The genetic patrimony of a major cereal is now known with a level of precision never before achieved. On December 18, 2002, in Tokyo, the international consortium for the sequencing of the rice genome (International Rice Genome Sequencing Project, IRGSP), of which Genoscope is a member, announced the completion of a high quality "draft" of the sequence of this genome. This is an important step toward a complete sequence of the rice genome, which should be finished at the end of 2004. The finishing step, which is now being carried out by IRGSP, is indispensible if we want to reap all the benefits of this genome sequence for studying rice and other cereals.

Rice will be the second plant to be completely sequenced, after Arabidopsis thaliana, the model plant for which the genome was published at the end of 2000. Rice was chosen as the cereal genome to be sequenced as a first priority. It has the smallest genome of all the cereals: 430 million nucleotides (Mb). The corn genome is five times larger, and wheat is 40 times larger! However, large blocks of homologous genes with a relatively conserved arrangement are shared by several cereals. Rice is therefore a good entry point for characterizing cereal genes, and for associating them with agronomic characteristics. More generally, rice can serve as a model for one of the two groups of flowering plants, the monocotyledons, just as Arabidopsis is a model for the dicotyledons.

The rice genome is also interesting in itself, because this cereal constitutes the daily nutritional base for more than half of humanity. A knowledge of its genome will be a boon to plant breeders who are trying to increase yields and create new varieties resistant to diseases, pests, and drought or saline conditions.

Formed in 1997 under the leadership of Japan, IRGSP chose to sequence a variety of the sub-species japonica, which is cultivated in Japan and in temperate regions, and for which important physical and genetic mapping resources were already available. The main sequencing work began in 2000. As has been the case for other "large" genomes, the 12 chromosomes of rice, or regions of them, were distributed between the sequencing institutes of the 10 participating countries (Brazil, China, Korea, UK, USA, France, India, Japan, Taiwan and Thailand). Genoscope - the French National Sequencing Center - sequenced chromosome 12, which is 30 Mb long.

Members of the consortium were financed from public funds, and they adopted the same principles as those that regulated the sequencing of the human genome : sequences obtained with a sufficient level of precision are deposited immediately in public databases, where they are freely accessible to scientists working in both the private and public sectors all over the world.

The consortium also adopted the same "clone by clone" sequencing strategy as was used for the sequencing of the human genome: the rice genome was broken into large fragments for which the order and position on the chromosomes were determined. Each large fragment was broken into small fragments which were then sequenced, so that the sequencing puzzle of the genome was first put together locally. In IRGSP's present genomic draft, the small overlapping fragments have been assembled into about 2000 blocks, or contigs, precisely ordered and oriented, and which are separated by an equal number of gaps. The finishing work involves "plugging" these gaps, which, in fact, is the most laborious part of the work. The quantity of sequence used for assembly of the draft represents 10 times the total length of the genome (10X coverage), and that translates into a precision of 99.99% (one error in 10,000 nucleotides).

The sequencing of the rice genome has already been announced several times. The agrobiotechnological firms Monsanto (2000) and Syngenta (2002) have each announced their own genomic "drafts". Monsanto has in fact completed a small volume of sequence of large genomic clones comparable to those of IRGSP, whereas Syngenta has used the method of "whole genome shotgun" sequencing, in which fragments from the entire genome are sequenced randomly, and then assembled based only on their overlaps. With a coverage of 6X, the result is a group of 42,000 contigs of which only a portion are anchored on the rice chromosomes. This "draft" is therefore very different from that presented on December 18 ; in the IRGSP version, the sequence data are more precise, and are organized into fewer blocks which are perfectly ordered and oriented. Above all, the contigs are anchored in relation to physical and genetic maps. This makes it possible to search for genes of agronomic interest localized in a genetic interval of the genome.

The data from Monsanto and Syngenta have not been deposited in public databases. However, IRGSP was able to access this data in order to perfect its own assembly, thereby gaining several months. This integration was possible because the two projects utilized the same variety of rice as the consortium. A Chinese academic group has produced a draft of the indica sub-species of rice, the most commonly cultivated, using whole genome shotgun sequencing with a coverage of 4X ; this data was not, however, utilized by IRGSP.

The draft sequence of the consortium has already proven useful for the scientific community working on rice. The localization and characterization of the rice genes, as well as the identification of numerous new genetic markers, will greatly accelerate varietal selection. However, the annotation work, which will lead to the identification of genes of other cereals, cannot be precisely achieved until the "finished" version of the rice genome is ready. The errors and discontinuities in the sequence have a major effect on gene prediction, such that the range of 32,000 to 50,000 genes announced by Syngenta should be considered simply as an order of magnitude. It is therefore important for the international effort to continue. Our capacity to feed the world in the 21st century will likely depend on this effort.