Legume Genomics and Genetics 2025, Vol.16, No.3, 100-107 http://cropscipublisher.com/index.php/lgg 104 yields, but also allow broad beans to be planted in places that could not be planted before. In order to make broad beans more adaptable to abiotic stresses such as cold, breeding work needs to pay special attention to selecting strains with strong resistance. This is critical to improving the planting effect in temperate regions (Duc et al., 2015; Rubiales, 2023). 6 Integration of Multi-Omics for Trait Mining 6.1 Use of transcriptomics and GWAS to identify yield QTLs Transcriptome and GWAS are commonly used methods to study yield and other complex traits in legumes. Combining these two types of data can find genes related to yield, seed composition, stress resistance, and quantitative trait loci (QTL). These methods have been used in soybeans and other crops. The results are also good. Not only have many important genes been found, but some useful molecular markers have also been developed (Chaudhary et al., 2015; Yang et al., 2021). Now, with multi-omics databases and platforms, it has become more convenient to find these candidate genes and understand their regulatory mechanisms. These tools have greatly improved the efficiency of identifying traits (Yang et al., 2023). 6.2 Genomic prediction models for complex traits If only one type of data is used for prediction, the effect is usually not ideal. But if several different data are put together, such as genome, transcriptome and methylation group data, the prediction results will be more accurate. This approach can help us understand more clearly which genes are affecting trait changes, how they interact and how they are regulated. Studies have also found that this multi-omics approach performs better than traditional methods in various environments (Wang et al., 2024). Now, many breeding projects have also begun to use machine learning and statistical methods to process these data. Methods such as random forests and multi-kernel learning can make predictions more stable and reliable (Acharjee et al., 2016; Briscik et al., 2024). 6.3 Combining metabolomic and phenomic data for precision breeding If we use metabolome and phenotypic data together with other omics data, we can see more clearly which molecules and physiological mechanisms affect yield and quality. In this way, a predictive model from genes to traits can be established to help us identify important metabolic pathways and biomarkers (Hu et al., 2021). This information can help us improve target traits faster, shorten breeding time, and make it easier to breed legumes with good performance. 7 Concluding Remarks Legumes have rich genetic diversity, which is the basis for their improvement. These genetic differences can help us increase yields, enhance disease resistance, and improve nutritional quality. In particular, wild species and local varieties carry many unique genes. These genes are very helpful for breeding. Using these variations, we can cultivate legumes that can adapt to climate change, resist pests and diseases, and adapt to different environments. This diversity also affects some important traits, such as some varieties have well-developed root systems and strong drought resistance; some can coexist with rhizobia to help plants fix nitrogen. These characteristics not only increase yields, but also protect soil and the environment. In the future, breeding should combine traditional methods with modern technologies. For example, genomic tools, phenotyping, and rapid breeding can be used to make breeding more efficient. Introducing good genes from wild species, that is, pre-breeding, can help increase genetic diversity. This can solve the problem of stagnant yield growth. There are many advanced tools now, such as high-throughput phenotyping, modeling software, gene editing, etc., which can help us find good traits faster and improve breeding efficiency. Breeders, agricultural experts and farmers also need to work together. They have to jointly decide the breeding goals, considering both yield and sustainable development. In terms of technology, digitizing germplasm resources is an important step. If combined with multi-omics data, artificial intelligence and some intelligent analysis tools, we can greatly change the way of breeding. Through these databases and models, we can find useful traits and excellent genes faster and make more scientific choices.
RkJQdWJsaXNoZXIy MjQ4ODYzNA==