Search

Although durum wheat (Triticum durum L. ssp. durum Desf.) has traditionally been used to make a range of food products, its use has been restricted due to the absence of the D-genome glutenin proteins, the relatively low variability in starch composition, and its very hard grain texture. This review focuses on the manipulation of the starch and protein composition and modification of the hardness of durum wheat in order to improve its technological and nutritional value and expand its utilization for application to a wider number of end products. Starch is composed of amylopectin and amylose in a 3:1 ratio, and their manipulation has been explored for achieving starch with modified composition. In particular, silencing of the genes involved in amylose and amylopectin synthesis has made it possible to isolate durum wheat lines with amylose content varying from 2–3% up to 75%. This has created opportunities for new products with different properties and enhanced nutritional value. Durum-made bread has generally inferior quality to bread made from common wheat. Attempts to introduce the Glu-D1 subunits 1Dx5 + 1Dy10 and 1Dx2 + 1Dy12 produced stronger dough, but the former produced excessively strong, inelastic doughs, and loaf volume was either inferior or not affected. In contrast, the 1Dx2 + 1Dy12 sometimes improved bread loaf volume (LV) depending on the glutenin subunit background of the genotype receiving these genes. Further breeding and selection are needed to improve the dough extensibility to allow higher LV and better texture. The versatility of durum wheat has been greatly expanded with the creation of soft-textured durum via non-GMO introgression means. This soft durum mills like soft hexaploid wheat and has similar baking properties. The pasta quality is also not diminished by the soft-textured kernels. The Glu-D1 locus containing the subunits 1Dx2 + 1Dy12 has also been introgressed to create higher quality soft durum bread.

Breeding for grain yield, biotic and abiotic stress resistance, and end-use quality are important goals of wheat breeding programs. Screening for end-use quality traits is usually secondary to grain yield due to high labor needs, cost of testing, and large seed requirements for phenotyping. Genomic selection provides an alternative to predict performance using genome-wide markers under forward and across location predictions, where a previous year’s dataset can be used to build the models. Due to large datasets in breeding programs, we explored the potential of the machine and deep learning models to predict fourteen end-use quality traits in a winter wheat breeding program. The population used consisted of 666 wheat genotypes screened for five years (2015–19) at two locations (Pullman and Lind, WA, USA). Nine different models, including two machine learning (random forest and support vector machine) and two deep learning models (convolutional neural network and multilayer perceptron) were explored for cross-validation, forward, and across locations predictions. The prediction accuracies for different traits varied from 0.45–0.81, 0.29–0.55, and 0.27–0.50 under cross-validation, forward, and across location predictions. In general, forward prediction accuracies kept increasing over time due to increments in training data size and was more evident for machine and deep learning models. Deep learning models were superior over the traditional ridge regression best linear unbiased prediction (RRBLUP) and Bayesian models under all prediction scenarios. The high accuracy observed for end-use quality traits in this study support predicting them in early generations, leading to the advancement of superior genotypes to more extensive grain yield trails. Furthermore, the superior performance of machine and deep learning models strengthens the idea to include them in large scale breeding programs for predicting complex traits.