Knowledge-based analysis of microarray gene expression data by using support vector machines

We introduce a method of functionally classifying genes by using gene expression data from DNA microarray hybridization experiments. The method is based on the theory of support vector machines (SVMs). SVMs are considered a supervised computer learning method because they exploit prior knowledge of gene function to identify unknown genes of similar function from expression data. SVMs avoid several problems associated with unsupervised clustering methods, such as hierarchical clustering and self-organizing maps. SVMs have many mathematical features that make them attractive for gene expression analysis, including their flexibility in choosing a similarity function, sparseness of solution when dealing with large data sets, the ability to handle large feature spaces, and the ability to identify outliers. We test several SVMs that use different similarity metrics, as well as some other supervised learning methods, and find that the SVMs best identify sets of genes with a common function using expression data. Finally, we use SVMs to predict functional roles for uncharacterized yeast ORFs based on their expression data.

Importance of replication in microarray gene expression studies: Statistical methods and evidence from repetitive cDNA hybridizations

We present statistical methods for analyzing replicated cDNA microarray expression data and report the results of a controlled experiment. The study was conducted to investigate inherent variability in gene expression data and the extent to which replication in an experiment produces more consistent and reliable findings. We introduce a statistical model to describe the probability that mRNA is contained in the target sample tissue, converted to probe, and ultimately detected on the slide. We also introduce a method to analyze the combined data from all replicates. Of the 288 genes considered in this controlled experiment, 32 would be expected to produce strong hybridization signals because of the known presence of repetitive sequences within them. Results based on individual replicates, however, show that there are 55, 36, and 58 highly expressed genes in replicates 1, 2, and 3, respectively. On the other hand, an analysis by using the combined data from all 3 replicates reveals that only 2 of the 288 genes are incorrectly classified as expressed. Our experiment shows that any single microarray output is subject to substantial variability. By pooling data from replicates, we can provide a more reliable analysis of gene expression data. Therefore, we conclude that designing experiments with replications will greatly reduce misclassification rates. We recommend that at least three replicates be used in designing experiments by using cDNA microarrays, particularly when gene expression data from single specimens are being analyzed.

The Stanford Microarray Database

The Stanford Microarray Database (SMD) stores raw and normalized data from microarray experiments, and provides web interfaces for researchers to retrieve, analyze and visualize their data. The two immediate goals for SMD are to serve as a storage site for microarray data from ongoing research at Stanford University, and to facilitate the public dissemination of that data once published, or released by the researcher. Of paramount importance is the connection of microarray data with the biological data that pertains to the DNA deposited on the microarray (genes, clones etc.). SMD makes use of many public resources to connect expression information to the relevant biology, including SGD [Ball,C.A., Dolinski,K., Dwight,S.S., Harris,M.A., Issel-Tarver,L., Kasarskis,A., Scafe,C.R., Sherlock,G., Binkley,G., Jin,H. et al. (2000) Nucleic Acids Res., 28, 77–80], YPD and WormPD [Costanzo,M.C., Hogan,J.D., Cusick,M.E., Davis,B.P., Fancher,A.M., Hodges,P.E., Kondu,P., Lengieza,C., Lew-Smith,J.E., Lingner,C. et al. (2000) Nucleic Acids Res., 28, 73–76], Unigene [Wheeler,D.L., Chappey,C., Lash,A.E., Leipe,D.D., Madden,T.L., Schuler,G.D., Tatusova,T.A. and Rapp,B.A. (2000) Nucleic Acids Res., 28, 10–14], dbEST [Boguski,M.S., Lowe,T.M. and Tolstoshev,C.M. (1993) Nature Genet., 4, 332–333] and SWISS-PROT [Bairoch,A. and Apweiler,R. (2000) Nucleic Acids Res., 28, 45–48] and can be accessed at http://genome-www.stanford.edu/microarray.

Amygdala-enriched genes identified by microarray technology are restricted to specific amygdaloid subnuclei

Microarray technology represents a potentially powerful method for identifying cell type- and regionally restricted genes expressed in the brain. Here we have combined a microarray analysis of differential gene expression among five selected brain regions, including the amygdala, cerebellum, hippocampus, olfactory bulb, and periaqueductal gray, with in situ hybridization. On average, 0.3% of the 34,000 genes interrogated were highly enriched in each of the five regions, relative to the others. In situ hybridization performed on a subset of amygdala-enriched genes confirmed in most cases the overall region-specificity predicted by the microarray data and identified additional sites of brain expression not examined on the microarrays. Strikingly, the majority of these genes exhibited boundaries of expression within the amygdala corresponding to cytoarchitectonically defined subnuclei. These results define a unique set of molecular markers for amygdaloid subnuclei and provide tools to genetically dissect their functional roles in different emotional behaviors.

Statistical modeling of large microarray data sets to identify stimulus-response profiles

A statistical modeling approach is proposed for use in searching large microarray data sets for genes that have a transcriptional response to a stimulus. The approach is unrestricted with respect to the timing, magnitude or duration of the response, or the overall abundance of the transcript. The statistical model makes an accommodation for systematic heterogeneity in expression levels. Corresponding data analyses provide gene-specific information, and the approach provides a means for evaluating the statistical significance of such information. To illustrate this strategy we have derived a model to depict the profile expected for a periodically transcribed gene and used it to look for budding yeast transcripts that adhere to this profile. Using objective criteria, this method identifies 81% of the known periodic transcripts and 1,088 genes, which show significant periodicity in at least one of the three data sets analyzed. However, only one-quarter of these genes show significant oscillations in at least two data sets and can be classified as periodic with high confidence. The method provides estimates of the mean activation and deactivation times, induced and basal expression levels, and statistical measures of the precision of these estimates for each periodic transcript.

Host microarray analysis reveals a role for the Salmonella response regulator phoP in human macrophage cell death

Bacterial pathogens manipulate host cells to promote pathogen survival and dissemination. We used a 22,571 human cDNA microarray to identify host pathways that are affected by the Salmonella enterica subspecies typhimurium phoP gene, a transcription factor required for virulence, by comparing the expression profiles of human monocytic tissue culture cells infected with either the wild-type bacteria or a phoP∷Tn10 mutant strain. Both wild-type and phoP∷Tn10 bacteria induced a common set of genes, many of which are proinflammatory. Differentially expressed genes included those that affect host cell death, suggesting that the phoP regulatory system controls bacterial genes that alter macrophage survival. Subsequent experiments showed that the phoP∷Tn10 mutant strain is defective for killing both cultured and primary human macrophages but is able to replicate intracellularly. These experiments indicate that phoP plays a role in Salmonella-induced human macrophage cell death.

Recursive partitioning for tumor classification with gene expression microarray data

Precise classification of tumors is critically important for cancer diagnosis and treatment. It is also a scientifically challenging task. Recently, efforts have been made to use gene expression profiles to improve the precision of classification, with limited success. Using a published data set for purposes of comparison, we introduce a methodology based on classification trees and demonstrate that it is significantly more accurate for discriminating among distinct colon cancer tissues than other statistical approaches used heretofore. In addition, competing classification trees are displayed, which suggest that different genes may coregulate colon cancers.

Parallel human genome analysis: microarray-based expression monitoring of 1000 genes.

Microarrays containing 1046 human cDNAs of unknown sequence were printed on glass with high-speed robotics. These 1.0-cm2 DNA "chips" were used to quantitatively monitor differential expression of the cognate human genes using a highly sensitive two-color hybridization assay. Array elements that displayed differential expression patterns under given experimental conditions were characterized by sequencing. The identification of known and novel heat shock and phorbol ester-regulated genes in human T cells demonstrates the sensitivity of the assay. Parallel gene analysis with microarrays provides a rapid and efficient method for large-scale human gene discovery.

Tissue-specific gene expression in soybean (Glycine max) detected by cDNA microarray analysis

We have constructed cDNA microarrays for soybean (Glycine max L. Merrill), containing approximately 4,100 Unigene ESTs derived from axenic roots, to evaluate their application and utility for functional genomics of organ differentiation in legumes. We assessed microarray technology by conducting studies to evaluate the accuracy of microarray data and have found them to be both reliable and reproducible in repeat hybridisations. Several ESTs showed high levels (>50 fold) of differential expression in either root or shoot tissue of soybean. A small number of physiologically interesting, and differentially expressed sequences found by microarray analysis were verified by both quantitative real-time RT-PCR and Northern blot analysis. There was a linear correlation (r(2) = 0.99, over 5 orders of magnitude) between microarray and quantitative real-time RT-PCR data. Microarray analysis of soybean has enormous potential not only for the discovery of new genes involved in tissue differentiation and function, but also to study the expression of previously characterised genes, gene networks and gene interactions in wild-type, mutant or transgenic; plants.

Microarray expression profiling in melanoma reveals a BRAF mutation signature

We have used microarray gene expression pro. ling and machine learning to predict the presence of BRAF mutations in a panel of 61 melanoma cell lines. The BRAF gene was found to be mutated in 42 samples (69%) and intragenic mutations of the NRAS gene were detected in seven samples (11%). No cell line carried mutations of both genes. Using support vector machines, we have built a classifier that differentiates between melanoma cell lines based on BRAF mutation status. As few as 83 genes are able to discriminate between BRAF mutant and BRAF wild-type samples with clear separation observed using hierarchical clustering. Multidimensional scaling was used to visualize the relationship between a BRAF mutation signature and that of a generalized mitogen-activated protein kinase ( MAPK) activation ( either BRAF or NRAS mutation) in the context of the discriminating gene list. We observed that samples carrying NRAS mutations lie somewhere between those with or without BRAF mutations. These observations suggest that there are gene-specific mutation signals in addition to a common MAPK activation that result from the pleiotropic effects of either BRAF or NRAS on other signaling pathways, leading to measurably different transcriptional changes.

A microarray study of post-mortem mRNA degradation in mouse brain tissue

Background: Although there is evidence that post-mortem interval (PMI) is not a major contributor to reduced overall RNA integrity, it may differentially affect a subgroup of gene transcripts that are susceptible to PMI-related degradation. This would particularly have ramifications for microarray studies that include a broad spectrum of genes. Method: Brain tissue was removed from adult mice at 0, 6, 12, 18, 24,36 and 48 h post-mortem. RNA transcript abundance was measured by hybridising RNA from the zero time point with test RNA from each PMI time point, and differential gene expression was assessed using cDNA microarrays. Sequence and ontological analyses were performed on the group of RNA transcripts showing greater than two-fold reduction. Results: Increasing PMI was associated with decreased tissue pH and increased RNA degradation as indexed by 28S/18S ribosomal RNA ratio. Approximately 12% of mRNAs detected on the arrays displayed more than a two-fold decrease in abundance by 48 It post-mortem. An analysis of nucleotide composition provided evidence that transcripts with the AUUUA motif in the 3' untranslated region (3'UTR) were more susceptible to PMI-related RNA degradation, compared to transcripts not carrying the 3'UTR AUUUA motif. Consistent with this finding, ontological analysis showed transcription factors and elements to be over-represented in the group of transcripts susceptible to degradation. Conclusion: A subgroup of mammalian mRNA transcripts are particularly susceptible to PMI-related degradation, and as a group, they are more likely to carry the YUTR AUUUA motif. PMI should be controlled for in human and animal model post-mortem brain studies, particularly those including a broad spectrum of mRNA transcripts. (c) 2005 Elsevier B.V. All rights reserved.

Microarray-based comparative genomic hybridisation of breast cancer patients receiving neoadjuvant chemotherapy

We analysed the molecular genetic profiles of breast cancer samples before and after neoadjuvant chemotherapy with combination doxorubicin and cyclophosphamide (AC). DNA was obtained from microdissected frozen breast core biopsies from 44 patients before chemotherapy. Additional samples were obtained before the second course of chemotherapy (D21) and after the completion of the treatment (surgical specimens) in 17 and 21 patients, respectively. Microarray-based comparative genome hybridisation was performed using a platform containing approx5800 bacterial artificial chromosome clones (genome-wide resolution: 0.9 Mb). Analysis of the 44 pretreatment biopsies revealed that losses of 4p, 4q, 5q, 12q13.11–12q13.12, 17p11.2 and 17q11.2; and gains of 1p, 2p, 7q, 9p, 11q, 19p and 19q were significantly associated with oestrogen receptor negativity. 16q21–q22.1 losses were associated with lobular and 8q24 gains with ductal types. Losses of 5q33.3–q4 and 18p11.31 and gains of 6p25.1–p25.2 and Xp11.4 were associated with HER2 amplification. No correlations between DNA copy number changes and clinical response to AC were found. Microarray-based comparative genome hybridisation analysis of matched pretreatment and D21 biopsies failed to identify statistically significant differences, whereas a comparison between matched pretreatment and surgical samples revealed a statistically significant acquired copy number gain on 11p15.2–11p15.5. The modest chemotherapy-driven genomic changes, despite profound loss of cell numbers, suggest that there is little therapeutic selection of resistant non-modal cell lineages.

Gene expression in human alcoholism: Microarray analysis of frontal cortex

Background: Changes in brain gene expression are thought to be responsible for the tolerance, dependence, and neurotoxicity produced by chronic alcohol abuse, but there has been no large scale study of gene expression in human alcoholism. Methods: RNA was extracted from postmortem samples of superior frontal cortex of alcoholics and nonalcoholics. Relative levels of RNA were determined by array techniques. We used both cDNA and oligonucleotide microarrays to provide coverage of a large number of genes and to allow cross-validation for those genes represented on both types of arrays. Results: Expression levels were determined for over 4000 genes and 163 of these were found to differ by 40% or more between alcoholics and nonalcoholics. Analysis of these changes revealed a selective reprogramming of gene expression in this brain region, particularly for myelin-related genes which were downregulated in the alcoholic samples. In addition, cell cycle genes and several neuronal genes were changed in expression. Conclusions: These gene expression changes suggest a mechanism for the loss of cerebral white matter in alcoholics as well as alterations that may lead to the neurotoxic actions of ethanol.

On the simultaneous use of clinical and microarray expression data in the cluster analysis of tissue samples

This paper considers a model-based approach to the clustering of tissue samples of a very large number of genes from microarray experiments. It is a nonstandard problem in parametric cluster analysis because the dimension of the feature space (the number of genes) is typically much greater than the number of tissues. Frequently in practice, there are also clinical data available on those cases on which the tissue samples have been obtained. Here we investigate how to use the clinical data in conjunction with the microarray gene expression data to cluster the tissue samples. We propose two mixture model-based approaches in which the number of components in the mixture model corresponds to the number of clusters to be imposed on the tissue samples. One approach specifies the components of the mixture model to be the conditional distributions of the microarray data given the clinical data with the mixing proportions also conditioned on the latter data. Another takes the components of the mixture model to represent the joint distributions of the clinical and microarray data. The approaches are demonstrated on some breast cancer data, as studied recently in van't Veer et al. (2002).

Microarray and proteomic analysis of the human alcoholic brain

The superior frontal cortex (SFC) is selectively damaged in chronic alcohol abuse, with localized neuronal loss and tissue atrophy. Regions such as motor cortex show little neuronal loss except in severe co-morbidity (liver cirrhosis or WKS). Altered gene expression was found in microarray comparisons of alcoholic and control SFC samples [1]. We used Western blots and proteomic analysis to identify the proteins that also show differential expression. Tissue was obtained at autopsy under informed, written consent from uncomplicated alcoholics and age- and sex-matched controls. Alcoholics had consumed 80 g ethanol/day chronically (often, 200 g/day for 20 y). Controls either abstained or were social drinkers ( 20 g/day). All subjects had pathological confirmation of liver and brain diagnosis; none had been polydrug abusers. Samples were homogenized in water and clarified by brief centrifugation (1000g, 3 min) before storage at –80°C. For proteomics the thawed suspensions were centrifuged (15000g, 50 min) to prepare soluble fractions. Aliquots were pooled from SFC samples from the 5 chronic alcoholics and 5 matched controls used in the previous microarray study [1]. 2-Dimensional electrophoresis was performed in triplicate using 18 cm format pH 4–7 and pH 6–11 immobilized pH gradients for firstdimension isoelectric focusing. Following second-dimension SDS-PAGE the proteins were fluorescently stained and the images collected by densitometry. 182 proteins differed by 2-fold between cases and controls. 141 showed lower expression in alcoholics, 33 higher, and 8 were new or had disappeared. To date 63 proteins have been identified using MALDI-MS and MS-MS. Western blots were performed on uncentrifuged individual samples from 76 subjects (controls, uncomplicated alcoholics and cirrhotic alcoholics). A common standard was run on every gel. After transfer, immunolabeling, and densitometry, the intensities of the unknown bands were compared to those of the standards. We focused on proteins from transcripts that showed clear differences in a series of microarray studies, classified into common sets including Regulators of G-protein Signaling and Myelin-associated proteins. The preponderantly lower level of differentially expressed proteins in alcoholics parallels the microarray mRNA analysis in the same samples. We found that mRNA and protein expression do not frequently correspond; this may help identify pathogenic processes acting at the level of transcription, translation, or post-translationally.