1 Introduction

Soybean (Glycine max) is actually not just the raw material of soy products that we are familiar with. Its role in the global agricultural system is more important than imagined. Rich in protein and high in oil content, along with some active substances beneficial to health, it not only makes its way onto human dining tables but also widely appears in animal feed and industrial products. Soybeans have "played a role" in global food security, farmers' income and the development of related industries (Azam et al., 2020).

When it comes to soybeans, one component that cannot be overlooked is isoflavones. This is a naturally occurring polyphenolic metabolite, which is particularly common in soybean seeds (Huang, 2024). There are mainly three types: genistein, soybean flavin and soybean flavin. They do not exist in just one form, but include various derivative forms such as aglycones, glycosides, acetyl glycosides, malonyl glycosides, among which malonyl glycosides are considered to have the highest content (Azam et al., 2024). These substances are also quite interesting-although they come from plants, they can imitate the effect of estrogen in the human body.

Many studies suggest that isoflavones are beneficial to health, such as helping to reduce the risk of certain cancers, relieve menopausal discomfort, improve bone condition, and even have positive effects on the cardiovascular and metabolic systems (Li et al., 2022). But their "functions" are not only useful to humans. In plants, isoflavones help cope with environmental stress and also promote mutualism with nitrogen-fixing bacteria (Ng et al., 2021). However, the content of isoflavones in different soybean germplasms varies greatly. Some literature indicates that this difference can reach 7 to 8 times. This variation is mainly influenced by genotype, environmental factors, and the stage of seed development.

This study will systematically review the currently known genetic mechanisms of isoflavone accumulation in soybean seeds, including QTL mapping, GWAS research progress, and strategies for integrating multi-omics data. It will also discuss the intervention effects of developmental and environmental factors. In addition, we will also explore the methods and challenges currently adopted at the breeding and biotechnology levels to increase the content of isoflavones, and put forward some thoughts and prospects on how to select and breed soybean varieties that are more suitable for health, nutrition or industrial uses.

2 Biosynthetic Pathway and Core Genes in Isoflavone Accumulation

2.1 Pathway from phenylpropanoids to isoflavones

The synthetic pathway of isoflavones does not start from the middle. It is actually a branch of the phenylalanine metabolic pathway, and the entire process starts with phenylalanine. This amino acid is converted into cinnamic acid under the action of PAL enzyme, and then C4H converts it into p-coumaric acid. Next comes the 4CL enzyme, which activates p-coumaric acid into p-coumaryl coenzyme A. At this point, Chalketone synthase (CHS) and chalketone reductase (CHR) are introduced. They cause the condensation of p-coumaryl-CoA with malonyl-CoA, eventually generating chalketones, such as naringenin chalketone or isorlicorice. Without chalcone isomerase (CHI), these intermediate products would not be able to be converted into naringenin or glycyrrhizin. But this is not the end. What truly determines whether one enters the isoflavone pathway is IFS, or isoflavone synthase. It can gradually convert flavanones into final products such as soy flavin, genistein and soybean flavin (Dastmalchi and Dhaubhadel, 2014).

2.2 Transport, storage, and glycosylation/malonylation

After isoflavones are synthesized, they do not simply "lie flat". The subsequent processing procedures are equally important, such as decoration, transportation and storage. The first step is glycosylation, which is carried out by UGT enzymes. This process converts them into glycoside forms, such as soy flavin glycosides, etc. However, if it only stays at this stage, it is still not stable enough. As a result, some glycosides will be further modified by malonyl transferase (MAT) to form malonyl glycosides. This kind of structurally stable form is the most dominant storage method in seeds (Kim et al., 2024b). The storage location is not arbitrary either. These isoflavones are eventually transported into the vacuole and are handled by specialized vacuole transport proteins. The expression of UGT and MAT is not necessarily stable. They are greatly influenced by genotype and developmental stage, which directly determines the accumulation degree of different varieties of isoflavones (Zhang et al., 2023).

2.3 Seed-stage specificity and tissue localization

The accumulation of isoflavones does not stably exist from the moment the seed is formed. It is actually very "time-dependent", growing the fastest between R5 (seed filling begins) and R7 (physiological maturity) (Figure 1) (Kim et al., 2021). However, the expression varies among different genotypes; some increase rapidly while others do so slowly. Inside the seeds, the distribution of these compounds also has a "preference". Although the hypocotyl, cotyledons and other parts are synthesized, the cotyledons have the highest content (Zernova et al., 2009). Furthermore, whether it is biosynthetic genes such as CHS and IFS, or UGT and MAT involved in modification, their expression time points are mostly concentrated in the later stage of seed development, precisely when the content of isoflavones begins to increase significantly.

3 Natural Variation and Genetic Architecture of Isoflavone Content in Soybean Seeds

3.1 QTL and GWAS signals for isoflavones across diverse germplasm

Different studies have employed QTL mapping or GWAS methods to identify tens of thousands of SNPS related to isoflavone content in various soybean populations. These research results are not exactly the same, but most of the QTLS with major effects are concentrated on chromosomes 5, 8, 11 and 19. For instance, a major QTL (qMGly_11) found on chromosome 11 can explain up to 79.9% of malonyl-soybean flavin variations in certain populations (Watanabe et al., 2019). qIF5-1 on chromosome 5 can explain 52.5% of the total isoflavone variation (Cai et al., 2018). Of course, not all QTLS have such a strong explanatory power. Some studies have also collated over 1,000 QTLS through meta-analysis and found that some of these loci are relatively stable, while others only function in specific environments (Chen et al., 2021). These differences indicate that the genetic signals exhibited may vary under different germplasms and conditions.

3.2 Allelic and haplotype diversity in key loci

The differences in isoflavone content are ultimately determined by the natural variations of core synthetic genes and regulatory genes. The repetitive isoflavone synthase genes such as IFS1 and IFS2, as well as the CHS family, MYB29, MYB77 and some bZIP class transcription factors, all demonstrated significant allelic and haplotype diversity (Chu et al., 2017). Not every haplotype brings high content, but some can indeed significantly increase the accumulation amount. For instance, a variant of GmMYB29 is associated with an increase in isoflavones, while the content of GmMYB77 may decrease (Liu et al., 2024). For instance, the allelic differences in the Glyma.11G108100 and GmMPK1 genes also show a significant association with the accumulation of isoflavones.

3.3 Effect sizes, epistasis, and polygenic background

Not all sites related to isoflavones can explain a large proportion of the phenotypic differences. Although some major QTLS have made significant contributions, on the whole, the trait of isoflavone content is controlled by many genes together. The interaction between genes should not be ignored either, and some may even affect the effects of existing QTLS. For instance, QTLS on chromosomes 5 and 6 have been confirmed to have additive effects, which can further increase the content of isoflavones (Zhao et al., 2022). However, such upper-level interactions can make the overall genetic structure more complex, and especially in different environments, this interaction effect can change (Kim et al., 2023; Pei et al., 2018). For this reason, high-density mapping and cross-environment experiments become very crucial. Only in this way can those stable sites suitable for label-assisted selection be screened out (Wang et al., 2015).

4 Regulatory Networks and Multi-Omics Insights in Isoflavone Accumulation

4.1 Transcriptional control by MYB-bHLH-WD40/NAC factors and co-expression modules

The synthesis of isoflavones is not a random process. In soybean seeds, it is controlled by a complex transcriptional regulatory network. In fact, MYB protein is a "key player", but it doesn't fight alone. Members of R2R3-MYB such as GmMYB3a, GmMYB29 and GmMYB68 not only activate key genes such as CHS7, CHS8 and IFS2, but also significantly increase the accumulation of isoflavones once they are overexpressed (Xu et al., 2025; Xu et al., 2024. However, MYB alone is not sufficient. These factors often "team up" with bHLH and WD40 proteins to form complexes, which conduct more refined regulation of structural genes in the phenylpropane pathway (Zhang et al., 2023). Co-expression network analyses like WGCNA have also helped us discover some modules containing MYB, bZIP and NAC factors, especially key genes such as Glyma.11G108100 and Glyma.11G107100. The relationship between them and isoflavone content is very close (Wu et al., 2024). These networks indicate that during the seed development stage, biosynthesis and modified genes actually cooperate quite closely.

4.2 eQTL/mQTL integration linking regulatory variants to metabolite levels

Sometimes, the source of regulatory differences is not obvious, but once the data of eQTL and mQTL are integrated, the problem becomes clear. For instance, there is a SNP in the 5 '-UTR of GmMYB29. It may seem like just a positional change, but it can affect gene expression and thereby influence the level of isoflavones. This kind of natural variation is actually not uncommon. After combining GWAS with co-expression analysis, it is possible to precisely identify regulatory factors such as MYB4 and bZIPs, and even regulatory variant points in structural genes GmOMT7 and GmIE3-1 (Azam et al., 2023a). This type of regulation can be either cis or trans-it has become increasingly clear that these different types of regulatory elements jointly participate in the changes in the accumulation of isoflavones among different germplasms.

4.3 Epigenetic layers

In addition to transcriptional regulation, epigenetic factors are also quite worthy of attention. For key genes like IFS1 and IFS2, the degree of cytosine methylation in their coding regions and promoter regions is often associated with high isoflavone content. In some highly expressed varieties, the methylation levels in these regions were significantly elevated (Gupta et al., 2019). Of course, epigenetics does not only look at methylation, whether chromatin is easy to open, and the state of histone modification, all of which will have an impact on regulation. At present, however, there are not many studies on these mechanisms during the development period of soybean seeds. But it's not completely at a loss. Through the integration of transcriptomic, epigenomic and metabolomic data, we have begun to gradually reveal the full regulatory picture behind isoflavone biosynthesis, including those previously overlooked "recessive levels" (Chen et al., 2023).

5 Environmental and Developmental Modulation of Isoflavone Content in Soybean Seeds

5.1 Genotype × environment

In addition to transcriptional regulation, epigenetic factors are also quite worthy of attention. For key genes like IFS1 and IFS2, the degree of cytosine methylation in their coding regions and promoter regions is often associated with high isoflavone content. In some highly expressed varieties, the methylation levels in these regions were significantly elevated (Gupta et al., 2019). Of course, epigenetics does not only look at methylation, whether chromatin is easy to open, and the state of histone modification, all of which will have an impact on regulation. At present, however, there are not many studies on these mechanisms during the development period of soybean seeds. But it's not completely at a loss. Through the integration of transcriptomic, epigenomic and metabolomic data, we have begun to gradually reveal the full regulatory picture behind isoflavone biosynthesis, including those previously overlooked "recessive levels" (Chen et al., 2023).

5.2 Seed developmental timing and hormonal cues

Isoflavones do not start to accumulate in large quantities as soon as the seeds emerge. It has its own habit of "picking the right time", usually reaching its peak during the grouting and maturation periods. This time point is affected by maturity type and photcycle. Some late-maturing varieties or varieties grown in cooler areas tend to accumulate more (Kim et al., 2024a). One more point that is often overlooked is the resource allocation within the seeds. If there is more carbohydrates, isoflavones are also likely to be high. However, if the protein and fat content is very high, isoflavones often do not increase. These changes are actually the result of the combined effect of multiple factors, including the developmental stage, environment, and the coordination among genotypes, which determine exactly how much isoflavones accumulate in the seeds in the end.

5.3 Stable vs. plastic loci

In fact, it is not that easy to steadily increase the content of isoflavones in the long term. Because this trait is a quantitative trait, it is not controlled by just one or two major genes. Some gene loci remain quite stable in different environments, but others do not. Their performance may change when the location is different or the climate varies. QTL and GWAS analyses conducted in multiple environments also revealed this phenomenon: some genes with relatively small effects, as well as the mutual influence among some genes, were particularly affected by the environment (Kim et al., 2020). Some varieties, no matter where they are grown, have a relatively high level of isoflavones. However, there are also some whose performance is inconsistent. This is why multiple years and multiple points of experiments have to be conducted repeatedly during breeding. Nowadays, many people use marker-assisted selection, aiming to superimpose those stable QTLS, select varieties that perform stably in different environments, and minimize the "interference" brought by the external environment (Gutierrez-Gonzalez et al., 2010).

6 Case Study

6.1 Discovery: multi-environment GWAS pinpoints a lead SNP near an IFS/UGT cluster; validation via NILs

Not all GWAS results can be stably replicated, but this discovery is indeed an exception. In multiple environments, large-scale GWAS analysis identified a large number of SNPS related to the content of soy isoflavones. Interestingly, several strong signals are concentrated in a familiar region-that is, the location containing IFS and UGT gene clusters (Kim et al., 2022). For instance, a SNP close to the bZIP transcription factor gene Glyma.11G108100 has repeatedly appeared in analyses of different environments and years, each time associated with higher levels of isoflavones. This result is not merely at the statistical level. After verifying the near-isogenic lines (NIL) containing different alleles at this locus, it was found that the lines with allele C had a much higher content of isoflavones compared to those with allele G, which also verified the reliability of the GWAS results.

6.2 Mechanism: expression/QTL colocalization shows cis-regulatory haplotype elevates seed isoflavones

Statistical signals alone are not enough. Without a clear understanding of the mechanism, it is often difficult to apply. After combining the expression data and QTL localization, the researchers further confirmed that the key SNP actually belongs to a cis-regulatory haplotype. This type of haplotype can increase the expression levels of Glyma.11G108100 and multiple biosynthetic genes around it during the seed development stage (Azam et al., 2023b). Not only that, the Co-expression Network (WGCNA) also indicates that this region is like a "hub" for expression control. Strains with dominant haplotypes have higher transcriptional levels and more obvious accumulation of isoflavones (Figure 2). Subsequent transgenic analysis of hairy roots also provided direct evidence for the connection between this regulation and metabolism, basically stringing together the entire mechanism chain.

6.3 Deployment: marker development, backcrossing into an elite background, and field validation

With the mechanism foundation and stable signals in place, the subsequent matters become practical. The research team developed this SNP into a molecular marker for label-assisted selection (MAS). Breeders introduced genes with dominant haplotypes into high-yield backgrounds through continuous backcross and then continuously tracked these target alleles using MAS technology (Yang et al., 2023). Of course, the key still lies in the field performance. In the trials conducted in multiple places, the results showed that the isoflavone content of the new variety was significantly increased, and the yield did not decrease. This also indicates that the entire route from GWAS localization to functional verification and then to field application is feasible, especially in the breeding of soybean varieties with high isoflavone content, which has practical significance.

7 Breeding and Engineering Strategies for Isoflavone Content in Soybean

7.1 Marker-assisted and genomic selection pipelines

Not every breeding strategy is suitable for dealing with complex traits controlled by multiple genes, such as isoflavone content, but some methods do perform well. For instance, in multiple environments and different years, some stable QTLS have been identified-mainly distributed on chromosomes 2, 4, 5, 6, 10, 12, 15, 19 and 20 (Yang et al., 2024). Some areas recur among different groups, indicating that they are indeed useful. During breeding, marker-assisted selection (MAS) utilizes these markers closely related to traits to introduce favorable genes into superior strains. However, genomic selection (GS) does not follow this line. It relies more on the marker data of the entire genome to predict the breeding value of the target trait, even not missing some loci with minor effects. Both of these methods can be applied to traits that are greatly influenced by the environment, especially when multiple alleles need to be superimposed.

7.2 Gene editing (CRISPR/Cas) to tune pathway/regulators

Of course, in addition to traditional selective breeding, gene editing technologies such as CRISPR/Cas have also gradually been put into use. Its advantage lies in its precise operation, allowing it to directly target regulatory factors or key metabolic genes (Wang et al., 2024). For instance, negative regulatory factors like GmMYB77, after being edited, saw a significant increase in the total amount of isoflavones and certain specific types of isoflavones, and the agronomic traits were not damaged. There were also studies attempting to knock out GmMPK1 and GmMT1, and the results were quite good. The isoflavone levels increased, and the stress resistance also improved. However, one should not blindly pursue a certain trait. When editing these genes, one must be cautious not to affect the yield or cause anti-nutritional problems. Therefore, after the editing is completed, the phenotypic verification step cannot be skipped.

7.3 Pangenome/haplotype-aware selection

When using traditional reference genomes, some structural variations are often missed, while pan-genome and haplotype-aware selection can well make up for this shortcoming. These strategies can identify some special variations, especially those haplotype combinations associated with high isoflavone content. Interestingly, after analyzing different germplasm resources, it was found that the range of natural variation is even larger than imagined. Some rare haplotypes are not only stable but also can maintain a high accumulation level of isoflavones for a long time (Bi et al., 2015). In addition, introducing some wild or semi-wild relatives is a good way to broaden the genetic basis and can also bring in new alleles. Nowadays, many breeding efforts have begun to identify and integrate these useful genes by leveraging haplotype-based markers and pan-genome resources.

8 Conclusion

At present, there are still many thorny problems in isoflavone research, among which the most prominent one is that the phenotypic analysis methods used by various laboratories are not uniform, and the results are naturally difficult to compare. Although high performance liquid chromatography (HPLC) and liquid chromatography-mass spectrometry (LC-MS) are commonly used to measure the content of isoflavones, these two methods are indeed quite accurate and can simultaneously quantify glycoside ligands and glycosides. But problems also arise-when there are many differences in details such as the extraction steps, the selection of standard substances, and the reporting methods, the results are likely to not match. The good news is that researchers are also making efforts. For instance, they have recently developed an HPLC method capable of rapidly analyzing up to 12 types of isoflavones, and at the same time introduced standard substances to reduce errors between laboratories. However, in the final analysis, a universally recognized standard operating procedure and a set of universal reference standards have not yet been truly established. This remains an urgent matter at hand.

On the other hand, germplasm resources have developed rapidly in recent years. The dataset at hand now is not only large in volume but also has clear features. This has made the work of gene mapping and screening of superior strains proceed much more smoothly than before. A study analyzed over 1,100 soybean germplasms from different ecological environments and found that the isoflavone content could vary by as much as seven times, with a very obvious difference. It also identified good materials suitable for breeding. By using GWAS, transcriptomics and metabolomics together, many key regulatory genes and haplotypes have also come to light. Pan-genome and haplotype analyses, on the other hand, have begun to uncover the secrets behind the diversity of isoflavones, such as structural variations and rare alleles. It can be said that these new resources and multi-omics platforms have laid a very solid foundation for precise breeding and in-depth research.

There are still quite a few things to do in the future. For instance, how can we identify those "good genes" that can still stably accumulate isoflavones under drought and high temperatures? For instance, under the premise of ensuring the balance between nutrition and agronomic traits, how can its synthesis be more precisely regulated? Gene editing and network regulation analysis are opening up new perspectives on these issues. Meanwhile, the idea of customizing the isoflavone content based on consumer demands has also been put on the agenda-not only for nutrition, but also for taste, flavor and even special functions. To achieve these goals, it is almost impossible without standardized phenotypic data, integrated databases or strict functional verification. These basic tasks still need to be continuously advanced.

Acknowledgments

Author extends the sincere thanks to two anonymous peer reviewers for their invaluable feedback on the manuscript of this paper.

Conflict of Interest Disclosure

The author affirms that this research was conducted without any commercial or financial relationships that could be construed as a potential conflict of interest.

References

Azam M., Zhang S., Abdelghany A., Shaibu A., Feng Y., Li Y., Tian Y., Hong H., Li B., and Sun J., 2020, Seed isoflavone profiling of 1168 soybean accessions from major growing ecoregions in China, Food Research International, 130: 108957.

https://doi.org/10.1016/j.foodres.2019.108957

Azam M., Zhang S., Huai Y., Abdelghany A., Shaibu A., Qi J., Feng Y., Liu Y., Li J., Qiu L., Li B., and Sun J., 2023a, Identification of genes for seed isoflavones based on bulk segregant analysis sequencing in soybean natural population, Theoretical and Applied Genetics, 136: 1-12.

https://doi.org/10.1007/s00122-023-04258-5

Azam M., Zhang S., Li J., Ahsan M., Agyenim-Boateng K., Qi J., Feng Y., Liu Y., Li B., Qiu L., and Sun J., 2023b, Identification of hub genes regulating isoflavone accumulation in soybean seeds via GWAS and WGCNA approaches, Frontiers in Plant Science, 14: 1120498.

https://doi.org/10.3389/fpls.2023.1120498

Azam M., Zhang S., Qi J., Abdelghany A., Shaibu A., Feng Y., Ghosh S., Agyenim-Boateng K., Liu Y., Yao L., Li J., Li B., Wang B., and Sun J., 2024, Effect of origin, seed coat color, and maturity group on seed isoflavones in diverse soybean germplasm, Plants, 13(13): 1774.

https://doi.org/10.3390/plants13131774

Bi Y., Xiao J., Lin H., Liu M., Liu M., Luan X., Zhang B., Xie X., Guo D., and Lai Y., 2015, Heterosis and combining ability estimates in isoflavone content using different parental soybean accessions: wild soybean, a valuable germplasm for soybean breeding, PLoS ONE, 10(1): e0114827.

https://doi.org/10.1371/journal.pone.0114827

Cai Z., Cheng Y., Ma Z., Liu X., Ma Q., Xia Q., Zhang G., Mu Y., and Nian H., 2018, Fine-mapping of QTLs for individual and total isoflavone content in soybean (Glycine max L.) using a high-density genetic map, Theoretical and Applied Genetics, 131: 555-568.

https://doi.org/10.1007/s00122-017-3018-x

Chen H., Liu C., Li Y., Wang X., Pan X., Wang F., and Zhang Q., 2023, Developmental dynamic transcriptome and systematic analysis reveal the major genes underlying isoflavone accumulation in soybean, Frontiers in Plant Science, 14: 1014349.

https://doi.org/10.3389/fpls.2023.1014349

Chen H., Pan X., Wang F., Liu C., Wang X., Li Y., and Zhang Q., 2021, Novel QTL and meta-QTL mapping for major quality traits in soybean, Frontiers in Plant Science, 12: 774270.

https://doi.org/10.3389/fpls.2021.774270

Chu S., Wang J., Zhu Y., Liu S., Zhou X., Zhang H., Wang C., Yang W., Tian Z., Cheng H., and Yu D., 2017, An R2R3-type MYB transcription factor, GmMYB29, regulates isoflavone biosynthesis in soybean, PLoS Genetics, 13(5): e1006770.

https://doi.org/10.1371/journal.pgen.1006770

Dastmalchi M., and Dhaubhadel S., 2014, Soybean seed isoflavonoids: biosynthesis and regulation, Phytochemicals-Biosynthesis, Function and Application, 1: 1-21.

https://doi.org/10.1007/978-3-319-04045-5_1

Gupta O., Dahuja A., Sachdev A., Jain P., Kumari S., T, V., and Praveen S., 2019, Cytosine methylation of isoflavone synthase gene in the genic region positively regulates its expression and isoflavone biosynthesis in soybean seeds, DNA and Cell biology, 38(6): 510-520.

https://doi.org/10.1089/dna.2018.4584

Gutierrez-Gonzalez J., Wu X., Gillman J., Lee J., Zhong R., Yu O., Shannon G., Ellersieck M., Nguyen H., and Sleper D., 2010, Intricate environment-modulated genetic networks control isoflavone accumulation in soybean seeds, BMC Plant Biology, 10: 105.

https://doi.org/10.1186/1471-2229-10-105

Huang D.D., 2024, Unravelling the biosynthesis of isoflavones in soybeans from a metabolic perspective, Journal of Energy Bioscience, 15(6): 378-387.

http://dx.doi.org/10.5376/jeb.2024.15.0032

Kim D., Lyu J., Lim Y., Kim J., Hung N., Eom S., Kim S., Kim J., Bae C., and Kwon S., 2021, Differential gene expression associated with altered isoflavone and fatty acid contents in soybean mutant diversity pool, Plants, 10(6): 1037.

https://doi.org/10.3390/plants10061037

Kim E., Jung J., Yu O., Lee S., Kim M., Lee S., Park H., Jo Y., Joo Y., and Oh S., 2024a, Natural variation in tocopherols, B vitamins, and isoflavones in seeds of 13 Korean conventional soybean varieties, Applied Biological Chemistry, 67: 1-12.

https://doi.org/10.1186/s13765-024-00896-5

Kim J., Lee J., Seo J., Ha B., and Kwon S., 2024b, Differentially expressed genes related to isoflavone biosynthesis in a soybean mutant revealed by a comparative transcriptomic analysis, Plants, 13(5): 584.

https://doi.org/10.3390/plants13050584

Kim J., Lyu J., Kim D., Hung N., Seo J., Ahn J., Lim Y., Eom S., Ha B., and Kwon S., 2022, Genome wide association study to detect genetic regions related to isoflavone content in a mutant soybean population derived from radiation breeding, Frontiers in Plant Science, 13: 968466.

https://doi.org/10.3389/fpls.2022.968466

Kim J., Seo J., Lee J., Lyu J., Ryu J., Eom S., Ha B., and Kwon S., 2023, QTL mapping reveals key factors related to the isoflavone contents and agronomic traits of soybean (Glycine max), BMC Plant Biology, 23: 517.

https://doi.org/10.1186/s12870-023-04519-x

Kim S., Song Y., Yi Y., Kim H., and Kim Y., 2020, Effect of biotic substances on isoflavone content in soybean germination, Korean Journal of Crop Science, 65(2): 84-92.

https://doi.org/10.7740/KJCS.2020.65.2.084

Kurosaki H., and Koyano S., 2023, Effects of shading, soil moisture, fertilizations and sowing time on isoflavone content of soybean seed in Hokkaido, Plant Production Science, 26: 364-377.

https://doi.org/10.1080/1343943X.2023.2262755

Li R., Zou J., Sun D., Jing Y., Wu D., Lian M., Teng W., Zhan Y., Li W., Zhao X., and Han Y., 2022, Fine-mapping and functional analyses of a candidate gene controlling isoflavone content in soybeans seed, Frontiers in Plant Science, 13: 865584.

https://doi.org/10.3389/fpls.2022.865584

Liu Y., Zhang S., Li J., Muhammad A., Feng Y., Qi J., Sha D., Hao Y., Li B., and Sun J., 2024, An R2R3‐type MYB transcription factor, GmMYB77, negatively regulates isoflavone accumulation in soybean [Glycine max (L.) Merr.], Plant Biotechnology Journal, 23(3): 824-838.

https://doi.org/10.1111/pbi.14541

Ng M., Ku Y., Yung W., Cheng S., Man C., Yang L., Song S., Chung G., and Lam H., 2021, MATE-type proteins are responsible for isoflavone transportation and accumulation in soybean seeds, International Journal of Molecular Sciences, 22(21): 12017.

https://doi.org/10.3390/ijms222112017

Pei R., Zhang J., Tian L., Zhang S., Han F., Yan S., Wang L., Li B., and Sun J., 2018, Identification of novel QTL associated with soybean isoflavone content, The Crop Journal, 6(3): 244-252.

https://doi.org/10.1016/J.CJ.2017.10.004

Wang X.M., Qi Y.X., Sun G.H., Zhang S., Li W., and Wang Y.P., 2024, Improving soybean breeding efficiency using marker-assisted selection, Molecular Plant Breeding, 15(5): 259-268.

http://dx.doi.org/10.5376/mpb.2024.15.0025

Wang S., Zhang Y., Zhu G., Shi X., Chen X., Herrera-Balandrano D., Liu F., and Laborda P., 2022, Occurrence of isoflavones in soybean sprouts and strategies to enhance their content: a review, Journal of Food Science, 87(5): 1961-1982.

https://doi.org/10.1111/1750-3841.16131

Wang Y., Han Y., Zhao X., Li Y., Teng W., Li D., Zhan Y., and Li W., 2015, Mapping isoflavone QTL with main, epistatic and QTL×environment effects in recombinant inbred lines of soybean, PLoS ONE, 10(3): e0118447.

https://doi.org/10.1371/journal.pone.0118447

Watanabe S., Yamada R., Kanetake H., Kaga A., and Anai T., 2019, Identification and characterization of a major QTL underlying soybean isoflavone malonylglycitin content, Breeding Science, 69: 564-572.

https://doi.org/10.1270/jsbbs.19027

Wu X., Yang Z., Zhu Y., Zhan Y., Li Y., Teng W., Han Y., and Zhao X., 2024, Bioinformatics identification and expression analysis of Acetyl-CoA Carboxylase reveal its role in isoflavone accumulation during soybean seed development, International Journal of Molecular Sciences, 25(18): 10221.

https://doi.org/10.3390/ijms251810221

Xu Z., Li J., Song X., Zhang Y., Wang Y., Zhu Y., Liu T., He Y., Liu Y., Wang Q., and Yan F., 2025, Overexpression of the R2R3‐MYB transcription factor GmMYB3a enhances isoflavone accumulation in soybean, Physiologia Plantarum, 177(1): e70120.

https://doi.org/10.1111/ppl.70120

Xu Z., Lui Y., Zhao Y., Song X., Zhu Y., Wang Y., He Y., Li J., Wang Q., and Yan F., 2024, R2R3-MYB transcription factor GmMYB68 is involved in the accumulation of soybean isoflavones, Plant Physiology and Biochemistry, 216: 109187.

https://doi.org/10.1016/j.plaphy.2024.109187

Yang S., Zhang M., Yao R., Chen L., Cong W., Yao D., Zhang J., Zhang J., and Li X., 2024, Linkage mapping and QTL analysis of isoflavones composition in soybean seeds, Phyton, 93(9): 2209-2225.

https://doi.org/10.32604/phyton.2024.055046

Yang Z., Wu X., Yang Y., Qu Y., Xu J., Wu D., Li D., Han Y., Zhao X., and Li Y., 2023, Identification of QTNs, QEIs interactions and genes for isoflavones in soybean seeds, Industrial Crops and Products, 197: 116631.

https://doi.org/10.1016/j.indcrop.2023.116631

Zernova O., Lygin A., Widholm J., and Lozovaya V., 2009, Modification of isoflavones in soybean seeds via expression of multiple phenolic biosynthetic genes, Plant Physiology and Biochemistry, 47(9): 769-777.

https://doi.org/10.1016/j.plaphy.2009.05.006

Zhang B., Zhao K., Ren H., Lamlom S., Liu X., Wang X., Zhang F., Yuan R., and Wang J., 2023, Comparative study of isoflavone synthesis genes in two wild soybean varieties using transcriptomic analysis, Agriculture, 13(6): 1164.

https://doi.org/10.3390/agriculture13061164

Zhao Q., Qin J., Li X., Liu B., Liu Y., Yang Q., Liu S., Zhao X., Ma N., Yan L., Zhang M., Yang C., and Liao H., 2022, Coordinate inheritance of seed isoflavone and protein in soybean, Agriculture, 12(8): 1178.

https://doi.org/10.3390/agriculture12081178