Genome-wide association studies for earliness, MYMIV resistance, and other associated traits in mungbean (Vigna radiata L. Wilczek) using genotyping by sequencing approach

Association mapping panel and phenotypic observations

The association mapping (AM) panel comprised of 132 diverse mungbean genotypes (37 released varieties, 37 advanced breeding lines, and 58 exotic germplasm lines) (Fig. S1) that were planted during kharif-2021 and kharif-2022 at Indian Agriculture Research Institute, Pusa, New Delhi (28.7041°N, 77.1025°E) in augmented design. The AM panel was derived from a larger set of nearly 300 genotypes previously studied for various traits based on the diversity of various studied traits. Standard agricultural practices were followed, except for insecticide treatments, to allow for the natural occurrence of whiteflies, as IARI is known to be a hotspot for MYMIV. The genotypes were raised on a 5.0 m plot with plant-to-plant and row-to-row spacing of 10 × 30 cm (50 plants/row) and observations were recorded for flowering time, SPAD value, leaf area, trichomes, and YMD reaction (Table 1).

To score the MYMIV infection, the AUDPC and coefficient of infection (CI) were measured at 30, 45, and 60 days after sowing (Ahmed, 1985). To maintain the sufficient disease load in the field, the infector row technique was used wherein PS16 (MYMIV susceptible) was used as a susceptible check after every 5 rows and also around the plots. AUPDC was calculated by counting the total number of leaves and also the total number of infected leaves from three infected plants of 132 genotypes constituting the AM panel (Fig. 1). CI (%) was calculated by multiplying the response value with the intensity of infection. The disease severity was calculated (30–60 days after sowing) using the following formula:

$\text{Disease severity}\left(\mathrm{DS}\right)=${Sum of all disease ratings/ (total number of plants × maximum grade)}×100 ${\displaystyle \mathrm{A}.\mathrm{U}.\mathrm{D}.\mathrm{P}.\mathrm{C}.=\sum _{\mathrm{i}}^{n-1}\left[\frac{Yi+Y\left(i+1\right)}{2}\mathrm{x}\left(t\left(i+1\right)-\mathrm{ti}\right)\right].}$

The flowering was recorded as days to first flowering (DFF), days to 50% flowering (DFPF), and days to 100% flowering (DHPF). SPAD values were recorded using SPAD 502 plus Chlorophyll Meter (Piekielek et al., 1995; Matsunaka et al., 1997) when plants were 45 days old. The trichome density was measured in the dorsal leave surface of a 30-day-old plant using a compound microscope with bright field optics (10x) and an ocular scale. The upper epidermis (middle region) of the leaf was meticulously painted with nail polish before mounting on a glass slide. Using a piece of clear sticky tape, the film was lifted from the leaf surface after five minutes. The 0–4 scale was used for the measurement of trichome density (in a unit area); where a value of 4 was marked as the highest trichome density, while 0 meant no trichome (Fig. S2). Leaf area was measured by LI-3100C-Leaf Area Meter (Engin & Engin, 2013) using 03 randomly selected leaves from a 30-day-old plant and expressed as cm². The principal component analysis (PCA) was performed using R program.

Confirmation of causal YMD agent

To confirm the causal agent of YMD, MYMIV and MYMV-specific PCR-based amplification was performed. The MYMIV AV1 gene-specific (BM925F; 5′-AGG TGT CCC TAC CAA CAT-3′; BM926R; 5′-CCA TGG ATT GTT CCT TAC AA-3′) and MYMV gene-specific primers (AVI1F: 5′-GGA AGT GTC CCT GCC AGC G-3′and AC1R: 5′-CCA CAG GTT GAA GAA AGC AC-3′) having 497 bp and 925 bp expected amplification respectively were used on a set of susceptible mungbean genotypes. Water was used as a negative control. The PCR reaction included preheating (94 °C; 2–3 min), followed by denaturation (94 °C, 30-sec), annealing (56 °C; 45-sec), and extension (72 °C; 40-sec), repeated for 35 cycles, followed by a final extension at 72 ° C for 10 min and the product was resolved on 1.0% agarose gel.

Genotyping by sequencing assay and annotation of identified SNPs

The mungbean genotypes were grown in germination paper at room temperature and DNA was isolated from 5–6 days old seedlings (Doyle & Doyle, 1987). The sequencing was done through Illumina Hiseq 4000 using two 144-plex GBS libraries and sequenced (2 × 150 bp PE) (Bastien, Sonah & Belzile, 2014; Reddy et al., 2020). The sequencing was done through Illumina Hiseq 4000 using two 144-plex genotyping by sequencing (GBS) libraries and sequenced (2 × 150 bp PE) (Bastien, Sonah & Belzile, 2014; Reddy et al., 2020) and a comprehensive genetic coverage (150,000 reads per sample) was targeted. DNA extraction was performed using the DNeasy Plant Mini Kit, and quality of extracted DNA was confirmed through Nanodrop spectrophotometry. The library preparation was carried out using the TruSeq^® DNA Nano LP kit (Illumina, San Diego, CA, USA) and was quantified using a Qubit 4.0 fluorometer with the DNA HS assay kit (ThermoFisher, Waltham, MA, USA). The sequence data is available at NCBI (PRJNA609409; Reddy et al., 2020). Then demultiplexing and mapping of FASTQ reads onto the mungbean reference genome of Kang et al. (2014) was carried out. The high-quality SNPs were detected using Bowtie v2.1.0 (Langmead & Salzberg, 2012) and reference-based GBS pipeline/genotyping approach of STACKS v1.01 (Kujur et al., 2015). The unmapped reads were analyzed using a de novo approach. The chromosomal locations of candidate genes have been identified using reference mungbean genome available at NCBI databases and data filtering using Beagle 5.4 (Browning, Zhou & Browning, 2018).

Genetic diversity, LD estimation, and marker-trait association analysis

MEGA11 (Tamura, Stecher & Kumar, 2021) was used to generate a phylogenetic tree using neighbor-joining (NJ) method and population structure was determined using STRUCTURE simulations with 20 replications at different levels of population numbers (K = 1 to 10) (STRUCTURE v2.3.4, Pritchard, Stephens & Donnelly, 2000). Tassel was used to calculate linkage disequilibrium (LD) by pooling the LD (r²) estimates and visualized through R-base. Interpretation of genome-wide SNPs was done with Q-matrix and K-matrix through general linear model (GLM) and MLM using TASSEL 5 (Saxena et al., 2014); and also using BLINK by ‘R’ program (Ihaka & Gentleman, 1996). The QQ plot was generated to determine the relative distribution of observed and expected log10(p) values for each associated SNP. Finally, Manhattan plot was generated through GWAS analysis for various studied traits like flowering time, leaf area, SPAD value, and AUDPC (Figs. S3, S4, S5).

Digital gene expression analysis for the identified candidate gene

The digital gene expression analysis was performed for the identified candidate genes using Arabidopsis orthologs. The gene expression pattern has been searched using ‘Expression Angler’, an online search tool (Austin et al., 2016) for the studied traits like flowering time, SPAD value, leaf area, MYMIV disease, etc.

Descriptive statistics

Pearson correlation coefficients were estimated using the STAR (Statistical Tool for Agricultural Research) 2.1.0 software (Gulles et al., 2014). The PCA and broad sense heritability (h²) were measured by R package (Singh & Chaudhary, 1977) using the formula Vg/Vg+Ve, where Vg and Ve are genotypic and environmental variances, respectively.

Viral confirmation and phenotypic variation in the AM panel for the studied traits

The MYMIV AV1 gene-specific PCR-based amplification (497 bp) of the infected mungbean samples has confirmed the identity of the virus as MYMIV under Delhi (North Indian) conditions (Fig. S6). The frequency distribution of 132 mungbean genotypes revealed a wide range for AUDPC (1.75 to 1266.98; mean = 391.21; h² = 99%) and CI (0.33 to 45.53; mean = 13.17; h² = 92.55%) (Fig. 2). The onset of flowering, as measured by days to first flower (DFF) varied between 24 and 42 days (mean = 36.18 days; h² = 68%) (Table 2). A significant positive correlation was observed between DFF and DFPF (r = 0.44); and DFF and DHPF (r = 0.3) (Fig. 3A). The DFPF ranged from 33 to 50 days (mean = 44.4 days; h² = 80%), while DHPF ranged from 43 to 60 days (mean = 52.41 days; h² = 82%). A negative correlation was observed between SPAD value and AUDPC (r = −0.57). Interestingly, PCA explained 47.4% of total variance from the first two PCs for the studied traits (Fig. 3B). A very broad range was recorded for the leaf area (99.42 to 221.30 cm²; mean = 146.10 cm²; h² = 84%), SPAD value (29.33 to 62.32; mean = 46.44; h² = 80%), and trichomes frequency (1 to 4; mean = 2.98; h² = 46%).

SNP discovery

The study identified 31,953 SNPs distributed across 11 chromosomes of mungbean genome, mapping approximately 71.8% of SNPs (Reddy et al., 2020). The highest number of SNPs (4,899; 15.3%) were mapped on chromosome 1, while the lowest (1,309; 4.09%) was on chromosome 3. The SNP density varied across the chromosomes, with the lowest on chromosome 10 (0.71/100 bp; 4.71%), and the highest on chromosome 11 (9.92/100 bp; 6.15%) (Table 3).

Linkage disequilibrium and population structure analysis

LD was calculated by pooling the LD estimates (r2) using 31,953 SNPs and plotting their average r2 against 500 kb interval (maximum = 3,000 kb). LD varied throughout the chromosomes and at r2 = 0.2, the mean LD decay was 39.59 kb across all the chromosomes. With the increasing physical distance between SNPs, the LD decreased consistently (Fig. 4A). The population genetic structure of AM panel was determined using STRUCTUREv2.3.4 and all the SNPs were mapped on 11 mungbean chromosomes. The results revealed a maximum ΔK and K as 2, confirming the classification of studied genotypes into two genetically distinct populations (POP I and II) (Figs. 4B, 4C; Table S1). Identified SNPs (31,953) revealed a very high level of polymorphism and the phylogenetic tree also showed two populations (Pop I and II) as two distinct clusters (Fig. 5).

GWAS for earliness and MYMIV infection in mungbean

GWAS was conducted using 31,953 GBS-based SNPs which were integrated with phenotyping data, high-resolution population genetic structure, and PCA data. The Manhattan plot was constructed to display the results (Fig. 6). Significant marker-trait associations (MTAs) were identified using threshold [−log(p) = 3.0] as a cutoff. Both MLM and BLINK models identified 117 and 221 significant SNPs, respectively. Of these, 44 SNPs were shared by both models and were found associated with the studied traits. MLM identified 57 associated SNPs for flowering (DFF (28), DFPF (eight), and DHPF (21)) (Fig. S4); while BLINK has detected 92 SNPs associated with flowering (DFF (37), DFPF (18), and DHPF (37)). The number of significant SNPs for AUDPC, LA, SPAD, and TRI, as identified by MLM was 19, 22, 12, and seven, respectively (Table 4); while for BLINK this was 40, 52, 19, and 18, respectively (Fig. 6). The quantile–quantile plot determined the relative distribution of observed and expected log10(p) values of each SNP with the studied traits (Fig. S5).

Candidate genes for the studied traits in mungbean

To investigate the candidate genes regulating earliness (flowering time) and the yellow mosaic disease (YMD and other traits) trait in mungbean, we used 117 (MLM) and 221 (BLINK) significant SNPs. A total of 44 SNPs shared by both models were identified to be linked with 33 candidate genes having a role in the regulation of flowering time, MYMIV disease, chlorophyll formation (SPAD), leaf area, and trichomes formation in mungbean (Table S2). Primarily Arabidopsis orthologous has been identified as having a role in the regulation of earliness (flowering time) and disease resistance (Table 5).

A candidate gene (VRADI09G06940) regulating YMD resistance was identified as belonging to the disease resistance protein family (TIR-NBS-LRR class) on chromosome 9 (region Vr09:10949253-10955046). Also, the Senescence/dehydration-associated protein-related gene (VRADI01G03610) on chromosome 1 (Vr01:5710582-5714237) and pentatricopeptide repeat (PPR) superfamily protein gene (Vradi05g01370) on chromosome 5 (Vr05:1506441-1510662) were found associated with the imposition of disease resistance. For flowering several candidate genes have been identified, like flowering time control protein FPA gene (VRADI10G01470; chr: 10; Vr10:4265856-4270038), EARLY FLOWERING 3 gene (VRADI05G11900; chr: 5; Vr05:20836591-20841432), E3 ubiquitin-protein ligase DRIP2 gene (VRADI03G01320; chr: 3; Vr03:1791098-1797025) and homeobox-leucine zipper protein HAT22-like gene (VRADI05G11940; chr: 5; Vr05:20891041-20891644). For leaf area, GWAS has identified candidate genes like WRKY family transcription factor (VRADI03G06560; chr: 3; Vr03:7968517-7970262), and gibberellin 20 oxidase 2-like (VRADI03G06690; chr: 3; Vr03:8092219-8095631). For SPAD, E3 ubiquitin protein ligase RIE1 gene (VRADI07G28100; chr: 11; Vr11:18925287-18934324) and for trichomes, LOB domain-containing protein 21 gene (VRADI06G04290; chr: 6; Vr06:5012124-5012669) and a receptor kinase 2 gene (VRADI08G19670; chr: 8; Vr08:41687381-41723765) were identified.

Validation of identified candidate genes through digital gene expression analysis

Digital gene expression analysis helped in the in-silico validation of candidate genes using Arabidopsis orthologous (compared with the Vigna radiata genome) by identifying which part of the plant gets affected by the candidate genes. Digital gene expression analysis showed highest expression of the genes like flowering time control protein FPA (AT4G16280), EARLY FLOWERING3 (AT3G21320), homeobox-leucine zipper protein HAT22 (AT4G17460), K-transporter 1 (AT4G23640), Ca-dependent protein kinase 6 (AT4G23650), and E3 ubiquitin-protein ligase DRIP2 (AT2G30580) in the flowering tissues (Fig. 7A). Genes having the highest digital expression in leaves during MYMIV infection include disease resistance protein (TIR-NBS-LRR class) family (AT5G36930), 30S ribosomal protein S31 (AT2G21290), pentatricopeptide repeat (PPR) superfamily protein (AT1G11290) and E3 ubiquitin-protein ligase (Fig. 7B). The genes like RIE1 (AT2G01735) for SPAD (leaf senescent), WRKY family transcription factor (AT1G30650) for leaf area, LOB domain-containing protein 21 (AT3G11090) for trichomes showed affecting flowering and other plant developmental stages when studied through digital gene expression analysis (Fig. S7).

Phenotypic variations in the studied genotypes

The results found a significant relationship between earliness and yellow mosaic disease (YMD) among 132 mungbean genotypes under field conditions. A wide AUDPC (30.15 (resistant) to 1,266.98 (susceptible)) and CI range (1.75 to 45.53%) have been recorded in the studied genotypes after 30, 45, and 60 days of sowing. These results were comparable to that of Sandhu & Dhaliwal (2019) who reported AUDPC in mungbean as 205.60 after 75 days of sowing. Similarly, Devi et al. (2017) reported AUDPC value in blackgram ranging from 105.09 (resistant) to 1,684.80 (susceptible). Bag et al. (2014) also reported a very similar range of AUDPC (347.8 to 3,000) and CI (3.6 to 69.2) in mungbean. Very high broad sense heritability for MYMIV disease (h² = 99%) was recorded in our study and a very similar result was also reported (h² = 99.4%) in blackgram (Thirumalai & Murugan, 2020). However, relatively low broad-sense heritability for MYMIV resistance in mungbean was reported at Gazipur (63.64%) and Madaripur (79.68%) (Alam, Somta & Srinives, 2014b). Similarly, Chen et al. (2013) observed 79% heritability for MYMIV resistance. The results suggested that the genetic variables are important in MYMIV resistance and breeding is possible using standard selection procedures.

The range of DFF (24 to 42 d), DFPF (31 to 51 d), and DHPF (43 to 63 d) was very similar to that reported by Sardar et al. (2019) in mungbean (DFF: 33 to 44 d; DFPF: 38 to 50 d; DHPF: 62 to 64 d). The broad sense heritability (h²) for DFF (68%), DFPF (80%), and DHPF (82%) was very similar to that recorded by Degefa, Petros & Andargie (2014) as DFF (87%), DFPF (97.79%) and DHPF (97.79%), respectively. Similarly, Begum et al. (2013) reported similar h² for DFF (88.75%), DFPF (97.79%) and DHPF (88.77%). Also, Tripathi et al. (2020) reported similar h² for DFF (62.5%), DFPF (83.6%) and DHPF (95%). Leaf area ranged for leaflets from 99.42 cm² (33.21/leaf) to 221.30 cm² (73.76/leaf), Similarly, Hossain et al. (2017) observed that the leaf area range concerning light in different (30, 40, and 50 days) showed range from minimum 96.29 cm²/leaflet to maximum 461.96 cm²/leaflet. Whereas Kaur & Kumar (2020) reported leaf area in mungbean ranging from 27 cm²/leaf to 30 cm²/leaf.

The h² for leaf area was 84%, whereas Sofia et al. (2017) recorded slightly higher h² (98.64%) for leaf area. The association between SPAD and yield was found positive, while a negative association was recorded between SPAD and MYMIV severity. Similar observations were also recorded by Alam et al. (2014). The range of SPAD value was 29.33 (highly infected) to 62.32 (low or no infection), which was comparable to Alam et al. (2014), where they have reported a range of 47 to 52 for low disease score, while 23 to 32 for high disease scoring genotypes. The h² for SPAD was recorded as 80%, while Yimram, Somta & Srinives (2009) reported the range from 21% to 92%. Similarly, Sofia et al. (2017) reported h² = 67.98% for SPAD in mungbean. Trichome frequency ranged from 1 to 4, and a similar trichome frequency range (3.01 to 5.31) was also reported in black gram (Taggar & Gill, 2012).

SNP discovery, linkage disequilibrium, and population structure analysis

The total length of the pseudomolecule assemblies was 332.95 Mb which accounted for 71.8% of the total mungbean genome. The SNPs mined were found dispersed in different parts of the mungbean genomes, which has resulted in varied SNP densities on the physical map of the genome. The identified SNPs (31,953) can be used for the QTL mapping for several traits, as well as comparative genome mapping involving mungbean and other legumes. The results are comparable to the number of SNPs (55,634) reported by Reddy et al. (2020) for mungbean. Furthermore, these chromosome-based SNPs can be used for marker-based mungbean breeding applications.

The identified SNPs were annotated which will accelerate the process of identifying the QTLs governing the traits of interest in mungbean. A high-resolution LD pattern was identified and several studies in various legumes have been conducted using GBS and WGRS to find the LD pattern and dissect the traits (Zhou et al., 2015; Varshney et al., 2019). This study demonstrated a higher LD estimate (r² = 0.5) and less extensive LD decay (500–1,000 kb). Similarly, Reddy et al. (2020) also reported a higher LD estimate (r² = 0.62) and less extensive LD decay (50–100 kb) in mungbean. Comparable LD decay was found (100 kb range) in cultivated genotypes which is slightly higher than 60 kb LD decay as reported for the wild mungbean genotypes (Noble et al., 2018). Liu et al. (2016) reported 200–250 kb as the range for LD decay in soybean; while Hyten et al. (2007) reported 90 to 574 kb LD decay. The structure analysis revealed two population groups, which corresponded with that of the PCA and phylogenetic analysis. Furthermore, this is the first high-resolution determination of LD decay in mungbean with such a large number of genome-wide markers (31,953 SNPs).

Candidate gene identification and validation

A large number of high-quality SNPs (31,953 SNPs) identified from the association mapping panel were used for the mapping of traits like earliness, MYMIV resistance, etc. MLM and BLINK models revealed 338 significant SNPs about the candidate genes involved in the regulation of the studied traits. Orthologous of Arabidopsis thaliana were used to discover the candidate genes involved primarily in the regulation of earliness (flowering time), YMD resistance, and other traits. A total of 44 SNPs shared by both models were found associated with 33 candidate genes have been identified. This study identified NBS-LRR-encoding genes (chr. 9) as having a role in YMD resistance. Similarly, Purwar et al. (2023) also reported the role of NBS-LRR-encoding genes in the imposition of YMD resistance in mungbean and two wild non-progenitors. The TIR-NBS-LRR class of disease resistance protein family (chr. 4 and 9) was found to regulate YMD resistance in mungbean as also reported by Purwar et al. (2023). Depuydt & Vandepoele (2021) also reported the involvement of this gene in a variety of developmental processes and molecular responses, including defense responses.

Similarly, in Vigna mungo, Maiti, Paul & Pal (2012) also reported TIR-NBS-LRR encoding candidate gene regulating MYMIV-resistance. These proteins are encoded by a class of defense-related R genes. Pathogens are recognized by nucleotide-binding site—leucine-rich repeats (NBS-LRRs) proteins, which trigger downstream signaling pathways that activate the defense response. Interestingly, methyl esterification of pectin regulates plant-pathogen interactions thereby regulating YMD resistance. The degree and pattern of methyl esterification alter cell wall characteristics which affects both plant physiological functions and resistance (Lionetti, Cervone & Bellincampi, 2012). In addition, a 30S ribosomal protein S31 in chloroplastic silenced plants showed a delay in expression in the nonhost pathogen (Nagaraj et al., 2016). Similarly, Wang et al. (2017) also observed an interaction between 30S ribosomal subunit protein S11 and Cucumber mosaic virus LS2b protein affecting viral replication, infection, and gene silencing suppressor activity.

For flowering time regulation, several candidate genes (EARLY FLOWERING3, FPA, E3 ubiquitin ligase, CPK32) located on different mungbean chromosomes have been identified. For flowering, an EARLY FLOWERING3 (AT3G21320) gene (chr. 5) was identified as having a role in the regulation of flowering time by encoding a novel protein in Arabidopsis (Hicks, Albertson & Wagner, 2001). The same gene was also reported in barley regulating flowering time by mediating gibberellin production and FLOWERING LOCUS T expression (Boden et al., 2014). Likewise, FPA (AT2G43410) gene (chr. 10) regulates the flowering time via a day-length independent pathway through protein-containing RNA-binding domains in plants (Yu et al., 2021; Macknight et al., 1997). Similarly, Yu et al. (2021) reported that the BDR proteins interact with FPA and the complex is crucial for repressing the flowering time gene FLOWERING LOCUS C (FLC) in Arabidopsis.

In addition, a flower initiation gene E3 ubiquitin-protein ligase DRIP2 gene (chr. 3) was identified which functions by increasing the transcription of WRKY71 gene and thus promoting the expression of Flowering Locus T (FT) and LEAFY (LFY) genes and ultimately counteracting the DELLA gene inhibition (Wang et al., 2022). In addition, several studies revealed the role of ubiquitin E3 ligases in the regulation of flowering time (Pavicic et al., 2017; Lazaro et al., 2012; Peng et al., 2013; Xia et al., 2013). Similarly, Yu et al. (2021) reported that the flowering time control protein FPA of negative transcription elongation factor regulates flowering time in Arabidopsis.

Ca-dependent protein kinase (CPK32) gene (chr. 7) regulates the flowering time via Ca-signaling (Li et al., 2022); while the K-transporter 6 gene (Negishi et al., 2018) and homeobox-leucine zipper protein (HAT22) also regulates the flowering time (Schena & Davis, 1994). These kinases likely contribute to the crossroads of HAT22/ABA and CK signaling during drought response, aiding in leaf senescence (Depuydt & Vandepoele, 2021). The details of different candidate genes and their functions are enumerated in Table 5.

When identified genes were compared with the Vigna radiata genome, some candidate genes were found regulating various other traits. WRKY family transcription factor (WRKY15) mainly regulates the leaf area (15–25%) and plant biomass (Bakshi & Oelmüller, 2014). The WRKY54 and WRKY70 genes were reportedly having negative regulation of leaf senescence in Arabidopsis thaliana (Besseau, Li & Palva, 2012). While, FAR1-RELATED SEQUENCE 4 protein gene regulates the light-mediated signal transduction, chloroplast division, and chlorophyll biosynthesis by regulating the cellular activities through transcriptional activation or suppression of numerous target genes (Ma & Li, 2018). The E3 ubiquitin-protein ligase RIE1 gene has a role in the regulation of leaf senescence. The gene fusion test of the RIE1 promoter with the β-glucuronidase (GUS) gene indicated that this is expressed in several plant tissues (Xu & Li, 2003). Likewise, LOB domain-containing protein 21 gene was found associated with the development of glandular trichomes in plants (Chen et al., 2020).

The role of various transcription factors, including MYB, NAC, GATA, LBD, and ZF on Populus CesA gene expression showed varied expression for the leaf trichomes (Takata & Taniguchi, 2015). Interestingly in cotton, CesA4 was expressed in trichomes, while in Arabidopsis, CesA4, CesA7, and CesA8 were not expressed in the trichomes (Wu, Hu & Liu, 2009; Betancur et al., 2010; Kim et al., 2011).

Thus, breeding for the varieties carrying the genes governing MYMIV resistance will enhance the overall yield. Furthermore, the candidate genes governing flowering time, chlorophyll content, leaf area, and trichome density, provide opportunities for targeted crop improvement. For example, flowering time regulation is crucial for adapting the mungbean crop to the changing climatic conditions. The early-maturing varieties can be developed by manipulating these genes, enabling the farmers to synchronize mungbean cultivation under seasonal variations. Additionally, the genes associated with chlorophyll content, leaf area, and trichome density can be leveraged to enhance photosynthetic efficiency and stress tolerance, contributing to improved crop performance.