Progress and perspectives on genomic selection models for crop breeding

Genomic selection, a molecular breeding technique, is playing an increasingly important role in improving the efficiency of artificial selection and genetic gain in modern crop breeding programs. A series of algorithms have been proposed to improve the prediction accuracy of genomic selection. In this review, we describe emerging genomic selection techniques and summarize methods for best linear unbiased prediction and Bayesian estimation of the traditional statistics used for prediction during genomic selection. Moreover, with the rapid development of artificial intelligence, several machine learning algorithms are increasingly being employed to capture the effects of more genes to further improve prediction accuracy, which we describe in this review. We also describe the advantages and disadvantages of traditional models and machine learning models and discuss several crucial factors that could affect prediction accuracy. We propose that additional artificial intelligence techniques will be required for big data management, feature processing, and model innovation to generate a comprehensive model to optimize the prediction accuracy of genomic selection. We believe that improvements in artificial intelligence could accelerate the arrival of Breeding 4.0, in which combining any known alleles into optimal combinations in crops will be fully customizable.