Enhancing Genetic Gain In A Wheat Breeding Program Using Genomics Phenomics Machine And Deep Learning Algorithms

Classical plant breeding has evolved considerably during the last century. However, the rate of genetic gain is insufficient to cope with a 2% annual increase in the human population, which is expected to reach 9.8 billion by 2050. Plant breeders and scientists are under pressure to develop new varieties and crops having higher yield, higher nutritional value, climate resilience, and disease and insect resistance. The solution requires the merging of new techniques like next-generation sequencing, genome-wide association studies, genomic selection, high throughput phenotyping, speed breeding, machine and deep learning, and CRISPR mediating gene editing with previously used tools and breeder's skills. The main goal of this research was to explore the potential of genomics, phenomics, machine and deep learning tools in a wheat (Triticum aestivum L.) breeding program. Grain yield and grain protein content (GPC) are two traits very important in hard red spring wheat breeding, yet difficult to select for due to their well-known negative correlation. A nested association mapping population was used to map the regions controlling the stability of grain protein content. This study also demonstrated that genome-wide prediction of GPC with ridge regression best linear unbiased (rrBLUP) estimates reached up to r = 0.69. Genomic selection (GS) is transforming the field of plant breeding and implementing models that improve prediction accuracy for complex traits is needed. Analytical methods for complex datasets traditionally used in other disciplines represent an opportunity for improving prediction accuracy. We predicted five different quantitative traits with varying genetic architecture using cross-validations, independent validations, and different sets of SNP markers. Deep learning models gave 0 to 5% higher prediction accuracy than rrBLUP model under both cross and independent validations for all five traits used in this study. 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 previous years dataset can be used to build the models. Nine different models, including two machine learning and two deep learning models, 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. Genomics and phenomics have the potential to revolutionize the field of plant breeding. Incorporation of secondary correlated traits in GS models has been demonstrated to improve accuracy. In another study, ability to predict GPC and grain yield was assessed using secondary traits, univariate, covariate, and multivariate GS models for within and across cycle predictions. Our results indicate that GS accuracy increased by an average of 12 for GPC and 20% for grain yield by including secondary traits in the models. An increased prediction ability for GPC and grain yield with the inclusion of secondary traits demonstrates the potential to improve the genetic gain per unit time and cost in wheat breeding.