Ancient diversity of Triticum aestivum subspecies as source of novel loci for bread wheat improvement

Ancient subspecies of hexaploid wheat, not yet subjected to intensive selection, harbor potentially valuable alternative genetic variability for the genetic improvement of modern cultivated bread wheat. To investigate these hitherto unexplored resources, we established a panel, currently unique, consisting of 190 accessions of <i>Triticum aestivum</i> belonging to five different neglected subspecies, <i>compactum</i>, <i>sphaerococcum</i>, <i>macha</i>, <i>spelta</i>, and <i>vavilovii</i>, with few <i>aestivum</i> references. The panel was genotyped through the iSelect Illumina arrays (20K and 25K) and phenotyped for 25 traits related to phenology, morphology, yield, and physiology for 4 years under field conditions. We found wide variability for all traits analyzed, both within and among subspecies, demonstrating the richness contained therein. Through a genome-wide association study (GWAS), we identified a total of 126 marker-trait associations (MTAs), including 4 for years, 58 for morphological traits, 39 related to yield, and 25 for physiological traits, some of them confirming loci previously published and others being novel. Fourteen MTAs were associated with multiple traits. Among them, one on chromosome 2D at 360.2 Mb was associated with spike density, length, and shape, and thus is of particular interest because it may underlie the compactum (<i>C</i>) gene, until now considered difficult to clone because of its centromeric position. The physical distance defined by this MTA is considerably smaller (1.7 Mb) than what is reported so far in the literature, paving the way toward physical mapping of the <i>C</i> gene. A potential candidate gene has been identified for the trait grain number per spike. This is <i>TraesCS6A03G0476500</i>, coding for a monosaccharide-sensing protein 2, located on chromosome 6A at 233 Mb and identified through an MTA that segregates exclusively in <i>compactum</i> accessions. The results obtained confirm the remarkable potential present in the panel of wheat subspecies analyzed in this study, which, being characterized by a very short linkage disequilibrium (LD) decay, allowed the definition of rather narrow ranges around key traits, such as those related to yield, providing new perspectives on transferring genes across subspecies for wheat improvement.