Sequencing of 6.7 Mb of the melon genome using a BAC pooling strategy

Vctor M Gonzlez 0 Andrej Benjak 1 Elizabeth Marie Hnaff 0 Gisela Mir 1 Josep M Casacuberta 0 Jordi Garcia-Mas 1 Pere Puigdomnech 0 0 Molecular Genetics Department, Center for Research in Agricultural Genomics CRAG (CSIC-IRTA-UAB) , Jordi Girona, 18-26, 08034 Barcelona , Spain 1 IRTA, Center for Research in Agricultural Genomics CRAG (CSIC-IRTA-UAB) , Carretera de Cabrils Km 2, 08348 (Barcelona) , Spain Background: Cucumis melo (melon) belongs to the Cucurbitaceae family, whose economic importance among horticulture crops is second only to Solanaceae. Melon has a high intra-specific genetic variation, morphologic diversity and a small genome size (454 Mb), which make it suitable for a great variety of molecular and genetic studies. A number of genetic and genomic resources have already been developed, such as several genetic maps, BAC genomic libraries, a BAC-based physical map and EST collections. Sequence information would be invaluable to complete the picture of the melon genomic landscape, furthering our understanding of this species' evolution from its relatives and providing an important genetic tool. However, to this day there is little sequence data available, only a few melon genes and genomic regions are deposited in public databases. The development of massively parallel sequencing methods allows envisaging new strategies to obtain long fragments of genomic sequence at higher speed and lower cost than previous Sanger-based methods. Results: In order to gain insight into the structure of a significant portion of the melon genome we set out to perform massive sequencing of pools of BAC clones. For this, a set of 57 BAC clones from a double haploid line was sequenced in two pools with the 454 system using both shotgun and paired-end approaches. The final assembly consists of an estimated 95% of the actual size of the melon BAC clones, with most likely complete sequences for 50 of the BACs, and a total sequence coverage of 39x. The accuracy of the assembly was assessed by comparing the previously available Sanger sequence of one of the BACs against its 454 sequence, and the polymorphisms found involved only 1.7 differences every 10,000 bp that were localized in 15 homopolymeric regions and two dinucleotide tandem repeats. Overall, the study provides approximately 6.7 Mb or 1.5% of the melon genome. The analysis of this new data has allowed us to gain further insight into characteristics of the melon genome such as gene density, average protein length, or microsatellite and transposon content. The annotation of the BAC sequences revealed a high degree of collinearity and protein sequence identity between melon and its close relative Cucumis sativus (cucumber). Transposon content analysis of the syntenic regions suggests that transposition activity after the split of both cucurbit species has been low in cucumber but very high in melon. Conclusions: The results presented here show that the strategy followed, which combines shotgun and BAC-end sequencing together with anchored marker information, is an excellent method for sequencing specific genomic regions, especially from relatively compact genomes such as that of melon. However, in agreement with other results, this map-based, BAC approach is confirmed to be an expensive way of sequencing a whole plant genome. Our results also provide a partial description of the melon genome's structure. Namely, our analysis shows that the melon genome is highly collinear with the smaller one of cucumber, the size difference being mainly due to the expansion of intergenic regions and proliferation of transposable elements. - Background During recent years an important effort has been made to increase the tools available for the genomic analysis of major plant crop species. Since the first genome sequence available of Arabidopsis thaliana [1], several others have been published. They include model plants such as Brachypodium [2] but, increasingly, species that have been chosen for their importance in agriculture. For example the rice [3], maize [4], sorghum [5] or soybean [6] genomes are complex but the wealth of genetic information matches their economic interest. Consequently, for both scientific and economic reasons an increasing number of plant genomes are being analyzed, providing important resources useful for their biological study and breeding. Several species of interest from both scientific and economic perspectives are of the Cucurbitaceae family. These include melon, cucumber, watermelon and squashes, all of which have been the object of biological and agricultural interest for centuries. In recent years various molecular tools have been established. For instance, the first assembly of the cucumber genome [7], as well as an increasing number of genetic and genomic resources developed for melon, a diploid species with a relatively compact (around 454 Mb [8]) genome [9]. These include tools such as a collection of more than 129,000 ESTs [10,11], BAC libraries [12,13], oligo-based microarrays [14,15], TILLING and EcoTILLING platforms [16,17], a set of near isogenic lines (NILs) [18] and several melon genetic maps [11,19-25]. Recently, we have built a physical map with 0.9x genomic coverage using both a BAC library and a genetic map previously developed in our laboratories [http://melonomics.upv.es/ public_files, [26]], the first report of such a genomic resource of a Cucurbitaceae species so far. This physical map has also been integrated with the genetic map by anchoring a number of physical contigs (representing 12% of the melon genome) to 175 known genetic markers. These tools have been useful in the study of interesting agronomical traits such as virus or fungi resistance [27,28], sex determination [29,30] or the control of ripening [31,32]. These results demonstrate that molecular genetic approaches can successfully be used in melon to address basic questions of biological or agronomic relevance. More extensive sequence information would be invaluable to complete the picture of the melon genomic landscape. Indeed, the sequences of only a few selected genomic regions have been published, totaling no more than 500 kb [29,33-35] and as of May 2010 no more than 173 melon genes can be found in GenBank [11], although a collection of ESTs probably representing more than 70% of the transcriptome is currently available [11]. The sequencing of the Sorghum genome has shown the feasibility of sequencing a plant genome larger than that of melon (730 Mb) using a Sangerbased whole genome shotgun approach [5]. However, the development of new massively parallel sequencing technologies allows envisaging a complete sequencing of the species at higher speed and at lower cost than previous Sanger-based methods. To this end, both whole genome sequencing approaches as well as map-based, BAC-to-BAC strategies have been proposed to sequence plant genomes [36,37]. A small number of research projects involving 454 sequencing of BAC (...truncated)