhibiting high alpha-fetoprotein (AFP) levels, high alanine aminotransferase (ALT) levels, multiple nodules and large tumors. Loss of ACSM3 expression was found to correlate with advanced HCC stages and a poor survival. In addition, HNF4α was found to positively regulate the expression of the ACSM3 gene, while PPARγ was found to transcriptionally repress it. Downregulation of ACSM3 expression was perceived upon activation of the TGFβ, WNT, AKT and MYC signalling pathways. In addition, we found that ACSM3 expression correlates with fatty acid oxidation in HCC.

CONCLUSION

Our data provide evidence for a differential expression and regulation of the ACSM3 gene in HCC, and may lay a foundation for therapeutically targeting fatty acid metabolism in these tumors.

Pathway Cross-Talk Analysis in Detecting Significant Pathways in Barrett's Esophagus Patients

Background

The pathological mechanism of Barrett’s esophagus (BE) is still unclear. In the present study, pathway cross-talks were analyzed to identify hub pathways for BE, with the purpose of finding an efficient and cost-effective detection method to discover BE at its early stage and take steps to prevent its progression.

Material/Methods

We collected and preprocessed gene expression profile data, original pathway data, and protein-protein interaction (PPI) data. Then, we constructed a background pathway cross-talk network (BPCN) based on the original pathway data and PPI data, and a disease pathway cross-talk network (DPCN) based on the differential pathways between the PPI data and the BE and normal control. Finally, a comprehensive analysis was conducted on these 2 networks to identify hub pathway cross-talks for BE, so as to better understand the pathological mechanism of BE from the pathway level.

Results

A total of 12 411 genes, 300 pathways (6919 genes), and 787 896 PPI interactions (16 730 genes) were separately obtained from their own databases. Then, we constructed a BPCN with 300 nodes (42 293 interactions) and a DPCN with 296 nodes (15 073 interactions). We identified 4 hub pathways: AMP signaling pathway, cGMP-PKG signaling pathway, natural killer cell-mediated cytotoxicity, and osteoclast differentiation. We found that these pathways might play important roles during the occurrence and development of BE.

Conclusions

We predicted that these pathways (such as AMP signaling pathway and cAMP signaling pathway) could be used as potential biomarkers for early diagnosis and therapy of BE.

Downregulation of paraoxonase 3 contributes to aggressive human hepatocellular carcinoma progression and associates with poor prognosis

Paraoxonase (PON) enzymes possess antioxidant properties and protect against cardiovascular diseases. As a member of PON family, PON3 is primarily synthesized in the liver and poorly investigated. This study aimed to examine the expression of PON3 in human hepatocellular carcinoma (HCC) and investigate the clinical significance and biological function of PON3 in HCC patients. We first analyzed PON3 expression in 50 paired HCC samples (HCC tissues vs matched para-cancerous tissues) and 160 clinical HCC specimens by using immunohistochemistry (IHC). Our results showed that the expression of PON3 was downregulated in HCC and significantly associated with tumor-node-metastasis (TNM) stage, tumor size, and tumor number. Kaplan-Meier survival and Cox regression analyses showed that PON3 was an independent prognostic factor for overall survival (OS) and time to recurrence (TTR). Finally, we aimed to reveal the biological function of PON3 in HCC growth and metastasis, and our results showed that overexpression of PON3 potently inhibited growth and metastasis of HCC. Collectively, our study demonstrated that PON3 exhibited tumor-suppressive effects toward HCC and it might serve as a novel prognostic marker in HCC.

Nonalcoholic fatty liver disease and hepatocellular carcinoma

The fastest growing cause of cancer-related death is hepatocellular carcinoma (HCC), which is at least partly attributable to the rising prevalence of non-alcoholic fatty liver disease. Non-alcoholic fatty liver disease (NAFLD) encompasses a broad spectrum of conditions, ranging from non-progressive bland steatosis to malignant transformation into hepatocellular cancer. The estimated annual HCC incidence in the progressive form of NAFLD - non-alcoholic steatohepatitis (NASH) - is about 0.3%. The risk of HCC development is higher in men and increases with age, more advanced fibrosis, progressive obesity, insulin resistance and diabetes mellitus. Studies on the molecular mechanism of HCC development in NAFLD have shown that hepatocarcinogenesis is associated with complex changes at the immunometabolic interface. In line with these clinical risk factors, administration of a choline-deficient high-fat diet to mice over a prolonged period results in spontaneous HCC development in a high percentage of animals. The role of altered insulin signaling in tumorigenesis is further supported by the observation that components of the insulin-signaling cascade are frequently mutated in hepatocellular cancer cells. These changes further enhance insulin-mediated growth and cell division of hepatocytes. Furthermore, studies investigating nuclear factor kappa B (NF-κB) signaling and HCC development allowed dissection of the complex links between inflammation and carcinogenesis. To conclude, NAFLD reflects an important risk factor for HCC, develops also in non-cirrhotic livers and is a prototypic cancer involving inflammatory and metabolic pathways. STRENGTHS/WEAKNESSES AND SUMMARY OF THE TRANSLATIONAL POTENTIAL OF THE MESSAGES IN THE PAPER: The systematic review summarizes findings from unbiased clinical and translational studies on hepatocellular cancer in non-alcoholic fatty liver disease. This provides a concise overview on the epidemiology, risk factors and molecular pathogenesis of the NAFL-NASH-HCC sequence. One limitation in the field is that few HCC studies stratify patients by underlying etiology, although the etiology of the underlying liver disease is an important co-determinant of clinical disease course and molecular pathogenesis. Molecular profiling of NAFL and associated HCC holds great translational potential for individualized surveillance, prevention and therapy.

Recurrence-associated pathways in hepatitis B virus-positive hepatocellular carcinoma

Background

Despite the recent identification of several prognostic gene signatures, the lack of common genes among experimental cohorts has posed a considerable challenge in uncovering the molecular basis underlying hepatocellular carcinoma (HCC) recurrence for application in clinical purposes. To overcome the limitations of individual gene-based analysis, we applied a pathway-based approach for analysis of HCC recurrence.

Results

By implementing a permutation-based semi-supervised principal component analysis algorithm using the optimal principal component, we selected sixty-four pathways associated with hepatitis B virus (HBV)-positive HCC recurrence (p < 0.01), from our microarray dataset composed of 142 HBV-positive HCCs. In relation to the public HBV- and public hepatitis C virus (HCV)-positive HCC datasets, we detected 46 (71.9%) and 18 (28.1%) common recurrence-associated pathways, respectively. However, overlap of recurrence-associated genes between datasets was rare, further supporting the utility of the pathway-based approach for recurrence analysis between different HCC datasets. Non-supervised clustering of the 64 recurrence-associated pathways facilitated the classification of HCC patients into high- and low-risk subgroups, based on risk of recurrence (p < 0.0001). The pathways identified were additionally successfully applied to discriminate subgroups depending on recurrence risk within the public HCC datasets. Through multivariate analysis, these recurrence-associated pathways were identified as an independent prognostic factor (p < 0.0001) along with tumor number, tumor size and Edmondson’s grade. Moreover, the pathway-based approach had a clinical advantage in terms of discriminating the high-risk subgroup (N = 12) among patients (N = 26) with small HCC (<3 cm).

Conclusions

Using pathway-based analysis, we successfully identified the pathways involved in recurrence of HBV-positive HCC that may be effectively used as prognostic markers.

Electronic supplementary material

The online version of this article (doi:10.1186/s12864-015-1472-x) contains supplementary material, which is available to authorized users.

Involvement of regucalcin as a suppressor protein in human carcinogenesis: insight into the gene therapy

Regucalcin, which its gene is located on the X chromosome, plays a multifunctional role as a suppressor protein in cell signal transduction in various types of cells and tissues. The suppression of regucalcin gene expression has been shown to involve in carcinogenesis. Regucalcin gene expression was uniquely downregulated in carcinogenesis of rat liver in vivo, although the expression of other many genes was upregulated, indicating that endogenous regucalcin plays a suppressive role in the development of hepatocarcinogenesis. Overexpression of endogenous regucalcin was found to suppress proliferation of rat cloned hepatoma cells in vitro. Moreover, the regucalcin gene and its protein levels were demonstrated specifically to downregulate in human hepatocellular carcinoma by analysis with multiple gene expression profiles and proteomics. Regucalcin gene expression was also found to suppress in human tumor tissues including kidney, lung, brain, breast and prostate, suggesting that repressed regucalcin gene expression leads to the development of carcinogenesis in various tissues. Regucalcin may play a role as a suppressor protein in carcinogenesis. Overexpression of endogenous regucalcin is suggested to reveal preventive and therapeutic effects on carcinogenesis. Delivery of the regucalcin gene may be a novel useful tool in the gene therapy of carcinogenesis. This review will discuss regarding to an involvement of regucalcin as a suppressor protein in human carcinogenesis in insight into the gene therapy.

Loss of CDC14B expression in clear cell renal cell carcinoma: meta-analysis of microarray data sets

OBJECTIVES

To discover significant differentially expressed genes (DEGs) in clear cell renal cell carcinoma (ccRCC) that might be unidentified by single microarray analysis.

METHODS

The effect sizes of five ccRCC microarray data sets were combined using a random-effects model. The most downregulated gene was validated in paired 80 ccRCC tissues by immunohistochemistry.

RESULTS

CDC14B was the most downregulated gene among 1,761 DEGs. CDC14B was strongly expressed in the apical proximal tubules in the nonneoplastic tissues, while it was completely absent in 10 (12.5%) of 80 or downregulated in 70 (87.5%) of 80 ccRCC cases. The complete loss of CDC14B correlated with high T stage (P = .038), advanced TNM stage (P = .027), tumor recurrence (P = .038), and shorter recurrence-free survival (P = .046) compared with the partial loss of CDC14B.

CONCLUSIONS

Microarray meta-analysis is a useful tool for pathologists. CDC14B expression is downregulated in ccRCC, suggesting its role in renal carcinogenesis.

Integrative Analysis of Cancer Diagnosis Studies with Composite Penalization

In cancer diagnosis studies, high-throughput gene profiling has been extensively conducted, searching for genes whose expressions may serve as markers. Data generated from such studies have the "large d, small n" feature, with the number of genes profiled much larger than the sample size. Penalization has been extensively adopted for simultaneous estimation and marker selection. Because of small sample sizes, markers identified from the analysis of single datasets can be unsatisfactory. A cost-effective remedy is to conduct integrative analysis of multiple heterogeneous datasets. In this article, we investigate composite penalization methods for estimation and marker selection in integrative analysis. The proposed methods use the minimax concave penalty (MCP) as the outer penalty. Under the homogeneity model, the ridge penalty is adopted as the inner penalty. Under the heterogeneity model, the Lasso penalty and MCP are adopted as the inner penalty. Effective computational algorithms based on coordinate descent are developed. Numerical studies, including simulation and analysis of practical cancer datasets, show satisfactory performance of the proposed methods.

Application of microRNA and mRNA expression profiling on prognostic biomarker discovery for hepatocellular carcinoma

Background

Hepatocellular carcinoma (HCC) is one of the most highly malignant and lethal cancers of the world. Its pathogenesis has been reported to be multi-factorial, and the molecular carcinogenesis of HCC can not be attributed to just a few individual genes. Based on the microRNA and mRNA expression profiling of normal liver tissues, pericancerous hepatocellular tissues and hepatocellular carcinoma tissues, we attempted to find prognosis related gene sets for HCC patients.

Results

We identified differentially expressed genes (DEG) from three comparisons: Cancer/Normal, Cancer/Pericancerous and Pericancerous/Normal. GSEA (gene set enrichment analysis) were performed. Based on the enriched gene sets of GO terms, pathways and transcription factor targets, it was found that the genome instability and cell proliferation increased while the metabolism and differentiation decreased in HCC tissues. The expression profile of DEGs in each enriched gene set was used to correlate to the postoperative survival time of HCC patients. Nine gene sets were found to prognostic correlation. Furthermore, after substituting DEG-targeting-microRNA for DEG members of each gene set, two gene sets with the microRNA expression profiles were obtained that had prognostic potential.

Conclusions

The malignancy of HCC could be represented by gene sets, and pericancerous liver exhibits important characteristics of liver cancer. The expression level of gene sets not only in HCC but also in the pericancerous liver showed potential for prognosis implying an option for HCC prognosis at an early stage. Additionally, the gene-targeting-microRNA expression profiles also showed prognostic potential, demonstrating that the multi-factorial molecular pathogenesis of HCC is contributed by various genes and microRNAs.

Electronic supplementary material

The online version of this article (doi:10.1186/1471-2164-15-S1-S13) contains supplementary material, which is available to authorized users.

Cross-Platform Microarray Data Integration Combining Meta-Analysis and Gene Set Enrichment Analysis

Integrative analysis of microarray data has been proven as a more reliable approach to deciphering molecular mechanisms underlying biological studies. Traditional integration such as meta-analysis is usually gene-centered. Recently, gene set enrichment analysis (GSEA) has been widely applied to bring gene-level interpretation to pathway-level. GSEA is an algorithm focusing on whether an a priori defined set of genes shows statistically significant differences between two biological states. However, GSEA does not support integrating multiple microarray datasets generated from different studies. To overcome this, the improved version of GSEA, ASSESS, is more applicable, after necessary modifications. By making proper combined use of meta-analysis, GSEA, and modified ASSESS, this chapter reports two workflow pipelines to extract consistent expression pattern change at pathway-level, from multiple microarray datasets generated by the same or different microarray production platforms, respectively. Such strategies amplify the advantage and overcome the disadvantage than if using each method individually, and may achieve a more comprehensive interpretation towards a biological theme based on an increased sample size. With further network analysis, it may also allow an overview of cross-talking pathways based on statistical integration of multiple gene expression studies. A web server where one of the pipelines is implemented is available at: http://lifecenter.sgst.cn/mgsea//home.htm.

Integrative prescreening in analysis of multiple cancer genomic studies

BACKGROUND

In high throughput cancer genomic studies, results from the analysis of single datasets often suffer from a lack of reproducibility because of small sample sizes. Integrative analysis can effectively pool and analyze multiple datasets and provides a cost effective way to improve reproducibility. In integrative analysis, simultaneously analyzing all genes profiled may incur high computational cost. A computationally affordable remedy is prescreening, which fits marginal models, can be conducted in a parallel manner, and has low computational cost.

RESULTS

An integrative prescreening approach is developed for the analysis of multiple cancer genomic datasets. Simulation shows that the proposed integrative prescreening has better performance than alternatives, particularly including prescreening with individual datasets, an intensity approach and meta-analysis. We also analyze multiple microarray gene profiling studies on liver and pancreatic cancers using the proposed approach.

CONCLUSIONS

The proposed integrative prescreening provides an effective way to reduce the dimensionality in cancer genomic studies. It can be coupled with existing analysis methods to identify cancer markers.

Protectors or Traitors: The Roles of PON2 and PON3 in Atherosclerosis and Cancer

Cancer and atherosclerosis are major causes of death in western societies. Deregulated cell death is common to both diseases, with significant contribution of inflammatory processes and oxidative stress. These two form a vicious cycle and regulate cell death pathways in either direction. This raises interest in antioxidative systems. The human enzymes paraoxonase-2 (PON2) and PON3 are intracellular enzymes with established antioxidative effects and protective functions against atherosclerosis. Underlying molecular mechanisms, however, remained elusive until recently. Novel findings revealed that both enzymes locate to mitochondrial membranes where they interact with coenzyme Q10 and diminish oxidative stress. As a result, ROS-triggered mitochondrial apoptosis and cell death are reduced. From a cardiovascular standpoint, this is beneficial given that enhanced loss of vascular cells and macrophage death forms the basis for atherosclerotic plaque development. However, the same function has now been shown to raise chemotherapeutic resistance in several cancer cells. Intriguingly, PON2 as well as PON3 are frequently found upregulated in tumor samples. Here we review studies reporting PON2/PON3 deregulations in cancer, summarize most recent findings on their anti-oxidative and antiapoptotic mechanisms, and discuss how this could be used in putative future therapies to target atherosclerosis and cancer.

Identification of cancer genomic markers via integrative sparse boosting

In high-throughput cancer genomic studies, markers identified from the analysis of single data sets often suffer a lack of reproducibility because of the small sample sizes. An ideal solution is to conduct large-scale prospective studies, which are extremely expensive and time consuming. A cost-effective remedy is to pool data from multiple comparable studies and conduct integrative analysis. Integrative analysis of multiple data sets is challenging because of the high dimensionality of genomic measurements and heterogeneity among studies. In this article, we propose a sparse boosting approach for marker identification in integrative analysis of multiple heterogeneous cancer diagnosis studies with gene expression measurements. The proposed approach can effectively accommodate the heterogeneity among multiple studies and identify markers with consistent effects across studies. Simulation shows that the proposed approach has satisfactory identification results and outperforms alternatives including an intensity approach and meta-analysis. The proposed approach is used to identify markers of pancreatic cancer and liver cancer.

Microarray meta-analysis defines global angiogenesis-related gene expression signatures in human carcinomas

Angiogenesis is a prerequisite for progression of cancers. The number of genes linked to angiogenesis suggests the existence of complex gene-networks, which remain to be elucidated. To identify angiogenesis genes deregulated in carcinomas, we performed a meta-profiling analysis of published gene expression microarray studies. Own microarray and quantitative RT-PCR data were obtained from a colorectal carcinoma cohort. Applying highly stringent inclusion criteria, 15 cancer array studies were suitable for our analysis. These studies provided 789 tumor specimens and 190 samples of healthy tissues yielding a total of approx. 1,000,000 gene expression measurements. Meta-analysis on the expression of 480 angiogenesis-related genes in 10 cancer types identified a characteristic, entity-independent "global" cancer expression signature of 25 angiogenesis-related genes showing high frequency down-regulation when compared to corresponding healthy tissues. Furthermore, we characterized 25 genes displaying frequent up-regulation, yet less often than the 25 down-regulated genes. Comparative inter-study cross-validation revealed that both signatures discriminate cancers from healthy tissues with high accuracy in independent test sets. Moreover, own microarray data of colorectal carcinomas confirmed the specific and sensitive discriminating potential of both signatures. These results were validated by quantitative RT-PCR for eight genes displaying the highest differences in the microarray analysis. Our study for the first time defines global gene expression signatures linked to angiogenesis in carcinomas. Our findings suggest that gene down-regulation may represent a central aspect of tumor angiogenesis.

Integrated analysis of the heterogeneous microarray data

Background

As the magnitude of the experiment increases, it is common to combine various types of microarrays such as paired and non-paired microarrays from different laboratories or hospitals. Thus, it is important to analyze microarray data together to derive a combined conclusion after accounting for heterogeneity among data sets. One of the main objectives of the microarray experiment is to identify differentially expressed genes among the different experimental groups. We propose the linear mixed effect model for the integrated analysis of the heterogeneous microarray data sets.

Results

The proposed linear mixed effect model was illustrated using the data from 133 microarrays collected at three different hospitals. Though simulation studies, we compared the proposed linear mixed effect model approach with the meta-analysis and the ANOVA model approaches. The linear mixed effect model approach was shown to provide higher powers than the other approaches.

Conclusions

The linear mixed effect model has advantages of allowing for various types of covariance structures over ANOVA model. Further, it can handle easily the correlated microarray data such as paired microarray data and repeated microarray data from the same subject.

Strengthening insights into host responses to mastitis infection in ruminants by combining heterogeneous microarray data sources

Background

Gene expression profiling studies of mastitis in ruminants have provided key but fragmented knowledge for the understanding of the disease. A systematic combination of different expression profiling studies via meta-analysis techniques has the potential to test the extensibility of conclusions based on single studies. Using the program Pointillist, we performed meta-analysis of transcription-profiling data from six independent studies of infections with mammary gland pathogens, including samples from cattle challenged in vivo with S. aureus, E. coli, and S. uberis, samples from goats challenged in vivo with S. aureus, as well as cattle macrophages and ovine dendritic cells infected in vitro with S. aureus. We combined different time points from those studies, testing different responses to mastitis infection: overall (common signature), early stage, late stage, and cattle-specific.

Results

Ingenuity Pathway Analysis of affected genes showed that the four meta-analysis combinations share biological functions and pathways (e.g. protein ubiquitination and polyamine regulation) which are intrinsic to the general disease response. In the overall response, pathways related to immune response and inflammation, as well as biological functions related to lipid metabolism were altered. This latter observation is consistent with the milk fat content depression commonly observed during mastitis infection. Complementarities between early and late stage responses were found, with a prominence of metabolic and stress signals in the early stage and of the immune response related to the lipid metabolism in the late stage; both mechanisms apparently modulated by few genes, including XBP1 and SREBF1.

The cattle-specific response was characterized by alteration of the immune response and by modification of lipid metabolism. Comparison of E. coli and S. aureus infections in cattle in vivo revealed that affected genes showing opposite regulation had the same altered biological functions and provided evidence that E. coli caused a stronger host response.

Conclusions

This meta-analysis approach reinforces previous findings but also reveals several novel themes, including the involvement of genes, biological functions, and pathways that were not identified in individual studies. As such, it provides an interesting proof of principle for future studies combining information from diverse heterogeneous sources.

Large-scale integration of microarray data reveals genes and pathways common to multiple cancer types

The global gene expression analysis of cancer and healthy tissues typically results in large numbers of genes that are significantly altered in cancer. Such data, however, has been difficult to interpret due to the high level of variation of gene lists across laboratories and the small sample sizes used in individual studies. In this investigation, we compiled microarray data obtained from the same platform family from 84 laboratories, resulting in a database containing 1,043 healthy tissue samples and 4,900 cancer samples for 13 different tissue types. The primary cancers considered included adrenal gland, brain, breast, cervix, colon, kidney, liver, lung, ovary, pancreas, prostate and skin tissues. We normalized the data together and analyzed subsets for the discovery of genes involved in normal to cancer transformation. Our integrated significance analysis of microarrays approach produced top 400 gene lists for each of the 13 cancer types. These lists were highly statistically enriched with genes already associated with cancer in research publications excluding microarray studies (p < 1.31 E - 12). The genes MTIM and RRM2 appeared in nine and TOP2A in eight lists of significantly altered genes in cancer. In total, there were 132 genes present in at least four gene lists, 11 of which were not previously associated with cancer. The list contains 17 metal ions and 15 adenyl ribonucleotide binding proteins, six kinases and six transcription factors. Our results point to the value of integrating microarray data in the study of combination drug therapies targeting metastasis.

Integrative analysis and variable selection with multiple high-dimensional data sets

In high-throughput -omics studies, markers identified from analysis of single data sets often suffer from a lack of reproducibility because of sample limitation. A cost-effective remedy is to pool data from multiple comparable studies and conduct integrative analysis. Integrative analysis of multiple -omics data sets is challenging because of the high dimensionality of data and heterogeneity among studies. In this article, for marker selection in integrative analysis of data from multiple heterogeneous studies, we propose a 2-norm group bridge penalization approach. This approach can effectively identify markers with consistent effects across multiple studies and accommodate the heterogeneity among studies. We propose an efficient computational algorithm and establish the asymptotic consistency property. Simulations and applications in cancer profiling studies show satisfactory performance of the proposed approach.

Meta-analysis of muscle transcriptome data using the MADMuscle database reveals biologically relevant gene patterns

Background

DNA microarray technology has had a great impact on muscle research and microarray gene expression data has been widely used to identify gene signatures characteristic of the studied conditions. With the rapid accumulation of muscle microarray data, it is of great interest to understand how to compare and combine data across multiple studies. Meta-analysis of transcriptome data is a valuable method to achieve it. It enables to highlight conserved gene signatures between multiple independent studies. However, using it is made difficult by the diversity of the available data: different microarray platforms, different gene nomenclature, different species studied, etc.

Description

We have developed a system tool dedicated to muscle transcriptome data. This system comprises a collection of microarray data as well as a query tool. This latter allows the user to extract similar clusters of co-expressed genes from the database, using an input gene list. Common and relevant gene signatures can thus be searched more easily. The dedicated database consists in a large compendium of public data (more than 500 data sets) related to muscle (skeletal and heart). These studies included seven different animal species from invertebrates (Drosophila melanogaster, Caenorhabditis elegans) and vertebrates (Homo sapiens, Mus musculus, Rattus norvegicus, Canis familiaris, Gallus gallus). After a renormalization step, clusters of co-expressed genes were identified in each dataset. The lists of co-expressed genes were annotated using a unified re-annotation procedure. These gene lists were compared to find significant overlaps between studies.

Conclusions

Applied to this large compendium of data sets, meta-analyses demonstrated that conserved patterns between species could be identified. Focusing on a specific pathology (Duchenne Muscular Dystrophy) we validated results across independent studies and revealed robust biomarkers and new pathways of interest. The meta-analyses performed with MADMuscle show the usefulness of this approach. Our method can be applied to all public transcriptome data.

Asymmetric microarray data produces gene lists highly predictive of research literature on multiple cancer types

Background

Much of the public access cancer microarray data is asymmetric, belonging to datasets containing no samples from normal tissue. Asymmetric data cannot be used in standard meta-analysis approaches (such as the inverse variance method) to obtain large sample sizes for statistical power enrichment. Noting that plenty of normal tissue microarray samples exist in studies not involving cancer, we investigated the viability and accuracy of an integrated microarray analysis approach based on significance analysis of microarrays (merged SAM) using a collection of data from separate diseased and normal samples.

Results

We focused on five solid cancer types (colon, kidney, liver, lung, and pancreas), where available microarray data allowed us to compare meta-analysis and integrated approaches. Our results from the merged SAM significantly overlapped gene lists from the validated inverse-variance method. Both meta-analysis and merged SAM approaches successfully captured the aberrances in the cell cycle that commonly occur in the different cancer types. However, the integrated SAM analysis replicated the known cancer literature (excluding microarray studies) with much more accuracy than the meta-analysis.

Conclusion

The merged SAM test is a powerful, robust approach for combining data from similar platforms and for analyzing asymmetric datasets, including those with only normal or only cancer samples that cannot be utilized by meta-analysis methods. The integrated SAM approach can also be used in comparing global gene expression between various subtypes of cancer arising from the same tissue.

Bimodal gene expression patterns in breast cancer

We identified a set of genes with an unexpected bimodal distribution among breast cancer patients in multiple studies. The property of bimodality seems to be common, as these genes were found on multiple microarray platforms and in studies with different end-points and patient cohorts. Bimodal genes tend to cluster into small groups of four to six genes with synchronised expression within the group (but not between the groups), which makes them good candidates for robust conditional descriptors. The groups tend to form concise network modules underlying their function in cancerogenesis of breast neoplasms.

Generation of expression clone set for functional proteomics of human gastric and liver cancers

Two thousand sixty-eight multi-purpose expression clones for the 326 candidate genes related to gastric or liver cancers were constructed using the Gateway system. These clones can be expressed as His, Glutathione-S-transferase (GST) or Enhanced version of the green fluorescent protein (EGFP) fusion proteins in E. coli, insect cells or mammalian cells. For the 246 E. coli expression clones, the GST fusion proteins had greater expression efficiency and solubility than the His fusion proteins. Approximately 20% of the expressed proteins had unexpected molecular weights. A detailed sequence analysis of these clones revealed frameshift mutations resulting from insertion, deletion or substitution of nucleotides. The results indicate that these changes in the candidate genes may affect the occurrence of gastric or liver cancers. In addition, when 105 proteins, which were expressed in E. coli at very low or undetectable levels, were expressed in insect cells, 76% of the proteins were expressed very well and most were soluble. We also found that most of the 30 proteins prepared using EGFP mammalian expression clones were localized to cellular compartments expected by Gene ontology (GO) and this localization was unaffected if the EGFP-fusion was at the N-terminal or C-terminal region of the protein. Antibody production and subcellular localization analysis of the candidate genes as well as a screen of genes involved in carcinogenesis pathways are currently in progress using these expression clones. These studies provide a valuable resource for developing a better understanding of the molecular mechanism of carcinogenesis in both gastric and liver cancer and would be very helpful in diagnosis and therapeutic predictions.

A global meta-analysis of microarray expression data to predict unknown gene functions and estimate the literature-data divide

MOTIVATION

Approximately 9334 (37%) of human genes have no publications documenting their function and, for those that are published, the number of publications per gene is highly skewed. Furthermore, for reasons not clear, the entry of new gene names into the literature has slowed in recent years. If we are to better understand human/mammalian biology and complete the catalog of human gene function, it is important to finish predicting putative functions for these genes based upon existing experimental evidence.

RESULTS

A global meta-analysis (GMA) of all publicly available GEO two-channel human microarray datasets (3551 experiments total) was conducted to identify genes with recurrent, reproducible patterns of co-regulation across different conditions. Patterns of co-expression were divided into parallel (i.e. genes are up and down-regulated together) and anti-parallel. Several ranking methods to predict a gene's function based on its top 20 co-expressed gene pairs were compared. In the best method, 34% of predicted Gene Ontology (GO) categories matched exactly with the known GO categories for approximately 5000 genes analyzed versus only 3% for random gene sets. Only 2.4% of co-expressed gene pairs were found as co-occurring gene pairs in MEDLINE.

CONCLUSIONS

Via a GO enrichment analysis, genes co-expressed in parallel with the query gene were frequently associated with the same GO categories, whereas anti-parallel genes were not. Combining parallel and anti-parallel genes for analysis resulted in fewer significant GO categories, suggesting they are best analyzed separately. Expression databases contain much unexpected genetic knowledge that has not yet been reported in the literature. A total of 1642 Human genes with unknown function were differentially expressed in at least 30 experiments.

AVAILABILITY

Data matrix available upon request.

A novel regulatory event-based gene set analysis method for exploring global functional changes in heterogeneous genomic data sets

BACKGROUND

Analyzing gene expression data by assessing the significance of pre-defined gene sets, rather than individual genes, has become a main approach in microarray data analysis and this has promisingly derive new biological interpretations of microarray data. However, the detection power of conventional gene list or gene set-based approaches is limited on highly heterogeneous samples, such as tumors.

RESULTS

We developed a novel method, the regulatory event-based Gene Set Analysis (eGSA), which considers not only the consistently changed genes but also every gene regulation (event) of each sample to overcome the detection limit. In comparison with conventional methods, eGSA can detect functional changes in heterogeneous samples more precisely and robustly. Furthermore, by utilizing eGSA, we successfully revealed novel functional characteristics and potential mechanisms of very early hepatocellular carcinoma (HCC).

CONCLUSION

Our study creates a novel scheme to directly target the major cellular functional changes in heterogeneous samples. All potential regulatory routines of a functional change can be further analyzed by the regulatory event frequency. We also provide a case study on early HCCs and reveal a novel insight at the initial stage of hepatocarcinogenesis. eGSA therefore accelerates and refines the interpretation of heterogeneous genomic data sets in the absence of gene-phenotype correlations.

An overview of hepatocellular carcinoma study by omics-based methods

Hepatocellular carcinoma (HCC) is one of the most deadly malignancies worldwide. Scientists have been studying the molecular mechanism of HCC for years, but the understanding of it remains incomplete and scattered across the literature at different molecular levels. Chromosomal aberrations, epigenetic abnormality and changes of gene expression have been reported in HCC. High-throughput omics technologies have been widely applied, aiming at the discovery of candidate biomarkers for cancer staging, prediction of recurrence and prognosis, and treatment selection. Large amounts of data on genetic and epigenetic abnormalities, gene expression profiles, microRNA expression profiles and proteomics have been accumulating, and bioinformatics is playing a more and more important role. In this paper, we review the current omics-based studies on HCC at the levels of genomics, transcriptomics and proteomics. Integrating observations from multiple aspects is an essential step toward the systematic understanding of the disease.

Regularized gene selection in cancer microarray meta-analysis

BACKGROUND

In cancer studies, it is common that multiple microarray experiments are conducted to measure the same clinical outcome and expressions of the same set of genes. An important goal of such experiments is to identify a subset of genes that can potentially serve as predictive markers for cancer development and progression. Analyses of individual experiments may lead to unreliable gene selection results because of the small sample sizes. Meta analysis can be used to pool multiple experiments, increase statistical power, and achieve more reliable gene selection. The meta analysis of cancer microarray data is challenging because of the high dimensionality of gene expressions and the differences in experimental settings amongst different experiments.

RESULTS

We propose a Meta Threshold Gradient Descent Regularization (MTGDR) approach for gene selection in the meta analysis of cancer microarray data. The MTGDR has many advantages over existing approaches. It allows different experiments to have different experimental settings. It can account for the joint effects of multiple genes on cancer, and it can select the same set of cancer-associated genes across multiple experiments. Simulation studies and analyses of multiple pancreatic and liver cancer experiments demonstrate the superior performance of the MTGDR.

CONCLUSION

The MTGDR provides an effective way of analyzing multiple cancer microarray studies and selecting reliable cancer-associated genes.

A resampling-based meta-analysis for detection of differential gene expression in breast cancer

Background

Accuracy in the diagnosis of breast cancer and classification of cancer subtypes has improved over the years with the development of well-established immunohistopathological criteria. More recently, diagnostic gene-sets at the mRNA expression level have been tested as better predictors of disease state. However, breast cancer is heterogeneous in nature; thus extraction of differentially expressed gene-sets that stably distinguish normal tissue from various pathologies poses challenges. Meta-analysis of high-throughput expression data using a collection of statistical methodologies leads to the identification of robust tumor gene expression signatures.

Methods

A resampling-based meta-analysis strategy, which involves the use of resampling and application of distribution statistics in combination to assess the degree of significance in differential expression between sample classes, was developed. Two independent microarray datasets that contain normal breast, invasive ductal carcinoma (IDC), and invasive lobular carcinoma (ILC) samples were used for the meta-analysis. Expression of the genes, selected from the gene list for classification of normal breast samples and breast tumors encompassing both the ILC and IDC subtypes were tested on 10 independent primary IDC samples and matched non-tumor controls by real-time qRT-PCR. Other existing breast cancer microarray datasets were used in support of the resampling-based meta-analysis.

Results

The two independent microarray studies were found to be comparable, although differing in their experimental methodologies (Pearson correlation coefficient, R = 0.9389 and R = 0.8465 for ductal and lobular samples, respectively). The resampling-based meta-analysis has led to the identification of a highly stable set of genes for classification of normal breast samples and breast tumors encompassing both the ILC and IDC subtypes. The expression results of the selected genes obtained through real-time qRT-PCR supported the meta-analysis results.

Conclusion

The proposed meta-analysis approach has the ability to detect a set of differentially expressed genes with the least amount of within-group variability, thus providing highly stable gene lists for class prediction. Increased statistical power and stringent filtering criteria used in the present study also make identification of novel candidate genes possible and may provide further insight to improve our understanding of breast cancer development.

Validation of putative reference genes for gene expression studies in human hepatocellular carcinoma using real-time quantitative RT-PCR

BACKGROUND

Reference genes, which are often referred to as housekeeping genes are frequently used to normalize mRNA levels between different samples in quantitative reverse transcription polymerase chain reaction (qRT-PCR). The selection of reference genes is critical for gene expression studies because the expression of these genes may vary among tissues or cells and may change under certain circumstances. Here, a systematic evaluation of six putative reference genes for gene expression studies in human hepatocellular carcinoma (HCC) is presented.

METHODS

Six genes, beta-2-microglobulin (B2M), glyceraldehyde-3-phosphate dehydrogenase (GAPDH), hydroxymethyl-bilane synthase (HMBS), hypoxanthine phosphoribosyl-transferase 1 (HPRT1), succinate dehydrogenase complex, subunit A (SDHA) and ubiquitin C (UBC), with distinct functional characteristics and expression patterns were evaluated by qRT-PCR. Inhibitory substances in RNA samples were quantitatively assessed and controlled using an external RNA control. The stability of selected reference genes was analyzed using both geNorm and NormFinder software.

RESULTS

HMBS and GAPDH were identified as the optimal reference genes for normalizing gene expression data between paired tumoral and adjacent non-tumoral tissues derived from patients with HCC. HMBS, GAPDH and UBC were identified to be suitable for the normalization of gene expression data among tumor tissues; whereas the combination of HMBS, B2M, SDHA and GAPDH was suitable for normalizing gene expression data among five liver cancer cell lines, namely Hep3B, HepG2, HuH7, SK-HEP-1 and SNU-182. The determined gene stability was increased after exclusion of RNA samples containing relatively higher inhibitory substances.

CONCLUSION

Of six genes studied, HMBS was found to be the single best reference gene for gene expression studies in HCC. The appropriate choice of combination of more than one reference gene to improve qRT-PCR accuracy depends on the kind of liver tissues or cells under investigation. Quantitative assessment and control of qRT-PCR inhibitors using an external RNA control can reduce the variation of qRT-PCR assay and facilitate the evaluation of gene stability. Our results may facilitate the choice of reference genes for expression studies in HCC.

Validation of putative reference genes for gene expression studies in human hepatocellular carcinoma using real-time quantitative RT-PCR

BACKGROUND

Reference genes, which are often referred to as housekeeping genes are frequently used to normalize mRNA levels between different samples in quantitative reverse transcription polymerase chain reaction (qRT-PCR). The selection of reference genes is critical for gene expression studies because the expression of these genes may vary among tissues or cells and may change under certain circumstances. Here, a systematic evaluation of six putative reference genes for gene expression studies in human hepatocellular carcinoma (HCC) is presented.

METHODS

Six genes, beta-2-microglobulin (B2M), glyceraldehyde-3-phosphate dehydrogenase (GAPDH), hydroxymethyl-bilane synthase (HMBS), hypoxanthine phosphoribosyl-transferase 1 (HPRT1), succinate dehydrogenase complex, subunit A (SDHA) and ubiquitin C (UBC), with distinct functional characteristics and expression patterns were evaluated by qRT-PCR. Inhibitory substances in RNA samples were quantitatively assessed and controlled using an external RNA control. The stability of selected reference genes was analyzed using both geNorm and NormFinder software.

RESULTS

HMBS and GAPDH were identified as the optimal reference genes for normalizing gene expression data between paired tumoral and adjacent non-tumoral tissues derived from patients with HCC. HMBS, GAPDH and UBC were identified to be suitable for the normalization of gene expression data among tumor tissues; whereas the combination of HMBS, B2M, SDHA and GAPDH was suitable for normalizing gene expression data among five liver cancer cell lines, namely Hep3B, HepG2, HuH7, SK-HEP-1 and SNU-182. The determined gene stability was increased after exclusion of RNA samples containing relatively higher inhibitory substances.

CONCLUSION

Of six genes studied, HMBS was found to be the single best reference gene for gene expression studies in HCC. The appropriate choice of combination of more than one reference gene to improve qRT-PCR accuracy depends on the kind of liver tissues or cells under investigation. Quantitative assessment and control of qRT-PCR inhibitors using an external RNA control can reduce the variation of qRT-PCR assay and facilitate the evaluation of gene stability. Our results may facilitate the choice of reference genes for expression studies in HCC.

Key issues in conducting a meta-analysis of gene expression microarray datasets

Adaikalavan Ramasamy and colleagues outline seven key issues and suggest a stepwise approach in conducting a meta-analysis of microarray datasets.

Combining transcriptional datasets using the generalized singular value decomposition

Background

Both microarrays and quantitative real-time PCR are convenient tools for studying the transcriptional levels of genes. The former is preferable for large scale studies while the latter is a more targeted technique. Because of platform-dependent systematic effects, simple comparisons or merging of datasets obtained by these technologies are difficult, even though they may often be desirable. These difficulties are exacerbated if there is only partial overlap between the experimental conditions and genes probed in the two datasets.

Results

We show here that the generalized singular value decomposition provides a practical tool for merging a small, targeted dataset obtained by quantitative real-time PCR of specific genes with a much larger microarray dataset. The technique permits, for the first time, the identification of genes present in only one dataset co-expressed with a target gene present exclusively in the other dataset, even when experimental conditions for the two datasets are not identical. With the rapidly increasing number of publically available large scale microarray datasets the latter is frequently the case. The method enables us to discover putative candidate genes involved in the biosynthesis of the (1,3;1,4)-β-D-glucan polysaccharide found in plant cell walls.

Conclusion

We show that the generalized singular value decomposition provides a viable tool for a combined analysis of two gene expression datasets with only partial overlap of both gene sets and experimental conditions. We illustrate how the decomposition can be optimized self-consistently by using a judicious choice of genes to define it. The ability of the technique to seamlessly define a concept of "co-expression" across both datasets provides an avenue for meaningful data integration. We believe that it will prove to be particularly useful for exploiting large, publicly available, microarray datasets for species with unsequenced genomes by complementing them with more limited in-house expression measurements.

An integrative approach to characterize disease-specific pathways and their coordination: a case study in cancer

Background

The most common application of microarray technology in disease research is to identify genes differentially expressed in disease versus normal tissues. However, it is known that, in complex diseases, phenotypes are determined not only by genes, but also by the underlying structure of genetic networks. Often, it is the interaction of many genes that causes phenotypic variations.

Results

In this work, using cancer as an example, we develop graph-based methods to integrate multiple microarray datasets to discover disease-related co-expression network modules. We propose an unsupervised method that take into account both co-expression dynamics and network topological information to simultaneously infer network modules and phenotype conditions in which they are activated or de-activated. Using our method, we have discovered network modules specific to cancer or subtypes of cancers. Many of these modules are consistent with or supported by their functional annotations or their previously known involvement in cancer. In particular, we identified a module that is predominately activated in breast cancer and is involved in tumor suppression. While individual components of this module have been suggested to be associated with tumor suppression, their coordinated function has never been elucidated. Here by adopting a network perspective, we have identified their interrelationships and, particularly, a hub gene PDGFRL that may play an important role in this tumor suppressor network.

Conclusion

Using a network-based approach, our method provides new insights into the complex cellular mechanisms that characterize cancer and cancer subtypes. By incorporating co-expression dynamics information, our approach can not only extract more functionally homogeneous modules than those based solely on network topology, but also reveal pathway coordination beyond co-expression.

Genome-wide differences in hepatitis C- vs alcoholism-associated hepatocellular carcinoma

AIM: To look at a comprehensive picture of etiology-dependent gene abnormalities in hepatocellular carcinoma in Western Europe.

METHODS: With a liver-oriented microarray, transcript levels were compared in nodules and cirrhosis from a training set of patients with hepatocellular carcinoma (alcoholism, 12; hepatitis C, 10) and 5 controls. Loose or tight selection of informative transcripts with an abnormal abundance was statistically valid and the tightly selected transcripts were next quantified by qRTPCR in the nodules from our training set (12 + 10) and a test set (6 + 7).

RESULTS: A selection of 475 transcripts pointed to significant gene over-representation on chromosome 8 (alcoholism) or -2 (hepatitis C) and ontology indicated a predominant inflammatory response (alcoholism) or changes in cell cycle regulation, transcription factors and interferon responsiveness (hepatitis C). A stringent selection of 23 transcripts whose differences between etiologies were significant in nodules but not in cirrhotic tissue indicated that the above dysregulations take place in tumor but not in the surrounding cirrhosis. These 23 transcripts separated our test set according to etiologies. The inflammation-associated transcripts pointed to limited alterations of free iron metabolism in alcoholic vs hepatitis C tumors.

CONCLUSION: Etiology-specific abnormalities (chromosome preference; differences in transcriptomes and related functions) have been identified in hepatocellular carcinoma driven by alcoholism or hepatitis C. This may open novel avenues for differential therapies in this disease.

A non-parametric meta-analysis approach for combining independent microarray datasets: application using two microarray datasets pertaining to chronic allograft nephropathy

Background

With the popularity of DNA microarray technology, multiple groups of researchers have studied the gene expression of similar biological conditions. Different methods have been developed to integrate the results from various microarray studies, though most of them rely on distributional assumptions, such as the t-statistic based, mixed-effects model, or Bayesian model methods. However, often the sample size for each individual microarray experiment is small. Therefore, in this paper we present a non-parametric meta-analysis approach for combining data from independent microarray studies, and illustrate its application on two independent Affymetrix GeneChip studies that compared the gene expression of biopsies from kidney transplant recipients with chronic allograft nephropathy (CAN) to those with normal functioning allograft.

Results

The simulation study comparing the non-parametric meta-analysis approach to a commonly used t-statistic based approach shows that the non-parametric approach has better sensitivity and specificity. For the application on the two CAN studies, we identified 309 distinct genes that expressed differently in CAN. By applying Fisher's exact test to identify enriched KEGG pathways among those genes called differentially expressed, we found 6 KEGG pathways to be over-represented among the identified genes. We used the expression measurements of the identified genes as predictors to predict the class labels for 6 additional biopsy samples, and the predicted results all conformed to their pathologist diagnosed class labels.

Conclusion

We present a new approach for combining data from multiple independent microarray studies. This approach is non-parametric and does not rely on any distributional assumptions. The rationale behind the approach is logically intuitive and can be easily understood by researchers not having advanced training in statistics. Some of the identified genes and pathways have been reported to be relevant to renal diseases. Further study on the identified genes and pathways may lead to better understanding of CAN at the molecular level.

Functional genomics of the chicken--a model organism

Since the sequencing of the genome and the development of high-throughput tools for the exploration of functional elements of the genome, the chicken has reached model organism status. Functional genomics focuses on understanding the function and regulation of genes and gene products on a global or genome-wide scale. Systems biology attempts to integrate functional information derived from multiple high-content data sets into a holistic view of all biological processes within a cell or organism. Generation of a large collection ( approximately 600K) of chicken expressed sequence tags, representing most tissues and developmental stages, has enabled the construction of high-density microarrays for transcriptional profiling. Comprehensive analysis of this large expressed sequence tag collection and a set of approximately 20K full-length cDNA sequences indicate that the transcriptome of the chicken represents approximately 20,000 genes. Furthermore, comparative analyses of these sequences have facilitated functional annotation of the genome and the creation of several bioinformatic resources for the chicken. Recently, about 20 papers have been published on transcriptional profiling with DNA microarrays in chicken tissues under various conditions. Proteomics is another powerful high-throughput tool currently used for examining the dynamics of protein expression in chicken tissues and fluids. Computational analyses of the chicken genome are providing new insight into the evolution of gene families in birds and other organisms. Abundant functional genomic resources now support large-scale analyses in the chicken and will facilitate identification of transcriptional mechanisms, gene networks, and metabolic or regulatory pathways that will ultimately determine the phenotype of the bird. New technologies such as marker-assisted selection, transgenics, and RNA interference offer the opportunity to modify the phenotype of the chicken to fit defined production goals. This review focuses on functional genomics in the chicken and provides a road map for large-scale exploration of the chicken genome.

Prediction potential of candidate biomarker sets identified and validated on gene expression data from multiple datasets

BACKGROUND

Independently derived expression profiles of the same biological condition often have few genes in common. In this study, we created populations of expression profiles from publicly available microarray datasets of cancer (breast, lymphoma and renal) samples linked to clinical information with an iterative machine learning algorithm. ROC curves were used to assess the prediction error of each profile for classification. We compared the prediction error of profiles correlated with molecular phenotype against profiles correlated with relapse-free status. Prediction error of profiles identified with supervised univariate feature selection algorithms were compared to profiles selected randomly from a) all genes on the microarray platform and b) a list of known disease-related genes (a priori selection). We also determined the relevance of expression profiles on test arrays from independent datasets, measured on either the same or different microarray platforms.

RESULTS

Highly discriminative expression profiles were produced on both simulated gene expression data and expression data from breast cancer and lymphoma datasets on the basis of ER and BCL-6 expression, respectively. Use of relapse-free status to identify profiles for prognosis prediction resulted in poorly discriminative decision rules. Supervised feature selection resulted in more accurate classifications than random or a priori selection, however, the difference in prediction error decreased as the number of features increased. These results held when decision rules were applied across-datasets to samples profiled on the same microarray platform.

CONCLUSION

Our results show that many gene sets predict molecular phenotypes accurately. Given this, expression profiles identified using different training datasets should be expected to show little agreement. In addition, we demonstrate the difficulty in predicting relapse directly from microarray data using supervised machine learning approaches. These findings are relevant to the use of molecular profiling for the identification of candidate biomarker panels.

An attempt for combining microarray data sets by adjusting gene expressions

PURPOSE

The diverse experimental environments in microarray technology, such as the different platforms or different RNA sources, can cause biases in the analysis of multiple microarrays. These systematic effects present a substantial obstacle for the analysis of microarray data, and the resulting information may be inconsistent and unreliable. Therefore, we introduced a simple integration method for combining microarray data sets that are derived from different experimental conditions, and we expected that more reliable information can be detected from the combined data set rather than from the separated data sets.

MATERIALS AND METHODS

This method is based on the distributions of the gene expression ratios among the different microarray data sets and it transforms, gene by gene, the gene expression ratios into the form of the reference data set. The efficiency of the proposed integration method was evaluated using two microarray data sets, which were derived from different RNA sources, and a newly defined measure, the mixture score.

RESULTS

The proposed integration method intermixed the two data sets that were obtained from different RNA sources, which in turn reduced the experimental bias between the two data sets, and the mixture score increased by 24.2%. A data set combined by the proposed method preserved the inter-group relationship of the separated data sets.

CONCLUSION

The proposed method worked well in adjusting systematic biases, including the source effect. The ability to use an effectively integrated microarray data set yields more reliable results due to the larger sample size and this also decreases the chance of false negatives.

Novel and simple transformation algorithm for combining microarray data sets

Background

With microarray technology, variability in experimental environments such as RNA sources, microarray production, or the use of different platforms, can cause bias. Such systematic differences present a substantial obstacle to the analysis of microarray data, resulting in inconsistent and unreliable information. Therefore, one of the most pressing challenges in the field of microarray technology is how to integrate results from different microarray experiments or combine data sets prior to the specific analysis.

Results

Two microarray data sets based on a 17k cDNA microarray system were used, consisting of 82 normal colon mucosa and 72 colorectal cancer tissues. Each data set was prepared from either total RNA or amplified mRNA, and the difference of RNA source between these two data sets was detected by ANOVA (Analysis of variance) model. A simple integration method was introduced which was based on the distributions of gene expression ratios among different microarray data sets. The method transformed gene expression ratios into the form of a reference data set on a gene by gene basis. Hierarchical clustering analysis, density and box plots, and mixture scores with correlation coefficients revealed that the two data sets were well intermingled, indicating that the proposed method minimized the experimental bias. In addition, any RNA source effect was not detected by the proposed transformation method. In the mixed data set, two previously identified subgroups of normal and tumor were well separated, and the efficiency of integration was more prominent in tumor groups than normal groups. The transformation method was slightly more effective when a data set with strong homogeneity in the same experimental group was used as a reference data set.

Conclusion

Proposed method is simple but useful to combine several data sets from different experimental conditions. With this method, biologically useful information can be detectable by applying various analytic methods to the combined data set with increased sample size.

Rapid screen of human genes for relevance to cancer using fission yeast

A total of 437 human full-length cDNAs isolated by microarray analysis of liver and/or gastric cancer tissues were evaluated for their relevance to cancer using the fission yeast Schizosaccharomyces pombe. Overexpression of 161 human cDNAs in S. pombe caused growth inhibition and/or morphological changes, which can be considered as cancer-related phenotypes of S. pombe. Sixteen genes causing growth defects and morphological changes at the same time were chosen to validate their ostensible oncogenic properties. They were highly expressed in liver and/or gastric cancer cell lines. Also, when the mouse embryonic fibroblast cell type NIH3T3 was transfected with these genes, the proliferation rates of cells were increased by 32% to 120%. This study demonstrates that fission yeast can be used as an advantageous and powerful tool for the rapid screening of human genes relevant to cancer. Furthermore, the human genes screened can be tested further as diagnostic markers and potential therapeutic targets for liver and stomach cancers. They also can be studied further for the elucidation of mechanisms involved in carcinogenesis.

Meta-analysis of several gene lists for distinct types of cancer: a simple way to reveal common prognostic markers

Background

Although prognostic biomarkers specific for particular cancers have been discovered, microarray analysis of gene expression profiles, supported by integrative analysis algorithms, helps to identify common factors in molecular oncology. Similarities of Ordered Gene Lists (SOGL) is a recently proposed approach to meta-analysis suitable for identifying features shared by two data sets. Here we extend the idea of SOGL to the detection of significant prognostic marker genes from microarrays of multiple data sets. Three data sets for leukemia and the other six for different solid tumors are used to demonstrate our method, using established statistical techniques.

Results

We describe a set of significantly similar ordered gene lists, representing outcome comparisons for distinct types of cancer. This kind of similarity could improve the diagnostic accuracies of individual studies when SOGL is incorporated into the support vector machine algorithm. In particular, we investigate the similarities among three ordered gene lists pertaining to mesothelioma survival, prostate recurrence and glioma survival. The similarity-driving genes are related to the outcomes of patients with lung cancer with a hazard ratio of 4.47 (p = 0.035). Many of these genes are involved in breakdown of EMC proteins regulating angiogenesis, and may be used for further research on prognostic markers and molecular targets of gene therapy for cancers.

Conclusion

The proposed method and its application show the potential of such meta-analyses in clinical studies of gene expression profiles.