|Home | About | Journals | Submit | Contact Us | Français|
MAGEST is a database for maternal gene expression information for an ascidian, Halocynthia roretzi. The ascidian has become an animal model in developmental biological research because it shows a simple developmental process, and belongs to one of the chordate groups. Various data are deposited into the MAGEST database, e.g. the 3′- and 5′-tag sequences from the fertilized egg cDNA library, the results of similarity searches against GenBank and the expression data from whole mount in situ hybridization. Over the last 2 years, the data retrieval systems have been improved in several aspects, and the tag sequence entries have increased to over 20 000 clones. Additionally, we constructed a database, translated MAGEST, for the amino acid fragment sequences predicted from the EST data sets. Using this information comprehensively, we should obtain new information on gene functions. The MAGEST database is accessible via the Internet at http://www.genome.ad.jp/magest/.
In the early development of many animals, maternal cytoplasmic factors are known to have various essential roles. Ascidian is a lower chordate, a group that includes humans, and is a good model system to investigate the characteristics of maternal factors because its developmental processes are shown to be largely dependent on the maternal factors (1,2), and its genome size is small (3). We are interested in the maternal mRNAs as candidates for cytoplasmic determinants, and thus initiated a cDNA project to collect mRNA expressed sequence tags (ESTs) and their localization or expression data. Thus, we are constructing a database, named MAGEST (Maboya Gene Expression patterns and Sequence Tags) to analyze the data gathered by our project. Here we present an update of the MAGEST database (4,5). We have improved the data retrieval systems, and have started work on the translated MAGEST, which is a database for predicting amino acid fragment sequences.
To date, we have determined the tag sequences for more than 20 000 cDNA clones, and investigated the localization or expression patterns for more than 2000 genes of Halocynthia roretzi cDNAs. All of this information is deposited in public databases.
Many computer programs for the prediction of gene function or protein structure require the amino acid sequence, not the nucleotide sequence. In some cases, the ESTs, which are nucleotide sequence fragments, may encode part of an amino acid sequence. Translated MAGEST is a database for the fragments of amino acid sequences that are translated from MAGEST EST entries. Each MAGEST EST entry undergoes a BLASTX similarity search against the SWISS-PROT protein database (6). We obtained the translated sequence to parse the result file of BLASTX. To date, about 3000 entries are registered in the Translated MAGEST.
MAGEST is implemented in the Sybase relational database management system. Access to data is provided primarily via World Wide Web-based query forms at http://www.genome.ad.jp/magest/. MAGEST has comprehensive links to other sequence resources using the DBGET/LinkDB database retrieval system (7). Using the MAGEST database search query form, one can rapidly identify the expressed tag sequences and the gene expression patterns from each cluster ID or clone ID, and select a distinct data set from keywords that include the predicted functions and motifs. In addition, one can execute sequence similarity searches for MAGEST entries using the BLASTN or the TBLASTX sequence similarity search programs (8). Using the MAGEST expression pattern search query form, one can locate the group of genes that share the same expression patterns. Using the MAGEST advanced search query form, one can clarify the cluster size of each gene, and obtain a description of predicted functions.
The functions of ~50% of clones are unknown when only sequence similarity searches are carried out. To proceed with analyses of the functions of these genes, we attempt to translate the EST nucleotide sequences into the encoded amino acid fragment sequences using this database. For analyses of these gene structures, we used the SOSUI program (Classification and Secondary Structure Prediction System for Membrane Proteins) (9) against the Translated MAGEST database. SOSUI predicted that 1% of the cloned entries are membrane proteins. Recently, genome analyses of ascidians including EST analyses have been undertaken on a global scale. Comparison of MAGEST data with the data obtained from various projects should shed some light on the gene networks that control early development in chordates.
We thank Prof. Nori Satoh at Kyoto University for generous use of the facilities, and Dr Susumu Goto at Kyoto University for support with database management. This work has been supported by Grants-in-Aid from the Ministry of Education, Science, Sports and Culture, Japan (12680714, 13045018 and 13202024) to K.W.M. and the ‘Research for the Future’ Program from the Japan Society for the promotion of Science (96L00404) to K.W.M. The computational resources are from the Super Computer Laboratory of the Institute for the Chemical Research, Kyoto University.