The objective of this project, funded by the Genomics initiative Génoplante, is to generate through Agrobacterium-mediated transformation a library of 50,000 insertion lines of rice cv. Nipponbare. In a 2 year time-frame, we produced more than 39,400 primary transformants, with an average efficiency of 4.2 independent transformation events per cocultured callus. Each line was found to harbour an average of 2.2 and 3.2 copies of the T-DNA and of new inserts of the endogenous ty1-copia retrotransposon Tos17 respectively. 89% of the 14,893 walk-PCR products of amplification of the regions flanking the left border of T-DNA inserts yielded a readable sequence. A first survey of 7,480 genomic sequences against the rice BAC/PAC sequences detected 5,791 hits (77%), consistent with the percentage of the rice genome sequenced at time of analysis (76%). 4,303 (56%) T-DNA insertion sites were assigned to at least one position on the rice genome. The T-DNA density averaged 17.9 insertions per Mb sequenced and ranged from 14 to 20 insertions per Mb across the 12 chromosomes. A more detailed examination of the distribution of T-DNA inserts along the Chromosome 1 showed that 668 (92.8%) of 720 T-DNA insertions are assigned to a unique location. The results also demonstrated that i. the centromeric region exhibits a lower insertion density whereas higher insertion density is observed in the subtelomeric regions, ii. T-DNA inserts are very rarely found in repetitive DNA, iii. T-DNA seems to preferentially integrate promoter and intron regions of the predicted rice genes. Frequencies of putative insertions in both chr1 genes predicted by TIGR annotation and transcription factors identified by the whole genome sequencing (Goff et al 2002) led to the same conclusion that about 7-10% of the rice gene complement is already covered by the 6,111 tagged lines characterized in the collection. (Texte intégral)