Exon Array Dataset Research Articles

Platform-Independent Classification System to Predict Molecular Subtypes of High-Grade Serous Ovarian Carcinoma.

PURPOSEMolecular cancer subtyping is an important tool in predicting prognosis and developing novel precision medicine approaches. We developed a novel platform-independent gene expression–based classification system for molecular subtyping of patients with high-grade serous ovarian carcinoma (HGSOC).METHODSUnprocessed exon array (569 tumor and nine normal) and RNA sequencing (RNA-seq; 376 tumor) HGSOC data sets, with clinical annotations, were downloaded from the Genomic Data Commons portal. Sample clustering was performed by non-negative matrix factorization by using isoform-level expression estimates. The association between the subtypes and overall survival was evaluated by Cox proportional hazards regression model after adjusting for the covariates. A novel classification system was developed for HGSOC molecular subtyping. Robustness and generalizability of the gene signatures were validated using independent microarray and RNA-seq data sets.RESULTSSample clustering recaptured the four known The Cancer Genome Atlas molecular subtypes but switched the subtype for 22% of the cases, which resulted in significant (P = .006) survival differences among the refined subgroups. After adjusting for covariate effects, the mesenchymal subgroup was found to be at an increased hazard for death compared with the immunoreactive subgroup. Both gene- and isoform-level signatures achieved more than 92% prediction accuracy when tested on independent samples profiled on the exon array platform. When the classifier was applied to RNA-seq data, the subtyping calls agreed with the predictions made from exon array data for 95% of the 279 samples profiled by both platforms.CONCLUSIONIsoform-level expression analysis successfully stratifies patients with HGSOC into groups with differing prognosis and has led to the development of robust, platform-independent gene signatures for HGSOC molecular subtyping. The association of the refined The Cancer Genome Atlas HGSOC subtypes with overall survival, independent of covariates, enhances the clinical annotation of the HGSOC cohort.

Simultaneous dimension reduction and adjustment for confounding variation

Dimension reduction methods are commonly applied to high-throughput biological datasets. However, the results can be hindered by confounding factors, either biological or technical in origin. In this study, we extend principal component analysis (PCA) to propose AC-PCA for simultaneous dimension reduction and adjustment for confounding (AC) variation. We show that AC-PCA can adjust for (i) variations across individual donors present in a human brain exon array dataset and (ii) variations of different species in a model organism ENCODE RNA sequencing dataset. Our approach is able to recover the anatomical structure of neocortical regions and to capture the shared variation among species during embryonic development. For gene selection purposes, we extend AC-PCA with sparsity constraints and propose and implement an efficient algorithm. The methods developed in this paper can also be applied to more general settings. The R package and MATLAB source code are available at https://github.com/linzx06/AC-PCA.

Meta-analysis of prostate cancer gene expression data identifies a novel discriminatory signature enriched for glycosylating enzymes.

BackgroundTumorigenesis is characterised by changes in transcriptional control. Extensive transcript expression data have been acquired over the last decade and used to classify prostate cancers. Prostate cancer is, however, a heterogeneous multifocal cancer and this poses challenges in identifying robust transcript biomarkers.MethodsIn this study, we have undertaken a meta-analysis of publicly available transcriptomic data spanning datasets and technologies from the last decade and encompassing laser capture microdissected and macrodissected sample sets.ResultsWe identified a 33 gene signature that can discriminate between benign tissue controls and localised prostate cancers irrespective of detection platform or dissection status. These genes were significantly overexpressed in localised prostate cancer versus benign tissue in at least three datasets within the Oncomine Compendium of Expression Array Data. In addition, they were also overexpressed in a recent exon-array dataset as well a prostate cancer RNA-seq dataset generated as part of the The Cancer Genomics Atlas (TCGA) initiative. Biologically, glycosylation was the single enriched process associated with this 33 gene signature, encompassing four glycosylating enzymes. We went on to evaluate the performance of this signature against three individual markers of prostate cancer, v-ets avian erythroblastosis virus E26 oncogene homolog (ERG) expression, prostate specific antigen (PSA) expression and androgen receptor (AR) expression in an additional independent dataset. Our signature had greater discriminatory power than these markers both for localised cancer and metastatic disease relative to benign tissue, or in the case of metastasis, also localised prostate cancer.ConclusionIn conclusion, robust transcript biomarkers are present within datasets assembled over many years and cohorts and our study provides both examples and a strategy for refining and comparing datasets to obtain additional markers as more data are generated.Electronic supplementary materialThe online version of this article (doi:10.1186/s12920-014-0074-9) contains supplementary material, which is available to authorized users.

Endothelial, epithelial, and fibroblast cells exhibit specific splicing programs independently of their tissue of origin

Alternative splicing is the main mechanism of increasing the proteome diversity coded by a limited number of genes. It is well established that different tissues or organs express different splicing variants. However, organs are composed of common major cell types, including fibroblasts, epithelial, and endothelial cells. By analyzing large-scale data sets generated by The ENCODE Project Consortium and after extensive RT-PCR validation, we demonstrate that each of the three major cell types expresses a specific splicing program independently of its organ origin. Furthermore, by analyzing splicing factor expression across samples, publicly available splicing factor binding site data sets (CLIP-seq), and exon array data sets after splicing factor depletion, we identified several splicing factors, including ESRP1 and 2, MBNL1, NOVA1, PTBP1, and RBFOX2, that contribute to establishing these cell type–specific splicing programs. All of the analyzed data sets are freely available in a user-friendly web interface named FasterDB, which describes all known splicing variants of human and mouse genes and their splicing patterns across several dozens of normal and cancer cells as well as across tissues. Information regarding splicing factors that potentially contribute to individual exon regulation is also provided via a dedicated CLIP-seq and exon array data visualization interface. To the best of our knowledge, FasterDB is the first database integrating such a variety of large-scale data sets to enable functional genomics analyses at exon-level resolution.

Tissue-specific isoform switch and DNA hypomethylation of the pyruvate kinase PKM gene in human cancers

The M2 isoform of pyruvate kinase (PKM2) plays an important role in aerobic glycolysis and is a mediator of the Warburg effect in tumors. It was previously thought that tumor cells switch expression of PKM from normal tissue-expressed PKM1 to tumor-specific PKM2 via an alternative splicing mechanism. This view was challenged by a recent report demonstrating that PKM2 is already the major PKM isoform expressed in many differentiated normal tissues. Here, through analyses on sixteen tumor types using the cancer genome atlas RNA-Seq and exon array datasets, we confirmed that isoform switch from PKM1 to PKM2 occurred in glioblastomas but not in other tumor types examined. Despite lacking of isoform switches, PKM2 expression was found to be increased in all cancer types examined, and correlated strongly to poor prognosis in head and neck cancers. We further demonstrated that elevated PKM2 expression correlated well with the hypomethylation status of intron 1 of the PKM gene in multiple cancer types, suggesting epigenetic regulation by DNA methylation as a major mechanism in controlling PKM transcription in tumors. Our study suggests that isoform switch of PKM1 to PKM2 in cancers is tissue-specific and targeting PKM2 activity in tumors remains a promising approach for clinical intervention of multiple cancer types.

Deep brain stimulation modulates nonsense-mediated RNA decay in Parkinson’s patients leukocytes

BackgroundNonsense-Mediated decay (NMD) selectively degrades mRNA transcripts that carry premature stop codons. NMD is often triggered by alternative splicing (AS) modifications introducing such codons. NMD plays an important regulatory role in brain neurons, but the in vivo dynamics of AS and NMD changes in neurological diseases and under treatment were scarcely explored.ResultsHere, we report exon arrays analysis of leukocyte mRNA AS events prior to and following Deep Brain Stimulation (DBS) neurosurgery, which efficiently improves the motor symptoms of Parkinson’s disease (PD), the leading movement disorder, and is increasingly applied to treat other diseases. We also analyzed publicly available exon array dataset of whole blood cells from mixed early and advanced PD patients. Our in-house exon array dataset of leukocyte transcripts was derived from advanced PD patients’ pre- and post-DBS stimulation and matched healthy control volunteers. The mixed cohort exhibited 146 AS changes in 136 transcripts compared to controls, including 9 NMD protein-level assessed events. In comparison, PD patients from our advanced cohort differed from healthy controls by 319 AS events in 280 transcripts, assessed as inducing 27 protein-level NMD events. DBS stimulation induced 254 AS events in 229 genes as compared to the pre-DBS state including 44 NMD inductions. A short, one hour electrical stimulus cessation caused 234 AS changes in 125 genes compared to ON-stimulus state, 22 of these were assessed for NMD. Functional analysis highlighted disease-induced DNA damage and inflammatory control and its reversal under ON and OFF stimulus as well as alternative splicing in all the tested states.ConclusionsThe study findings indicate a potential role for NMD both in PD and following electrical brain stimulation. Furthermore, our current observations entail future implications for developing therapies for PD, and for interfering with the impaired molecular mechanisms that underlie PD and other neurodegenerative and neurological disorders, as well as DBS-treatable conditions in general.

Isoform level expression profiles provide better cancer signatures than gene level expression profiles

BackgroundThe majority of mammalian genes generate multiple transcript variants and protein isoforms through alternative transcription and/or alternative splicing, and the dynamic changes at the transcript/isoform level between non-oncogenic and cancer cells remain largely unexplored. We hypothesized that isoform level expression profiles would be better than gene level expression profiles at discriminating between non-oncogenic and cancer cellsgene level.MethodsWe analyzed 160 Affymetrix exon-array datasets, comprising cell lines of non-oncogenic or oncogenic tissue origins. We obtained the transcript-level and gene level expression estimates, and used unsupervised and supervised clustering algorithms to study the profile similarity between the samples at both gene and isoform levels.ResultsHierarchical clustering, based on isoform level expressions, effectively grouped the non-oncogenic and oncogenic cell lines with a virtually perfect homogeneity-grouping rate (97.5%), regardless of the tissue origin of the cell lines. However, gene levelthis rate was much lower, being 75% at best based on the gene level expressions. Statistical analyses of the difference between cancer and non-oncogenic samples identified the existence of numerous genes with differentially expressed isoforms, which otherwise were not significant at the gene level. We also found that canonical pathways of protein ubiquitination, purine metabolism, and breast-cancer regulation by stathmin1 were significantly enriched among genes thatshow differential expression at isoform level but not at gene level.ConclusionsIn summary, cancer cell lines, regardless of their tissue of origin, can be effectively discriminated from non-cancer cell lines at isoform level, but not at gene level. This study suggests the existence of an isoform signature, rather than a gene signature, which could be used to distinguish cancer cells from normal cells.

High serum vascular endothelial growth factor level is an adverse prognostic factor for high‐risk diffuse large B‐cell lymphoma patients treated with dose‐dense chemoimmunotherapy

To determine whether serum vascular endothelial growth factor (s-VEGF) levels and VEGF gene expression in tumor tissue predict survival of diffuse large B-cell lymphoma (DLBCL) patients treated with chemoimmunotherapy. VEGF levels were measured in serum samples from 102 patients <65 yrs with high-risk DLBCL using a quantitative sandwich enzyme immunoassay technique. Exon array data set of tumor tissues from 32 patients was concurrently used to determine VEGF-A exon and gene expression. All patients were treated in a Nordic phase II study with six dose-dense chemoimmunotherapy courses followed by systemic central nervous system prophylaxis. After a median follow-up time of 40 months, 3-yr progression-free survival (PFS) was inferior in patients with high s-VEGF levels compared to those with low levels (59% vs. 83%, P = 0.005). The relative risk of progression or relapse was 3.1-fold (95% confidence interval 1.34-6.91, P = 0.008). The predictive capacity of s-VEGF levels on PFS was most pronounced in the DLBCLs of non-germinal center subtype. In contrast to serum data, VEGF mRNA expression in the lymphoma tissue did not predict outcome, and no correlation was found between s-VEGF levels and lymphoma VEGF expression. Pretreatment s-VEGF level is a predictor of PFS after chemoimmunotherapy and may help to further stratify high-risk DLBCL patients into low- and high-risk groups.

BEAT: Bioinformatics Exon Array Tool to store, analyze and visualize Affymetrix GeneChip Human Exon Array data from disease experiments.

BackgroundIt is known from recent studies that more than 90% of human multi-exon genes are subject to Alternative Splicing (AS), a key molecular mechanism in which multiple transcripts may be generated from a single gene. It is widely recognized that a breakdown in AS mechanisms plays an important role in cellular differentiation and pathologies. Polymerase Chain Reactions, microarrays and sequencing technologies have been applied to the study of transcript diversity arising from alternative expression. Last generation Affymetrix GeneChip Human Exon 1.0 ST Arrays offer a more detailed view of the gene expression profile providing information on the AS patterns. The exon array technology, with more than five million data points, can detect approximately one million exons, and it allows performing analyses at both gene and exon level. In this paper we describe BEAT, an integrated user-friendly bioinformatics framework to store, analyze and visualize exon arrays datasets. It combines a data warehouse approach with some rigorous statistical methods for assessing the AS of genes involved in diseases. Meta statistics are proposed as a novel approach to explore the analysis results. BEAT is available at http://beat.ba.itb.cnr.it.ResultsBEAT is a web tool which allows uploading and analyzing exon array datasets using standard statistical methods and an easy-to-use graphical web front-end. BEAT has been tested on a dataset with 173 samples and tuned using new datasets of exon array experiments from 28 colorectal cancer and 26 renal cell cancer samples produced at the Medical Genetics Unit of IRCCS Casa Sollievo della Sofferenza.To highlight all possible AS events, alternative names, accession Ids, Gene Ontology terms and biochemical pathways annotations are integrated with exon and gene level expression plots. The user can customize the results choosing custom thresholds for the statistical parameters and exploiting the available clinical data of the samples for a multivariate AS analysis.ConclusionsDespite exon array chips being widely used for transcriptomics studies, there is a lack of analysis tools offering advanced statistical features and requiring no programming knowledge. BEAT provides a user-friendly platform for a comprehensive study of AS events in human diseases, displaying the analysis results with easily interpretable and interactive tables and graphics.

Tumor-specific usage of alternative transcription start sites in colorectal cancer identified by genome-wide exon array analysis

BackgroundApproximately half of all human genes use alternative transcription start sites (TSSs) to control mRNA levels and broaden the transcriptional output in healthy tissues. Aberrant expression patterns promoting carcinogenesis, however, may arise from alternative promoter usage.ResultsBy profiling 108 colorectal samples using exon arrays, we identified nine genes (TCF12, OSBPL1A, TRAK1, ANK3, CHEK1, UGP2, LMO7, ACSL5, and SCIN) showing tumor-specific alternative TSS usage in both adenoma and cancer samples relative to normal mucosa. Analysis of independent exon array data sets corroborated these findings. Additionally, we confirmed the observed patterns for selected mRNAs using quantitative real-time reverse-transcription PCR. Interestingly, for some of the genes, the tumor-specific TSS usage was not restricted to colorectal cancer. A comprehensive survey of the nine genes in lung, bladder, liver, prostate, gastric, and brain cancer revealed significantly altered mRNA isoform ratios for CHEK1, OSBPL1A, and TCF12 in a subset of these cancer types.To identify the mechanism responsible for the shift in alternative TSS usage, we antagonized the Wnt-signaling pathway in DLD1 and Ls174T colorectal cancer cell lines, which remarkably led to a shift in the preferred TSS for both OSBPL1A and TRAK1. This indicated a regulatory role of the Wnt pathway in selecting TSS, possibly also involving TP53 and SOX9, as their transcription binding sites were enriched in the promoters of the tumor preferred isoforms together with their mRNA levels being increased in tumor samples.Finally, to evaluate the prognostic impact of the altered TSS usage, immunohistochemistry was used to show deregulation of the total protein levels of both TCF12 and OSBPL1A, corresponding to the mRNA levels observed. Furthermore, the level of nuclear TCF12 had a significant correlation to progression free survival in a cohort of 248 stage II colorectal cancer samples.ConclusionsAlternative TSS usage in colorectal adenoma and cancer samples has been shown for nine genes, and OSBPL1A and TRAK1 were found to be regulated in vitro by Wnt signaling. TCF12 protein expression was upregulated in cancer samples and correlated with progression free survival.

Abstract 31: Isoform level gene expression signature is a better discriminator of cancer from normal cells than gene level expression signature

Abstract The majority of mammalian genes generate multiple transcript variants and protein isoforms, which usually are differentially expressed in different cellular and disease conditions. We hypothesize that the isoform level gene expression profiles can better discriminate normal and cancer cells with different sub-types than gene-level expression profiles. We analyzed Affymetrix exon array data sets, collected from public domain (ENCODE and TCGA datasets from NCBI GEO database), comprising 160 cell lines of either normal or tumor tissues of origin. The unsupervised clustering of genome-wide gene expression estimates at isoform-level effectively grouped the normal and cancer cell lines with 97.5% homogeneity, regardless of the tissue origins of the cell lines, with only 4 exceptions out of 160 samples. However, at gene level, this homogeneity grouping rate was much lower, and was about 75% at best. The supervised method revealed genes and isoforms that are differentially expressed between pairs of tumor and normal samples, which can serve as signatures to effectively differentiate tumor and normal cell lines. Finally, the pathway analyses both at gene and isoforms-level indicate that there are subtle differences of derived pathways. The study reveals that there exists a common isoform-level signature that can be used not only for effective discrimination of oncogenic cell lines from non-oncogenic cell lines, but may also have important biological implications for a better understanding of common defining characteristics of tumors of varied types. Finally, we present the identified isoform-level gene signatures and associated pathways for melanoma, breast cancer and glioblastoma. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 31. doi:10.1158/1538-7445.AM2011-31

Alternative Splicing of SLC39A14 in Colorectal Cancer is Regulated by the Wnt Pathway

Alternative splicing is a crucial step in the generation of protein diversity and its misregulation is observed in many human cancer types. By analyzing 143 colorectal samples using exon arrays, SLC39A14, a divalent cation transporter, was identified as being aberrantly spliced in tumor samples. SLC39A14 contains two mutually exclusive exons 4A and 4B and the exon 4A/4B ratio was significantly altered in adenomas (p = 3.6 × 10(-10)) and cancers (p = 9.4 × 10(-11)), independent of microsatellite stability status. The findings were validated in independent exon array data sets and by quantitative real-time reverse-transcription PCR (qRT-PCR). Aberrant Wnt signaling is a hallmark of colorectal tumorigenesis and is characterized by nuclear β-catenin. Experimental inactivation of Wnt signaling in DLD1 and Ls174T cells by knockdown of β-catenin or overexpression of dominant negative TCFs (TCF1 and TCF4) altered the 4A/4B ratio, indicating that SLC39A14 splicing is regulated by the Wnt pathway. An altered 4A/4B ratio was also observed in gastric and lung cancer where Wnt signaling is also known to be aberrantly activated. The splicing factor SRSF1 and its regulator, the kinase SRPK1, were found to be deregulated upon Wnt inactivation in colorectal carcinoma cells. SRPK1 was also found up-regulated in both adenoma samples (p = 1.5 × 10(-5)) and cancer samples (p = 5 × 10(-4)). In silico splicing factor binding analysis predicted SRSF1 to bind predominantly to the cancer associated exon 4B, hence, it was hypothesized that SRPK1 activates SRSF1 through phosphorylation, followed by SRSF1 binding to exon 4B and regulation of SLC39A14 splicing. Indeed, siRNA-mediated knockdown of SRPK1 and SRSF1 in DLD1 and SW480 colorectal cancer cells led to a change in the 4A/4B isoform ratio, supporting a role of these factors in the regulation of SLC39A14 splicing. In conclusion, alternative splicing of SLC39A14 was identified in colorectal tumors and found to be regulated by the Wnt pathway, most likely through regulation of SRPK1 and SRSF1.

Exon Array Analysis using re-defined probe sets results in reliable identification of alternatively spliced genes in non-small cell lung cancer

BackgroundTreatment of non-small cell lung cancer with novel targeted therapies is a major unmet clinical need. Alternative splicing is a mechanism which generates diverse protein products and is of functional relevance in cancer.ResultsIn this study, a genome-wide analysis of the alteration of splicing patterns between lung cancer and normal lung tissue was performed. We generated an exon array data set derived from matched pairs of lung cancer and normal lung tissue including both the adenocarcinoma and the squamous cell carcinoma subtypes. An enhanced workflow was developed to reliably detect differential splicing in an exon array data set. In total, 330 genes were found to be differentially spliced in non-small cell lung cancer compared to normal lung tissue. Microarray findings were validated with independent laboratory methods for CLSTN1, FN1, KIAA1217, MYO18A, NCOR2, NUMB, SLK, SYNE2, TPM1, (in total, 10 events) and ADD3, which was analysed in depth. We achieved a high validation rate of 69%. Evidence was found that the activity of FOX2, the splicing factor shown to cause cancer-specific splicing patterns in breast and ovarian cancer, is not altered at the transcript level in several cancer types including lung cancer.ConclusionsThis study demonstrates how alternatively spliced genes can reliably be identified in a cancer data set. Our findings underline that key processes of cancer progression in NSCLC are affected by alternative splicing, which can be exploited in the search for novel targeted therapies.

Genome-wide prediction of cis-acting RNA elements regulating tissue-specific pre-mRNA alternative splicing

BackgroundHuman genes undergo various patterns of pre-mRNA splicing across different tissues. Such variation is primarily regulated by trans-acting factors that bind on exonic and intronic cis-acting RNA elements (CAEs). Here we report a computational method to mechanistically identify cis-acting RNA elements that contribute to the tissue-specific alternative splicing pattern. This method is an extension of our previous model, SplicingModeler, which predicts the significant CAEs that contribute to the splicing differences between two tissues. In this study, we introduce tissue-specific functional levels estimation step, which allows evaluating regulatory functions of predicted CAEs that are involved in more than two tissues.ResultsUsing a publicly available Affymetrix Genechip® Human Exon Array dataset, our method identifies 652 cis-acting RNA elements (CAEs) across 11 human tissues. About one third of predicted CAEs can be mapped to the known RBP (RNA binding protein) binding sites or match with other predicted exonic splicing regulator databases. Interestingly, the vast majority of predicted CAEs are in intronic regulatory regions. A noticeable exception is that many exonic elements are found to regulate the alternative splicing between cerebellum and testes. Most identified elements are found to contribute to the alternative splicing between two tissues, while some are important in multiple tissues. This suggests that genome-wide alternative splicing patterns are regulated by a combination of tissue-specific cis-acting elements and "general elements" whose functional activities are important but differ across multiple tissues.ConclusionIn this study, we present a model-based computational approach to identify potential cis-acting RNA elements by considering the exon splicing variation as the combinatorial effects of multiple cis-acting regulators. This methodology provides a novel evaluation on the functional levels of cis-acting RNA elements by estimating their tissue-specific functions on various tissues.

Probe Selection and Expression Index Computation of Affymetrix Exon Arrays

BackgroundThere is great current interest in developing microarray platforms for measuring mRNA abundance at both gene level and exon level. The Affymetrix Exon Array is a new high-density gene expression microarray platform, with over six million probes targeting all annotated and predicted exons in a genome. An important question for the analysis of exon array data is how to compute overall gene expression indexes. Because of the complexity of the design of exon array probes, this problem is different in nature from summarizing gene-level expression from traditional 3′ expression arrays.Methodology/Principal FindingsIn this manuscript, we use exon array data from 11 human tissues to study methods for computing gene-level expression. We showed that for most genes there is a subset of exon array probes having highly correlated intensities across multiple samples. We suggest that these probes could be used as reliable indicators of overall gene expression levels. We developed a probe selection algorithm to select such a subset of highly correlated probes for each gene, and computed gene expression indexes using the selected probes.Conclusions/SignificanceOur results demonstrate that probe selection improves gene expression estimates from exon arrays. The selected probes can be used in future analyses of other exon array datasets to compute gene expression indexes.