The Gene Ontology (GO) database and informatics resource

The Gene Ontology (GO) project (http://www. geneontology.org/) provides structured, controlled vocabularies and classifications that cover several domains of molecular and cellular biology and are freely available for community use in the annotation of genes, gene products and sequences. Many model organism databases and genome annotation groups use the GO and contribute their annotation sets to the GO resource. The GO database integrates the vocabularies and contributed annotations and provides full access to this information in several formats. Members of the GO Consortium continually work collectively, involving outside experts as needed, to expand and update the GO vocabularies. The GO Web resource also provides access to extensive documentation about the GO project and links to applications that use GO data for functional analyses.

Genome-wide scan of bipolar disorder in 65 pedigrees: supportive evidence for linkage at 8q24, 18q22, 4q32, 2p12, and 13q12

The purpose of this study was to assess 65 pedigrees ascertained through a Bipolar I (BPI) proband for evidence of linkage, using nonparametric methods in a genome-wide scan and for possible parent of origin effect using several analytical methods. We identified 15 loci with nominally significant evidence for increased allele sharing among affected relative pairs. Eight of these regions, at 8q24, 18q22, 4q32, 13q12, 4q35, 10q26, 2p12, and 12q24, directly overlap with previously reported evidence of linkage to bipolar disorder. Five regions at 20p13, 2p22, 14q23, 9p13, and 1q41 are within several Mb of previously reported regions. We report our findings in rank order and the top five markers had an NPL>2.5. The peak finding in these regions were D8S256 at 8q24, NPL 3.13; D18S878 at 18q22, NPL 2.90; D4S1629 at 4q32, NPL 2.80; D2S99 at 2p12, NPL 2.54; and D13S1493 at 13q12, NPL 2.53. No locus produced statistically significant evidence for linkage at the genome-wide level. The parent of origin effect was studied and consistent with our previous findings, evidence for a locus on 18q22 was predominantly from families wherein the father or paternal lineage was affected. There was evidence consistent with paternal imprinting at the loci on 13q12 and 1q41.

Three cell wall mannoproteins facilitate the uptake of iron in Saccharomyces cerevisiae

Analysis of iron-regulated gene expression in Saccharomyces cerevisiae using cDNA microarrays has identified three putative cell wall proteins that are directly regulated by Aft1p, the major iron-dependent transcription factor in yeast. FIT1, FIT2, and FIT3 (for facilitator of iron transport) were more highly expressed in strains grown in low concentrations of iron and in strains in which AFT1-1(up), a constitutively active allele of AFT1, was expressed. Northern blot analysis confirmed that FIT1, FIT2, and FIT3 mRNA transcript levels were increased 60-230-fold in response to iron deprivation in an Aft1p-dependent manner. Fit1p was localized exclusively to the cell wall by indirect immunofluorescence. Deletion of the FIT genes, individually or in combination, resulted in diminished uptake of iron bound to the siderophores ferrioxamine B and ferrichrome, without diminishing the uptake of ferric iron salts, or the siderophores triacetylfusarinine C and enterobactin. FIT-deletion strains exhibited increased expression of Aft1p target genes as measured by a FET3-lacZ reporter gene or by Arn1p Western blotting, indicating that cells respond to the absence of FIT genes by up-regulating systems of iron uptake. Aft1p activation in FIT-deleted strains occurred when either ferrichrome or ferric salts were used as sources of iron during growth, suggesting that the FIT genes enhance uptake of iron from both sources. Enzymatic digestion of the cell wall resulted in the release of significant amounts of iron from cells, and the relative quantity of iron released was reduced in FIT-deletion strains. Fit1p, Fit2p, and Fit3p may function by increasing the amount of iron associated with the cell wall and periplasmic space.

Relation of gene expression phenotype to immunoglobulin mutation genotype in B cell chronic lymphocytic leukemia

The most common human leukemia is B cell chronic lymphocytic leukemia (CLL), a malignancy of mature B cells with a characteristic clinical presentation but a variable clinical course. The rearranged immunoglobulin (Ig) genes of CLL cells may be either germ-line in sequence or somatically mutated. Lack of Ig mutations defined a distinctly worse prognostic group of CLL patients raising the possibility that CLL comprises two distinct diseases. Using genomic-scale gene expression profiling, we show that CLL is characterized by a common gene expression "signature," irrespective of Ig mutational status, suggesting that CLL cases share a common mechanism of transformation and/or cell of origin. Nonetheless, the expression of hundreds of other genes correlated with the Ig mutational status, including many genes that are modulated in expression during mitogenic B cell receptor signaling. These genes were used to build a CLL subtype predictor that may help in the clinical classification of patients with this disease.

Dominant-lethal alpha-tubulin mutants defective in microtubule depolymerization in yeast

The dynamic instability of microtubules has long been understood to depend on the hydrolysis of GTP bound to beta-tubulin, an event stimulated by polymerization and necessary for depolymerization. Crystallographic studies of tubulin show that GTP is bound by beta-tubulin at the longitudinal dimer-dimer interface and contacts particular alpha-tubulin residues in the next dimer along the protofilament. This structural arrangement suggests that these contacts could account for assembly-stimulated GTP hydrolysis. As a test of this hypothesis, we examined, in yeast cells, the effect of mutating the alpha-tubulin residues predicted, on structural grounds, to be involved in GTPase activation. Mutation of these residues to alanine (i.e., D252A and E255A) created poisonous alpha-tubulins that caused lethality even as minor components of the alpha-tubulin pool. When the mutant alpha-tubulins were expressed from the galactose-inducible promoter of GAL1, cells rapidly acquired aberrant microtubule structures. Cytoplasmic microtubules were largely bundled, spindle assembly was inhibited, preexisting spindles failed to completely elongate, and occasional, stable microtubules were observed unattached to spindle pole bodies. Time-lapse microscopy showed that microtubule dynamics had ceased. Microtubules containing the mutant proteins did not depolymerize, even in the presence of nocodazole. These data support the view that alpha-tubulin is a GTPase-activating protein that acts, during microtubule polymerization, to stimulate GTP hydrolysis in beta-tubulin and thereby account for the dynamic instability of microtubules.

The glycine allele of a glycine/arginine polymorphism in the beta2-adrenergic receptor gene is associated with essential hypertension in a population of Chinese origin

BACKGROUND

Several studies implicate polymorphisms in the human beta-adrenergic receptor gene (ADRB2) in the susceptibility to hypertension. We sought to replicate these results in a population of Chinese origin primarily from Taiwan and the San Francisco Bay area.

METHODS

We genotyped >800 hypertensive subjects and individuals with low-normal blood pressure that were derived largely from the same families as the hypertensive patients for three polymorphisms in the ADRB2 gene: a C/T transition at position 47 (C-47T) in the 5' leader cistron; another C/T transition that results in a glycine/ arginine substitution at codon 16 (Gly16Arg), and a G/C transversion that causes a glutamate/glutamine substitution at codon 27 (Glu27Gln).

RESULTS

The Gly16Arg was significantly associated with hypertension (P < .03). Under a dominant model, for hypertension the relative risk for the Gly/Gly and Gly/Arg genotypes versus the Arg/Arg genotype was 1.35 (95% confidence limits [CL] 1.08, 1.70); for low-normal blood pressure the relative risk was 0.79 (95% CL 0.66, 0.94). This polymorphism explained approximately 1% of the variance in systolic and diastolic blood pressures in our study population. There was no evidence of association between the C-47T and Glu27Gln polymorphisms and hypertension in this population.

CONCLUSIONS

The Glyl6 allele in the beta2-adrenergic receptor gene is a susceptibility allele for essential hypertension in a population of Chinese origin.

Diversity of gene expression in adenocarcinoma of the lung

The global gene expression profiles for 67 human lung tumors representing 56 patients were examined by using 24,000-element cDNA microarrays. Subdivision of the tumors based on gene expression patterns faithfully recapitulated morphological classification of the tumors into squamous, large cell, small cell, and adenocarcinoma. The gene expression patterns made possible the subclassification of adenocarcinoma into subgroups that correlated with the degree of tumor differentiation as well as patient survival. Gene expression analysis thus promises to extend and refine standard pathologic analysis.

RERG is a novel ras-related, estrogen-regulated and growth-inhibitory gene in breast cancer

Using microarray analysis, we identified a unique ras superfamily gene, termed RERG (ras-related and estrogen-regulated growth inhibitor), whose expression was decreased or lost in a significant percentage of primary human breast tumors that show a poor clinical prognosis. Importantly, high RERG expression correlated with expression of a set of genes that define a breast tumor subtype that is estrogen receptor-positive and associated with a slow rate of tumor cell proliferation and a favorable prognosis for these cancer patients. RERG mRNA expression was induced rapidly in MCF-7 cells stimulated by beta-estradiol and repressed by tamoxifen treatment. Like Ras, RERG protein exhibited intrinsic GDP/GTP binding and GTP hydrolysis activity. Unlike Ras proteins, RERG lacks a known recognition signal for COOH-terminal prenylation and was localized primarily in the cytoplasm. Expression of RERG protein in MCF-7 breast carcinoma cells resulted in a significant inhibition of both anchorage-dependent and anchorage-independent growth in vitro and inhibited tumor formation in nude mice. These features of RERG are strikingly different from most Ras superfamily GTP-binding pro-teins and suggest that the loss of RERG expression may contribute to breast tumorigenesis.

Genetic variation in aldosterone synthase predicts plasma glucose levels

The mineralocorticoid hormone, aldosterone, is known to play a role in sodium homeostasis. We serendipitously found, however, highly significant association between single-nucleotide polymorphisms in the aldosterone synthase gene and plasma glucose levels in a large population of Chinese and Japanese origin. Two polymorphisms--one in the putative promoter (T-344C) and another resulting in a lysine/arginine substitution at amino acid 173, which are in complete linkage disequilibrium in this population--were associated with fasting plasma glucose levels (P = 0.000017) and those 60 (P = 0.017) and 120 (P = 0.0019) min after an oral glucose challenge. A C/T variant in intron 1, between these polymorphisms, was not associated with glucose levels. Arg-173 and -344C homozygotes were most likely to be diabetic [odds ratio 2.51; 95% confidence interval (C.I.) 1.39-3.92; P = 0.0015] and have impaired fasting glucose levels (odds ratio 3.53; 95% C.I. 2.02-5.5; P = 0.0000036). These results suggest a new role for aldosterone in glucose homeostasis.

Genomic expression responses to DNA-damaging agents and the regulatory role of the yeast ATR homolog Mec1p

Eukaryotic cells respond to DNA damage by arresting the cell cycle and modulating gene expression to ensure efficient DNA repair. The human ATR kinase and its homolog in yeast, MEC1, play central roles in transducing the damage signal. To characterize the role of the Mec1 pathway in modulating the cellular response to DNA damage, we used DNA microarrays to observe genomic expression in Saccharomyces cerevisiae responding to two different DNA-damaging agents. We compared the genome-wide expression patterns of wild-type cells and mutants defective in Mec1 signaling, including mec1, dun1, and crt1 mutants, under normal growth conditions and in response to the methylating-agent methylmethane sulfonate (MMS) and ionizing radiation. Here, we present a comparative analysis of wild-type and mutant cells responding to these DNA-damaging agents, and identify specific features of the gene expression responses that are dependent on the Mec1 pathway. Among the hundreds of genes whose expression was affected by Mec1p, one set of genes appears to represent an MEC1-dependent expression signature of DNA damage. Other aspects of the genomic responses were independent of Mec1p, and likely independent of DNA damage, suggesting the pleiotropic effects of MMS and ionizing radiation. The complete data set as well as supplemental materials is available at http://www-genome.stanford.edu/mec1.

Genetic variation in the human urea transporter-2 is associated with variation in blood pressure

The kidney, by regulating the volume of fluid in the body, plays a key role in regulating blood pressure (BP). The kidney uses primarily sodium and, to a lesser extent, urea to maintain the appropriate volume of fluid. Genetic variation in proteins that determine sodium reabsorption and excretion is known to significantly influence BP. However, the influence of genetic variation in urea transporters on BP has not been examined. We determined therefore whether nucleotide variation in the kidney-specific human urea transporter, HUT2, is associated with variation in BP. After determining the genomic structure of the coding sequence, seven single nucleotide polymorphisms (SNPs) were identified. Two of the SNPs result in Val/Ile and Ala/Thr amino acid substitutions at positions 227 and 357 in the HUT2 open reading frame, respectively. Another SNP is silent and four others are in introns or the 3' untranslated region. Over 1000 hypertensive and low-normotensive individuals of Chinese origin were typed for five of these SNPs using a high-throughput genotyping method. The Ile227 and Ala357 alleles were associated with low diastolic BP in men but not women, with odds ratios 2.1 [95% confidence interval (CI) 1.5-2.7, P < 0.001] and 1.5 (95% CI 1.2-1.8, P < 0.001), respectively. There was a similar trend for systolic BP, and odds ratios for the Ile227 and Ala357 alleles were 1.7 (95% CI 1.2-2.3, P = 0.002) and 1.3 (95% CI 1.1-1.6, P = 0.007), respectively, in men.

Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications

The purpose of this study was to classify breast carcinomas based on variations in gene expression patterns derived from cDNA microarrays and to correlate tumor characteristics to clinical outcome. A total of 85 cDNA microarray experiments representing 78 cancers, three fibroadenomas, and four normal breast tissues were analyzed by hierarchical clustering. As reported previously, the cancers could be classified into a basal epithelial-like group, an ERBB2-overexpressing group and a normal breast-like group based on variations in gene expression. A novel finding was that the previously characterized luminal epithelial/estrogen receptor-positive group could be divided into at least two subgroups, each with a distinctive expression profile. These subtypes proved to be reasonably robust by clustering using two different gene sets: first, a set of 456 cDNA clones previously selected to reflect intrinsic properties of the tumors and, second, a gene set that highly correlated with patient outcome. Survival analyses on a subcohort of patients with locally advanced breast cancer uniformly treated in a prospective study showed significantly different outcomes for the patients belonging to the various groups, including a poor prognosis for the basal-like subtype and a significant difference in outcome for the two estrogen receptor-positive groups.

Microarrays in primary breast cancer--lessons from chemotherapy studies

Current development in molecular techniques has extended the opportunities to explore genetic alterations in malignant tissue. There is a need to improve prognostication and, in particular, to understand the mechanisms of treatment resistance in different tumours. Gene analyses by microarrays allow concomitant analyses of several genes in concert, providing new opportunities for tumour classification and understanding of key biological disturbances. This paper outlines our continuing studies exploring prognostic and, we hope, predictive factors in breast cancer therapy.

Promoter-specific binding of Rap1 revealed by genome-wide maps of protein-DNA association

We determined the distribution of repressor-activator protein 1 (Rap1) and the accessory silencing proteins Sir2, Sir3 and Sir4 in vivo on the entire yeast genome, at a resolution of 2 kb. Rap1 is central to the cellular economy during rapid growth, targeting 294 loci, about 5% of yeast genes, and participating in the activation of 37% of all RNA polymerase II initiation events in exponentially growing cells. Although the DNA sequence recognized by Rap1 is found in both coding and intergenic sequences, the binding of Rap1 to the genome was highly specific to intergenic regions with the potential to act as promoters. This global phenomenon, which may be a general characteristic of sequence-specific transcriptional factors, indicates the existence of a genome-wide molecular mechanism for marking promoter regions.

High-throughput genotyping with single nucleotide polymorphisms

To make large-scale association studies a reality, automated high-throughput methods for genotyping with single-nucleotide polymorphisms (SNPs) are needed. We describe PCR conditions that permit the use of the TaqMan or 5' nuclease allelic discrimination assay for typing large numbers of individuals with any SNP and computational methods that allow genotypes to be assigned automatically. To demonstrate the utility of these methods, we typed >1600 individuals for a G-to-T transversion that results in a glutamate-to-aspartate substitution at position 298 in the endothelial nitric oxide synthase gene, and a G/C polymorphism (newly identified in our laboratory) in intron 8 of the 11-beta hydroxylase gene. The genotyping method is accurate-we estimate an error rate of fewer than 1 in 2000 genotypes, rapid-with five 96-well PCR machines, one fluorescent reader, and no automated pipetting, over one thousand genotypes can be generated by one person in one day, and flexible-a new SNP can be tested for association in less than one week. Indeed, large-scale genotyping has been accomplished for 23 other SNPs in 13 different genes using this method. In addition, we identified three "pseudo-SNPs" (WIAF1161, WIAF2566, and WIAF335) that are probably a result of duplication.

Missing value estimation methods for DNA microarrays

MOTIVATION

Gene expression microarray experiments can generate data sets with multiple missing expression values. Unfortunately, many algorithms for gene expression analysis require a complete matrix of gene array values as input. For example, methods such as hierarchical clustering and K-means clustering are not robust to missing data, and may lose effectiveness even with a few missing values. Methods for imputing missing data are needed, therefore, to minimize the effect of incomplete data sets on analyses, and to increase the range of data sets to which these algorithms can be applied. In this report, we investigate automated methods for estimating missing data.

RESULTS

We present a comparative study of several methods for the estimation of missing values in gene microarray data. We implemented and evaluated three methods: a Singular Value Decomposition (SVD) based method (SVDimpute), weighted K-nearest neighbors (KNNimpute), and row average. We evaluated the methods using a variety of parameter settings and over different real data sets, and assessed the robustness of the imputation methods to the amount of missing data over the range of 1--20% missing values. We show that KNNimpute appears to provide a more robust and sensitive method for missing value estimation than SVDimpute, and both SVDimpute and KNNimpute surpass the commonly used row average method (as well as filling missing values with zeros). We report results of the comparative experiments and provide recommendations and tools for accurate estimation of missing microarray data under a variety of conditions.

Missing value estimation methods for DNA microarrays

MOTIVATION

Gene expression microarray experiments can generate data sets with multiple missing expression values. Unfortunately, many algorithms for gene expression analysis require a complete matrix of gene array values as input. For example, methods such as hierarchical clustering and K-means clustering are not robust to missing data, and may lose effectiveness even with a few missing values. Methods for imputing missing data are needed, therefore, to minimize the effect of incomplete data sets on analyses, and to increase the range of data sets to which these algorithms can be applied. In this report, we investigate automated methods for estimating missing data.

RESULTS

We present a comparative study of several methods for the estimation of missing values in gene microarray data. We implemented and evaluated three methods: a Singular Value Decomposition (SVD) based method (SVDimpute), weighted K-nearest neighbors (KNNimpute), and row average. We evaluated the methods using a variety of parameter settings and over different real data sets, and assessed the robustness of the imputation methods to the amount of missing data over the range of 1--20% missing values. We show that KNNimpute appears to provide a more robust and sensitive method for missing value estimation than SVDimpute, and both SVDimpute and KNNimpute surpass the commonly used row average method (as well as filling missing values with zeros). We report results of the comparative experiments and provide recommendations and tools for accurate estimation of missing microarray data under a variety of conditions.

Role of thioredoxin reductase in the Yap1p-dependent response to oxidative stress in Saccharomyces cerevisiae

The Saccharomyces cerevisiae Yap1p transcription factor is required for the H2O2-dependent activation of many antioxidant genes including the TRX2 gene encoding thioredoxin 2. To identify factors that regulate Yap1p activity, we carried out a genetic screen for mutants that show elevated expression of a TRX2-HIS3 fusion in the absence of H2O2. Two independent mutants isolated in this screen carried mutations in the TRR1 gene encoding thioredoxin reductase. Northern blot and whole-genome expression analysis revealed that the basal expression of most Yap1p targets and many other H2O2-inducible genes is elevated in Deltatrr1 mutants in the absence of external stress. In Deltatrr1 mutants treated with H2O2, the Yap1p targets, as well as genes comprising a general environmental stress response and genes encoding protein-folding chaperones, are hyperinduced. However, despite the elevated expression of genes encoding antioxidant enzymes, Deltatrr1 mutants are extremely sensitive to H2O2. The results suggest that cells lacking thioredoxin reductase have diminished capacity to detoxify oxidants and/or to repair oxidative stress-induced damage and that the thioredoxin system is involved in the redox regulation of Yap1p transcriptional activity.

Genomic binding sites of the yeast cell-cycle transcription factors SBF and MBF

Proteins interact with genomic DNA to bring the genome to life; and these interactions also define many functional features of the genome. SBF and MBF are sequence-specific transcription factors that activate gene expression during the G1/S transition of the cell cycle in yeast. SBF is a heterodimer of Swi4 and Swi6, and MBF is a heterodimer of Mbpl and Swi6 (refs 1, 3). The related Swi4 and Mbp1 proteins are the DNA-binding components of the respective factors, and Swi6 mayhave a regulatory function. A small number of SBF and MBF target genes have been identified. Here we define the genomic binding sites of the SBF and MBF transcription factors in vivo, by using DNA microarrays. In addition to the previously characterized targets, we have identified about 200 new putative targets. Our results support the hypothesis that SBF activated genes are predominantly involved in budding, and in membrane and cell-wall biosynthesis, whereas DNA replication and repair are the dominant functions among MBF activated genes. The functional specialization of these factors may provide a mechanism for independent regulation of distinct molecular processes that normally occur in synchrony during the mitotic cell cycle.

Saccharomyces Genome Database provides tools to survey gene expression and functional analysis data

Upon the completion of the SACCHAROMYCES: cerevisiae genomic sequence in 1996 [Goffeau,A. et al. (1997) NATURE:, 387, 5], several creative and ambitious projects have been initiated to explore the functions of gene products or gene expression on a genome-wide scale. To help researchers take advantage of these projects, the SACCHAROMYCES: Genome Database (SGD) has created two new tools, Function Junction and Expression Connection. Together, the tools form a central resource for querying multiple large-scale analysis projects for data about individual genes. Function Junction provides information from diverse projects that shed light on the role a gene product plays in the cell, while Expression Connection delivers information produced by the ever-increasing number of microarray projects. WWW access to SGD is available at genome-www.stanford. edu/Saccharomyces/.

Supervised harvesting of expression trees

BACKGROUND

We propose a new method for supervised learning from gene expression data. We call it 'tree harvesting'. This technique starts with a hierarchical clustering of genes, then models the outcome variable as a sum of the average expression profiles of chosen clusters and their products. It can be applied to many different kinds of outcome measures such as censored survival times, or a response falling in two or more classes (for example, cancer classes). The method can discover genes that have strong effects on their own, and genes that interact with other genes.

RESULTS

We illustrate the method on data from a lymphoma study, and on a dataset containing samples from eight different cancers. It identified some potentially interesting gene clusters. In simulation studies we found that the procedure may require a large number of experimental samples to successfully discover interactions.

CONCLUSIONS

Tree harvesting is a potentially useful tool for exploration of gene expression data and identification of interesting clusters of genes worthy of further investigation.

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.

Genomic expression programs in the response of yeast cells to environmental changes

We explored genomic expression patterns in the yeast Saccharomyces cerevisiae responding to diverse environmental transitions. DNA microarrays were used to measure changes in transcript levels over time for almost every yeast gene, as cells responded to temperature shocks, hydrogen peroxide, the superoxide-generating drug menadione, the sulfhydryl-oxidizing agent diamide, the disulfide-reducing agent dithiothreitol, hyper- and hypo-osmotic shock, amino acid starvation, nitrogen source depletion, and progression into stationary phase. A large set of genes (approximately 900) showed a similar drastic response to almost all of these environmental changes. Additional features of the genomic responses were specialized for specific conditions. Promoter analysis and subsequent characterization of the responses of mutant strains implicated the transcription factors Yap1p, as well as Msn2p and Msn4p, in mediating specific features of the transcriptional response, while the identification of novel sequence elements provided clues to novel regulators. Physiological themes in the genomic responses to specific environmental stresses provided insights into the effects of those stresses on the cell.

DNA microarray analysis of gene expression in response to physiological and genetic changes that affect tryptophan metabolism in Escherichia coli

We investigated the global changes in mRNA abundance in Escherichia coli elicited by various perturbations of tryptophan metabolism. To do so we printed DNA microarrays containing 95% of all annotated E. coli ORFs. We determined the expression profile that is predominantly dictated by the activity of the tryptophan repressor. Only three operons, trp, mtr, and aroH, exhibited appreciable expression changes consistent with this profile. The quantitative changes we observed in mRNA levels for the five genes of the trp operon were consistent within a factor of 2, with expectations based on established Trp protein levels. Several operons known to be regulated by the TyrR protein, aroF-tyrA, aroL, aroP, and aroG, were down-regulated on addition of tryptophan. TyrR can be activated by any one of the three aromatic amino acids. Only one operon, tnaAB, was significantly activated by the presence of tryptophan in the medium. We uncovered a plethora of likely indirect effects of changes in tryptophan metabolism on intracellular mRNA pools, most prominent of which was the sensitivity of arginine biosynthetic operons to tryptophan starvation.

Ongoing immunoglobulin somatic mutation in germinal center B cell-like but not in activated B cell-like diffuse large cell lymphomas

B cell diffuse large cell lymphoma (B-DLCL) is a heterogeneous group of tumors, based on significant variations in morphology, clinical presentation, and response to treatment. Gene expression profiling has revealed two distinct tumor subtypes of B-DLCL: germinal center B cell-like DLCL and activated B cell-like DLCL. In a separate study, we determined that B-DLCL can also be subdivided into two groups based on the presence or absence of ongoing Ig gene hypermutation. Here, we evaluated the correlation between these B-DLCL subtypes established by the two different methods. Fourteen primary B-DLCL cases were studied by gene expression profiling using DNA microarrays and for the presence of ongoing mutations in their Ig heavy chain gene. All seven cases classified as germinal center B cell-like DLCL by gene expression showed the presence of ongoing mutations in the Ig genes. Five of the seven cases classified by gene expression as activated B cell-like DLCL had no ongoing somatic mutations, whereas, in the remaining two cases, a single point mutation was observed in only 2 of 15 and 21 examined molecular clones of variable heavy (V(H)) chain gene, respectively. These two cases were distantly related to the rest of the activated B cell-like DLCL tumors by gene expression. Our findings validate the concept that lymphoid malignancies are derived from cells at discrete stages of normal lymphocyte maturation and that the malignant cells retain the genetic program of those normal cells.

Singular value decomposition for genome-wide expression data processing and modeling

We describe the use of singular value decomposition in transforming genome-wide expression data from genes x arrays space to reduced diagonalized "eigengenes" x "eigenarrays" space, where the eigengenes (or eigenarrays) are unique orthonormal superpositions of the genes (or arrays). Normalizing the data by filtering out the eigengenes (and eigenarrays) that are inferred to represent noise or experimental artifacts enables meaningful comparison of the expression of different genes across different arrays in different experiments. Sorting the data according to the eigengenes and eigenarrays gives a global picture of the dynamics of gene expression, in which individual genes and arrays appear to be classified into groups of similar regulation and function, or similar cellular state and biological phenotype, respectively. After normalization and sorting, the significant eigengenes and eigenarrays can be associated with observed genome-wide effects of regulators, or with measured samples, in which these regulators are overactive or underactive, respectively.

Molecular portraits of human breast tumours

Human breast tumours are diverse in their natural history and in their responsiveness to treatments. Variation in transcriptional programs accounts for much of the biological diversity of human cells and tumours. In each cell, signal transduction and regulatory systems transduce information from the cell's identity to its environmental status, thereby controlling the level of expression of every gene in the genome. Here we have characterized variation in gene expression patterns in a set of 65 surgical specimens of human breast tumours from 42 different individuals, using complementary DNA microarrays representing 8,102 human genes. These patterns provided a distinctive molecular portrait of each tumour. Twenty of the tumours were sampled twice, before and after a 16-week course of doxorubicin chemotherapy, and two tumours were paired with a lymph node metastasis from the same patient. Gene expression patterns in two tumour samples from the same individual were almost always more similar to each other than either was to any other sample. Sets of co-expressed genes were identified for which variation in messenger RNA levels could be related to specific features of physiological variation. The tumours could be classified into subtypes distinguished by pervasive differences in their gene expression patterns.

Analysis of topoisomerase function in bacterial replication fork movement: use of DNA microarrays

We used DNA microarrays of the Escherichia coli genome to trace the progression of chromosomal replication forks in synchronized cells. We found that both DNA gyrase and topoisomerase IV (topo IV) promote replication fork progression. When both enzymes were inhibited, the replication fork stopped rapidly. The elongation rate with topo IV alone was 1/3 of normal. Genetic data confirmed and extended these results. Inactivation of gyrase alone caused a slow stop of replication. Topo IV activity was sufficient to prevent accumulation of (+) supercoils in plasmid DNA in vivo, suggesting that topo IV can promote replication by removing (+) supercoils in front of the chromosomal fork.

Two yeast forkhead genes regulate the cell cycle and pseudohyphal growth

There are about 800 genes in Saccharomyces cerevisiae whose transcription is cell-cycle regulated. Some of these form clusters of co-regulated genes. The 'CLB2' cluster contains 33 genes whose transcription peaks early in mitosis, including CLB1, CLB2, SWI5, ACE2, CDC5, CDC20 and other genes important for mitosis. Here we find that the genes in this cluster lose their cell cycle regulation in a mutant that lacks two forkhead transcription factors, Fkh1 and Fkh2. Fkh2 protein is associated with the promoters of CLB2, SWI5 and other genes of the cluster. These results indicate that Fkh proteins are transcription factors for the CLB2 cluster. The fkh1 fkh2 mutant also displays aberrant regulation of the 'SIC1' cluster, whose member genes are expressed in the M-G1 interval and are involved in mitotic exit. This aberrant regulation may be due to aberrant expression of the transcription factors Swi5 and Ace2, which are members of the CLB2 cluster and controllers of the SIC1 cluster. Thus, a cascade of transcription factors operates late in the cell cycle. Finally, the fkh1 fkh2 mutant displays a constitutive pseudohyphal morphology, indicating that Fkh1 and Fkh2 may help control the switch to this mode of growth.

Genome-wide characterization of the Zap1p zinc-responsive regulon in yeast

The Zap1p transcription factor senses cellular zinc status and increases expression of its target genes in response to zinc deficiency. Previously known Zap1p-regulated genes encode the Zrt1p, Zrt2p, and Zrt3p zinc transporter genes and Zap1p itself. To allow the characterization of additional genes in yeast important for zinc homeostasis, a systematic study of gene expression on the genome-wide scale was used to identify other Zap1p target genes. Using a combination of DNA microarrays and a computer-assisted analysis of shared motifs in the promoters of similarly regulated genes, we identified 46 genes that are potentially regulated by Zap1p. Zap1p-regulated expression of seven of these newly identified target genes was confirmed independently by using lacZ reporter fusions, suggesting that many of the remaining candidate genes are also Zap1p targets. Our studies demonstrate the efficacy of this combined approach to define the regulon of a specific eukaryotic transcription factor.

Lack of evidence for an association between alpha-adducin and blood pressure regulation in Asian populations

Recent studies have found the tryptophan allele of a glycine to tryptophan polymorphism at position 460 (G460W) of the alpha-adducin protein to be associated with essential hypertension in European populations. We examined whether the tryptophan allele is associated with hypertension in a different population, comprised of subjects of Chinese origin from Taiwan, and Chinese and Japanese origin from the San Francisco Bay area and Hawaii. We adapted the 5' allelic discrimination assay or TaqMan to type individuals for the G460W polymorphism, and using this method we typed more than 1000 individuals. The frequency of the W allele was slightly increased in the treated subjects in the Chinese population (0.458 v 0.423) but not the Japanese population (0.549 v 0.558). We considered dominant, recessive, and additive models in our analysis. There was a significant result for a recessive model for systolic blood pressure in the Chinese population (chi2 6.84, df = 2, P < .05), but only suggestive evidence for diastolic blood pressure (chi2 3.30). In contrast, in the Japanese population, there was no evidence for a positive association under any model. For the combined Chinese and Japanese samples, the evidence for association with alpha-adducin was not significant.

Gene ontology: tool for the unification of biology. The Gene Ontology Consortium

Genomic sequencing has made it clear that a large fraction of the genes specifying the core biological functions are shared by all eukaryotes. Knowledge of the biological role of such shared proteins in one organism can often be transferred to other organisms. The goal of the Gene Ontology Consortium is to produce a dynamic, controlled vocabulary that can be applied to all eukaryotes even as knowledge of gene and protein roles in cells is accumulating and changing. To this end, three independent ontologies accessible on the World-Wide Web (http://www.geneontology.org) are being constructed: biological process, molecular function and cellular component.

Structure-function relationships in yeast tubulins

A comprehensive set of clustered charged-to-alanine mutations was generated that systematically alter TUB1, the major alpha-tubulin gene of Saccharomyces cerevisiae. A variety of phenotypes were observed, including supersensitivity and resistance to the microtubule-destabilizing drug benomyl, lethality, and cold- and temperature-sensitive lethality. Many of the most benomyl-sensitive tub1 alleles were synthetically lethal in combination with tub3Delta, supporting the idea that benomyl supersensitivity is a rough measure of microtubule instability and/or insufficiency in the amount of alpha-tubulin. The systematic tub1 mutations were placed, along with the comparable set of tub2 mutations previously described, onto a model of the yeast alpha-beta-tubulin dimer based on the three-dimensional structure of bovine tubulin. The modeling revealed a potential site for binding of benomyl in the core of beta-tubulin. Residues whose mutation causes cold sensitivity were concentrated at the lateral and longitudinal interfaces between adjacent subunits. Residues that affect binding of the microtubule-binding protein Bim1p form a large patch across the exterior-facing surface of alpha-tubulin in the model. Finally, the positions of the mutations suggest that proximity to the alpha-beta interface may account for the finding of synthetic lethality of five viable tub1 alleles with the benomyl-resistant but otherwise entirely viable tub2-201 allele.

Large-scale identification of secreted and membrane-associated gene products using DNA microarrays

Membrane-associated and secreted proteins are an important class of proteins and include receptors, transporters, adhesion molecules, hormones and cytokines. Although algorithms have been developed to recognize potential amino-terminal membrane-targeting signals or transmembrane domains in protein sequences, their accuracy is limited and they require knowledge of the entire coding sequence, including the N terminus, which is not currently available for most of the genes in most organisms, including human. Several experimental approaches for identifying secreted and membrane proteins have been described, but none have taken a comprehensive genomic approach. Furthermore, none of these methods allow easy classification of clones from arrayed cDNA libraries, for which large-scale gene-expression data are now becoming available through the use of DNA microarrays. We describe here a rapid and efficient method for identifying genes that encode secreted or membrane proteins. mRNA species bound to membrane-associated polysomes were separated from other mRNAs by sedimentation equilibrium or sedimentation velocity. The distribution of individual transcripts in the 'membrane-bound' and 'cytosolic' fractions was quantitated for thousands of genes by hybridization to DNA microarrays. Transcripts known to encode secreted or membrane proteins were enriched in the membrane-bound fractions, whereas those known to encode cytoplasmic proteins were enriched in the fractions containing mRNAs associated with free and cytoplasmic ribosomes. On this basis, we identified over 275 human genes and 285 yeast genes that are likely to encode previously unrecognized secreted or membrane proteins.

Gene ontology: tool for the unification of biology. The Gene Ontology Consortium

Genomic sequencing has made it clear that a large fraction of the genes specifying the core biological functions are shared by all eukaryotes. Knowledge of the biological role of such shared proteins in one organism can often be transferred to other organisms. The goal of the Gene Ontology Consortium is to produce a dynamic, controlled vocabulary that can be applied to all eukaryotes even as knowledge of gene and protein roles in cells is accumulating and changing. To this end, three independent ontologies accessible on the World-Wide Web (http://www.geneontology.org) are being constructed: biological process, molecular function and cellular component.

Desferrioxamine-mediated iron uptake in Saccharomyces cerevisiae. Evidence for two pathways of iron uptake

In the yeast Saccharomyces cerevisiae, uptake of iron is largely regulated by the transcription factor Aft1. cDNA microarrays were used to identify new iron and AFT1-regulated genes. Four homologous genes regulated as part of the AFT1-regulon (ARN1-4) were predicted to encode members of a subfamily of the major facilitator superfamily of transporters. These genes were predicted to encode proteins with 14 membrane spanning domains and were from 26 to 53% identical at the amino acid level. ARN3 is identical to SIT1, which is reported to encode a ferrioxamine B permease. Deletion of ARN3 did not prevent yeast from using ferrioxamine B as an iron source; however, deletion of ARN3 and FET3, a component of the high affinity ferrous iron transport system, did prevent uptake of ferrioxamine-bound iron and growth on ferrioxamine as an iron source. The siderophore-mediated transport system and the high affinity ferrous iron transport system were localized to separate cellular compartments. Epitope-tagged Arn3p was expressed in intracellular vesicles that co-sediment with the endosomal protein Pep12. In contrast, Fet3p was expressed on the plasma membrane and was digested by extracellular proteases. These data indicate that S. cerevisiae has two pathways for ferrrioxamine-mediated iron uptake, one occurring at the plasma membrane and the other occurring in an intracellular compartment.

Systematic variation in gene expression patterns in human cancer cell lines

We used cDNA microarrays to explore the variation in expression of approximately 8,000 unique genes among the 60 cell lines used in the National Cancer Institute's screen for anti-cancer drugs. Classification of the cell lines based solely on the observed patterns of gene expression revealed a correspondence to the ostensible origins of the tumours from which the cell lines were derived. The consistent relationship between the gene expression patterns and the tissue of origin allowed us to recognize outliers whose previous classification appeared incorrect. Specific features of the gene expression patterns appeared to be related to physiological properties of the cell lines, such as their doubling time in culture, drug metabolism or the interferon response. Comparison of gene expression patterns in the cell lines to those observed in normal breast tissue or in breast tumour specimens revealed features of the expression patterns in the tumours that had recognizable counterparts in specific cell lines, reflecting the tumour, stromal and inflammatory components of the tumour tissue. These results provided a novel molecular characterization of this important group of human cell lines and their relationships to tumours in vivo.

A gene expression database for the molecular pharmacology of cancer

We used cDNA microarrays to assess gene expression profiles in 60 human cancer cell lines used in a drug discovery screen by the National Cancer Institute. Using these data, we linked bioinformatics and chemoinformatics by correlating gene expression and drug activity patterns in the NCI60 lines. Clustering the cell lines on the basis of gene expression yielded relationships very different from those obtained by clustering the cell lines on the basis of their response to drugs. Gene-drug relationships for the clinical agents 5-fluorouracil and L-asparaginase exemplify how variations in the transcript levels of particular genes relate to mechanisms of drug sensitivity and resistance. This is the first study to integrate large databases on gene expression and molecular pharmacology.

Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling

Diffuse large B-cell lymphoma (DLBCL), the most common subtype of non-Hodgkin's lymphoma, is clinically heterogeneous: 40% of patients respond well to current therapy and have prolonged survival, whereas the remainder succumb to the disease. We proposed that this variability in natural history reflects unrecognized molecular heterogeneity in the tumours. Using DNA microarrays, we have conducted a systematic characterization of gene expression in B-cell malignancies. Here we show that there is diversity in gene expression among the tumours of DLBCL patients, apparently reflecting the variation in tumour proliferation rate, host response and differentiation state of the tumour. We identified two molecularly distinct forms of DLBCL which had gene expression patterns indicative of different stages of B-cell differentiation. One type expressed genes characteristic of germinal centre B cells ('germinal centre B-like DLBCL'); the second type expressed genes normally induced during in vitro activation of peripheral blood B cells ('activated B-like DLBCL'). Patients with germinal centre B-like DLBCL had a significantly better overall survival than those with activated B-like DLBCL. The molecular classification of tumours on the basis of gene expression can thus identify previously undetected and clinically significant subtypes of cancer.

Integrating functional genomic information into the Saccharomyces genome database

The Saccharomyces Genome Database (SGD) stores and organizes information about the nearly 6200 genes in the yeast genome. The information is organized around the 'locus page' and directs users to the detailed information they seek. SGD is endeavoring to integrate the existing information about yeast genes with the large volume of data generated by functional analyses that are beginning to appear in the literature and on web sites. New features will include searches of systematic analyses and Gene Summary Paragraphs that succinctly review the literature for each gene. In addition to current information, such as gene product and phenotype descriptions, the new locus page will also describe a gene product's cellular process, function and localization using a controlled vocabulary developed in collaboration with two other model organism databases. We describe these developments in SGD through the newly reorganized locus page. The SGD is accessible via the WWW at http://genome-www.stanford.edu/Saccharomyces/

'Gene shaving' as a method for identifying distinct sets of genes with similar expression patterns

BACKGROUND

Large gene expression studies, such as those conducted using DNA arrays, often provide millions of different pieces of data. To address the problem of analyzing such data, we describe a statistical method, which we have called 'gene shaving'. The method identifies subsets of genes with coherent expression patterns and large variation across conditions. Gene shaving differs from hierarchical clustering and other widely used methods for analyzing gene expression studies in that genes may belong to more than one cluster, and the clustering may be supervised by an outcome measure. The technique can be 'unsupervised', that is, the genes and samples are treated as unlabeled, or partially or fully supervised by using known properties of the genes or samples to assist in finding meaningful groupings.

RESULTS

We illustrate the use of the gene shaving method to analyze gene expression measurements made on samples from patients with diffuse large B-cell lymphoma. The method identifies a small cluster of genes whose expression is highly predictive of survival.

CONCLUSIONS

The gene shaving method is a potentially useful tool for exploration of gene expression data and identification of interesting clusters of genes worth further investigation.