Chips around the world: Proceedings from the Nature Genetics Microarray Meeting
2000; American Physical Society; Volume: 2; Issue: 2 Linguagem: Inglês
10.1152/physiolgenomics.2000.2.2.53
ISSN1531-2267
Autores Tópico(s)Gene expression and cancer classification
ResumoNews and ReportsChips around the world: Proceedings from the Nature Genetics Microarray MeetingFERNANDO DANGONDFERNANDO DANGONDCenter for Neurologic Diseases, Harvard Institutes of Medicine, Brigham and Women's Hospital, Boston, Massachusetts 02115Published Online:13 Mar 2000https://doi.org/10.1152/physiolgenomics.2000.2.2.53MoreSectionsPDF (58 KB)Download PDF ToolsExport citationAdd to favoritesGet permissionsTrack citations ShareShare onFacebookTwitterLinkedInEmail an overwhelming response to the invitation by the journal Nature Genetics for the submission of DNA microarray work was evident in the recent “Microarray Meeting: Technology, Application and Analysis,” which took place in Scottsdale, AZ, September 22–25, 1999. A reflection of a rapidly expanding field, the meeting was attended by many of the outstanding contributors and pioneers of this technology. The meeting was sponsored by Affymetrix, the National Cancer Institute (NCI), and the National Human Genome Research Institute (NHGRI), as well as multiple other companies dedicated to supporting the DNA microarray biotechnology field (see end of this report). The meeting also featured over 200 poster presentations from international scientists, covering a wide range of functional genomics-related applications. The following is a report of the conference highlights.The group was welcomed by Barbara Cohen (Editor, Nature Genetics), who gave a general summary of some of the experimental difficulties encountered with DNA microarrays. Dr. Cohen also mentioned that there is a need to establish standards for experimental design and data validation, to understand the magnitude of the bioinformatic challenge of large data sets, and to make microarray data comparable for publication.Edwin Southern (University of Oxford) began his talk by explaining how there is considerable room for improvement of current technologies. He discussed the physical characteristics of DNA arrays and how the density or “crowding” of probes binding to a solid surface may affect the results. He also explained that helically configured bases of RNA with minimal perturbation of their conformation may bind better to the probe by forming a stable structure, with the secondary and tertiary structures being the main determinants of the extent of hybridization, as opposed to the base composition or sequence. Fragmenting the target cRNA into short lengths solves some of the problems caused by conformational changes. He acknowledged that cross-hybridization continues to be an unsolved problem with this technology. Dr. Southern also discussed how microarrays can be used to select antisense reagents, helping solve the current problems of having poor predictability for the efficacy of an antisense sequence.Hans Lerach (Max Plank-Institut fur Molekulare Genetik, Germany) explained that the study of selected genes should include information gathering about DNA and mRNA sequence, expression patterns at the protein and RNA levels, protein-interacting characteristics, protein structure, and phenotype analysis. He described some of his work with the generation of nonredundant clone libraries and their spotting covalently to glass using piezoelectric technology, to compare diseased vs. normal tissues. Finally, he described the cloning of cDNA libraries routinely into expression vectors as a preliminary step to examine protein expression and adaptation of “dished out” proteins to high-throughput protein assays, including the use of antibodies for protein microarrays.The representation of sample to sample comparisons in multidimensional scaling graphs was presented by Michael Bittner (Laboratory of Cancer Genetics, NHGRI). He emphasized that cooperative efforts to expedite microarray technology data processing was the key to success. Clone purification and sequence verification are necessary but do not guarantee the same level of confidence measurement for individual clones spotted in the arrays. For categorization of coregulated genes as potential functional partners, one must also take into account that genes are expressed differently in different tissues and undergo quantitative and qualitative modulation. In addition, the signal-to-noise ratio calculation is more complex for minimally expressed genes that approach background intensity levels.Walter Klimecki (Motorola BioChip Systems) discussed hydrogel-based microarrays that incorporate a primer extension reaction of amplicons. This technology has been applied to examining mitochondrial single-nucleotide polymorphisms (SNPs) with high accuracy and family linkage from maternal lines in groups characterized by excessive obesity and coming from a founder population.The use of thermal inkjet technology for fabricating microarrays was presented by Douglas Amorese (Hewlett-Packard). Picoliter amounts are delivered “on the fly” at rapid firing rates (10,000 fires/s at a 10 m/s velocity) without apparent shearing of DNA. He also addressed issues related to residual DNA in the apparatus. The speed and efficiency of microarray manufacturing may be greatly enhanced by this technology.Joseph L. DeRisi (Department of Biochemistry and Biophysics, Univ. of California) focused his talk on the coregulated proteasome-encoding genes. Searching for similarities in promoter regions and using the one-hybrid screen, he was able to identify a zinc finger transcription factor that coprecipitates as a component of the proteasome. Yeast mutants for the transcription factor lead to massive downregulation of proteasome-encoding genes. Future directions include using transcription factor-binding DNA fragments as probes in a yeast intergenic array. He acknowledged that the high degeneracy of motifs that a certain transcription factor can bind still constitutes a limitation. DNA mobility shifts are complementary but do not work in all cases, as there are different factors with opposing functions that may only bind one site.An update on the use of Affymetrix GeneChips was given by David Lockhart. He also explained the process of GeneChip manufacturing, as a light-directed, spatially specific, massively parallel DNA synthesis that utilizes alternate phases of light and chemical manipulation. Oligo probe sequences are carefully selected (based on most uniqueness and using general principles of primer selection), and their 3′ end is covalently attached to a flexible linker/spacer. Truncated short mers that lie in the solid surface are a byproduct of imperfect synthesis but do not seem to interfere significantly with the results. He explained that a newly released protocol using total RNA is available and works very well. He showed the extremely high reproducibility of the GeneChips when comparing in vivo-derived murine samples. Because of their high sensitivity, GeneChips lose linearity of expression levels at the highest values, thus underestimating relative changes. Also, the denominator needs to be above zero value to calculate fold change. He described experiments of cell cycle regulation in human fibroblasts. Samples were collected every 2 h for 24 h following arrest. His group identified five clusters clearly related to the cell cycle using highly conservative means for selection. In addition to bioinformatics, he stressed the importance of human intuition and understanding of biologic principles. He gave an interesting example of how a collaborator's insight led to the identification of commonly induced expression patterns between seemingly unrelated (cAMP vs. alcohol) stimuli.Francis S. Collins (NHGRI) spoke about the next 50 years and the consequences of the completion of the Human Genome Project. The advent of predictive genetic testing for a broad range of conditions, genetic interventions and the corresponding effective legislation, knowledge of common motifs in all human proteins, precise targeting of cancer and mental illness, safe germ line gene therapy, cataloguing of aging genes, computer modeling of the cell function, and debates about humans taking charge of their own evolution will take place, as comprehensive genomic-based health care revolutionizes human health. Dr. Collins also predicts the surging of major antitechnology movements throughout the world. In addition, he discussed how the study of SNPs in chimpanzees and gorillas using the Affymetrix GeneChips has helped identify common ancestral human alleles. The Human Genome Project will be completed by the year 2002/2003, with a rough draft released by the Spring of the year 2000 by the joint effort of the “G5”: Whitehead/MIT Group, the Joint Genome Institute, the Sanger Center, the Human Genome Center, and the Genome Sequencing Center at Washington University. The best account of the progress can be followed at http://www.ncbi.nlm.nih.gov/genome/seq. However, the genome is a lot more than sequence, as cataloguing human variation is another challenging task. The SNIP Consortium's goal is to discover 300,000 SNPs within the next 3 years. He announced the creation of Future NHGRI Centers of Excellence in Genomic Analysis to bring together scientists with interests in computational genomics, functional genomics, and population genetics, similar to the National Science Foundation Science and Technology Centers. He also announced the Mammalian Gene Collection (MGC) Initiative, an NHGRI- and NCI-led project that aims to collect all full-length human cDNAs.The importance, accessibility, and richness of gene expression variation information over that of sequence allelic variation was stressed by Patrick O. Brown (Howard Hughes Medical Institute and Department of Biochemistry, Stanford University School of Medicine). In addition to abundance of transcripts, variation in translation rates over time can be observed in a genomic scale. Proteins working in stoichiometric complexes tend to be tightly coregulated and thus represent physiological motifs or subroutines of a given physiological process. The systematic collection of such data remains a challenge, and the investigator needs the assistance of computerized programs that translate thousands of experimental results into a picture with biological meaning. Coupling our knowledge of coding sequences with the regulatory elements that control their site-specific, temporal, and quantitative expression requires optimization.Stuart K. Kim (Department of Developmental Biology, Stanford University School of Medicine) discussed his work on Caenorhabditis elegans development using gene microarrays. His group focuses on gene expression profiling for identifying enrichment of germ line stem cell genes that account for their totipotent and immortal functions. The identification of genes that have an impact on mitotic proliferation and on the differentiation of sperm and oocytes can also be identified using the arrays.Several investigators discussed their studies on yeast cell cycle gene regulation. Paul Spellman explained the benefits of clustering his yeast experimental data. Genes commonly induced by various stressors (i.e., heat shock, redox, starvation, and osmotic shock) and an RNA-processing cluster can be identified. Shared regulatory sequences and genes clustered as members of the translation apparatus can also be ascertained. Anti-correlated homologous pairs (highly homologous genes with opposite regulation) over a temperature range were also found.Joseph Mangan (St. George's Hospital Medical School, London, UK) spoke about using microarrays for the study of alveolar macrophages infected with Mycobacterium tuberculosis. Technical problems regarding RNA stabilization during extraction have now been properly addressed with a technique that lyses macrophages but not the mycobacteria. His group uses random-primed cDNA approaches, as TB mRNA lacks poly-A tails. In addition, 98% is made up of ribosomal RNA.A fundamental aspect of DNA profiling is its applicability to the elucidation of regulatory pathways that modulate the transcriptional apparatus. Richard A. Young (Whitehead Institute for Biomedical Research, Cambridge, MA) spoke about the complexity of regulator recruitment of RNA polymerase II and the use of DNA microarrays to discover regulatory circuits. A limited number of transcription factors may activate the expression of thousands of genes. He gave examples of repressor genes that, if turned off during the diauxic shift, result in the expression of hundreds of genes. He discussed the identification of genes mediating ribosomal gene expression and the genomic reprogramming that occurs in response to hydrogen peroxide. Some critical regulator mRNAs may exist at 1 copy per cell and still serve as critical regulators. The amplification of message that occurs with translation accounts for the effects seen in the absence of detectable mRNA. The use of antibodies against cross-linked DNA and protein followed by antibody retrieval can be used to develop microarrays to study protein-DNA interactions. Finally, he discussed the derepression of telomere-proximal genes upon histone depletion, suggesting that nucleosome density may play a role in maintenance of repression, and gave examples of silencer mediator genes.Jonathan Pevsner (Kennedy Krieger Institute and Johns Hopkins University) mentioned results of his studies of Rett's syndrome and autism with DNA microarrays. By using pool subtraction, RETT-enriched and control-enriched RNAs were amplified independently and subjected to DNA microarray analysis. He was able to identify and confirm differential expression of genes in Rett's disease brain. He discussed the issues regarding collection of postmortem tissues for RNA isolation. These include the agonal state, the postmortem interval, and viral contamination. He emphasized the use of microarrays to find signatures of disease processes without necessarily revealing the particular gene defect underlying such processes.DNA microarrays are also being used to study gene expression in plants. Asaph Aharoni (Center for Plant Breeding and Reproduction Research, Netherlands) talked about DNA microarrays for the study of strawberry development. Genes involved in quality traits (texture, color, firmness) were investigated. Genes associated with development from the green to red stages were also identified. Finally, he gave examples of cross-hybridization of some strawberry RNAs with those of other plants.The application of DNA microarrays to study developmental pathways is gaining ground as a highly informative approach. Yasuchi Okasaki (Genomic Sciences Center, RIKEN, Tsukuba, Japan) utilized mouse full-length cDNA arrays with up to 20,000 genes to examine gene expression in murine central nervous system development. This group has constructed over 80 libraries from different stages of development.The rapid processing of data obtained from the DNA microarrays remains a new frontier for investigators in the field. David Lipman [National Center for Biotechnology Information (NCBI), Bethesda, MD] talked about challenges encountered with genomic information, with emphasis in its static and dynamic nature, with functional assignments changing over time. In addition, he reminded the audience that complete genomes, including that of yeast, have incomplete experimentally verified gene functions. Current challenges include selection and annotation, quality of processing software, data normalization and analysis, functional annotation of hypothetical genes, pathway reconstruction, the need for using a common expression database or data format, and the need for correlative clinical and expression databases. He gave a brief discussion of OMIM (curated functional descriptions and references of human gene loci), COG (clusters of orthologous gene groups), CDD (conserved domain database, which uses statistical models tied to structural data), and the new GEO (gene expression omnibus, a repository for submission of works that share interest or themes, including SAGE and microarray data). He also discussed issues related to using Unigene, LocusLink, and RefSeq. The latter is a nonredundant, curated list of reference sequences that can be used to improve the quality of Unigene data.Another Bioinformatics field pioneer, Mark Boguski (Computational Biology Branch, NCBI), spoke about his group's efforts to summarize information from textual databases using document “neighboring” and to identify functionally related (clustered) genes that share neighbored references. While attempting to produce “executive summaries” that cluster reference documents rather than gene expression levels, he has found that the main limitation for this approach lies in the actual structural quality and consistency of archival database information.Hidemasa Bono (Institute for Chemical Research, Kyoto University, Japan) discussed the use of the Kyoto Encyclopedia of Genes and Genomes (KEGG) and gave a lively introduction to the web site and a detailed explanation of its contents. An advantage is the additional annotation provided by the GENES database from KEGG. This database provides EC numbers that make it possible to map the enzyme on metabolic pathway diagrams. This mapping is used to evaluate several clustering analysis methods, which include the nearest-neighbor, furthest-neighbor, and group average methods.The importance of studies correlating the expression patterns of genes with promoter region sequences was discussed by Michael Eisen (Stanford University). He reviewed hierarchical clustering and referred to K-means, self-organizing map algorithms, and how it is more useful to think of these methods as instruments to measure gene expression, keeping in mind that some genes with only 1.2-fold level changes may still have a functional significance.Terry Gasterland (Rockefeller University, New York, NY) talked about constructing explanations for gene expression clusters. This can be approached by looking at cotranscription in another genome, common upstream motifs, common subcellular localization, and common physiological roles. Questions such as “are homologs in the same cluster?” or “are homologs clustered the same in other species?” should be asked routinely.The accumulation of large amounts of useful data in a comprehensive manner was discussed by Roland Stoughton (Rosetta Inpharmatics). He explained how compensatory transcriptional responses of a specific pathway could be the most sensitive indicators of inhibition at the protein level and explained the need for a chemical compendium, to lay down landmark profiles for querying or mining. A genetic compendium also provides pathway reporter keys. Large functional data sets of genes and compounds are being amassed at Rosetta Inpharmatics.Jeffrey Trent (NHGRI Director) talked about mining melanoma with microarrays. He applied multidimensional scaling and showed how to plot 38 experiments from the raw data matrix in a multidimensional space. Rotation provides means to avoid false clustering and results in a weighted list of genes that discriminates clusters. This, combined with the dendrogram analysis and plots, is used to provide a more detailed view of genes whose expression levels are not overlapping. His group has performed microarray analysis of metastatic melanoma tissues (26 cell cultures and 5 patient biopsies). Correlation of gene expression levels to an in vitro scratch wound response test was found.DNA microarray customization for specific purposes was one of the highlights of the talk by Louis Staudt (NCI, Bethesda, MD). He discussed his studies on the gene expression of lymphoma and the importance of somatic Ig gene hypermutation in the B cell germinal centers. Germinal center B cell libraries were created, and sequences were blasted and kept in a relational database. This was used to create a “lymphochip” with 18,500 spots, which included genes of immunologic and oncologic importance. His group has performed cell sorting of tonsillar and peripheral blood cells and were able to distinguish clusters that are characteristic of resting and activated B cells, follicular cells, and chronic lymphocytic leukemia (CLL) cells.Jonathan Pollack (Stanford Medical Center) used comparative genomic hybridization (CGH) on cDNA microarrays. He showed how fluorescence ratios can be directly comparable to chromosomal copy numbers. For example, he showed how CGH performs when comparing 45 XO to 46 XX karyotypes, with a coefficient of variation of less than 10%.Peter Lichter (Deutsches Krebsforschungszentrum, Germany) discussed his findings on lymphomas and how to use microarrays as predictive markers for use in clinical trials, in combination with disease-specific matrix CGH. He stressed the difference between genomic and expression profiles. The use of a matrix of contiguous DNA fragments derived from one chromosomal region can be enhanced with redundant clones to increase sensitivity and specificity. Patient survival correlated with changes in gene expression.Retinoic acid-induced differentiation has also been subjected to DNA array analysis. Geoffrey Childs (Albert Einstein College of Medicine, New York) described a stem cell model for ES/EC cell growth and differentiation, using F9 embryonal carcinoma cells with and without retinoic acid treatment. Using DNA microarrays, his group identified multiple clones involved in differentiation. Adding cycloheximide as a protein synthesis inhibitor gave further optimization of his results.Ash Alizadeh (Stanford University School of Medicine) talked of clusters as fingerprints to define subgroups within diffuse large cell lymphoma and asked the question of whether expression levels could be related to ontogeny phases and associated with survival curves. He showed how diffuse large-cell lymphoma (DLCL), follicular lymphoma, and CLL can be differentiated and further subdivided, leading to the concept of “diseases within diseases.”Studies on colorectal cancer by Raju Kucherlapati (Albert Einstein College of Medicine) take advantage of DNA microarray technology to examine tumor-related genes. He first discussed the known genetic differences between sporadic colon cancer, hereditary nonpolyposis, and familial adenomatous polyposis and explained the effect of mismatch repair gene family members on the rate of APC mutations (MLH1, PMS1–2, and MSH2, MSH3, and MSH6). The animal model shows complete loss of APC gene expression by colonic immunohistochemistry. APC homozygous mice develop embryonic lethality. He studied embryonic stem (ES) cells that only have mutations in APC as an advantageous model, circumventing the problems posed by early lethality. They showed that ES APC −/− have an increase in β-catenin (known to bind APC). They compared normal ES to APC-mutated ES cells and found changes in the expression of multiple genes.Olli-P. Kallionemi (NHGRI) spoke about the challenge of extrapolating large volumes of pathological data into clinical studies. To develop better and faster means, his group came up with “tissue chips.” Samples from embedded tumor tissues (pathological samples) are positioned into arrays that contain thousands of different tumor samples and can be probed with antibody staining. This can be correlated to survival curves as well, as a “prognosis chip.” The tissue chip system is ideal for molecular evaluation at the population level and may be used to validate DNA microarray data. His group has also performed chromosomal fluorescence in situ hybridization (FISH) using microarrays.The study of human polymorphisms in the context of disease association was addressed by Aravinda Chakravarti (Case Western Reserve University, Cleveland, OH). He discussed the strengths of genetic effects and explained how Mendelian disorders are of recent evolutionary origin and thus have a small population incidence but a high risk among proband relatives with the gene defect being in a single locus, which facilitates the studies. In contrast, complex disorders are of ancient origin and thus have a more global population incidence but a low risk for proband relatives, with the gene defects in multiple locations. To detect genes of interest, he described the use of linkage analysis and association analysis in cases and controls. In collaboration with Affymetrix, they scanned 75 selected genes and found 15 polymorphic variants per 10 kb examined, with some genes having dramatically more allelic variants than others. An interesting question is: What are the selective pressures that lead to one gene being more variable than others? He also discussed the importance of polymorphisms in untranslated regions.Leonid Kruglyak (Fred Hutchinson Cancer Research Center, Seattle, Washington) talked about direct (cataloguing and testing all functional variants of a gene) and indirect (using dense SNP maps and testing for linkage dysequilibrium) methods for studying population risk. As an example, he mentioned ApoE, which exists as ApoE2, 3, and 4. It is known that ApoE4 is found in 15% of the general population but is present in as much as 40% of Alzheimer's disease patients. He discussed linkage dysequilibrium in relation to ancestral human populations and how one must study thousands of individuals for linkage dysequilibrium studies: the lesser the number of founders, the higher the linkage dysequilibrium. He also mentioned that there is more linkage dysequilibrium when there is a slow period of growth following the founder phase. Whereas association studies detect higher frequency and lower effect alleles, linkage studies detect lower frequency and higher effect alleles. He predicts that the cataloguing of several hundred thousand SNPs will be necessary for meaningful disease association results.A good example of how quickly the applications and techniques of DNA chips are evolving was the talk by Ann-Christine Syvanen (National Public Health Institute, Finland). She has performed minisequencing by primer extension for genotyping using microarrays and also a variation termed allele-specific extension (using extra primers that carry the mutation as well). Studies in Finland can be unique in that the Finnish population has a few selected genetic diseases that do not exist in other parts of the world. This group of researchers uses a method of DNA fragmentation with DNase and alkaline phosphatase.Rolf Krahe (Division of Human Cancer Genetics, Ohio State University) talked about global expression analysis in myotonic dystrophy, a disease characterized by amplification of unstable CTG repeats in the 3′-untranslated region (3′-UTR) of the myotonic dystrophy protein kinase (DMPK) gene. The abnormal CTG repeat allele has abnormal splicing. His studies suggest that nuclear aggregates that do not leave the nucleus lead to loss of function. Microarray analysis showed the downregulation of certain genes that correlated with the number of repeats and global dysregulation of genes as part of a general defect in nuclear export of some mRNAs.The study of human infectious diseases in vitro with DNA chips is also gaining ground. Charlie Xiang (US Army Medical Research Institute of Infectious Diseases, Frederick, MD) discussed time-course experiments with microarrays to examine Ebola virus (Zaire and Reston strains) infection of human monocytes. Changes in expression levels were observed for cytokines, chemokines, cell cycle, apoptosis, and signal transduction genes. These studies were able to suggest that Ebola Zaire but not Reston may prevent macrophage apoptosis and facilitate sustained viral production.Another speaker that addressed the issue of gene expression profiling during viral infection was Thomas E. Shenk (Howard Hughes Medical Institute, Princeton University). He explained how a constituent of the cytomegalovirus (CMV) particle induces most, if not all, of the mRNA expression changes. His group has studied the effect of CMV mutations. He explained how several phases that begin with viral entry into the cell lead to changes in gene expression, following the induction of intracellular signals. His talk was also characterized by a good example of how data from a few microarrays can be correlated with the literature to delineate a wide set of new experiments.David Duggan (Cancer Genetics Branch, NHGRI) talked about BRCA1- and BRCA2-associated breast cancer tissue identification by microarrays. He showed how a cell cycle gene was shown to segregate with one type but not the other in a statistically significant manner. He also discussed the rotation of multidimensional scaling images to demonstrate how some genes cluster together. The use of weighted gene analysis (center-to-center distance between clusters, for example) followed by dendrogram visualization can be incorporated in the quantitative analysis.Robert Lipshutz (Affymetrix, Santa Clara, CA) talked about the challenges of linkage dysequilibrium and association studies, and the use of multiplex biotin-labeled PCR applications for allele detection. These applications have reproducibility and reliability for the study of selected polymorphisms. He also discussed the single base primer extension-based arrays, yeast genome typing for SNP discovery, and use of these techniques for speciation of mycobacteria. His experiments demonstrate the effectiveness of Affymetrix GeneChips as powerful tools to deal with the massive amounts of data generated by the Human Genome Project.The predictability of whether a particular gene profile represents a signature of a tumor type was discussed by Eric Lander (Whitehead Institute for Biomedical Research). He stated that morphology assessment, antibody-based and cytogenetic assays preceded the use of microarray technology in tumor type prediction. He discussed mathematical ways of predicting strength of association [discovery of excess density of genes that correlate with either acute myeloid leukemia (AML) or acute lymphoblastic leukemia (ALL)] or isolated genes that correlate with remission. Median prediction strengths were very low for genes randomly selected from various clusters. Two subclasses of medulloblastoma could be predicted with high accuracy. His group also examined the expression of genes that correlated with the presence of specific translocations.Abdel Elkahloun (Research Genetics, Huntsville, AL) talked about the use of laser capture microdissection, DNA arrays, and quantitative real-time PCR for studying breast cancer progression. Nanogram quantities of RNA from laser capture microdissection and oligo-dT reverse transcription and second-strand synthesis using radioactive nucleotides were used to show the expression of genes that correlate with several morphological stages of human breast cancer.Patrick Brown delivered the keynote speech entitled “Observing the Living Genome.” Understanding human genetic variation is a fundamental challenge. There is little knowledge of the patterns of distribution of reciprocal meiotic recombination events across human genomes. In yeast, microarrays have helped determine SPO11 cleavage sites in meiosis which can be used as a match for yeast recombination frequencies. He also discussed how the type of cell and the internal and external environment may influence gene expression. He showed how, despite significant variability in handling, timing, and collection of samples, the expression picture is very detailed and unexpectedly predictable (i.e., tissue from the same tumor in vivo can be identified at repeat sampling times as originating from the same source compared with tumor samples derived from other individuals). Patterns of gene expression are very complex, and reductions of data, in an effort to draw conclusions, come at a price. He showed some progress with his antibody arrays (i.e., immobilized specific antibodies are used as probes while a mixture of dye-labeled proteins serve as target). Finally, he stressed the need for systematic curated databases of multiple fields (i.e., Genomics, Physiology, Cell Biology, Anatomy, Pathology). The absence of such databases are perceived as obstacles to scientific progress. Dr. Brown encouraged Journal editors and the audience to embrace the idea of a “PubMed Central” which would help bring together different fields and remove such obstacles.Other sponsors of the meeting included: AppliedPrecision.com, Amersham Pharmacia Biotech, Axon Instruments, Hewlett-Packard, Incyte, Motorola, PE Biosystems, BioRobotics, Clontech, Display Systems Biotech, Genicon Sciences, and Molecularware. The meeting was produced by BioEdge.net. For information regarding poster presentations by the meeting delegates, visit the Nature Genetics web page at http://www.nature.com/ng/microarray.This article has no references to display. Previous Back to Top Next FiguresReferencesRelatedInformationCited ByAutomated Real-Time Spotting System for DNA/Protein Microarray ApplicationsAutomated liquid dispensing pin for DNA microarray applicationsIEEE Transactions on Automation Science and Engineering, Vol. 3, No. 2Comparative Studies Using cDNA vs. Oligonucleotide Arrays14 January 2010Applications to Cancer Research of “Lab-on-a-chip” Devices Based on Dielectrophoresis (DEP)23 June 2016 | Technology in Cancer Research & Treatment, Vol. 2, No. 1Paradoxical overexpression and translocation of connexin43 in homocysteine-treated endothelial cellsHong Li, Sergey Brodsky, Sindu Kumari, Virginijus Valiunas, Peter Brink, Jun-Ichi Kaide, Alberto Nasjletti, and Michael S. Goligorsky1 June 2002 | American Journal of Physiology-Heart and Circulatory Physiology, Vol. 282, No. 6The Use of DNA Array Technology in Studies of Ocular Viral PathogenesisDNA and Cell Biology, Vol. 21, No. 5-6Physiological Genomics: Who we are and where we’re goingVictor J. Dzau, Editor-in-Chief, and Susan Glueck, Deputy Editor21 December 2001 | Physiological Genomics, Vol. 7, No. 2 More from this issue > Volume 2Issue 2March 2000Pages 53-58 Copyright & PermissionsCopyright © 2000 the American Physiological Societyhttps://doi.org/10.1152/physiolgenomics.2000.2.2.53PubMed11015582History Published online 13 March 2000 Published in print 13 March 2000 PDF download Metrics Downloaded 114 times
Referência(s)