Recently, the maize team of Huazhong Agricultural University published a paper at Genome Biology entitled "Target-Oriented Prioritization: Targeted Selection Strategy by integrating organismal and molecular traits through predictive analytics in breeding" Research papers.
Based on a non-complete two-row hybrid population of maize containing 5820 hybrids designed in combination with genetic research and breeding applications, a set of machine learning algorithms based on ideal target material recognition: TOP, target-oriented prioritization. The algorithm can integrate omics data to achieve collaborative selection of multiple traits, and on the basis of ensuring the overall consistency of breeding goals, the specific traits can be achieved better, which provides technical support for intelligent crop design and breeding.
Crop breeding technology is facing new transformation and upgrading. The use of gene editing and synthetic biology technology, with the help of biological and environmental big data and artificial intelligence technology, rapid aggregation of favorable alleles, to achieve the targeted intelligent breeding of new crop varieties is considered to be the future development direction of breeding technology.
Synergistic improvement of multiple traits is currently the key to restricting breeding efficiency. In crop breeding, breeders usually expect to improve multiple traits at the same time, but there are often chains of burdens in different traits, such as high-yielding varieties that are often not disease-resistant, and varieties suitable for mechanized harvesting have rapid grain dehydration, but yields will be affected. Choosing two or more traits at the same time is often more difficult than choosing a single trait to breed. In actual breeding, there are three main methods of multi-trait selection: one is the item-by-item selection method, which selects only one trait in a breeding cycle; the other is the independent elimination level method, which selects multiple traits at the same time in a breeding cycle and intersects the materials that meet the conditions; and the third is the exponential selection method, which is weighted according to the economic importance of the traits or the extent of the expected improvement. Although the exponential selection method is more effective than the item-by-item selection method and the independent elimination level method, the assignment of trait importance depends on the experience of breeders, and the ideal selection index must be established for each specific population and breeding target, which is very difficult to operate and difficult to generalize.
Dr. Wenyu Yang, the first author of the article, has developed a set of DNA profiling techniques for crops, targeting specific varieties (commercial varieties or district test control materials), conducting "phenotype portraits" of materials through genomic information in breeding resources, and searching for materials that are most similar to the "target portrait" as a whole. The method is named target-oriented prioritization (TOP). The study used four independent sets of different datasets to test the effect of TOP selection, including 5820 F1 maize hybrids, 368 maize inbred lines, 282 maize inbred lines, and 210 rice inbred lines. The results of the study show that the TOP method has wide applicability in multiple species and multiple data, which can effectively balance the complex correlation between multiple traits, and screen out candidate materials with better specific traits under the premise of overall similarity with specific target varieties. If other omics big data is further added, the selection accuracy of TOP can be further greatly improved. Taking the large-scale promotion of maize variety "Zhengdan 958" in mainland production as the target material, 86 (selection rate of 0.25%) were selected from 34188 hybrid combinations that can be combined theoretically, and the field experiment verification of these selected hybrid combinations was further carried out, and the results showed that 10 hybrid combinations achieved a yield increase of 0.75% to 8.66% on the basis of the overall traits similar to "Zhengdan 958", which provided excellent material resources for subsequent accurate variety improvement, compared with conventional hybrid breeding , greatly reducing the workload.
Maize genome breeding selection TOP algorithm flow
Dr. Yang Wenyu and Professor Yan Jianbing's team carried out collaborative research, using the CUBIC population and various data built by the team in the early stage, combined with their own understanding of biological data and mathematical professional advantages, and strived to learn the knowledge of genetics and biological breeding, through the intersection and integration of disciplines, made a series of research results: developed a identity-by-descent (IBD) inference method based on the hidden Markov model. The genetic recombination events of 24 parents in the CUBIC population can be accurately estimated, and the accuracy reaches 95%, which lays the foundation for subsequent association analysis and gene mining, and the corresponding results are published in Genome Biology as a common work. Participated in the development of hybrid advantage prediction algorithm for NCII hybrid population derived from CUBIC, and independently developed a set of machine learning algorithm TOP based on ideal target material recognition, providing strong technical support for intelligent breeding of crop genomes.
Using the maize CUBIC population, it has published three consecutive series of papers at Genome Biology, from genetic analysis of agronomic traits to analysis of heterosis mechanisms to intelligent design breeding.
Dr. Yang Wenyu, a young teacher in the College of Science of Huazhong Agricultural University, is the first author of the paper. Professor Yan Jianbing and Professor Xiao Yingjie of the State Key Laboratory of Crop Genetic Improvement and Hubei Hongshan Laboratory are co-corresponding authors. Professor Guo Tingting and postdoctoral fellow Luo Liangyun of Huazhong Agricultural University, Dr. Marilyn Warburton of the UsDa, Researcher Zhao Jiuran and Associate Researcher Zhang Ruyang of beijing Academy of Agriculture and Forestry Sciences also participated in the study.
Original link:
https://genomebiology.biomedcentral.com/articles/10.1186/s13059-022-02650-w
Scientific research services (cooperative investment)
Patent Application | Rapid Detection of Transgenic Plants| Tissue Culture| Yeast Double Hybrid | Rice Mutant Library| Plant Genetic Transformation, Vector Construction, BIFC, RACE, Subcellular Localization, Genome Editing | Genetics technical services, SNP typing, high homologous segment SNP typing, target interval resequencing, methylation sequencing | Plant antibodies |
The forefront of plant science, focusing on the frontier progress of plant science, information, the release of recruitment information and method software sharing. For submission and recruitment, please reply to the "submission" in the background, all of which are free; for business cooperation, please contact WeChat ID: zwkxqy;