Mungbean [Vigna radiata (L.) Wilczek var. radiata] is a warm season legume, central to sustainable agriculture and nutrition in Asia and globally. It has a short growth cycle of 60-75 days and adapts well to tropical and subtropical regions, making it ideal for integration into cereal-based systems like the rice-wheat crop rotations in South and Southeast Asia (Liu et al., 2018; Chauhan and Wiiliams, 2018; Nayak et al., 2025). The global mungbean area spans approximately 7.5-8.0 million/ha, with 80-90% cultivated in Asia (Liu et al., 2018; Kahlon et al., 2023). India leads in production with around 4.5 Mha, followed by Myanmar at approximately 1.2 Mha (Yan et al., 2013; Silva et al., 2025). Mungbean also enhances sustainability through deep rooting and nitrogen fixation (Liu et al., 2018). Despite its importance, yield remains low-typically under 1.2 t/ha (Adsule et al., 1986; Chauhan and Wiiliams, 2018; Proulx, 2021). Resistance from gene pools exists but often fails under field conditions (EFSA Panel on Nutrition, Novel Foods and Food Allergens, 2021). Abiotic stresses (Salinity and water logging) also hinder nodulation and growth. Wide-ranging climate change only worsens these pressures. Moreover, the genetic base remains narrow: although gene banks conserve ~14,000 accessions (second to soyabean), breeding largely depends on a handful of elite lines (Silva et al., 2025; Liu et al., 2018).
       
Conventional breeding methods (phenotypic selection in field trials) are time-consuming and hampered by the seasonal nature of mungbean and low trait heritabilities under stress. The focus is on the transition from marker discovery to marker validation and deployment, differences between major- and complex-effect traits that are best suited to MAS and GS, respectively and the potential use of GWAS and multi-omics approaches to identify candidate genes which can then be validated using functional genomics and genome editing. Cutting-edge genome editing offers new opportunities beyond marker-based selection. The CRISPR/Cas 9 system has been successfully deployed in mungbean, albeit in early studies. Notably, Talakayala et al., (2022) used CRISPR/Cas9 to knock out two viral genes (AV1 and AC1) in mungbean, yielding transgenic lines with strong resistance to yellow mosaic virus (Talakayala et al., 2022). Meanwhile, other omics technologies (transcriptomics, proteomics, metabolomics) are shedding light on the molecular basis of stress responses and complex trait regulation. For example, proteomic studies of MYMV-infected plants revealed defense-related proteins that could be targeted by reverse-genetic screens (Wang et al., 2024). Such systems-level data will guide trait dissection and candidate gene identification. Marker-assisted selection (MAS) has already demonstrated benefits in mungbean: for example, SSR markers linked to resistance QTLs allow early-generation screening for MYMD resistance, greatly speeding up the breeding cycle (Kaur et al., 2018). The recent publication of the mungbean reference genome (~579 Mb) and high-density linkage maps has transformed genetic analysis (Kang et al., 2014; Silva et al., 2025). Next-generation sequencing (NGS) and genotyping-by-sequencing (GBS) can now generate thousands of SNP markers across the genome (Liu et al., 2018). These high-throughput marker platforms enable genome-wide association studies (GWAS) and high-resolution QTL mapping for complex traits. For instance, recent GWAS in mungbean have identified multiple loci for seed yield, flowering time and stress tolerances (Huppertz et al., 2023). In parallel, genomic selection (GS)-which uses genome-wide markers to predict breeding values-has emerged as a powerful approach in other legumes and is now being piloted in mungbean breeding programs. Consequently, the field currently lacks a synthesis that consolidates post-genome-assembly marker and QTL discoveries.
       
This review also critically assesses the existing constraints that hinder the implementation of such technologies in mungbean. Consequently, there is still a need for an integrated assessment of how these resources can be translated into breeding decisions. In particular, relatively limited attention has been given to comparing the practical utility of marker-assisted selection (MAS), genome-wide association studies (GWAS), genomic selection (GS), multi-omics, phenomics, epigenetic approaches and genome editing according to trait architecture, validation status, breeding population requirements and deployment readiness. The main objective of this review is to evaluate the progress and practical utility of MAS, QTL mapping and GWAS for improvement in mungbean breeding.
 
Molecular breeding strategies in mungbean
 
Molecular breeding integrates genomic tools into traditional improvement schemes to accelerate mungbean cultivar development. It encompasses marker-assisted selection (MAS) and marker-assisted backcrossing (MABC) for introgressing desired alleles, as well as genome-wide approaches like genomic selection (GS) (Liu et al., 2018). In mungbean, these strategies have been piloted with encouraging results. For example, marker-assisted backcrossing has successfully transferred specific resistance genes into elite varieties: VrPGIP2 from a bruchid-resistant accession (V2802) was introgressed for seed bruchid resistance and VrMLO12 from a powdery-mildew-resistant parent (V4718) was introgressed for foliar disease resistance (Wu et al., 2022). The line V2984 was backcrossed into VC1973A as shown in (Fig 1). Such MAS campaigns have enhanced resistance to major pests and pathogens in recipient lines.  Table 1 summarizes key marker-assisted breeding examples in mungbean.




       
Traditional MAS is based on individual trait loci, whereas GS uses genome-wide marker profiles to predict performance. In principle, GS could accelerate improvement of complex traits (e.g. yield) by capturing small-effect QTLs, but in mungbean it remains largely unexplored due to limited training populations (Liu et al., 2018). Nonetheless, the availability of high-density SNP data and phenotypic records is laying the groundwork for future GS schemes. Recent genomic advances also enrich molecular breeding: for instance, the near-complete genome of mungbean has enabled precise QTL mapping. A high-quality assembly of cultivar KUML4 (468.1 Mb, N50 = 40.75 Mb) identified 27,667 genes and revealed expansions in defense-related families (Khanbo et al., 2024).
       
Using such resources, QTLs for agronomic traits have been fine-mapped; for example, QTLs for synchronous pod maturity (synchronous flowering/pod set) were found to span VrDet1 (a TFL1 ortholog) and homologs of soybean maturity genes E3 and J (ELF3) (Xia et al., 2021). These findings illustrate how genomic tools sharpen MAS: identified markers (e.g. linked to VrPGIP2, VrMLO12, VrDet1) facilitate precise selection of target alleles (Wu et al., 2022). Future strategies will combine MAS with GS and even transgenic or speed-breeding methods to stack multiple traits.
       
In summary, these techniques should all be seen as complementary, not competitive, technologies.MAS is best suited for validated major-effect loci, GWAS is very useful for discovery and fine mapping and GS may be more suitable for highly polygenic traits. GWAS and multi-omics could be used for the discovery of candidate loci, validate major-effect loci for MAS and genome-wide prediction of complex traits in an efficient breeding program. This integration would enable the breeders to invest resources in the target traits in the light of their genetic architecture, instead of following a single genomic strategy for all breeding objectives.
 
Genomic resources and molecular markers
 
The genomic toolkit for mungbean has expanded rapidly. The first reference genome (cultivar VC1973A) provided an initial framework and more complete assemblies have followed. A very recent assembly of cultivar KUML4 yielded 468.08/ Mb with scaffold N50 of 40.75/ Mb, covering over 96.9% of the genome in 11 chromosomes and annotating approximately 27,667 genes (Khanbo et al., 2024). This high-quality genome revealed specific expansions in defense-related gene families and serves as a robust reference for breeding. Leveraging these genomes, researchers have mapped important traits; for example, major QTLs for synchronous pod maturity included VrDet1 (a TFL1-like gene) and homologs of soybean E3 and J (ELF3), illuminating the genetic control of determinate growth (Xia et al., 2021). Together, these assemblies and annotations form a critical genomic resource for mungbean breeders (Liu et al., 2018; Khanbo et al., 2024).
       
Marker development has also progressed from early low-density tools to genome-wide systems. Initially, breeders used a few dozen SSRs (often transferred from the related azuki bean) and other markers, but now genotyping-by-sequencing (GBS) and SNP arrays provide dense coverage. For instance, Manjunatha et al., (2023) obtained ~55,634 raw SNPs by GBS (15,926 high-quality SNPs after filtering) (Manjunatha et al., 2023), as shown in Fig 2. Other studies have used thousands of SNPs: Liu et al., (2018) reported 2,912 trait-associated SNPs for 33 traits and large GWAS panels have been genotyped at up to ~3.6 million SNPs (Manjunatha et al., 2023). Such marker resources underpin modern mapping and selection. Ongoing efforts include development of mungbean-specific SNP chips and KASP assays for key genes. Table 2 summarizes genomic resources and marker platforms available to the mungbean research community.




 
Comparative utility of MAS, GWAS and genomic selection
 
An advantage of genomic selection (GS) over MAS is the advantage that many of the loci are included, many small effect loci, which may not have reached the level of significance to be used in MAS. This is important to mungbean traits with complex genetic architecture like grain yield, yield stability, drought tolerance, heat tolerance etc. that are quantitative in nature. However, marker density is not the only measure of prediction accuracy. It relies on the heritability of the traits, the number of independent chromosome segments within the training and breeding populations, the degree of accuracy of the phenotypes, the structure of the population, the relationship between the training and breeding populations, the quality of the markers and how much the genotype × environment interaction is accounted for by the prediction models.
       
Based on the current evidence, availability of ample numbers and representative training populations with reliable phenotypic data from diverse environments is the greatest constraint in the implementation of GS in mungbean. Theoretical predictions in this review demonstrated that it is possible to significantly increase the accuracy of the predictions for low-heritability traits by increasing the number of individuals in the training set. Thus, coordinated multi-environment training populations should be made a priority in future mungbean breeding program. A wide genetic diversity in these populations is desirable but not at the detriment of adequate representation of elite breeding material so that prediction models continue to be relevant to selection candidates.
       
The evidence that has been examined suggests that there is no one best genomic method for mungbean improvement. Rather, the technology used should be based on the genetic structure of the target trait, magnitude of the effects of the associated genes, trait heritability, environmental sensitivity and availability of markers with proven value. Marker-assisted selection has proven to be very successful in situations where a trait can be influenced by one or a few genes with relatively large and consistent effects. The use of markers associated with VrPGIP2 for bruchid resistance and VrMLO12 for powdery mildew resistance is an example of the use of MAS and marker-assisted backcrossing in traits in which diagnostic or tightly linked markers are available. When there are multiple small effect genes and/or genotype × environment interactions, however, MAS becomes less efficient.
       
The complementary approach is provided by the GWAS approach, which takes advantage of historical recombination in a variety of germplasm for marker-trait associations at relatively high resolution. GWAS has been used to identify loci related to phenology, yield components, leaf morphology and abiotic-stress response in mungbean. However, it is not a sufficient understanding of statistical association together with the absence of cause-and-effect. GWAS identified loci thus need to be validated in other populations, growing conditions and genetic background before they are routinely used in plant breeding. This distinction is important especially in the case of complex stress traits, which may show variation in expression and apparent effect of individual loci in response to environmental variation.
       
The difference between MAS and genomic selection is that markers with large effects have to be identified previously, which is impractical in most cases. Rather, across- the-board marker data is utilized to calculate breeding values and GS is particularly appealing for polygenetic characteristics like yield, adaptation and stress performance, which are controlled by various small-effect markers. However, its effectiveness is highly dependent on the size and composition of the training population, on phenotypic quality, on the density of markers, on the genetic similarity of the training population and the population to be selected, on the heritability of the trait and on the genotype × environment interaction. Therefore, the limited application of GS in mungbean is not a limitation imposed by the statistical concept, but rather the current requirement of large size of the training population, well-phenotyped and encompassing multiple environments.
       
Together, these should be considered complementary technologies rather than competing. MAS is best suited to validated major loci, GWAS can be very informative for discovery and fine mapping and GS may be more suitable for highly polygenetic traits. GWAS and multi-omics approaches could be applied in an efficient breeding program to identify candidate loci, validate major-effect loci for MAS and take advantage of genome-wide prediction for complex traits. This would enable breeders to prioritize their resources based on the underlying genetic structure of the trait on which they are working, instead of using the same genomic strategy for all traits.
 
Trait dissection: Biotic stress resistance
 
Numerous pests and diseases afflict mungbean and recent molecular studies have identified their genetic bases. Seed bruchid beetles (Callosobruchus spp.) are controlled by a well-characterized Br resistance locus. This locus comprises two tandem genes, VrPGIP1 and VrPGIP2, encoding polygalacturonase-inhibiting proteins that defend against insect digestive enzymes (Wu et al., 2022). These genes lie on Chr5 and are tightly linked to SSR markers (e.g. SSR013, SSR158), enabling MAS for bruchid resistance (Wu et al., 2022). Powdery mildew resistance has been attributed largely to VrMLO12, an MLO family gene in clade II. Loss-of-function variants of VrMLO12 (identified in resistant lines like RUM5) disrupt host susceptibility pathways, providing mlo-type resistance (Yundaeng et al., 2020). Cercospora leaf spot resistance QTLs (e.g. qCLS) have been mapped on chromosome 3, with candidate genes like a TBP-associated factor (VrTAF5 homolog) under study (Hoang et al., 2023). Viral disease resistance (e.g., mungbean yellow mosaic virus, MYMV) is quantitative: GWAS identified major QTLs such as qMYMIV4.1 on Chr4 and qMYMIV2.1 on Chr2 (Kitsanachandee et al., 2013). Other pests-e.g. bean aphid-are linked to genes like VrBG7 on Chr6 (a lectin-like receptor) identified in wild mungbean, suggesting an aphid-resistance locus (Chiteri et al., 2023). These discoveries facilitate breeding: for example, MABC has been used to stack VrPGIP2 and VrMLO12 into susceptible varieties, significantly improving bruchid and mildew resistance (Wu et al., 2022).
 
Trait dissection: Agronomic and abiotic traits
 
Important agronomic and abiotic traits in mungbean are genetically dissected. For example, the transition from indeterminate to determinate (synchronous pod maturity) growth habit is regulated by VrDet1, a TFL1 ortholog; cultivated mungbeans possess promoter mutations that silence VrDet1, leading to terminal flowering (Kang et al., 2014). Yield-related and stress-associated traits have been extensively mapped using QTL and GWAS approaches. A recent meta-analysis of approximately 660 published QTLs for yield traits identified 72 meta-QTLs (MQTLs) across all 11 linkage groups, with around 20 MQTLs overlapping GWAS signals (Wang et al., 2022). Physiological traits like leaf chlorophyll content and trichome density are controlled by major QTLs on chromosome 7, namely qSPAD-7-1 and qTric-7-2, respectively. These loci explain a significant portion of phenotypic variation and harbor candidate genes such as transcription factors (MYB102, ZFP6) and enzymes like S-Adenosyl-L-Methionine Synthase and FTSH1, which are involved in chlorophyll metabolism (Xia et al., 2021). Moreover, GWAS for salt tolerance revealed VrFR08, a fatty acid desaturase gene upregulated under saline stress, while root architecture-related QTLs on chromosome 5 include DREB2 transcription factors linked to drought resistance (Niu et al., 2022). These findings offer valuable genetic markers for breeding and functional validation (Table 3).


       
G×E interactions are equally significant for mungbean improvement, given that the crop is cultivated across highly diverse agroecological zones in South and Southeast Asia. Stress tolerance traits-such as heat, drought, or salinity resistance-frequently show environment-specific performance. A genotype that thrives in one set of climatic or soil conditions may perform inconsistently in another (Pattanayak et al., 2019). Incorporating G×E into genomic prediction models improves selection accuracy by identifying genotypes that combine broad adaptability with stable performance, while also helping to tailor varieties for specific environments. By integrating insights from both epistasis and G×E into marker-assisted selection (MAS) and genomic selection (GS), modern breeding programs can more effectively accelerate the development of mungbean cultivars with enhanced resilience to biotic and abiotic stresses.
 
Genome-wide approaches: GWAS and genomic selection
 
Genome-wide association studies (GWAS) have become a powerful tool to dissect complex traits in mungbean. By exploiting historical recombination in diverse panels, GWAS provides high-resolution mapping. In recent years, several GWAS in mungbean have revealed many trait-linked SNPs.
       
For example, Manjunatha et al., (2023) conducted GWAS on 153 genotypes (15,926 SNPs via genotyping-by-sequencing) and identified 50 significant SNPs influencing phenology and yield-related traits, including days to flowering, SPAD, plant height, pod count, seeds per pod, 100-seed weight and yield per plant. In a larger USDA panel comprising 482 accessions genotyped with ~264k SNPs. A comprehensive meta-analysis further integrated hundreds of published yield QTLs and validated 20 meta-QTLs that overlap independent GWAS signals.
       
GWAS has also pinpointed stress-related loci. Notably, multiple QTLs for MYMV resistance were found, including a major one, qMYMV_4_1 on chromosome 4. Similarly, seven QTLs for salt tolerance were associated with the candidate gene VrFR08. Furthermore, Chiteri et al., (2023) used over 5,000 leaf images from 484 accessions to derive morphological traits like leaf area and shape for GWAS applications (Table 4).


       
The parallel strategy of genomic selection (GS)-using genome-wide SNP profiles to predict breeding values-holds promise for mungbean, especially for polygenic traits. Although practical GS applications in mungbean remain limited, theoretical models indicate GS could accelerate breeding progress by capturing small-effect loci genome-wide. As genotyping costs decline and large training populations become available, GS is likely to complement marker-assisted selection (MAS) in future mungbean improvement programs (Fig 3).


 
Omics and phenomics integration
 
The integration of multi omics data and high throughput phenotyping (‘phenomics’) is transforming mungbean trait analysis. Transcriptomic and metabolomic profiling have elucidated stress response pathways. For instance, combined transcriptomics and metabolomics of mungbean sprouts under mild salinity revealed coordinated upregulation of the phenylpropanoid-flavonoid pathway; key enzymes DFR, COMT, CHS, CHI were transcriptionally induced, leading to higher levels of antioxidants like catechin and ferulic acid in leaves; transcription factors (MYB, WRKY, bHLH) were identified as likely regulators and targets for breeding salt tolerant sprouts (Lim et al., 2022).
       
Although proteomic studies in mungbean are fewer, Li et al., (2020) used proteomics to show differential expression of stress protective proteins (such as dehydrins and heat shock proteins) under drought conditions.
       
RNA interference (RNAi) has emerged as a powerful tool in modern plant biology, allowing scientists to explore gene functions and precisely reduce the expression of specific genes. Techniques like co-suppression, post transcriptional gene silencing and virus induced gene silencing are commonly used for targeted traits. RNAi has gained widespread attention and quickly replaced older methods like antisense technology due to its efficiency and accuracy (Nair et al., 2019).
       
In crops like mungbean, RNAi has proven especially valuable: it has helped uncover roles in fertility, somatic embryogenesis and resistance to both biotic and abiotic stresses; and it has contributed to improvements in oil quality, yield and maturity traits. As a result, RNAi is now a key tool in functional genomics and crop improvement worldwide (Nair et al., 2019). RNAi technology also represents a novel and powerful approach for improving plant resistance and tolerance to both biotic and abiotic stresses by delivering small non coding RNA molecules for sequence specific silencing of target genes (Nair et al., 2019) (Fig 4).


       
On the phenomics side, image-based platforms enable precise trait measurement at scale. Chiteri et al., (2023) collected over 5,000 leaf images from a diverse mungbean panel under field conditions and used digital image analysis to quantify traits like leaf area and shape. These phenotypes were then used in GWAS, uncovering genetic loci controlling leaf architecture. High-throughput root phenotyping (e.g. rhizobox imaging) has also been applied to map root traits under drought (Afonso et al., 2025). Future integration will leverage machine learning to combine genotypic, transcriptomic and phenomic data for predictive breeding. Table 5 highlights recent omics and phenomics studies in mungbean.


       
Equally important is the visualization of comparative genomic synteny between mungbean and related legumes such as soybean, cowpea and adzuki bean. Synteny maps highlight conserved QTL regions for agronomic traits (e.g., synchronous pod maturity, disease resistance) and reveal lineage-specific rearrangements. For example, VrDet1 (a TFL1 ortholog) in mungbean aligns with maturity loci (E3, J) in soybean, underscoring shared regulatory modules across legumes (Kang et al., 2014, Kaur et al., 2026). Such figures would contextualize mungbean improvement within the broader Vigna and Glycine gene pools. Together, gene network diagrams and synteny maps would enrich the manuscript by visually linking genomic insights with functional breeding applications.
       
Mungbean small RNA profiles, including miRNAs, have been described (e.g., miRNA identification), targeted gene editing of miR828/TAS4 in mungbean with confirmed results, However cotyledon colour alterations, has not been documented (Paul and Pal, 2017). The mungbean yellow mosaic virus genome in greengram (a cultivated Vigna radiata setting) has been targeted by CRISPR/Cas9 to lessen viral symptoms, proving that genome editing is feasible for viral resistance in a close legume backdrop (Talakayala et al., 2022). Since functional CRISPR disruption or enhancement in mungbean has not yet undergone peer review, VrPGIP2 and VrPGIP1-mapped potential resistance genes linked to bruchid resistance in mungbean-remain speculative editing targets. Clear punctuation and verification of phenotypic outcomes are essential in future assertions.
 
CRISPR applications in mungbean and related legumes
 
CRISPR/Cas9 genome editing is poised to revolutionize functional genomics and breeding in mungbean. To date, the most notable CRISPR achievement in mungbean has been targeting the viral genome rather than the host: Talakayala et al., (2022) engineered greengram plants expressing Cas9 and guide RNAs against multiple segments of the mungbean yellow mosaic India virus (MYMIV) genome. Transgenic lines showed >95% reduction in viral DNA and negligible mosaic symptoms compared to controls, demonstrating CRISPR’s utility for viral resistance in Vigna. Additionally, a recent GWAS study identified miR828 as a regulator of anthocyanin content; CRISPR knockout of miR828 (and its partner TAS4) confirmed its role, leading to deep purple cotyledons in edited plants. These examples show how CRISPR can validate candidate genes and create novel alleles in mungbean.
       
While direct gene-editing in mungbean is still emerging, lessons from other legumes inform its potential. In cowpea, CRISPR knockout of symbiosis genes (e.g. VuSYMRK, VuSPO11) produced nodulation mutants, proving the technique’s feasibility. Soybean researchers have edited genes controlling flowering and plant architecture (e.g. GmRIN, GmLHY, GmSD1) to alter flowering time and plant height. These successes suggest that candidate genes identified in mungbean (like VrMLO12 or VrPGIP2) could be edited to engineer disease or insect resistance. Lists reported CRISPR applications in mungbean and closely related legumes, highlighting targets and outcomes (Table 6). Continued improvements in mungbean transformation and editing efficiency will be critical to realize these prospects.


       
CRISPR/Cas research has progressed from laboratory experiments to producing stable phenotypes under field conditions in several major crops. In rice, targeted knockouts and allelic modifications of susceptibility genes have been shown to reduce the incidence of blast and bacterial blight. Importantly, these genome-edited rice lines retained their agronomic performance when tested under disease pressure in small-scale field environments (Zhou et al., 2022). In wheat, disruption of the asparagine biosynthesis pathway generated low-asparagine lines, which were evaluated in multi-season field trials across Europe. These trials confirmed both the stability of the trait and its significance for food safety (Raffan et al., 2023).
       
Beyond disease resistance, CRISPR-based strategies have also been applied to improve yield-related and grain quality traits. Such modifications have demonstrated beneficial effects in open-field assessments and replicated trials, allowing breeders to accelerate trait deployment compared with traditional methods. Recent reviews highlight a growing number of field trials and note that supportive regulatory decisions are beginning to align with the rapid progress of genome editing. Together, these developments indicate that CRISPR-edited crops have strong potential to deliver agronomically useful and stable phenotypes in real-world farming systems, though broader multi-location testing and harmonized policies remain critical for commercialization (Chen et al., 2024).
 
Applications of epigenetics for stress tolerance in mungbean
 
The epigenetic control is an extra level of regulation between sequences and phenotype. Epigenetic regulation can be related to changes in gene activity, via DNA methylation, histone modification, chromatin remodeling and small-RNA mediated regulation, in contrast to conventional genetic variation, which is caused by changes in the DNA sequence. These mechanisms are important in stress adaptation as environmental factors can affect gene expression without changing the DNA sequence.
       
Compared to the literature on QTL mapping and GWAS, evidence of epigenetic regulation in mungbean is still much less developed. Hence, epigenetic associations need to be distinguished from regular genetic associations. The identification of transcription factors involved in auxin-, abscisic-acid- and ethylene-related pathways offers proof for the regulatory control of developmental or stress-responsive traits, but this should not be automatically considered as evidence for an epigenetic mechanism. Direct evidence involves showing an epigenetic modification, link with altered gene expression and preferably with the phenotype.
       
Through short RNA, translation of miRNA or siRNA for post-transcriptional gene silencing, DNA methylation and histone modifications, epigenetic modifications provide heritable changes in the activities of genes and play a significant role in gene expression to cope with environmental challenges (Thiebaut et al., 2019, Wojtyla et al., 2020, Latutrie et al., 2019). Although proteins, such as transcriptional factors and others, play a mediating role in the action of the enzymes in gene regulation to achieve the objective, the enzymes that cause epigenetic alterations are not directly involved (Santos et al., 2011). Mungbean QTL research have not been widely reported (Kitsanachandee et al., 2013). qCLS found no connection with QTL (genes) previously identified in mungbean for resistance to powdery mildew disease (caused by) and insects azuki bean weevil (Callosobruchus chinensis L.) and podsucking bug (Riptortusclavatus Thunberg)and to powdery mildew disease (caused by fungus Erysiphe polygoni D.C.). The genes conferring resistance to CLS and powdery mildew disease inherited independently. Three genes, Vradi07g03990, Vradi08g06860 and Vradi0284s00060, encoding bZIP, AP2 and ARF transcription factors, respectively, were mapped to auxin, abscisic acid and ethylene-related signal transduction pathways, respectively, because “plant hormone signal transduction” was one of the enriched KEGG pathways (Kinoshita and Seki, 2014). This suggests that these plant hormones are involved in the regulation of SPM in mungbean. Nucleotide-based study will miss the epigenetic mechanisms influencing SPM in mungbean, which were found based on pure epialleles independent of genetic variation. The data consistently indicated that various transcription factors and receptor proteins regulate SPM in mungbean via plant hormone (especially gibberellin)-mediated signaling pathways (Kapazoglou et al., 2018).
 
Challenges of mungbean cultivation and breeding progress
 
Although advanced breeding techniques hold significant promise for accelerating genetic improvement in mungbean, their widespread adoption remains limited due to several technical and biological constraints. One major bottleneck is the lack of validated, tightly linked markers across diverse genetic backgrounds, especially for traits governed by multiple genes. Phenotyping for traits such as drought and heat tolerance remains labor-intensive and environment-dependent, reducing the efficiency of QTL validation.
       
Another challenge is the limited availability of genomic resources in under-researched pulse crops like grass pea, bambara groundnut and moth bean. Additionally, breeding programs in developing countries often lack access to advanced genotyping platforms, bioinformatics tools and capacity-building initiatives (Table 7). These constraints have hindered the mainstream adoption of MAS in national agricultural research systems.


       
Several QTLs and candidate genes have been mapped in mungbean for biotic and abiotic stress tolerance (Table 8 and 3). However, their relative contributions to trait variation are not obvious without visualization. Fig 5 summarizes major QTLs reported in recent studies, showing their phenotypic variance explained (PVE). Notably, VrPGIP2 (Bruchid resistance, Chr5) and VrFR08 (salt tolerance) account for >25% of observed variance, whereas polygenic traits like chlorophyll content (qSPAD-7-1) or trichome density (qTric-7-2) explain ~12-15%. These data highlight why MAS is highly effective for major-effect loci (e.g. VrPGIP2) but less so for complex polygenic traits.




               
To illustrate the potential of Genomic Selection (GS) for low-heritability traits (e.g. yield under stress, h² ≈0.20), the expected prediction accuracy across training population sizes using the Daetwyler approximation. Accuracy improves from ~0.15 at N=100 to ~0.40 at N=1,000 (assuming Me = 1,000). These results emphasize that GS in mungbean will require large, well-phenotyped training populations or multi-environment models to be effective.

Moving from technology-specific breeding toward integrated data-driven selection should be a future goal of mungbean improvement. Establishment of diverse well-characterized breeding populations of elite cultivars, landraces and underutilized germplasm should be the first priority. High density SNP genotyping platforms should be used in these populations and they should be phenotyped under standardized conditions in various environments. Validation of genomic associations should be the second priority. Ideally, candidate QTLs and GWAS signals should be evaluated in separate populations and environments before proceeding to making them available to breeders as, for example, KASP or low-cost assays. Loci with major effects and consistent expression can then be used in MAS and/or marker assisted backcrossing.

The convergence of advanced molecular tools is revolutionizing mungbean improvement by dramatically boosting yields, nutritional value and stress resilience. High-throughput genotyping (e.g. GBS) has generated dense marker maps frontiersin.org, facilitating both linkage and association studies of key traits. Marker-assisted selection has emerged as a cornerstone of modern pulse breeding, offering accuracy, speed and reliability in trait improvement. Its success in enhancing disease resistance, drought tolerance and nutritional quality has been well documented in chickpea, pigeonpea and lentil. The integration of MAS with next-generation genomics and precision phenotyping holds great promise for unlocking the genetic potential of pulses. To sustain and amplify its impact, greater investment in capacity building, infrastructure and multi-disciplinary collaborations is needed. As the global demand for plant-based protein and climate-resilient agriculture intensifies, MAS-driven pulse improvement will be pivotal to achieving food and nutritional security.

The authors would like to sincerely thank all co-authors for their valuable contributions to this review on mungbean improvement. Each member played an important role in collecting relevant literature, analyzing the findings and shaping the manuscript. The continuous support, meaningful discussions and shared knowledge of the team greatly helped in completing this work successfully.
 
Disclaimers
 
The views expressed in this article are those of the authors and may not necessarily reflect the opinions of their respective institutions. While every effort has been made to present accurate and reliable information, the authors do not take responsibility for any outcomes resulting from the use of this content.
 
Informed consent
 
This study is based entirely on previously published research and does not involve any new experiments on humans or animals. Therefore, ethical approval and informed consent were not required.

The authors declare that there are no conflicts of interest related to this publication. The study was carried out independently, without any external funding or influence on the research process or manuscript preparation.