Inclusive Evaluation of the Tetraploid Wheat Germplasm based on the Svevo Durum Wheat Genome Sequence Assembly

Triticum turgidum genetic resources are a wide reservoir of diversity, valuable for both durum wheat and common wheat pre-breeding, as highlighted by the success of synthetic wheats. However, a comprehensive analysis of these resources is lacking. As part of the Svevo durum wheat genome sequencing, we assembled a Global Tetraploid wheat Collection (GTC) of 1,856 accessions from 11 tetraploid wheat taxa, including wild and domesticated emmer (WEW and DEW) durum landraces (DWL) and modern durum wheat cultivars (DWC). We used the iSelect 90K SNP array anchored to the Svevo genome to investigate population structure and selection/demography signatures. We traced: i) the inheritance of haplotype blocks from either the North-eastern or Southern Levantine Fertile Crescent (Turkey) WEW populations to whole DEW germplasm, with a 0.65/0.35 overall inheritance ration, ii) the two subsequent independent but similar star-like dispersal patterns associated to DEW and DWL evolution, with six main populations each. Durum wheat most probably originated from the Southern-levant DEW. Ethiopian emmer and durum, T. turanicum and T. carthlicum were the most differentiated with minimal contribution to modern durum. Modern durum originated mostly from the North-African and Transcaucasian DWL populations. WEW, with the highest genome-wide diversity, provided the reference for assessing the diversity reductions associated to domestication and breeding. Numerous strong diversity depletions signals were observed primarily for the WEW-to-DEW transition. These signals progressively consolidated through domestication and breeding. Specific DEW-to-DWL signals were also observed. Diversity reduction index (DRI), Fst, haplotype-based XP-EHH and hapFLK and spatial pattern of site frequency spectrum (XP-CLR) metrics were considered. Frequently two or more indexes occurred in overlapping regions (selection clusters). In total, 104 pericentromeric (average size of 107.7 Mb) and 350 non-pericentromeric (average size of 11.4 Mb) clusters were identified. WEW-to-DEW and DEW-to-DWL transitions mostly involved extended pericentromeric regions tagged by DRI and Fst, while the DEW-to-DWL and DWL-to-DWC transitions were mostly associated to numerous XP-EHH and XP-CLR signals. We carried out a comprehensive projection of the published tetraploid QTLome and investigated the genome-wide QTL distribution and co-location between meta-QTLs and demography/selection signals. The usefulness of the GTC to elucidate the evolutionary patterns associated to loci, haplotype blocks and causative sequence variants for traits of breeding interest like cadmium-cumulation in grains, grain size and disease resistance loci are reported.