Databases and information integration for the Medicago truncatula genome and transcriptome

Abstract

An international consortium is sequencing the euchromatic genespace of Medicago truncatula. Extensive bioinformatic and database resources support the marker-anchored bacterial artificial chromosome (BAC) sequencing strategy. Existing physical and genetic maps and deep BAC-end sequencing help to guide the sequencing effort, while EST databases provide essential resources for genome annotation as well as transcriptome characterization and microarray design. Finished BAC sequences are joined into overlapping sequence assemblies and undergo an automated annotation process that integrates ab initio predictions with EST, protein, and other recognizable features. Because of the sequencing project's international and collaborative nature, data production, storage, and visualization tools are broadly distributed. This paper describes databases and Web resources for the project, which provide support for physical and genetic maps, genome sequence assembly, gene prediction, and integration of EST data. A central project Web site at medicago.org/genome provides access to genome viewers and other resources project-wide, including an Ensembl implementation at medicago.org, physical map and marker resources at mtgenome.ucdavis.edu, and genome viewers at the University of Oklahoma (www.genome.ou.edu), the Institute for Genomic Research (www.tigr.org), and Munich Information for Protein Sequences Center (mips.gsf.de).

Figure 1.

BAC sequence assembly viewer. The database and Web interface at http://medicago.org/genome includes a variety of visualization tools and data formats, including this sequence assembly viewer. The viewer accesses database information for each BAC, shown in the popup menu. Links from the popup menu lead to other database reports and directly to other data sources and views: GenBank records, contig and marker reports, contig images (such as one shown in Fig. 2), and genome viewers.

Figure 2.

Image of FPC contig. The database and Web interface at http://mtgenome.ucdavis.edu provides access and visualizations for the Medicago physical and genetic map. This image is typical of FPC contigs in the project, showing fingerprint-based BAC overlaps, BACs being sequenced, BACs with and without BAC end sequence, markers, and map position. BES, BAC end sequence; MTGS, M. truncatula genome sequence.

Figure 3.

Ensembl and MtDB EST pipelines and information resources. The EnsEBML annotation and the MtDB and Nimbus EST processing pipelines are supported by an automated system for downloading and maintaining up-to-date consistent target information (such as BLAST, HMM, and other targets on the right side of figure), and associated metadata. Core database resources, including BAC and EST sequences, BLAST hits, gene predictions, and other features (left and center of figure) are generally maintained in a relational format, and are accessible to all aspects of the projects. Software resources are generally deployed across compute clusters, and are managed by a scheduling system that directs loading and submission of work to cluster queues.