SplicerAV: a tool for mining microarray expression data for changes in RNA processing.

BACKGROUND: Over the past two decades more than fifty thousand unique clinical and biological samples have been assayed using the Affymetrix HG-U133 and HG-U95 GeneChip microarray platforms. This substantial repository has been used extensively to characterize changes in gene expression between biological samples, but has not been previously mined en masse for changes in mRNA processing. We explored the possibility of using HG-U133 microarray data to identify changes in alternative mRNA processing in several available archival datasets. RESULTS: Data from these and other gene expression microarrays can now be mined for changes in transcript isoform abundance using a program described here, SplicerAV. Using in vivo and in vitro breast cancer microarray datasets, SplicerAV was able to perform both gene and isoform specific expression profiling within the same microarray dataset. Our reanalysis of Affymetrix U133 plus 2.0 data generated by in vitro over-expression of HRAS, E2F3, beta-catenin (CTNNB1), SRC, and MYC identified several hundred oncogene-induced mRNA isoform changes, one of which recognized a previously unknown mechanism of EGFR family activation. Using clinical data, SplicerAV predicted 241 isoform changes between low and high grade breast tumors; with changes enriched among genes coding for guanyl-nucleotide exchange factors, metalloprotease inhibitors, and mRNA processing factors. Isoform changes in 15 genes were associated with aggressive cancer across the three breast cancer datasets. CONCLUSIONS: Using SplicerAV, we identified several hundred previously uncharacterized isoform changes induced by in vitro oncogene over-expression and revealed a previously unknown mechanism of EGFR activation in human mammary epithelial cells. We analyzed Affymetrix GeneChip data from over 400 human breast tumors in three independent studies, making this the largest clinical dataset analyzed for en masse changes in alternative mRNA processing. The capacity to detect RNA isoform changes in archival microarray data using SplicerAV allowed us to carry out the first analysis of isoform specific mRNA changes directly associated with cancer survival.

Robust test method for time-course microarray experiments.

BACKGROUND: In a time-course microarray experiment, the expression level for each gene is observed across a number of time-points in order to characterize the temporal trajectories of the gene-expression profiles. For many of these experiments, the scientific aim is the identification of genes for which the trajectories depend on an experimental or phenotypic factor. There is an extensive recent body of literature on statistical methodology for addressing this analytical problem. Most of the existing methods are based on estimating the time-course trajectories using parametric or non-parametric mean regression methods. The sensitivity of these regression methods to outliers, an issue that is well documented in the statistical literature, should be of concern when analyzing microarray data. RESULTS: In this paper, we propose a robust testing method for identifying genes whose expression time profiles depend on a factor. Furthermore, we propose a multiple testing procedure to adjust for multiplicity. CONCLUSIONS: Through an extensive simulation study, we will illustrate the performance of our method. Finally, we will report the results from applying our method to a case study and discussing potential extensions.

Expression signatures of TP53 mutations in serous ovarian cancers.

BACKGROUND: Mutations in the TP53 gene are extremely common and occur very early in the progression of serous ovarian cancers. Gene expression patterns that relate to mutational status may provide insight into the etiology and biology of the disease. METHODS: The TP53 coding region was sequenced in 89 frozen serous ovarian cancers, 40 early stage (I/II) and 49 advanced stage (III/IV). Affymetrix U133A expression data was used to define gene expression patterns by mutation, type of mutation, and cancer stage. RESULTS: Missense or chain terminating (null) mutations in TP53 were found in 59/89 (66%) ovarian cancers. Early stage cancers had a significantly higher rate of null mutations than late stage disease (38% vs. 8%, p < 0.03). In advanced stage cases, mutations were more prevalent in short term survivors than long term survivors (81% vs. 30%, p = 0.0004). Gene expression patterns had a robust ability to predict TP53 status within training data. By using early versus late stage disease for out of sample predictions, the signature derived from early stage cancers could accurately (86%) predict mutation status of late stage cancers. CONCLUSIONS: This represents the first attempt to define a genomic signature of TP53 mutation in ovarian cancer. Patterns of gene expression characteristic of TP53 mutation could be discerned and included several genes that are known p53 targets or have been described in the context of expression signatures of TP53 mutation in breast cancer.

Expression in aneuploid Drosophila S2 cells.

Extensive departures from balanced gene dose in aneuploids are highly deleterious. However, we know very little about the relationship between gene copy number and expression in aneuploid cells. We determined copy number and transcript abundance (expression) genome-wide in Drosophila S2 cells by DNA-Seq and RNA-Seq. We found that S2 cells are aneuploid for >43 Mb of the genome, primarily in the range of one to five copies, and show a male genotype ( approximately two X chromosomes and four sets of autosomes, or 2X;4A). Both X chromosomes and autosomes showed expression dosage compensation. X chromosome expression was elevated in a fixed-fold manner regardless of actual gene dose. In engineering terms, the system "anticipates" the perturbation caused by X dose, rather than responding to an error caused by the perturbation. This feed-forward regulation resulted in precise dosage compensation only when X dose was half of the autosome dose. Insufficient compensation occurred at lower X chromosome dose and excessive expression occurred at higher doses. RNAi knockdown of the Male Specific Lethal complex abolished feed-forward regulation. Both autosome and X chromosome genes show Male Specific Lethal-independent compensation that fits a first order dose-response curve. Our data indicate that expression dosage compensation dampens the effect of altered DNA copy number genome-wide. For the X chromosome, compensation includes fixed and dose-dependent components.

The characterization of twenty sequenced human genomes.

We present the analysis of twenty human genomes to evaluate the prospects for identifying rare functional variants that contribute to a phenotype of interest. We sequenced at high coverage ten "case" genomes from individuals with severe hemophilia A and ten "control" genomes. We summarize the number of genetic variants emerging from a study of this magnitude, and provide a proof of concept for the identification of rare and highly-penetrant functional variants by confirming that the cause of hemophilia A is easily recognizable in this data set. We also show that the number of novel single nucleotide variants (SNVs) discovered per genome seems to stabilize at about 144,000 new variants per genome, after the first 15 individuals have been sequenced. Finally, we find that, on average, each genome carries 165 homozygous protein-truncating or stop loss variants in genes representing a diverse set of pathways.

Comparison of pattern detection methods in microarray time series of the segmentation clock.

While genome-wide gene expression data are generated at an increasing rate, the repertoire of approaches for pattern discovery in these data is still limited. Identifying subtle patterns of interest in large amounts of data (tens of thousands of profiles) associated with a certain level of noise remains a challenge. A microarray time series was recently generated to study the transcriptional program of the mouse segmentation clock, a biological oscillator associated with the periodic formation of the segments of the body axis. A method related to Fourier analysis, the Lomb-Scargle periodogram, was used to detect periodic profiles in the dataset, leading to the identification of a novel set of cyclic genes associated with the segmentation clock. Here, we applied to the same microarray time series dataset four distinct mathematical methods to identify significant patterns in gene expression profiles. These methods are called: Phase consistency, Address reduction, Cyclohedron test and Stable persistence, and are based on different conceptual frameworks that are either hypothesis- or data-driven. Some of the methods, unlike Fourier transforms, are not dependent on the assumption of periodicity of the pattern of interest. Remarkably, these methods identified blindly the expression profiles of known cyclic genes as the most significant patterns in the dataset. Many candidate genes predicted by more than one approach appeared to be true positive cyclic genes and will be of particular interest for future research. In addition, these methods predicted novel candidate cyclic genes that were consistent with previous biological knowledge and experimental validation in mouse embryos. Our results demonstrate the utility of these novel pattern detection strategies, notably for detection of periodic profiles, and suggest that combining several distinct mathematical approaches to analyze microarray datasets is a valuable strategy for identifying genes that exhibit novel, interesting transcriptional patterns.

Diagnosis of partial body radiation exposure in mice using peripheral blood gene expression profiles.

In the event of a terrorist-mediated attack in the United States using radiological or improvised nuclear weapons, it is expected that hundreds of thousands of people could be exposed to life-threatening levels of ionizing radiation. We have recently shown that genome-wide expression analysis of the peripheral blood (PB) can generate gene expression profiles that can predict radiation exposure and distinguish the dose level of exposure following total body irradiation (TBI). However, in the event a radiation-mass casualty scenario, many victims will have heterogeneous exposure due to partial shielding and it is unknown whether PB gene expression profiles would be useful in predicting the status of partially irradiated individuals. Here, we identified gene expression profiles in the PB that were characteristic of anterior hemibody-, posterior hemibody- and single limb-irradiation at 0.5 Gy, 2 Gy and 10 Gy in C57Bl6 mice. These PB signatures predicted the radiation status of partially irradiated mice with a high level of accuracy (range 79-100%) compared to non-irradiated mice. Interestingly, PB signatures of partial body irradiation were poorly predictive of radiation status by site of injury (range 16-43%), suggesting that the PB molecular response to partial body irradiation was anatomic site specific. Importantly, PB gene signatures generated from TBI-treated mice failed completely to predict the radiation status of partially irradiated animals or non-irradiated controls. These data demonstrate that partial body irradiation, even to a single limb, generates a characteristic PB signature of radiation injury and thus may necessitate the use of multiple signatures, both partial body and total body, to accurately assess the status of an individual exposed to radiation.

Interaction of Cryptococcus neoformans Rim101 and protein kinase A regulates capsule.

Cryptococcus neoformans is a prevalent human fungal pathogen that must survive within various tissues in order to establish a human infection. We have identified the C. neoformans Rim101 transcription factor, a highly conserved pH-response regulator in many fungal species. The rim101 multiply sign in circle mutant strain displays growth defects similar to other fungal species in the presence of alkaline pH, increased salt concentrations, and iron limitation. However, the rim101 multiply sign in circle strain is also characterized by a striking defect in capsule, an important virulence-associated phenotype. This capsular defect is likely due to alterations in polysaccharide attachment to the cell surface, not in polysaccharide biosynthesis. In contrast to many other C. neoformans capsule-defective strains, the rim101 multiply sign in circle mutant is hypervirulent in animal models of cryptococcosis. Whereas Rim101 activation in other fungal species occurs through the conserved Rim pathway, we demonstrate that C. neoformans Rim101 is also activated by the cAMP/PKA pathway. We report here that C. neoformans uses PKA and the Rim pathway to regulate the localization, activation, and processing of the Rim101 transcription factor. We also demonstrate specific host-relevant activating conditions for Rim101 cleavage, showing that C. neoformans has co-opted conserved signaling pathways to respond to the specific niche within the infected host. These results establish a novel mechanism for Rim101 activation and the integration of two conserved signaling cascades in response to host environmental conditions.

Genome-wide mRNA expression correlates of viral control in CD4+ T-cells from HIV-1-infected individuals.

There is great interindividual variability in HIV-1 viral setpoint after seroconversion, some of which is known to be due to genetic differences among infected individuals. Here, our focus is on determining, genome-wide, the contribution of variable gene expression to viral control, and to relate it to genomic DNA polymorphism. RNA was extracted from purified CD4+ T-cells from 137 HIV-1 seroconverters, 16 elite controllers, and 3 healthy blood donors. Expression levels of more than 48,000 mRNA transcripts were assessed by the Human-6 v3 Expression BeadChips (Illumina). Genome-wide SNP data was generated from genomic DNA using the HumanHap550 Genotyping BeadChip (Illumina). We observed two distinct profiles with 260 genes differentially expressed depending on HIV-1 viral load. There was significant upregulation of expression of interferon stimulated genes with increasing viral load, including genes of the intrinsic antiretroviral defense. Upon successful antiretroviral treatment, the transcriptome profile of previously viremic individuals reverted to a pattern comparable to that of elite controllers and of uninfected individuals. Genome-wide evaluation of cis-acting SNPs identified genetic variants modulating expression of 190 genes. Those were compared to the genes whose expression was found associated with viral load: expression of one interferon stimulated gene, OAS1, was found to be regulated by a SNP (rs3177979, p = 4.9E-12); however, we could not detect an independent association of the SNP with viral setpoint. Thus, this study represents an attempt to integrate genome-wide SNP signals with genome-wide expression profiles in the search for biological correlates of HIV-1 control. It underscores the paradox of the association between increasing levels of viral load and greater expression of antiviral defense pathways. It also shows that elite controllers do not have a fully distinctive mRNA expression pattern in CD4+ T cells. Overall, changes in global RNA expression reflect responses to viral replication rather than a mechanism that might explain viral control.

Two genes on A/J chromosome 18 are associated with susceptibility to Staphylococcus aureus infection by combined microarray and QTL analyses.

Although it has recently been shown that A/J mice are highly susceptible to Staphylococcus aureus sepsis as compared to C57BL/6J, the specific genes responsible for this differential phenotype are unknown. Using chromosome substitution strains (CSS), we found that loci on chromosomes 8, 11, and 18 influence susceptibility to S. aureus sepsis in A/J mice. We then used two candidate gene selection strategies to identify genes on these three chromosomes associated with S. aureus susceptibility, and targeted genes identified by both gene selection strategies. First, we used whole genome transcription profiling to identify 191 (56 on chr. 8, 100 on chr. 11, and 35 on chr. 18) genes on our three chromosomes of interest that are differentially expressed between S. aureus-infected A/J and C57BL/6J. Second, we identified two significant quantitative trait loci (QTL) for survival post-infection on chr. 18 using N(2) backcross mice (F(1) [C18A]xC57BL/6J). Ten genes on chr. 18 (March3, Cep120, Chmp1b, Dcp2, Dtwd2, Isoc1, Lman1, Spire1, Tnfaip8, and Seh1l) mapped to the two significant QTL regions and were also identified by the expression array selection strategy. Using real-time PCR, 6 of these 10 genes (Chmp1b, Dtwd2, Isoc1, Lman1, Tnfaip8, and Seh1l) showed significantly different expression levels between S. aureus-infected A/J and C57BL/6J. For two (Tnfaip8 and Seh1l) of these 6 genes, siRNA-mediated knockdown of gene expression in S. aureus-challenged RAW264.7 macrophages induced significant changes in the cytokine response (IL-1 beta and GM-CSF) compared to negative controls. These cytokine response changes were consistent with those seen in S. aureus-challenged peritoneal macrophages from CSS 18 mice (which contain A/J chromosome 18 but are otherwise C57BL/6J), but not C57BL/6J mice. These findings suggest that two genes, Tnfaip8 and Seh1l, may contribute to susceptibility to S. aureus in A/J mice, and represent promising candidates for human genetic susceptibility studies.

Use of 16S ribosomal RNA gene analyses to characterize the bacterial signature associated with poor oral health in West Virginia.

BACKGROUND: West Virginia has the worst oral health in the United States, but the reasons for this are unclear. This pilot study explored the etiology of this disparity using culture-independent analyses to identify bacterial species associated with oral disease. METHODS: Bacteria in subgingival plaque samples from twelve participants in two independent West Virginia dental-related studies were characterized using 16S rRNA gene sequencing and Human Oral Microbe Identification Microarray (HOMIM) analysis. Unifrac analysis was used to characterize phylogenetic differences between bacterial communities obtained from plaque of participants with low or high oral disease, which was further evaluated using clustering and Principal Coordinate Analysis. RESULTS: Statistically different bacterial signatures (P<0.001) were identified in subgingival plaque of individuals with low or high oral disease in West Virginia based on 16S rRNA gene sequencing. Low disease contained a high frequency of Veillonella and Streptococcus, with a moderate number of Capnocytophaga. High disease exhibited substantially increased bacterial diversity and included a large proportion of Clostridiales cluster bacteria (Selenomonas, Eubacterium, Dialister). Phylogenetic trees constructed using 16S rRNA gene sequencing revealed that Clostridiales were repeated colonizers in plaque associated with high oral disease, providing evidence that the oral environment is somehow influencing the bacterial signature linked to disease. CONCLUSIONS: Culture-independent analyses identified an atypical bacterial signature associated with high oral disease in West Virginians and provided evidence that the oral environment influenced this signature. Both findings provide insight into the etiology of the oral disparity in West Virginia.

Differential Apaf-1 levels allow cytochrome c to induce apoptosis in brain tumors but not in normal neural tissues.

Brain tumors are typically resistant to conventional chemotherapeutics, most of which initiate apoptosis upstream of mitochondrial cytochrome c release. In this study, we demonstrate that directly activating apoptosis downstream of the mitochondria, with cytosolic cytochrome c, kills brain tumor cells but not normal brain tissue. Specifically, cytosolic cytochrome c is sufficient to induce apoptosis in glioblastoma and medulloblastoma cell lines. In contrast, primary neurons from the cerebellum and cortex are remarkably resistant to cytosolic cytochrome c. Importantly, tumor tissue from mouse models of both high-grade astrocytoma and medulloblastoma display hypersensitivity to cytochrome c when compared with surrounding brain tissue. This differential sensitivity to cytochrome c is attributed to high Apaf-1 levels in the tumor tissue compared with low Apaf-1 levels in the adjacent brain tissue. These differences in Apaf-1 abundance correlate with differences in the levels of E2F1, a previously identified activator of Apaf-1 transcription. ChIP assays reveal that E2F1 binds the Apaf-1 promoter specifically in tumor tissue, suggesting that E2F1 contributes to the expression of Apaf-1 in brain tumors. Together, these results demonstrate an unexpected sensitivity of brain tumors to postmitochondrial induction of apoptosis. Moreover, they raise the possibility that this phenomenon could be exploited therapeutically to selectively kill brain cancer cells while sparing the surrounding brain parenchyma.

Transcription factors MYOCD, SRF, Mesp1 and SMARCD3 enhance the cardio-inducing effect of GATA4, TBX5, and MEF2C during direct cellular reprogramming.

Transient overexpression of defined combinations of master regulator genes can effectively induce cellular reprogramming: the acquisition of an alternative predicted phenotype from a differentiated cell lineage. This can be of particular importance in cardiac regenerative medicine wherein the heart lacks the capacity to heal itself, but simultaneously contains a large pool of fibroblasts. In this study we determined the cardio-inducing capacity of ten transcription factors to actuate cellular reprogramming of mouse embryonic fibroblasts into cardiomyocyte-like cells. Overexpression of transcription factors MYOCD and SRF alone or in conjunction with Mesp1 and SMARCD3 enhanced the basal but necessary cardio-inducing effect of the previously reported GATA4, TBX5, and MEF2C. In particular, combinations of five or seven transcription factors enhanced the activation of cardiac reporter vectors, and induced an upregulation of cardiac-specific genes. Global gene expression analysis also demonstrated a significantly greater cardio-inducing effect when the transcription factors MYOCD and SRF were used. Detection of cross-striated cells was highly dependent on the cell culture conditions and was enhanced by the addition of valproic acid and JAK inhibitor. Although we detected Ca(2+) transient oscillations in the reprogrammed cells, we did not detect significant changes in resting membrane potential or spontaneously contracting cells. This study further elucidates the cardio-inducing effect of the transcriptional networks involved in cardiac cellular reprogramming, contributing to the ongoing rational design of a robust protocol required for cardiac regenerative therapies.

A host transcriptional signature for presymptomatic detection of infection in humans exposed to influenza H1N1 or H3N2.

There is great potential for host-based gene expression analysis to impact the early diagnosis of infectious diseases. In particular, the influenza pandemic of 2009 highlighted the challenges and limitations of traditional pathogen-based testing for suspected upper respiratory viral infection. We inoculated human volunteers with either influenza A (A/Brisbane/59/2007 (H1N1) or A/Wisconsin/67/2005 (H3N2)), and assayed the peripheral blood transcriptome every 8 hours for 7 days. Of 41 inoculated volunteers, 18 (44%) developed symptomatic infection. Using unbiased sparse latent factor regression analysis, we generated a gene signature (or factor) for symptomatic influenza capable of detecting 94% of infected cases. This gene signature is detectable as early as 29 hours post-exposure and achieves maximal accuracy on average 43 hours (p = 0.003, H1N1) and 38 hours (p-value = 0.005, H3N2) before peak clinical symptoms. In order to test the relevance of these findings in naturally acquired disease, a composite influenza A signature built from these challenge studies was applied to Emergency Department patients where it discriminates between swine-origin influenza A/H1N1 (2009) infected and non-infected individuals with 92% accuracy. The host genomic response to Influenza infection is robust and may provide the means for detection before typical clinical symptoms are apparent.

Gene selection using iterative feature elimination random forests for survival outcomes.

Although many feature selection methods for classification have been developed, there is a need to identify genes in high-dimensional data with censored survival outcomes. Traditional methods for gene selection in classification problems have several drawbacks. First, the majority of the gene selection approaches for classification are single-gene based. Second, many of the gene selection procedures are not embedded within the algorithm itself. The technique of random forests has been found to perform well in high-dimensional data settings with survival outcomes. It also has an embedded feature to identify variables of importance. Therefore, it is an ideal candidate for gene selection in high-dimensional data with survival outcomes. In this paper, we develop a novel method based on the random forests to identify a set of prognostic genes. We compare our method with several machine learning methods and various node split criteria using several real data sets. Our method performed well in both simulations and real data analysis.Additionally, we have shown the advantages of our approach over single-gene-based approaches. Our method incorporates multivariate correlations in microarray data for survival outcomes. The described method allows us to better utilize the information available from microarray data with survival outcomes.