Genomic Selection As a Breeding Tool for Improving Wheat Grain Yield through the Introgression of Exotic Alleles from the D-Genome Progenitor Aegilops tauschii

Hexaploid wheat, Triticum aestivum, is one of the three most important cereal crops in the world. In order to meet the estimated world demand by 2050, a rapid and precise method is needed to incorporate genetic variation from wild relatives into adapted germplasm. Limited hybridization events between the D-genome progenitor, Aegilops tauschii, and tetraploid wheat have occurred since polyploidization. An advanced backcross (AB) population of 407 inbred backcross lines (IBL) was created through direct hybridization of seven Ae. tauschii accessions and a hard white recurrent parent, 'KS05HW14', to restore normal chromosome number and adapt lines for yield evaluation. An AB population incorporates wild alleles into elite germplasm while allowing for QTL mapping of important agronomic traits. An alternative to QTL mapping is genomic selection (GS). GS estimates breeding values (GEBVs) of individuals using a model built with marker and phenotypic data from a training population. As the correlation between the GEBV and the phenotype increases, the effectiveness of GS increases. IBL’s were phenotyped for grain yield in eight environments with one replicate each: 2015 and 2016 in Manhattan and Hays, Kansas and Richville, Michigan and 2016 in Champaign, Illinois and Marianna, Arkansas. Entry-mean heritability ranged from 0.35 to 0.66. Genotyping-by-sequencing produced 29,679 markers. Accuracies for within environment predictions ranged from 0.23 to 0.53 using three GS models, while across environments ranged from 0.30 to 0.50. Marker-by-environment models improved prediction accuracies for nine environmental combinations. These results indicate GS is useful for integrating exotic alleles into current varieties.