Cluster analysis and display of genome-wide expression patterns

A system of cluster analysis for genome-wide expression data from DNA microarray hybridization is described that uses standard statistical algorithms to arrange genes according to similarity in pattern of gene expression. The output is displayed graphically, conveying the clustering and the underlying expression data simultaneously in a form intuitive for biologists. We have found in the budding yeast Saccharomyces cerevisiae that clustering gene expression data groups together efficiently genes of known similar function, and we find a similar tendency in human data. Thus patterns seen in genome-wide expression experiments can be interpreted as indications of the status of cellular processes. Also, coexpression of genes of known function with poorly characterized or novel genes may provide a simple means of gaining leads to the functions of many genes for which information is not available currently.

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.

Isolation and characterization of a pseudoautosomal region-specific genetic marker in C57BL/6 mice using genomic representational difference analysis.

Representational difference analysis was used to identify strain-specific differences in the pseudoautosomal region (PAR) of mouse X and Y chromosomes. One second generation (C57BL/6 x Mus spretus) x Mus spretus interspecific backcross male carrying the C57BL/6 (B6) PAR was used for tester DNA. DNA from five backcross males from the same generation that were M. spretus-type for the PAR was pooled for the driver. A cloned probe designated B6-38 was recovered that is B6-specific in Southern analysis. Analysis of genomic DNA from several inbred strains of laboratory mice and diverse Mus species and subspecies identified a characteristic Pst I pattern of fragment sizes that is present only in the C57BL family of strains. Hybridization was observed with sequences in DBA/2J and to a limited extent with Mus musculus (PWK strain) and Mus castaneus DNA. No hybridization was observed in DNA of different Mus species, M. spretus, M. hortulanus, and M. caroli. Genetic analyses of B6-38 was conducted using C57BL congenic males that carry M. spretus alleles for distal X chromosome loci and the PAR and outcrosses of heterozygous congenic females with M. spretus. These analyses demonstrated that the B6-38 sequences were inherited with both the X and Y chromosome. B6-38 sequences were genetically mapped as a locus within the PAR using two interspecific backcrosses. The locus defined by B6-38 is designated DXYRp1. Preliminary analyses of recombination between the distal X chromosome gene amelogenin (Amg) and the PAR loci for either TelXY or sex chromosome association (Sxa) suggest that the locus DXYRp1 maps to the distal portion of the PAR.

Nodulating strains of Rhizobium loti arise through chromosomal symbiotic gene transfer in the environment.

Rhizobia were isolated from nodules off a stand of Lotus corniculatus established with a single inoculant strain, ICMP3153, 7 years earlier in an area devoid of naturalized Rhizobium loti. The isolates showed diversity in growth rate, Spe I fingerprint of genomic DNA, and hybridization pattern to genomic DNA probes. The 19% of isolates that grew at the same rate as strain ICMP3153 were the only isolates that had the same fingerprint as strain ICMP3153. Sequencing of part of the 16S rRNA gene of several diverse isolates confirmed that they were not derived from the inoculant strain. Nevertheless, all non-ICMP3153 strains gave EcoRI and Spe I hybridization patterns identical to ICMP3153 when hybridized to nodulation gene cosmids. Hybridization of digests generated by the very rare cutting enzyme Swa I revealed that the symbiotic DNA region (at least 105 kb) was chromosomally integrated in the strains. The results suggest that the diverse strains arose by transfer of chromosomal symbiotic genes from ICMP3153 to nonsymbiotic rhizobia in the environment.

Types of Listeria monocytogenes predicted by the positions of EcoRI cleavage sites relative to ribosomal RNA sequences.

By using taxonomic characters derived from EcoRI restriction endonuclease digestion of genomic DNA and hybridization with a labeled rRNA operon from Escherichia coli, a polymorphic structure of Listeria monocytogenes, characterized by fragments with different frequencies of occurrence, was observed. This structure was expanded by creating predicted patterns through a recursive process of observation, expectation, prediction, and assessment of completeness. This process was applied, in turn, to normalized strain patterns, fragment bands, and positions of EcoRI recognition sites relative to rRNA regions. Analysis of 1346 strains provided observed patterns, fragment sizes, and their frequencies of occurrence in the patterns. Fragment size statistics led to the creation of unobserved combinations of bands, predicted pattern types. The observed fragment bands revealed positions of EcoRI sites relative to rRNA sequences. Each EcoRI site had a frequency of occurrence, and unobserved fragment sizes were postulated on the basis of knowing the restriction site locations. The result of the recursion process applied to the components of the strain data was an extended classification with observed and predicted members.

Subtraction hybridization identifies a transformation progression-associated gene PEG-3 with sequence homology to a growth arrest and DNA damage-inducible gene

Cancer is a progressive multigenic disorder characterized by defined changes in the transformed phenotype that culminates in metastatic disease. Determining the molecular basis of progression should lead to new opportunities for improved diagnostic and therapeutic modalities. Through the use of subtraction hybridization, a gene associated with transformation progression in virus- and oncogene-transformed rat embryo cells, progression elevated gene-3 (PEG-3), has been cloned. PEG-3 shares significant nucleotide and amino acid sequence homology with the hamster growth arrest and DNA damage-inducible gene gadd34 and a homologous murine gene, MyD116, that is induced during induction of terminal differentiation by interleukin-6 in murine myeloid leukemia cells. PEG-3 expression is elevated in rodent cells displaying a progressed-transformed phenotype and in rodent cells transformed by various oncogenes, including Ha-ras, v-src, mutant type 5 adenovirus (Ad5), and human papilloma virus type 18. The PEG-3 gene is transcriptionally activated in rodent cells, as is gadd34 and MyD116, after treatment with DNA damaging agents, including methyl methanesulfonate and γ-irradiation. In contrast, only PEG-3 is transcriptionally active in rodent cells displaying a progressed phenotype. Although transfection of PEG-3 into normal and Ad5-transformed cells only marginally suppresses colony formation, stable overexpression of PEG-3 in Ad5-transformed rat embryo cells elicits the progression phenotype. These results indicate that PEG-3 is a new member of the gadd and MyD gene family with similar yet distinct properties and this gene may directly contribute to the transformation progression phenotype. Moreover, these studies support the hypothesis that constitutive expression of a DNA damage response may mediate cancer progression.

Hybridization of single-stranded DNA targets to immobilized complementary DNA probes: comparison of hairpin versus linear capture probes

A microtiter-based assay system is described in which DNA hairpin probes with dangling ends and single-stranded, linear DNA probes were immobilized and compared based on their ability to capture single-strand target DNA. Hairpin probes consisted of a 16 bp duplex stem, linked by a T2-biotin·dT-T2 loop. The third base was a biotinylated uracil (UB) necessary for coupling to avidin coated microtiter wells. The capture region of the hairpin was a 3′ dangling end composed of either 16 or 32 bases. Fundamental parameters of the system, such as probe density and avidin adsorption capacity of the plates were characterized. The target DNA consisted of 65 bases whose 3′ end was complementary to the dangling end of the hairpin or to the linear probe sequence. The assay system was employed to measure the time dependence and thermodynamic stability of target hybridization with hairpin and linear probes. Target molecules were labeled with either a 5′-FITC, or radiolabeled with [γ-33P]ATP and captured by either linear or hairpin probes affixed to the solid support. Over the range of target concentrations from 10 to 640 pmol hybridization rates increased with increasing target concentration, but varied for the different probes examined. Hairpin probes displayed higher rates of hybridization and larger equilibrium amounts of captured targets than linear probes. At 25 and 45°C, rates of hybridization were better than twice as great for the hairpin compared with the linear capture probes. Hairpin–target complexes were also more thermodynamically stable. Binding free energies were evaluated from the observed equilibrium constants for complex formation. Results showed the order of stability of the probes to be: hairpins with 32 base dangling ends > hairpin probes with l6 base dangling ends > 16 base linear probes > 32 base linear probes. The physical characteristics of hairpins could offer substantial advantages as nucleic acid capture moieties in solid support based hybridization systems.

Chloroplast DNA evidence of colonization, adaptive radiation, and hybridization in the evolution of the Macaronesian flora.

Most evolutionary studies of oceanic islands have focused on the Pacific Ocean. There are very few examples from the Atlantic archipelagos, especially Macaronesia, despite their unusual combination of features, including a close proximity to the continent, a broad range of geological ages, and a biota linked to a source area that existed in the Mediterranean basin before the late Tertiary. A chloroplast DNA (cpDNA) restriction site analysis of Argyranthemum (Asteraceae: Anthemideae), the largest endemic genus of plants of any volcanic archipelago in the Atlantic Ocean, was performed to examine patterns of plant evolution in Macaronesia. cpDNA data indicated that Argyranthemum is a monophyletic group that has speciated recently. The cpDNA tree showed a weak correlation with the current sectional classification and insular distribution. Two major cpDNA lineages were identified. One was restricted to northern archipelagos--e.g., Madeira, Desertas, and Selvagens--and the second comprised taxa endemic to the southern archipelago--e.g., the Canary Islands. The two major radiations identified in the Canaries are correlated with distinct ecological habitats; one is restricted to ecological zones under the influence of the northeastern trade winds and the other to regions that are not affected by these winds. The patterns of phylogenetic relationships in Argyranthemum indicate that interisland colonization between similar ecological zones is the main mechanism for establishing founder populations. This phenomenon, combined with rapid radiation into distinct ecological zones and interspecific hybridization, is the primary explanation for species diversification.

Real-time detection of DNA hybridization and melting on oligonucleotide arrays by using optical wave guides.

The challenge of the Human Genome Project is to increase the rate of DNA sequence acquisition by two orders of magnitude to complete sequencing of the human genome by the year 2000. The present work describes a rapid detection method using a two-dimensional optical wave guide that allows measurement of real-time binding or melting of a light-scattering label on a DNA array. A particulate label on the target DNA acts as a light-scattering source when illuminated by the evanescent wave of the wave guide and only the label bound to the surface generates a signal. Imaging/visual examination of the scattered light permits interrogation of the entire array simultaneously. Hybridization specificity is equivalent to that obtained with a conventional system using autoradiography. Wave guide melting curves are consistent with those obtained in the liquid phase and single-base discrimination is facile. Dilution experiments showed an apparent lower limit of detection at 0.4 nM oligonucleotide. This performance is comparable to the best currently known fluorescence-based systems. In addition, wave guide detection allows manipulation of hybridization stringency during detection and thereby reduces DNA chip complexity. It is anticipated that this methodology will provide a powerful tool for diagnostic applications that require rapid cost-effective detection of variations from known sequences.

Mutagenicity in Escherichia coli of the major DNA adduct derived from the endogenous mutagen malondialdehyde

The spectrum of mutations induced by the naturally occurring DNA adduct pyrimido[1,2-α]purin-10(3H)-one (M1G) was determined by site-specific approaches using M13 vectors replicated in Escherichia coli. M1G was placed at position 6256 in the (−)-strand of M13MB102 by ligating the oligodeoxynucleotide 5′-GGT(M1G)TCCG-3′ into a gapped-duplex derivative of the vector. Unmodified and M1G-modified genomes containing either a cytosine or thymine at position 6256 of the (+)-strand were transformed into repair-proficient and repair-deficient E. coli strains, and base pair substitutions were quantitated by hybridization analysis. Modified genomes containing a cytosine opposite M1G resulted in roughly equal numbers of M1G→A and M1G→T mutations with few M1G→C mutations. The total mutation frequency was ≈1%, which represents a 500-fold increase in mutations compared with unmodified M13MB102. Transformation of modified genomes containing a thymine opposite M1G allowed an estimate to be made of the ability of M1G to block replication. The (−)-strand was replicated >80% of the time in the unadducted genome but only 20% of the time when M1G was present. Correction of the mutation frequency for the strand bias of replication indicated that the actual frequency of mutations induced by M1G was 18%. Experiments using E. coli with different genetic backgrounds indicated that the SOS response enhances the mutagenicity of M1G and that M1G is a substrate for repair by the nucleotide excision repair complex. These studies indicate that M1G, which is present endogenously in DNA of healthy human beings, is a strong block to replication and an efficient premutagenic lesion.

Exploring drug-induced alterations in gene expression in Mycobacterium tuberculosis by microarray hybridization

Tuberculosis is a chronic infectious disease that is transmitted by cough-propelled droplets that carry the etiologic bacterium, Mycobacterium tuberculosis. Although currently available drugs kill most isolates of M. tuberculosis, strains resistant to each of these have emerged, and multiply resistant strains are increasingly widespread. The growing problem of drug resistance combined with a global incidence of seven million new cases per year underscore the urgent need for new antituberculosis therapies. The recent publication of the complete sequence of the M. tuberculosis genome has made possible, for the first time, a comprehensive genomic approach to the biology of this organism and to the drug discovery process. We used a DNA microarray containing 97% of the ORFs predicted from this sequence to monitor changes in M. tuberculosis gene expression in response to the antituberculous drug isoniazid. Here we show that isoniazid induced several genes that encode proteins physiologically relevant to the drug’s mode of action, including an operonic cluster of five genes encoding type II fatty acid synthase enzymes and fbpC, which encodes trehalose dimycolyl transferase. Other genes, not apparently within directly affected biosynthetic pathways, also were induced. These genes, efpA, fadE23, fadE24, and ahpC, likely mediate processes that are linked to the toxic consequences of the drug. Insights gained from this approach may define new drug targets and suggest new methods for identifying compounds that inhibit those targets.

Fluorescence in situ hybridization analysis of keratinocyte growth factor gene amplification and dispersion in evolution of great apes and humans

Keratinocyte growth factor (KGF) is a member of the fibroblast growth factor family. Portions of the gene encoding KGF were amplified during primate evolution and are present in multiple nonprocessed copies in the human genome. Nucleotide analysis of a representative sampling of these KGF-like sequences indicated that they were at least 95% identical to corresponding regions of the KGF gene. To localize these sequences to specific chromosomal sites in human and higher primates, we used fluorescence in situ hybridization. In human, using a cosmid probe encoding KGF exon 1, we assigned the location of the KGF gene to chromosome 15q15–21.1. In addition, copies of KGF-like sequences hybridizing only with a cosmid probe encoding exons 2 and 3 were localized to dispersed sites on chromosome 2q21, 9p11, 9q12–13, 18p11, 18q11, 21q11, and 21q21.1. The distribution of KGF-like sequences suggests a role for alphoid DNA in their amplification and dispersion. In chimpanzee, KGF-like sequences were observed at five chromosomal sites, which were each homologous to sites in human, while in gorilla, a subset of four of these homologous sites was identified; in orangutan two sites were identified, while gibbon exhibited only a single site. The chromosomal localization of KGF sequences in human and great ape genomes indicates that amplification and dispersion occurred in multiple discrete steps, with initial KGF gene duplication and dispersion taking place in gibbon and involving loci corresponding to human chromosomes 15 and 21. These findings support the concept of a closer evolutionary relationship of human and chimpanzee and a possible selective pressure for such dispersion during the evolution of higher primates.

Highly conservative reciprocal translocations formed by apparent joining of exchanged DNA double-strand break ends

Chromosomal translocations induced by ionizing radiation and radiomimetic drugs are thought to arise by incorrect joining of DNA double-strand breaks. To dissect such misrepair events at a molecular level, large-scale, bleomycin-induced rearrangements in the aprt gene of Chinese hamster ovary D422 cells were mapped, the breakpoints were sequenced, and the original non-aprt parental sequences involved in each rearrangement were recovered from nonmutant cells. Of seven rearrangements characterized, six were reciprocal exchanges between aprt and unrelated sequences. Consistent with a mechanism involving joining of exchanged double-strand break ends, there was, in most cases, no homology between the two parental sequences, no overlap in sequences retained at the two newly formed junctions, and little or no loss of parental sequences (usually ≤2 bp) at the breakpoints. The breakpoints were strongly correlated (P < 0.0001) with expected sites of bleomycin-induced, double-strand breaks. Fluorescence in situ hybridization indicated that, in six of the mutants, the rearrangement was accompanied by a chromosomal translocation at the aprt locus, because upstream and downstream flanking sequences were detected on separate chromosomes. The results suggest that repair of free radical-mediated, double-strand breaks in confluence-arrested cells is effected by a conservative, homology-independent, end-joining pathway that does not involve single-strand intermediate and that misjoining of exchanged ends by this pathway can directly result in chromosomal translocations.

PCR-based subtractive hybridization and differences in gene content among strains of Helicobacter pylori

Genes that are characteristic of only certain strains of a bacterial species can be of great biologic interest. Here we describe a PCR-based subtractive hybridization method for efficiently detecting such DNAs and apply it to the gastric pathogen Helicobacter pylori. Eighteen DNAs specific to a monkey-colonizing strain (J166) were obtained by subtractive hybridization against an unrelated strain whose genome has been fully sequenced (26695). Seven J166-specific clones had no DNA sequence match to the 26695 genome, and 11 other clones were mixed, with adjacent patches that did and did not match any sequences in 26695. At the protein level, seven clones had homology to putative DNA restriction-modification enzymes, and two had homology to putative metabolic enzymes. Nine others had no database match with proteins of assigned function. PCR tests of 13 unrelated H. pylori strains by using primers specific for 12 subtracted clones and complementary Southern blot hybridizations indicated that these DNAs are highly polymorphic in the H. pylori population, with each strain yielding a different pattern of gene-specific PCR amplification. The search for polymorphic DNAs, as described here, should help identify previously unknown virulence genes in pathogens and provide new insights into microbial genetic diversity and evolution.

Conserved characteristics of heterochromatin-forming DNA at the 15q11-q13 imprinting center

Nuclear matrix binding assays (NMBAs) define certain DNA sequences as matrix attachment regions (MARs), which often have cis-acting epigenetic regulatory functions. We used NMBAs to analyze the functionally important 15q11-q13 imprinting center (IC). We find that the IC is composed of an unusually high density of MARs, located in close proximity to the germ line elements that are proposed to direct imprint switching in this region. Moreover, we find that the organization of MARs is the same at the homologous mouse locus, despite extensive divergence of DNA sequence. MARs of this size are not usually associated with genes but rather with heterochromatin-forming areas of the genome. In contrast, the 15q11-q13 region contains multiple transcribed genes and is unusual for being subject to genomic imprinting, causing the maternal chromosome to be more transcriptionally silent, methylated, and late replicating than the paternal chromosome. We suggest that the extensive MAR sequences at the IC are organized as heterochromatin during oogenesis, an organization disrupted during spermatogenesis. Consistent with this model, multicolor fluorescence in situ hybridization to halo nuclei demonstrates a strong matrix association of the maternal IC, whereas the paternal IC is more decondensed, extending into the nuclear halo. This model also provides a mechanism for spreading of the imprinting signal, because heterochromatin at the IC on the maternal chromosome may exert a suppressive position effect in cis. We propose that the germ line elements at the 15q11-q13 IC mediate their effects through the candidate heterochromatin-forming DNA identified in this study.