Molecular Plant Breeding 2024, Vol.15, No.1, 15-26 http://genbreedpublisher.com/index.php/mpb 18 With the rapid development of data analysis and artificial intelligence, significant progress has been made in applying these technologies in the field of plant breeding. Khan et al. (2022) discussed how AI is reshaping modern crop breeding by integrating “omics” approaches to better understand crops’ responses to environmental stress. They emphasized the role of AI in high-throughput phenotypic analysis and gene function analysis, highlighting how it improves the accuracy of crop phenotype, genotype, and environmental data. Hilli (2022) emphasized the application of ground and aerial platforms, as well as sensors, in characterizing crop phenotypes under different stress factors. The paper discussed how AI-based technologies enhance traditional breeding programs to meet the growing demands of agriculture. Xing et al. (2022) focused on the role of AI and computer vision (CV) in enhancing phenotype feature analysis in soybean breeding. They discussed how CV provides a high-resolution, cost-effective method for analyzing crop phenotypes, which is crucial for designing breeding programs. Sagan et al. (2022) proposed a data-driven approach for calibrating hyperspectral big data, which is crucial for high-throughput plant phenotypic analysis and breeding. Their work focuses on the automated calibration workflow for near-ground hyperspectral data. By comprehensively utilizing genetic information integrating, genome editing, high-throughput phenotypic analysis, and data analysis technologies, Breeding 4.0 provides breeders with more powerful and efficient tools and methods. This accelerates the process of crop improvement, achieving more precise and sustainable breeding goals. 3 Innovations and Advances in Breeding 4.0 3.1 Development of genomic prediction and genomic selection An important innovation in Breeding 4.0 is the development of genomic prediction and genomic selection. Genomic prediction utilizes genomic and phenotypic data to forecast the genetic value and expression of traits in individual crops. By establishing prediction models, breeders can rapidly and accurately assess the genetic potential of a large number of crop individuals, thereby making better selections for superior varieties. Genomic selection, on the other hand, employs genomic data to guide the breeding selection process. Through genomic selection, breeders can directly choose individuals within the genome that possess the desired traits, accelerating the breeding process and improving selection efficiency. Recent studies indicate that significant progress and applications of genomic prediction and genomic selection in the field of plant breeding. Buntaran et al. (2022) utilized simulations to assess the response to genomic selection, emphasizing the necessity of selecting entries in plant breeding projects to maximize the genetic gains for the traits of interest. They highlighted the role of genomic prediction in improving prediction accuracy and accelerating the breeding cycle. Montesinos-López et al. (2022a) demonstrated that using incomplete block designs for series allocation in genomic selection can improve predictions in plant breeding. Compared to random allocation, this method performs better, enhancing the efficiency of resource optimization in breeding programs. Weiß et al. (2022) explored phenotype selection in maize, and used near-infrared spectroscopy data for prediction. They found that phenotype prediction was minimally influenced by population structure, especially when dealing with diverse germplasm, making it a promising tool in practical breeding. Montesinos-López et al. (2022b) compared gradient boosting machines and Bayesian threshold BLUP for genomic prediction of categorical traits in wheat breeding. They found that gradient boosting machines generally outperform Bayesian models, indicating the need for further research on this approach in the context of genomic selection. Feldmann et al. (2022) discussed the impact of different methods for calculating genomic relationship matrices on genomic variance and heritability estimates, proposing a novel matrix that produces accurate estimates in both plants and animals.
RkJQdWJsaXNoZXIy MjQ4ODYzMg==