Résumé
A composite durum wheat population (Triticum turgidum, 2n=4X=28, A and B genomes) was created by crossing elite durum lines with wild (T. dicoccoides) and primitive emmer (T. dicoccum) accessions. A recessive male sterility gene was used in order to monitor the outcrossing rate in the population. Since 1997, this population is experiencing each year natural and anthropic selection with an outcrossing rate of 10 %. The EPO project aims at identifying SNPs from 100 fixed lines extracted from this important resource of diversity and to evaluate the potential of RNA seq for Genotyping by sequencing purposes. [br/] We obtained transcriptomics data for 50 individuals and built contigs using those data and standard NGS pipelines. As expected for a allotetraploid species, this first rough analysis demonstrated that for a large number of detected SNPs an excess of heterozygosity was observed. This seems to indicate that the simultaneous expression of the two A and B homeologous copies of many genes resulted in an erroneously merge of paralogous/homeologous sequences. This clearly pinpoints that such standard assembly protocols are not adapted do deal with polyploid species. [br/] We thus developed a dedicated assembly strategy that exploits strict and accession specific assemblies and Triticum monococcum and Aegilops speltoides diploids genomes to identify paralogs and homeologs in the durum wheat A and B genomes. This approach significantly increases the number of unambiguous within genome polymorphisms and offer promising perspective for genotyping by sequencing durum wheat and other polyploids of agronomic importance.