Gkh135 323.325

Nucleic Acids Research, 2004, Vol. 32, Database issue D323±D325 yMGV: a cross-species expression data mining toolGaeÈlle Lelandais1,3, SteÂphane Le Crom2, FreÂdeÂric Devaux1, SteÂphane Vialette1,George M. Church4, Claude Jacq1 and Philippe Marc4,* 1Laboratoire de GeÂneÂtique MoleÂculaire, CNRS UMR8541 and 2Laboratoire de Biologie MoleÂculaire du DeÂveloppement, INSERM U368, Ecole Normale SupeÂrieure, 46 Rue d'Ulm, 75005 Paris, France, 3Equipe de Bioinformatique GeÂnomique et MoleÂculaire, INSERM E346 Universite Paris 7, case 7113, 2 place Jussieu, 75005 Paris, France and 4Lipper Center for Computational Genetics and Department of Genetics, Harvard Medical School, 77 Louis Pasteur Avenue, Boston, MA 02115, USA Received September 14, 2003; Revised and Accepted October 27, 2003 The yeast Microarray Global Viewer (yMGV @ http:// New data have been added to the database on a regular basis transcriptome.ens.fr/ymgv) was created 3 years ago since the release of the original version, which contained as a database that houses a collection of 39 microarray data sets. Today, the yMGV database contains Saccharomyces cerevisiae and Schizosaccharo- data from 1544 genome-wide expression experiments, repre- myces pombe microarray data sets published in 82 senting 82 microarray publications. Importantly, expression different articles. yMGV couples data mining tools data sets from the S.pombe Sanger Institute project (5) have with a user-friendly web interface so that, with a few been included in version 2, enabling inter-organism queries.
mouse clicks, one can identify the conditions that The database architecture is designed to allow the addition of affect the expression of a gene or list of genes regu- data pertaining to other organisms in the near future.
lated in a set of experiments. One of the major new features we present here is a set of tools that allows for inter-organism comparisons. This should enable yMGV is under continuous development, and version 2 has the ®ssion yeast community to take advantage of been available since April 2003. Recent improvements allow the large amount of available information on bud- the user to critically assess the published data, e.g. summary ding yeast transcriptome. New tools and ongoing statistics, such as the mean and standard deviation of the developments are also presented here.
log2(ratio) distribution of a given microarray data set, are reported along with the log2(ratios) in that data set. New links to external databases have been added, and their connections improved (see Supplementary Material for current URLs). It is now possible to directly post a list of genes generated using Although several databases have been created to manage yMGV to other online tools, e.g. to KEGG for metabolic published microarray data, many of the associated tools are mapping (6), to RSA tools for cis-regulatory motif discovery underutilized due to cumbersome user interfaces and non- (7) or to SGD for Gene Ontology (GO) term mapping (8).
intuitive output. This is the most common dif®culty confront- Several additional features are entirely new to the database.
ing the mining and visualization of the vast amount of data Two of them, cross-species transcriptome comparison and produced by genomic technologies. The yeast Microarray compendium modules, are detailed below.
Global Viewer (yMGV) is a data mining tool coupled to a multi-organism database that currently houses Saccharomyces cerevisiae and Schizosaccharomyces pombe expression data.
The philosophy of yMGV is to empower biologists by Since its inception, a major goal of yMGV has been to providing a straightforward data mining interface, and by incorporate data originating from different organisms (9), and generating easily interpretable, mostly graphical, output. This the database schema has been designed to accommodate any tool has matured since its creation in 2001, and is now genome described using GO formalisms (8) (see logical recognized as an exemplary approach to the retrieval and scheme in Supplementary Material). Toward this goal, the interpretation of valuable biological information (1±3).
second organism to have been added to yMGV is the ®ssion The basic features of yMGV have been described previ- ously (4); here we present recent improvements to the data set Intra-species data analyses carried out by yMGV have been and interface, and plans for the development of future extended to incorporate inter-species data, allowing compari- sons of gene expression between orthologs. To facilitate *To whom correspondence should be addressed. Tel: +1 617 432 4136; Fax: +1 617 432 7266; Email: pmarc@genetics.med.harvard.edu D324 Nucleic Acids Research, 2004, Vol. 32, Database issue Figure 1. Using yMGV to compare gene expression in two organisms.
Figure 2. Using yMGV to ®nd genes co-expressed in a subset of conditions.
yMGV allows comparison of gene expression of two organisms (A and B) The user can enter a `seed' gene and choose one of the hand-curated (currently S.cerevisiae and S.pombe only) using an orthology table (C) con- microarray sets (A). yMGV computes the similarity between the expression structed using sequence information. The user can apply ®lters to the tran- pro®le of the seed and those of all other genes in this organism across the scriptome of one or both organisms (D and E) and get the list of microarray set (B). Highly correlated (or anti-correlated) genes are selected orthologous gene pairs that ®t the required expression pro®le and satisfy the (C) and a graphical representation shows their expression across ®lter parameters (F). GO description and links to organism-speci®c microarrays of the set (D). Gene-speci®c links to external databases are databases are provided for each gene (G).
provided, and the user can also post the whole gene list (E) to other databases in order to map them onto metabolic networks (KEGG) or the comparisons, a S.cerevisiae to S.pombe orthology table based GO tree (SGD), or to try to ®nd common cis-regulatory elements (RSA on sequence similarity is stored in the database (the table was created by the Sanger Institute). The web interface allows users to retrieve genes based on their log facilitate identi®cation of biological meaningful groups speci®ed experiments. If the experiments are from different (10,11). Since then, this approach has been used frequently organisms, the corresponding orthology tables are used and and with great success, e.g. clustering was used to isolate only orthologs meeting speci®ed thresholds are displayed (see interesting groups from a S.cerevisiae RNA data set of nearly 300 unrelated deletions or conditions (12). More recently, The evolutionary distance between S.cerevisiae and however, it has been shown that standard clustering methods S.pombe (at least 400 million years), and the absence of a are usually less effective when applied to large numbers of direct relationship between the sequence similarity and data sets that are biologically unrelated (13). Therefore, the functional similarity of two proteins, in¯uence the conclusions microarray experiments in yMGV were hand curated and that can be drawn from cross-species comparison.
classi®ed into 17 biologically coherent categories. We created Uncontrolled use of a module analysis based on orthology a module that lists genes that are signi®cantly co-expressed can yield misleading results. Accordingly, usage recommen- with respect to a user-selected reference gene according to a dations are associated with the module and can also be found chosen metric and a chosen biological category (see Fig. 2).
in the Supplementary Material of this article. When used with This proved to be very ef®cient for isolating genes co- discrimination, this tool should help the ®ssion yeast com- regulated only in speci®c conditions.
munity to easily take advantage of the huge amount of A list of biological categories and some examples of available information on the budding yeast transcriptome.
usage are provided in the tutorial available at http://www.
yMGV is, to our knowledge, the ®rst tool to allow this kind of transcriptome.ens.fr/ymgv/tutorial/.
A tutorial explaining the use of the cross-species tran- scriptome comparison module is available at http://www.
transcriptome.ens.fr/ymgv/tutorial/.
The major dif®culty in maintaining a database like yMGV is data retrieval and curation. Thanks to the genomics commu- nity, standardization of microarray data sets (14) has facili- tated the creation of central repositories for microarray data Several years ago, it was shown that the application of various (15,16). We plan to incorporate deposited data sets into yMGV clustering algorithms to large microarray data sets can in order to maximize its utility to the biology community.
Nucleic Acids Research, 2004, Vol. 32, Database issue D325 We also plan to add cis-regulatory elements to the yMGV output. This is essential, as phylogenetic footprinting has 1. Gasch,A.P. (2002) Yeast genomic expression studies using DNA proved to be a very powerful technique that will become more microarrays. Methods Enzymol., 350, 393±414.
and more ef®cient with increasing numbers of sequenced 2. Ulrich,R. and Friend,S.H. (2002) Toxicogenomics and drug discovery: genomes, thus giving a more accurate description of the motifs will new technologies help us produce better drugs? Nature Rev. Drug involved in transcriptome regulation.
We are also planning to give users the option to upload their 3. Wood,V. and Bahler,J. (2002) Website Review: How to get the best from ®ssion yeast genome data. Comp. Funct. Genomics, 3, 282±288.
own data sets, and to use these data sets like any other data set 4. Le Crom,S., Devaux,F., Jacq,C. and Marc,P. (2002) yMGV: helping biologists with yeast microarray data mining. Nucleic Acids Res., 30, Finally, one of our long-term goals is to create a module that captures properties (expression regulation, GO annotations, 5. Lyne,R., Burns,G., Mata,J., Penkett,C.J., Rustici,G., Chen,D., Langford,C., Vetrie,D. and Bahler,J. (2003) Whole-genome microarrays cis-regulatory motif) from an input gene list and retrieves of ®ssion yeast: characteristics, accuracy, reproducibility and processing genes sharing similar or partially similar properties.
6. Kanehisa,M., Goto,S., Kawashima,S. and Nakaya,A. (2002) The KEGG databases at GenomeNet. Nucleic Acids Res., 30, 42±46.
7. van Helden,J. (2003) Regulatory sequence analysis tools. Nucleic Acids The interface has been written in PHP and data are stored in a 8. Ashburner,M., Ball,C.A., Blake,J.A., Botstein,D., Butler,H., Cherry,J.M., Davis,A.P., Dolinski,K., Dwight,S.S., Eppig,J.T. et al. The Gene PostgreSQL relational database (logical scheme is available Ontology Consortium (2000) Gene ontology: tool for the uni®cation of in Supplementary Material). yMGV uses data provided by external databases, namely GO descriptions from SGD (17) 9. Marc,P., Devaux,F. and Jacq,C. (2001) yMGV: a database for and GeneDB (www.genedb.org), and the orthology table visualization and data mining of published genome-wide yeast expression data. Nucleic Acids Res., 29, E63.
10. Tavazoie,S., Hughes,J.D., Campbell,M.J., Cho,R.J. and Church,G.M.
(1999) Systematic determination of genetic network architecture. Nature 11. Eisen,M.B., Spellman,P.T., Brown,P.O. and Botstein,D. (1998) Cluster analysis and display of genome-wide expression patterns. Proc. Natl The Supplementary Material, available at NAR Online, contains: the database relational scheme, the yMGV data set 12. Hughes,T.R., Marton,M.J., Jones,A.R., Roberts,C.J., Stoughton,R., contributors 2001±2003, the list of URLs to other databases Armour,C.D., Bennett,H.A., Coffey,E., Dai,H., He,Y.D. et al. (2000) Functional discovery via a compendium of expression pro®les. Cell, 102, and tools used in yMGV, and a description of limitations and potential problems associated with ortholog expression 13. Gasch,A.P. and Eisen,M.B. (2002) Exploring the conditional coregulation of yeast gene expression through fuzzy k-means clustering.
14. Spellman,P.T., Miller,M., Stewart,J., Troup,C., Sarkans,U., Chervitz,S., Bernhart,D., Sherlock,G., Ball,C., Lepage,M. et al. (2002) Design and implementation of microarray gene expression markup language (MAGE-ML). Genome Biol., 3, RESEARCH0046.
The authors are grateful to the scientists who have supplied 15. Brazma,A., Parkinson,H., Sarkans,U., Shojatalab,M., Vilo,J., expression data and genome annotation (especially Valerie Abeygunawardena,N., Holloway,E., Kapushesky,M., Kemmeren,P., Wood), to Allegra Adele Petti for suggestions about the Lara,G.G. et al. (2003) ArrayExpressÐa public repository for microarray manuscript and to the following open source projects: Apache, gene expression data at the EBI. Nucleic Acids Res., 31, 68±71.
16. Edgar,R., Domrachev,M. and Lash,A.E. (2002) Gene Expression Debian, PHP and PostgreSQL. The yMGV project was funded Omnibus: NCBI gene expression and hybridization array data repository.
by the Programme Bioinformatique Inter-EPST-CNRS 2003.
P.M. is supported by the French Therapeutical Research 17. Dwight,S.S., Harris,M.A., Dolinski,K., Ball,C.A., Binkley,G., Association (AFRT) and the PhRMA foundation Center of Christie,K.R., Fisk,D.G., Issel-Tarver,L., Schroeder,M., Sherlock,G. et al.
Excellence in Integration of Genomics and Informatics (2002) Saccharomyces Genome Database (SGD) provides secondary gene annotation using the Gene Ontology (GO). Nucleic Acids Res., 30,

Source: http://slecrom.free.fr/download/2004_Lelandais_NAR_yMGV.pdf