Résumé
The efficiency of the Bulk Segregant Analysis (BSA) had clearly been demonstrated to detect genomic regions and genes involved in diverse traits. It allows large experiments reducing the cost and time and preserving the power of full individual's population analysis. These past few years the combination of BSA and New Generation Sequence (NGS) data (BSA-Seq) gave a new accuracy and depth to the discovery on many traits of interest, mainly on crop and model species. In our study, we applied the BSA-Seq to narrow down Populus genomic regions involved in the resistance to Melampsora larici populina (Mlp) leaf rust. We worked on the 1417 progenies derived from an interspecific cross Populus deltoides clone 73028-62 (Pd) x Populus trichocarpa clone 101-74 (Pt) in which segregates qualitative and quantitative Mlp resistances. Four bulks were constituted based on (1) the phenotypes for uredinias size (bulk1: large, bulk2 and bulk3: intermediate, bulk4: small) and (2) the Pt genotypes at the RUS locus governing the uredinia size (bulk1 and bulk3 : [RUS], bulk2 and bulk4 : [rUS]). We used independently the soft masked genomes of Populus trichocarpa Nisqually v3.0 (Ptv3) and Populus deltoides v2.0 (Pdv2) as references to map parents and bulks Illumina reads with the BWAmem/0.7.15 suite and to detect DNA variations with Freebayes/0.9.21. For each common variant position between parents and bulks, we tracked the specific alleles of Pt and/or Pd in the bulks. Respectively from Ptv3 and Pdv2, we identified 13 and 11 regions. We first evaluated the strategy retrieving the previously finally mapped RUS gene, designed new genomic markers from Ptv3 to better characterize this locus and performed the in silico validation. Then we proceed with it to fine map the other regions. Most of them co-locate with QTL and could explain the resistance to Mlp. So we demonstrated that, in our context, even if the reference genome is different to the studied genomes, BSA-Seq allowed us to detect quickly and costly, regions which may be involved in a complex trait and improve the fine mapping of a major gene. We think it can be a promising method on a high heterozygous diploid genome as Poplar, to decipher complex trait. Next step is to identify candidate genes within the regions and better describe the mechanisms of resistance.