Assessment of reliability of microarray data and estimation of signal thresholds using mixture modeling

Automatic segmentation of corpus callosum using Gaussian mixture modeling and Fuzzy C means methods

This paper presents a comparative study of the success and performance of the Gaussian mixture modeling and Fuzzy C means methods to determine the volume and cross-sectionals areas of the corpus callosum (CC) using simulated and real MR brain images. The Gaussian mixture model (GMM) utilizes weighted sum of Gaussian distributions by applying statistical decision procedures to define image classes. In the Fuzzy C means (FCM), the image classes are represented by certain membership function according to fuzziness information expressing the distance from the cluster centers. In this study, automatic segmentation for midsagittal section of the CC was achieved from simulated and real brain images. The volume of CC was obtained using sagittal sections areas. To compare the success of the methods, segmentation accuracy, Jaccard similarity and time consuming for segmentation were calculated. The results show that the GMM method resulted by a small margin in more accurate segmentation (midsagittal section segmentation accuracy 98.3% and 97.01% for GMM and FCM); however the FCM method resulted in faster segmentation than GMM. With this study, an accurate and automatic segmentation system that allows opportunity for quantitative comparison to doctors in the planning of treatment and the diagnosis of diseases affecting the size of the CC was developed. This study can be adapted to perform segmentation on other regions of the brain, thus, it can be operated as practical use in the clinic.

Microarray-based comparison of genetic differences between strains of Streptomyces turgidiscabies with focus on the pathogenicity island

The areas of the pathogenicity island (PAI) designated as 'colonization region' (CR) and 'toxicogenic region' (TR) [Lerat et al. (2009) Mol. Plant Pathol. 10, 579-585] contain genes required for virulence and phytoxin production, respectively, in Streptomyces spp. causing common scab on potatoes. The PAI was tested for genetic variability by microarray analysis in strains of S. turgidiscabies isolated from potatoes in Finland. The data revealed four types of PAI based on divergent CR and TR which occurred in different combinations. Only one PAI type was highly similar to S. scabies (strains 87.22 and ATTC49173). Using probes designed for the predicted genes of S. scabies, two gene clusters in S. scabies appeared to be similar to most strains of S. turgidiscabies and contained PAI genes corresponding to CR and TR. They were located approximately 5 Mb apart in the S. scabies genome, as compared with only 0.3 Mb in S. turgidiscabies Car8. Data from comparative genomic hybridization with probes designed for S. scabies genes and for the PAI of S. turgidiscabies were compared by multilocus cluster analysis, which revealed two strains of S. turgidiscabies that were very closely related at the whole-genome level, but contained distinctly different PAIs. The type strain of S. reticuliscabiei (DSM41804; synonymous to S. turgidiscabies) was clustered with S. turgidiscabies. Taken together, the data indicate wide genetic variability of PAIs among strains of S. turgidiscabies, and demonstrate that PAI is made up of a mosaic of regions which may undergo independent evolution.

Large-scale analysis of network bistability for human cancers

Protein-protein interaction and gene regulatory networks are likely to be locked in a state corresponding to a disease by the behavior of one or more bistable circuits exhibiting switch-like behavior. Sets of genes could be over-expressed or repressed when anomalies due to disease appear, and the circuits responsible for this over- or under-expression might persist for as long as the disease state continues. This paper shows how a large-scale analysis of network bistability for various human cancers can identify genes that can potentially serve as drug targets or diagnosis biomarkers.

The RNA-binding zinc-finger protein tristetraprolin regulates AU-rich mRNAs involved in breast cancer-related processes

Tristetraprolin (TTP or ZFP36) is a tandem CCCH zinc-finger RNA-binding protein that regulates the stability of certain AU-rich element (ARE) mRNAs. Recent work suggests that TTP is deficient in cancer cells when compared with normal cell types. In this study we found that TTP expression was lower in invasive breast cancer cells (MDAMB231) compared with normal breast cell lines MCF12A and MCF-10. TTP targets were probed using a novel approach by expressing the C124R zinc-finger TTP mutant that functions as dominant negative and increases target mRNA expression. In contrast to wild-type TTP, C124R TTP was able to increase certain ARE-mRNA expressions in serum-stimulated breast cancer cells. Using an ARE-gene microarray, novel targets of TTP regulation were identified, namely, urokinase plasminogen activator (uPA), uPA receptor and matrix metalloproteinase-1, all known to have prominent roles in breast cancer invasion and metastasis. Expression of these targets was upregulated in tumorigenic types, particularly in highly invasive MDAMB231. The mRNA half-lives of these TTP-regulated genes were increased in TTP-knockout embryonic mouse fibroblasts, as assessed using real-time polymerase chain reaction, whereas forced restoration of TTP by transfection led to a reduction in their mRNA levels. RNA immunoprecipitation confirmed an association of TTP, but not C124R, with these target transcripts. Moreover, TTP reduced, whereas the mutant C124R TTP increased, the activity of reporter constructs fused to target ARE. As a result of TTP regulation, invasiveness of MDAMB231 cells was reduced. The data suggest that TTP, in a 3' untranslated region-and ARE-dependent manner, regulates an important subset of cancer-related genes that are involved in cellular growth, invasion and metastasis.

Differential translocation of nuclear factor-kappaB in a cardiac muscle cell line under gravitational changes

Microgravity (micro-g) environments have been shown to elicit dysregulation of specific genes in a wide assay of cell types. It is known that the activation of transcription factors and molecular signaling pathways influence various physiological outcomes associated with stress and adaptive responses. Nuclear factor-kappa B (NF-kappaB) is one of the most prevailing oxidation-sensitive transcription factors. It is hypothesized that simulated microgravity would activate NF-kappaB and its downstream transcriptional networks, thus suggesting a role for NF-kappaB in microgravity induced muscle atrophy. To investigate the activation of NF-kappaB in a rat cardiac cell line (H9c2) under micro-g, rotating wall vessel bioreactors were used to simulate micro-g conditions. Western blotting revealed that mean nuclear translocation of NF-kappaB p65 subunit was 69% for micro-g and 46% for unit-g dynamic control as compared with a 30 min TNF-alpha positive control (p<0.05, n=3). The results from western blots were confirmed by enzyme-linked immunosorbent assay, which showed 66% for micro-g and 45% for dynamic control as compared with positive control (p<0.05, n=3). These results show significant differential translocation of NF-kappaB p65 under simulated micro-g. These results may be expanded upon to explain physiological changes such as muscle atrophy and further identify the regulatory pathways and effector molecules activated under exposure to micro-g.

Signal deconvolution based expression-detection and background adjustment for microarray data

Background adjustment is an essential stage in analyzing DNA microarrays. Discriminating expressed genes from unexpressed ones (expression detection), and estimating the expression levels of weakly expressed genes, critically depend on accurate treatment of the background intensity. Current methods for background adjustment either do not deal with nonspecific hybridization or strongly depend on the reliability of control probes. Existing model-based methods have limited accuracy. A new platform-independent background adjustment algorithm is presented. The algorithm relies on the deconvoluted experimental signal distribution for evaluating the expression probability and adjusting the background of each probe. Considering expression detection, it is shown, for two-channels cDNA arrays and for the Affymetrix GeneChip platform, that the algorithm performs at least as good or better than control-probes-based algorithms. For the Affymetrix GeneChip arrays, it is further shown that the algorithm outperforms the robust multiarray (RMA) expression measure in estimating genomewide expression levels.

Distinguishing bacterial pathogens of potato using a genome-wide microarray approach

A set of 9676 probes was designed for the most harmful bacterial pathogens of potato and tested in a microarray format. Gene-specific probes could be designed for all genes of Pectobacterium atrosepticum, c. 50% of the genes of Streptomyces scabies and c. 30% of the genes of Clavibacter michiganensis ssp. sepedonicus utilizing the whole-genome sequence information available. For Streptomyces turgidiscabies, 226 probes were designed according to the sequences of a pathogenicity island containing important virulence genes. In addition, probes were designed for the virulence-associated nip (necrosis-inducing protein) genes of P. atrosepticum, P. carotovorum and Dickeya dadantii and for the intergenic spacer (IGS) sequences of the 16S-23S rRNA gene region. Ralstonia solanacearum was not included in the study, because it is a quarantine organism and is not presently found in Finland, but a few probes were also designed for this species. The probes contained on average 40 target-specific nucleotides and were synthesized on the array in situ, organized as eight sub-arrays with an identical set of probes which could be used for hybridization with different samples. All bacteria were readily distinguished using a single channel system for signal detection. Nearly all of the c. 1000 probes designed for C. michiganensis ssp. sepedonicus, c. 50% and 40% of the c. 4000 probes designed for the genes of S. scabies and P. atrosepticum, respectively, and over 100 probes for S. turgidiscabies showed significant signals only with the respective species. P. atrosepticum, P. carotovorum and Dickeya strains were all detected with 110 common probes. By contrast, the strains of these species were found to differ in their signal profiles. Probes targeting the IGS region and nip genes could be used to place strains of Dickeya to two groups, which correlated with differences in virulence. Taken together, the approach of using a custom-designed, genome-wide microarray provided a robust means for distinguishing the bacterial pathogens of potato.

Gene expression in lung and basal forebrain during influenza infection in mice

Inbred mice develop strain-dependent changes in sleep during the first few days after inoculation with influenza virus. To identify genes with the potential to differentially modulate sleep under this condition, we performed complementary DNA microarray analysis of both lung and basal forebrain (BF) of infected (I) and uninfected BALB/cByJ (C) and C57BL/6J (B6) mice. This analysis showed significant variation in the expression pattern of 667 and 1217 of the surveyed genes in BF and lung, respectively (P < 0.01). Applying the additional criterion of an effect size >or=2, 495 genes differed in expression in lung compared with 204 in BF. In BF, more genes were differentially expressed as a function of mouse strain, whereas in lung, more genes were differentially expressed as a function of health status. Significant alterations in expression after infection were more numerous and robust in BALB/cByJ vs. C57BL/6J mice. Some genes showed significant variation in both tissues as a function of strain or condition, but the changes in general were not parallel. Genes that showed significant and robust variation as a function of strain, health status or tissue included those related to immune function, metabolism, signal transduction, cell cycle regulation, apoptosis and other miscellaneous categories. Different patterns of gene expression in BF of uninfected mice suggest the possibility of fundamental mechanistic differences in pathways that modulate vigilance in these strains, whereas differences in expression of lung of infected mice suggests different peripherally generated sleep-modulatory stimuli in the two strains.

Competing interactions between micro-RNAs determine neural progenitor survival and proliferation after ethanol exposure: evidence from an ex vivo model of the fetal cerebral cortical neuroepithelium

The fetal brain is sensitive to a variety of teratogens, including ethanol. We showed previously that ethanol induced mitosis and stem cell maturation, but not death, in fetal cerebral cortex-derived progenitors. We tested the hypothesis that micro-RNAs (miRNAs) could mediate the teratogenic effects of ethanol in a fetal mouse cerebral cortex-derived neurosphere culture model. Ethanol, at a level attained by alcoholics, significantly suppressed the expression of four miRNAs, miR-21, -335, -9, and -153, whereas a lower ethanol concentration, attainable during social drinking, induced miR-335 expression. A GABA(A) receptor-dependent mechanism mediated miR-21, but not miR-335 suppression, suggesting that divergent mechanisms regulate ethanol-sensitive miRNAs. Antisense-mediated suppression of miR-21 expression resulted in apoptosis, suggesting that miR-21 is an antiapoptotic factor. miR-335 knockdown promoted cell proliferation and prevented death induced by concurrently suppressing miR-21, indicating that miR-335 is a proapoptotic, antimitogenic factor whose actions are antagonistic to miR-21. Computational analyses identified two genes, Jagged-1, a Notch-receptor ligand, and embryonic-lethal abnormal vision, Drosophila-like 2 (ELAVL2), a brain-specific regulator of RNA stability, as presumptive targets of three of four ethanol-sensitive micro-RNAs. Combined knockdown of miR-335, -21, and -153 significantly increased Jagged-1 mRNA. Furthermore, ethanol induced both Jagged-1 and ELAVL2 mRNA. The collective suppression of micro-RNAs is consistent with ethanol induction of cell cycle and neuroepithelial maturation in the absence of apoptosis. These data identify a role for micro-RNAs as epigenetic intermediaries, which permit teratogens to shape complex, divergent developmental processes, and additionally demonstrate that coordinately regulated miRNAs exhibit both functional synergy and antagonism toward each other.

Novel algorithm for transcriptome analysis

A growing body of evidence implicates the oocyte as a key regulator of ovarian folliculogenesis and early embryonic development. We have screened bovine cDNA microarrays (containing expressed sequence tags representing >15,000 unique genes) with Cy3- and Cy5-labeled cDNA derived from bovine oocyte samples collected at two different stages of meiotic maturation (germinal vesicle vs. metaphase II; n = 3 samples per group). Here, we present a novel data analysis approach that uses all available information from above experiments to obtain and index the transcriptome of bovine oocytes and changes in transcriptome composition in response to meiotic maturation. Signal intensities (Fg) for all housekeeping genes were omitted prior to analysis. A local threshold for gene expression was computed as background intensity (Bg) plus 2 times the standard deviation of background and foreground signals. Within each array, data were normalized by the LOWESS procedure. Subsequently, a two-stage mixed model was fitted to remove systematic variations. In the first stage, the response was the LOWESS normalized Fg with treatment as a fixed effect. In stage 2, the residuals from stage 1 were analyzed in a gene-specific model that included treatment group and spots nested within patch and array. A test for the difference between least squares means for the treatment effect was performed. A false discovery rate (FDR) adjustment on the p values for the difference was carried out. This novel algorithm was compared with approaches that ignore the FDR and the threshold described herein and stark differences obtained.

Effect of acute aldosterone administration on gene expression profile in the heart

Aldosterone is known to have a number of direct adverse effects on the heart, including fibrosis and myocardial inflammation. However, genetic mechanisms of aldosterone action on the heart remain unclear. This paper describes an investigation of temporal changes in gene expression profile of the whole heart induced by acute administration of a physiologic dose of aldosterone in the mouse. mRNA levels of 34,000 known mouse genes were measured at eight time points after aldosterone administration using oligonucleotide microarrays and compared with those of the control animals who underwent a sham injection. A novel software tool (CAGED) designed for analysis of temporal microarray experiments using a Bayesian approach was used to identify genes differentially expressed between the aldosterone-injected and control group. CAGED analysis identified 12 genes as having significant differences in their temporal profiles between aldosterone-injected and control groups. All of these genes exhibited a decrease in expression level 1-3 h after aldosterone injection followed by a brief rebound and a return to baseline. These findings were validated by quantitative RT-PCR. The differentially expressed genes included phosphatases, regulators of steroid biosynthesis, inactivators of reactive oxygen species, and structural proteins. Several of these genes are known to functionally mediate biochemical phenomena previously observed to be triggered by aldosterone administration, such as phosphorylation of ERK1/2. These results provide the first description of cardiac genetic response to aldosterone and identify several potential mediators of known biochemical sequelae of aldosterone administration in the heart.

An inexpensive gel electrophoresis-based polymerase chain reaction method for quantifying mRNA levels

In order to engage their students in a core methodology of the new genomics era, an ever-increasing number of faculty at primarily undergraduate institutions are gaining access to microarray technology. Their students are conducting successful microarray experiments designed to address a variety of interesting questions. A next step in these teaching and research laboratory projects is often validation of the microarray data for individual selected genes. In the research community, this usually involves the use of real-time polymerase chain reaction (PCR), a technology that requires instrumentation and reagents that are prohibitively expensive for most undergraduate institutions. The results of a survey of faculty teaching undergraduates in classroom and research settings indicate a clear need for an alternative approach. We sought to develop an inexpensive and student-friendly gel electrophoresis-based PCR method for quantifying messenger RNA (mRNA) levels using undergraduate researchers as models for students in teaching and research laboratories. We compared the results for three selected genes measured by microarray analysis, real-time PCR, and the gel electrophoresis-based method. The data support the use of the gel electrophoresis-based method as an inexpensive, convenient, yet reliable alternative for quantifying mRNA levels in undergraduate laboratories.

Microarray analysis of gene expression: considerations in data mining and statistical treatment

DNA microarray represents a powerful tool in biomedical discoveries. Harnessing the potential of this technology depends on the development and appropriate use of data mining and statistical tools. Significant current advances have made microarray data mining more versatile. Researchers are no longer limited to default choices that generate suboptimal results. Conflicting results in repeated experiments can be resolved through attention to the statistical details. In the current dynamic environment, there are many choices and potential pitfalls for researchers who intend to incorporate microarrays as a research tool. This review is intended to provide a simple framework to understand the choices and identify the pitfalls. Specifically, this review article discusses the choice of microarray platform, preprocessing raw data, differential expression and validation, clustering, annotation and functional characterization of genes, and pathway construction in light of emergent concepts and tools.

The expression of proapoptosis genes is increased in bipolar disorder, but not in schizophrenia

Post-mortem studies conducted over the past 15 years suggest that apoptosis could play a role in the pathophysiology of bipolar disorder (BD) and, to a lesser degree, schizophrenia (SZ). To test this hypothesis, we have performed a post hoc analysis of an extant gene expression profiling database obtained from the hippocampus using a novel methodology with improved sensitivity. Consistent with the working hypothesis, BDs showed a marked upregulation of 19 out of 44 apoptosis genes; however, contrary to the hypothesis, the SZ group showed a downregulation of genes associated with apoptotic injury and death. These changes in the regulation of apoptosis genes were validated using quantitative RT-PCR. Additionally, antioxidant genes showed a marked downregulation in BDs, suggesting that accumulation of free radicals might occur in the setting of a previously reported decrease of the electron transport chain in this disorder. Overall, the changes seen in BDs and SZs do not appear to be related to exposure to either neuroleptics or mood stabilizers. We conclude that fundamental differences in the genetic regulation of apoptosis and antioxidant genes may help discriminate between the pathophysiology of BD and SZ and potentially point to new treatment strategies that are specific for each disorder.

Proteinase and growth factor alterations revealed by gene microarray analysis of human diabetic corneas

PURPOSE

To identify proteinases and growth factors abnormally expressed in human corneas of donors with diabetic retinopathy (DR), additional to previously described matrix metalloproteinase (MMP)-10 and -3 and insulin-like growth factor (IGF)-I.

METHODS

RNA was isolated from 35 normal, diabetic, and DR autopsy human corneas ex vivo or after organ culture. Amplified cRNA was analyzed using 22,000-gene microarrays (Agilent Technologies, Palo Alto, CA). Gene expression in each diabetic corneal cRNA was assessed against pooled cRNA from 7 to 9 normal corneas. Select differentially expressed genes were validated by quantitative real-time RT-PCR (QPCR) and immunohistochemistry. Organ cultures were treated with a cathepsin inhibitor, cystatin C, or MMP-10.

RESULTS

More than 100 genes were upregulated and 2200 were downregulated in DR corneas. Expression of cathepsin F and hepatocyte growth factor (HGF) genes was increased in ex vivo and organ-cultured DR corneas compared with normal corneas. HGF receptor c-met, fibroblast growth factor (FGF)-3, its receptor FGFR3, tissue inhibitor of metalloproteinase (TIMP)-4, laminin alpha4 chain, and thymosin beta(4) genes were downregulated. The data were corroborated by QPCR and immunohistochemistry analyses; main changes of these components occurred in corneal epithelium. In organ-cultured DR corneas, cystatin C increased laminin-10 and integrin alpha(3)beta(1), whereas in normal corneas MMP-10 decreased laminin-10 and integrin alpha(3)beta(1) expression.

CONCLUSIONS

Elevated cathepsin F and the ability of its inhibitor to produce a more normal phenotype in diabetic corneas suggest increased proteolysis in these corneas. Proteinase changes may result from abnormalities of growth factors, such as HGF and FGF-3, in DR corneas. Specific modulation of proteinases and growth factors could reduce diabetic corneal epitheliopathy.

Patterns of coordinate down-regulation of ARE-containing transcripts following immune cell activation

We evaluated the expression of over 900 AU-rich element (ARE)-containing transcripts in primary human T lymphocytes following stimulation with anti-CD3 and anti-CD28 antibodies and found that approximately 48% of these transcripts were regulated following T cell activation. We identified approximately 145 ARE-containing transcripts that were rapidly induced and then rapidly disappeared within 1 h after activation. Another 250 ARE-containing transcripts expressed in resting T cells were rapidly turned off within 30 min after activation. The rates of transcript disappearance correlated well with rapid mRNA decay measured following transcriptional arrest with actinomycin D. We identified a subset of ARE-containing transcripts that were rapidly induced following T cell activation that were also induced following lipopolysaccharide stimulation of THP-1 monocytes, and these transcripts exhibited rapid decay in both cell types. Our results suggest that ARE-mediated mRNA decay plays an important role in the precisely coordinated down-regulation of gene expression following immune cell activation.

Reliability analysis of microarray data using fuzzy c-means and normal mixture modeling based classification methods

MOTIVATION

A serious limitation in microarray analysis is the unreliability of the data generated from low signal intensities. Such data may produce erroneous gene expression ratios and cause unnecessary validation or post-analysis follow-up tasks. Therefore, the elimination of unreliable signal intensities will enhance reproducibility and reliability of gene expression ratios produced from microarray data. In this study, we applied fuzzy c-means (FCM) and normal mixture modeling (NMM) based classification methods to separate microarray data into reliable and unreliable signal intensity populations.

RESULTS

We compared the results of FCM classification with those of classification based on NMM. Both approaches were validated against reference sets of biological data consisting of only true positives and true negatives. We observed that both methods performed equally well in terms of sensitivity and specificity. Although a comparison of the computation times indicated that the fuzzy approach is computationally more efficient, other considerations support the use of NMM for the reliability analysis of microarray data.

AVAILABILITY

The classification approaches described in this paper and sample microarray data are available as Matlab( TM ) (The MathWorks Inc., Natick, MA) programs (mfiles) and text files, respectively, at http://rc.kfshrc.edu.sa/bssc/staff/MusaAsyali/Downloads.asp. The programs can be run/tested on many different computer platforms where Matlab is available.

CONTACT

asyali@kfshrc.edu.sa.

Microarray and EST database estimates of mRNA expression levels differ: the protein length versus expression curve for C. elegans

Background

Various methods for estimating protein expression levels are known. The level of correlation between these methods is only fair, and systematic biases in each of the methods cannot be ruled out. We here investigate systematic biases in the estimation of gene expression rates from microarray data and from abundance within the Expressed Sequence Tag (EST) database. We suggest that length is a significant factor in biases to measured gene expression rates.

As a specific example of the importance of the bias of expression rate with length, we address the following evolutionary question: Does the average C. elegans protein length increase or decrease with expression level? Two different answers to this question have been reported in the literature, one method using expression levels estimated by abundance within the EST database and another using microarrays. We have investigated this issue by constructing the full protein length versus expression curve for C. elegans, using both methods for estimating expression levels.

Results

The microarray data show a monotonic decrease of length with expression level, whereas the abundance within the EST database data show a non-monotonic behavior. Furthermore, the ratio of the expression level estimated by the EST database to that measured by microarrays is not constant, but rather systematically biased with gene length.

Conclusions

It is suggested that the length bias may lie primarily in the abundance within the EST database method, being not ameliorated by internal standards as it is in the microarray data, and that this bias should be removed before data interpretation. When this is done, both the microarray and the abundance within the EST database give a monotonic decrease of spliced length with expression level, and the correlation between the EST and microarray data becomes larger. We suggest that standard RNA controls be used to normalize for length bias in any method that measures expression.