Medicago truncatula, a close relative of alfalfa, is a preeminent model for legume genomics and has been the target of an international sequencing initiative. The Medicago truncatula sequencing project began in 2003 with the initial goal to decipher the sequences from the euchromatic portion of the medicago genome through BAC tiling path. The clone-based assembly (Mt version 3.5) contains a total of ~250 Mb of high quality DNA sequences distributed across 8 chromosomes, each with sequence contigs ranging in size from a few hundred Kb to 5-10 Mb. Since the initial sequencing of the BACs, additional whole genome shotgun sequences (WGS) were obtained using 454 and Illumina technology. The WGS contigs are generally smaller in size, but contains sequences that do not yet exist on the clone-based assembly. Our current work aims to integrate, manage and improve the structure and annotation of the Medicago genome. We plan to create the best possible sequence-based representation of the M. truncatula genome, by anchoring the unplaced BACs and WGS contigs through sequence overlaps, mate-pair linkages and optical map alignments. We will continue to support the IMGAG (International Medicago Genome Annotation Group) annotation pipeline that represents a consensus annotation process for the entire Medicago (and legume) community. We will further enrich the annotation of the M. truncatula genome by overlaying other data types including expression data, proteomic data, locations of transposon and fast-neutron induced mutations etc., and will host a Medicago genome browser and data repository at http://www.jcvi.org/medicago. An improved version of the Medicago genome assembly (Mt version 4) and the associated gene annotations will be released in the first half of 2012.