Artificial Intelligence-Assisted Genomic Prediction for Plant Breeding

Genomic prediction (GP) is a breeding strategy that leverages genome-wide genetic markers to predict complex traits, making it a favored method for selecting optimal breeding candidates in modern plant and animal science. Traditional GP models such as genomic best linear unbiased prediction, ridge regression best linear unbiased prediction, Bayesian approaches, least absolute shrinkage and selection operator, elastic net, penalized linear models, and reproducing kernel hilbert spaces regression have been widely employed for genotype-to-phenotype predictions. However, the high dimensionality of marker data and the intricate patterns they encode often limit the performance of these traditional methodologies. To overcome these challenges, artificial intelligence (AI) has introduced innovative solutions that offer new perspectives on longstanding issues in plant breeding. AI algorithms such as convolutional neural networks, recurrent neural networks, and multi-layer perceptrons are capable of processing genomic and phenomic datasets with remarkable speed and precision, enabling the accurate prediction of complex traits. For instance, genomic estimated breeding values can be derived solely from genotype data, facilitating the prediction of breeding values for target populations and thereby accelerating the breeding cycle. This chapter articulates the integration of AI technologies, particularly machine learning and deep learning techniques, into genomic crop prediction. It highlights the transformative role of these methodologies in advancing plant breeding and underscores their potential to revolutionize genotype-to-phenotype predictions.