Genomic selection (GS) has provided a comprehensive framework for efficient breeding by linking phenotypes to genome-wide markers. However, research on GS has predominantly focused on improving genotype-to-phenotype prediction models, often overlooking optimal cross design, which determines the potential of progeny selection and plays a critical role in crop breeding. In this study, an efficient GS framework, EMLGP (ensemble machine-learning for genomic prediction), was proposed for optimal cross design in crop breeding. EMLGP first employs machine-learning algorithms to train precise genotype-to-phenotype prediction models in a germplasm population and then integrates with genome simulations to predict optimal crosses in a breeding population. GS model training of 14 soybean traits demonstrated that EMLGP achieved superior performance, with the highest prediction accuracy (correlation coefficient) reaching 0.92. The prediction accuracy showed a maximum improvement of 35.85% over the classical GBLUP method. Further simulation studies confirmed that EMLGP exhibited robust performance under conditions of small-to-moderate sample sizes (300–5000), low-to-moderate trait heritabilities (0.4–0.6), and complex genetic architectures (100 causal loci). Validation using real data of rice, maize, cotton, sorghum, and switchgrass consistently affirmed EMLGP’s superiority, outperforming GBLUP and deep learning methods. Among the 14 soybean traits analyzed, 13 traits exhibited transgressive segregation potential in the progeny. Specifically, seed linolenic acid content in the northern China showed the highest recombination potential, exceeding the maximum parental value by 16.89%. In conclusion, EMLGP optimizes parental selection and phenotypic prediction, offering a robust framework for efficient, intelligence-driven crop breeding.

The soybean cropping system involves its distribution across the country, the lighting time, accumulated temperature and cropping system of the varieties, the rotation system, as well as the monocropping, intercropping and relay intercropping methods, serves as the foundation for soybean production, breeding, introduction, and technology innovation. Optimizing the soybean cropping system is of decisive significance for enhancing the comprehensive production capacity and benefits of soybeans in China. Since the founding of the People's Republic of China (PRC) 70 years ago, the area planted with soybeans in regions with one crop per year system has expanded, while the area in regions that have shifted from triple crops per two years system to double crops per year system has decreased. In areas that have transitioned from double crops per year and then to triple crops per year, the area planted with soybeans has remained stable with a slight increase. From a national perspective, the soybean cultivation region has expanded to the northern part of Northeast China, and the soybean cultivation region in the South and Southwest has remained stable with a slight increase. The Northwest region has performed a new high-yield area for soybeans. Historically, the division of soybean cultivation regions was based on the basic data, investigations and experiments of the planting system at that time. In the recent 30 years, there have been significant advancements in soybean production, breeding and cultivation techniques, especially in the changes of soybean cultivation areas. The division of ecological cultivation region is a fundamental task closely related to soybean cultivation, resource utilization, introduction and breeding for cultivars. Based on the review of the changes in soybean cultivation region in China since the PRC establishment, including the northward expansion and southward shift of cultivation region, the renewal and upgrading of varieties, the improvement of mechanization levels, the comprehensive progress of cultivation techniques, and the promotion of intercropping system, especially the emphasis on developing the soybean industry as a national policy in China since 2000, this review comprehensively analyzed the dynamic characteristics of the soybean cropping system and technical system in PRC and thus proposed suggestions for adjusting the ecological cultivation region divisions of soybeans. From which a new soybean ecological cultivation region system is proposed. The main results comprise the changes in soybean cropping regions and the advances in cropping system, the environmental cultivation regions and changes of soybeans, the ecology of modern soybeans in China, and discussion and prospect on ecological cultivation region of soybeans in China. Influenced by updates of soybean cultivars, advancements in cultivation and farming technology, and requirements on food security, the soybean cropping system has undergone significant changes. The new six ecological cultivation regions were suggested as Northeast Spring Planting Soybean Ecological Cultivation Region, Northwest Spring Planting Soybean Ecological Cultivation Region, Huang-Huai-Hai Summer Planting Soybean Ecological Cultivation Region, Changjiang Valleys Spring-Summer-Autumn Planting Soybean Ecological Cultivation Region, Southwest Plateau Spring-Summer Planting Soybean Ecological Cultivation Region, and South China All Season Planting Soybean Ecological Cultivation Region. This division and naming system is considered as consistent as that of the national crop cultivation system, and also pays attention to the connection with previous ecological cultivation region division systems in soybean.

Shade tolerance is essential for soybeans in inter/relay cropping systems. A genome-wide association study (GWAS) integrated with transcriptome sequencing was performed to identify genes and construct a genetic network governing the trait in a set of recombinant inbred lines derived from two soybean parents with contrasting shade tolerance. An improved GWAS procedure, restricted two-stage multi-locus genome-wide association study based on gene/allele sequence markers (GASM-RTM-GWAS), identified 140 genes and their alleles associated with shade-tolerance index (STI), 146 with relative pith cell length (RCL), and nine with both. Annotation of these genes by biological categories allowed the construction of a protein–protein interaction network by 187 genes, of which half were differentially expressed under shading and non-shading conditions as well as at different growth stages. From the identified genes, three ones jointly identified for both traits by both GWAS and transcriptome and two genes with maximum links were chosen as beginners for entrance into the network. Altogether, both STI and RCL gene systems worked for shade-tolerance with genes interacted each other, this confirmed that shade-tolerance is regulated by more than single group of interacted genes, involving multiple biological functions as a gene network.

"Breeding by design" for pure lines may be achieved by construction of an additive QTL-allele matrix in a germplasm panel or breeding population, but this option is not available for hybrids, where both additive and dominance QTL-allele matrices must be constructed. In this study, a hybrid-QTL identification approach, designated PLSRGA, using partial least squares regression (PLSR) for model fitting integrated with a genetic algorithm (GA) for variable selection based on a multi-locus, multi-allele model is described for additive and dominance QTL-allele detection in a diallel hybrid population (DHP). The PLSRGA was shown by simulation experiments to be superior to single-marker analysis and was then used for QTL-allele identification in a soybean DPH yield experiment with eight parents. Twenty-eight main-effect QTL with 138 alleles and nine QTL × environment QTL with 46 alleles were identified, with respective contributions of 61.8% and 23.5% of phenotypic variation. Main-effect additive and dominance QTL-allele matrices were established as a compact form of the DHP genetic structure. The mechanism of heterosis superior-to-parents (or superior-to-parents heterosis, SPH) was explored and might be explained by a complementary locus-set composed of OD+ (showing positive over-dominance, most often), PD+ (showing positive partial-to-complete dominance, less often) and HA+ (showing positive homozygous additivity, occasionally) loci, depending on the parental materials. Any locus-type, whether OD+, PD + and HA+, could be the best genotype of a locus. All hybrids showed various numbers of better or best genotypes at many but not necessarily all loci, indicating further SPH improvement. Based on the additive/dominance QTL-allele matrices, the best hybrid genotype was predicted, and a hybrid improvement approach is suggested. PLSRGA is powerful for hybrid QTL-allele detection and cross-SPH improvement.

Soybean mosaic virus (SMV) is a member of the genus Potyvirus that extensively impairs global soybean production. The full-length coding sequence of the MADS-box transcription factor GmCAL was cloned from the SMV-resistant soybean cultivar Kefeng 1. SMV-induced expression analysis indicated that GmCAL responded quickly to SMV-SC8 infection in Kefeng 1 but not in NN1138-2. GmCAL was expressed at high levels in flowers and pods but at lower levels in leaves. The gene was localized to the nucleus by subcellular localization assay. Virus-induced gene silencing did not increase the accumulation of SMV in GmCAL-silenced Kefeng 1 plants (with silencing efficiency ∼ 80%) after SC8 inoculation. GmCAL-silencing plants still conferred resistance to SC8 that might be owing to incomplete silencing of genes with lower expression. SMV content decreased significantly in GmCAL-overexpressing NN1138-2 plants after SMV-SC3, SMV-SC7, and SMV-SC8 inoculation in comparison with a vector control, showing that overexpression of GmCAL conferred broad-spectrum resistance to multiple SMV strains. These results confirm that GmCAL, a key regulator but not a specific SC8 resistance gene (Rsc8), is a positive regulatory transcription factor involved in soybean resistance to SMV.

Soil flooding stress, including seed-flooding, is a key issue in soybean production in high-rainfall and poorly drained areas. A nested association mapping (NAM) population comprising 230 lines of two recombinant inbred line (RIL) populations with a common parent was established and tested for seed-flooding tolerance using relative seedling length as indicator in two environments. The population was genotyped using RAD-seq (restriction site-associated DNA sequencing) to generate 6137 SNPLDB (SNP linkage disequilibrium block) markers. Using RTM-GWAS (restricted two-stage multi-locus multi-allele genome-wide association study), 26 main-effect QTL with 63 alleles and 12 QEI (QTL × environment) QTL with 27 alleles in a total of 33 QTL with 78 alleles (12 dual-effect alleles) were identified, explaining respectively 50.95% and 14.79% of phenotypic variation. The QTL-alleles were organized into main-effect and QEI matrices to show the genetic architecture of seed-flooding tolerance of the three parents and the NAM population. From the main-effect matrix, the best genotype was predicted to have genotypic value 1.924, compared to the parental value range 0.652–1.069, and 33 candidate genes involved in six biological processes were identified and confirmed by χ² test. The results may provide a way to match the breeding by design strategy.

A fast-growing protein and oil crop, soybean was domesticated in ancient China and disseminated early in Asia and afterwards to other continents, in particular the Americas in recent centuries. After adaptation, locally developed landraces and cultivars formed a diversity of geographic-populations. In an investigation of their phylogeographic features, marker-derived traits were combined with geography-related photo- and temperature-sensitive traits to study 13 geographic-populations comprising 371 accessions. Extreme differentiation among geographic-populations was observed for flowering date (33–94 days), maturity date (79–181 days), and main stem node number (6–25 nodes). Restriction-site associated DNA sequencing revealed strong genetic differentiation among these geographic-populations, including genetic richness (alleles, 35,242–44,986) and specific-present alleles (SPAs, 0–67). More SPAs (28–67) emerged in some secondary and tertiary centers than in centers of origin (8–11). Phenotypic and genotypic clustering divided 11 of the 13 geographic-populations into the same five sets of sensitivity-similar geographic-populations and grouped the populations of northeast China and northern North America rather than center-of-origin populations as secondary centers, indicating the importance of geography-related traits in determining genetic differences among geographic-populations. A model of four soybean dissemination paths is presented: from the center of origin to the north, east, and south in Asia and from northeast China to Europe and the Americas. These findings provide a detailed phylogeographic understanding of worldwide soybeans.