Introduction

Soybean [Glycine max (L.) Merr.] is an important protein and oil crop in the world. Soybean mosaic virus (SMV) belongs to the genus Potyvirus in the family of Potyviridae (Adams et al. 2005; Lin et al. 2020b). Plant infected with SMV not only presents typical symptoms such as leaf shrinkage, necrosis, plant dwarfing and seed mottling, but also leads to a serious reduction in seed yield and deterioration in seed quality (Wrather et al. 2001). The host defenses to the SMV could be divided into two broad categories, resistance and tolerance, based on the responses to viral infection (Lin et al. 2020b). In the USA, according to the disease reactions of two sensitive and six resistant soybean cultivars, 98 SMV isolates were classified into seven strains, namely G1 to G7 (Cho and Goodman 1979, 1982). In China, based on SMV isolates reaction of ten soybean differentials, SMV isolates collected from 2003 to 2010 were grouped into 22 strains (SC1 to SC22) (Guo et al. 2005; Li et al. 2010; Wang et al. 2003; Zhan et al. 2006). Among these strains, SC3 is a prime virus strain, and SC7 is an avirulent strain in Huang-Huai-Hai eco-region in China (Wang et al. 2003).

Up to now, there is no effective chemical method to control virus disease and the most cost-effective and eco-friendly strategy is to cultivate resistant varieties. Many resistance loci to SMV different strains have been reported. In America, three distinct SMV resistance loci (Rsv1, Rsv3 and Rsv4) have been identified and mapped on soybean chromosomes 13, 14, and 2, respectively (Gore et al. 2002; Hayes et al. 2000; Hayes and SaghaiMaroof 2000; Jeong and SaghaiMaroof 2004; SaghaiMaroof et al. 2010; Yu et al. 1994). Recently, a new Rsv5 locus was designated in soybean germplasm “York” as the resistance gene to replace the original Rsv1-y allele allocation on chromosome 13 (Klepadlo et al. 2017). In China, most SMV resistance loci have been positioned on chromosomes 2, 13, 14, and 6 (Fu et al. 2006; Li et al. 2006a; Ma et al. 2011; Wang et al. 2011a, b; Yang and Gai 2011; Yang et al. 2013). Karthikeyan et al. (2017, 2018) reported resistance loci to SMV strains SC5 and SC20 on chromosome 2 and chromosome 13, respectively. Yang et al. (2013) mapped resistance loci to SMV strains SC3, SC6 and SC7 in the neighbor of Rsv1 locus. Wang et al. (2011a) identified resistance loci to SC8 near Rsv4 locus. Yan et al. (2015) described resistance loci to SC7 on chromosome 2 in Kefeng 1. Meanwhile, Wang et al. (2018) summarized the QTL related to SMV resistance in soybean, and found that the resistance loci of different SMV strains were mainly located on five chromosomes (Gm02, Gm06, Gm10, Gm13 and Gm14). Further analysis revealed that there were ten loci (Rsv1, RSC3Q and RSC12, etc.) on chromosome 13, eight loci (Rsv4, RSC6 and RSC8, etc.) on chromosome 2, two loci (RSC15 and RSC18) on chromosome 6 and two loci (Rsv3 and RSC4) on chromosome 14.

In contrast to specific resistance to SMV, horizontal resistance (often referred to as tolerance) was also reported (Zhi et al. 2005), which was typically more durable and achieved by multiple genes. Zhi et al. (2005) revealed one additive major gene plus additive-dominant polygenes responsible for the horizontal resistance to Sa strain. Lin et al. (2020b) detected one QTL (qTsmv-13) for resistance and two QTL (qTsmv-2 and qTsmv-3) for tolerance to one recombinant SMV strain. Besides the linkage mapping strategy to mine the genetic loci for SMV, GWAS is also an important approach that has been successfully applied to identify loci associated with SMV. Yan et al. (2015) identified 19 SNPs associated with disease rate of SC7 in a natural population. Che et al. (2017) reported 104 SNPs significantly associated with resistance to SC7 in a soybean mutant panel. Che et al. (2020) also found 24 SNPs significantly associated with resistance to SC3 in a natural population.

A lot of work has been done on the detecting of SMV resistance loci and the prediction of candidate genes. However, only a few SMV resistance related loci and genes have been applied to date. The reason is partially due to the diversity of SMV strains in soybean production and the complicated infection of various SMV strains. So, more resistance loci and genes should be explored in soybean breeding program. In the present study, QTL mapping for SMV resistance was performed using a population consisting 193 RIL lines with SoySNP6K array genotyping data and phenotyping results of disease index (DI) inoculated with SMV strains SC3 and SC7. Meanwhile, GWAS was also performed using a natural population of 379 germplasms genotyped with the same SNP array and phenotyped with DI of SC3. The objectives were to utilize the linkage and genome-wide association mapping strategies to identify genetic loci and candidate genes that were associated with resistance to SC3 and SC7. The results will enrich useful SC3 and SC7 resistance genes and offer markers for soybean molecular breeding.

Materials and methods

Soybean materials and SMV strains

For QTL mapping, one recombinant inbred line (RIL) population was used in this study. The RIL population was derived from a cross between “C813” × “Kennong7” and consisted of 193 F6:8 RILs. “C813” is susceptible to SC3 and SC7 with DI 83.6 and 69.9, respectively, while “Kennong7” is resistant to SC3 and SC7 with DI 3.6 and 14.8, respectively (Table 1).

Table 1 Disease index to SC3 and SC7 strains in RIL population
Full size table

The GWAS panel consists of 379 diverse soybean germplasms mainly from Huang-Huai-Hai eco-region in China, including 302 cultivars and 77 landraces (with “Kennong7” and “C813” included). Detailed information of the GWAS panel was given in Appendix A.

SMV SC3 and SC7 strains were provided by the National Center for Soybean Improvement, Nanjing Agricultural University, China, and were cultured on SMV-susceptible host Nannong 1138-2.

Methods

Population planting and SMV inoculation

The seedlings of the RIL population and the GWAS panel were planted in round plastic pots (diameter × depth: 30 cm × 25 cm) which were filled with vermiculite and nutrient soil (3 : 1) in an aphid-free greenhouse with 14 h of supplemental light (~ 10,000 lx) per day, and room temperature 26 °C ± 1 °C at Hebei Agricultural University. For each line, 30 seedlings were used for each replicate. Three biological replicates were conducted for the RIL population inoculated with SC3 and SC7, and two biological replicates were used for GWAS panel inoculated with SC3.

The Nannong 1138-2 fresh leaves infected by SC3 or SC7 were grinded in 0.01 mol/L sodium phosphate buffer (approx.10mL per g leaf tissue, pH 7.4), using mortars and pestles. Seedlings were manually inoculated with the inoculum by rubbing unfolded primary leaves at the V1 stage and were inoculated again on the first trifoliate leaf at the V2 stage.

Resistance evaluation to SC3 and SC7 strains

Resistances to SC3 and SC7 strains were evaluated by scoring disease index one month after inoculation as described in Zhi et al. (2005). Disease index under 35 was defined as resistant accession and disease index over 35 was defined as susceptible accession. The statistical descriptive analysis was performed using SPSS V19.0 software. The disease index was transformed via the arcsine function, and the analysis of variance was conducted by using the transformed data via the SPSS V19.0 software. The broad-sense heritability was estimated via the transformed data according to the method described by Nyquist and Baker (1991).

$${\text{DI = }}\left[ {\sum {{\text{f}}_{{\text{i}}} {\text{S}}_{{\text{i}}} /\left( {{\text{n}} \times {\text{S}}_{{\max }} } \right)} } \right] \times 100$$

where Si was disease severity; fi was the number of plants with the Si; Smax was the highest Si; n was total number of surveyed plants.

Construction of linkage map and QTL mapping

The RIL population used in this study was genotyped by SoySNP6K array (the core subset selected from the SoySNP50K Illumina Bead Chip array), which contained 5403 SNPs and could represent major linkage disequilibrium (LD) blocks across the soybean genome (Akond et al. 2013, 2015; Wen et al. 2014).

The polymorphic SNP markers were imported into Joinmap software (v4.0) for construction of linkage map according to the previously described procedures (Ooijen 2006). Maximum likelihood algorithm and an independence LOD = 3.0 was used to calculate the distance between markers.

IciMapping version 4.0 (Wang et al. 2014) was used for inclusive composite interval mapping of additive (ICIM-ADD). The threshold of LOD score was 4.0 for statistical significance of QTL effects. The MapChart 2.3 software (Voorrips 2002) was used to graph the linkage map and analyze QTLs.

Population genetic analysis of GWAS panel

The natural population was genotyped by SoySNP6K array as above. The GWAS panel structure was analyzed by STRUCTURE 2.3.4 software (Evanno et al. 2005). Three independent runs were conducted and the K value (number of sub-groups) was from 1 to 12. Burn-in time and MCMC (Markov chain Monte Carlo) replication number were set both to 100,000 according to the previously described procedures (Sonah et al. 2015; Sun et al. 2017). Principal component and neighbor-joining tree analysis were performed using TASSEL 5.0 based on 3863 SNPs with minor allele frequency (MAF) > 20 %. TASSEL 5.0 was used to analyze linkage disequilibrium. The extent of LD decay was chosen by using the method of decreasing r2 to half of the maximum value.

Genome-wide association analysis

Mixed linear model was used to test associations between the SNPs (MAF > 5 %) and DI to SC3. In addition to the SNP and PCA matrix being tested, relative kinship matrix (K) was included as fixed and random effects, respectively. Top five principal components were used to build the P matrix for population structure correction. Analysis was performed with the software TASSEL 5.0. False discovery rate (FDR) ≤ 0.05 was used to identify significant associations between the SNP and DI.

Expression analysis of candidate genes in associated regions

To further verify the genetic loci and candidate genes found in this study controlling SMV resistance, the annotation of the candidate genes were analyzed. Gene annotation and candidate gene identification were based on the GlymaWm82.a2.v1 soybean sequence annotation database (https://www.soybase.org). The expressions of candidate genes were analyzed by using the transcriptome sequencing data of two soybean varieties (Zheng 92116 and Qihuang 30) at 0, 3, 10, 24, 120 and 240 h post inoculation (hpi) of the SC7 strain, which was described in our previous study (Chu et al. 2021).

Meanwhile, the expression analysis of candidate genes at 0, 4, 8, 12, 24, 48, 72 h post-inoculation (inoculated with SC3, mock inoculated with sodium phosphate buffer as control) was performed in resistant accessions (Kennong7, Han5408chun, Yudou12) and susceptible accessions (C813, Huangdou, Weidou267) via real-time qPCR (qRT-PCR) method. The qRT-PCR primers of the candidate genes and housekeeping gene were listed in Appendix B, and methods were similar as our previous study (Kong et al. 2018), with three technical replicates for each sample.

The virus contents were analyzed by detecting the relative expressions of CP gene of SMV as the description in Yuan et al. (2020).

Sequence analysis of candidate gene Glyma.13G184200

The amplification primers of Glyma.13G184200 were designed via the Primer Premier 5.0 software (Premier Biosoft International, Palo Alto, CA) (Appendix B). The coding sequences (CDS) were amplified by using the leaf cDNAs of Kennong7 and C813. PCR products were purified and ligated into the pMD19-T vector (Takara) according to the manufacturer’s protocol. The ligated product was transformed into DH5α competent cells, and the positive clones were sequenced by GENEWIZ (Tianjin, China). The CDS sequences and deduced amino acid sequences were aligned via the DNAMAN software.

Results

Genetic variation of resistance to SC3 and SC7 in the RIL population

The RIL population showed relative wide genetic variations to SC3 and SC7 with the range of DI 0 ~ 97.9 and 0 ~ 100, and the variation coefficient was up to 70.6 and 77.0 %, respectively (Table 1). Histogram analysis showed that the segregation of DI to SC3 and SC7 in the RIL population were both bi-modal distribution, suggesting that the resistance to SC3 and SC7 might be controlled by a major effect gene/QTL (Appendix C). In addition, the broad sense heritability of the resistance to SC3 and SC7 in the RIL population were 84.14 and 79.90 %, respectively (Table 1).

Genetic linkage map construction of the RIL population

The high density molecular genetic linkage map was constructed by the polymorphic SNP markers, which contained a total of 2,234 markers, and were grouped into 23 linkage groups (Appendix D). Chromosome 13 showed the most quantity of markers (193 SNPs) in the linkage groups. The whole map covered a total of 4,229.01 cM genetic distance and the average distance of the individual chromosome ranged from 1.21 cM (chromosome 3) to 3.38 cM (chromosome 4) and the average distance of the whole genetic map was 1.89 cM (Appendix E).

QTL mapping of disease index to SC3 and SC7 in the RIL population

After QTLs analysis basing on disease index, a co-located QTL with LOD values of 51.89 and 64.73 for SC3 and SC7, respectively, designated as qSMV13, was mapped in a narrow region between 142.39 ~ 142.74 cM on chromosome 13. The QTL qSMV13 accounted for 71.21 and 76.59 % of phenotypic variations to SC3 and SC7, respectively. The desirable allele of the QTL for SC3 and SC7 were both derived from the resistant parent “Kennong7”.

Further analysis revealed that qSMV13 were flanked by markers ss715614844 and ss715614864. Marker ss715614844 was nearby the reported SMV resistance genetic loci Rsv1-h (Ma et al. 2016), and marker ss715614864 was just within the Rsv1-h locus (Fig. 1).

Fig. 1
figure 1

QTL qSMV13 and candidate genes detected on chromosome 13 for resistance to SC3 and SC7 strains A, Associated marker on chromosome 13 for resistance to SC3 in GWAS panel B, QTL qSMV13 on chromosome 13 for resistance to SC3 and SC7 in RIL population C, Expressions of the candidate genes in the qSMV13 region and coat protein gene of SMV between resistant and sensitive varieties after inoculation with SC7

Full size image

Haplotype analysis of the qSMV13 flanking markers ss715614844 and ss715614864

To further verify the QTL obtained in the present study, the haplotype of the qSMV13 flanking markers ss715614844 and ss715614864 were analyzed. The RIL population could be divided into two haplotype groups, CT and TG, according to the SNP calls of the linkage markers. The disease index of CT-group (same genotype as resistance accession “Kennong7”) were 11.8 and 7.9, which were resistance to SC3 and SC7, while the TG-group (genotype as susceptible accession “C813”) were 71.0 and 63.1, which showed sensitive to SC3 and SC7, respectively. These results not only verified the QTL mapping results in this study, but also offered effective favorite allelic variation for marker-assisted selection of anti-SMV in soybean.

Genetic structure and linkage disequilibrium analysis of GWAS panel

The GWAS panel was genotyped by SoySNP6K beadchip as indicated above, which contained 5,403 SNPs. SNPs with MAF of > 0.05 and call rate > 80 % were selected for further analysis and 5,046 high-performing SNPs were obtained.

The GWAS panel was divided into two major subgroups (group1 and group2) according to the STRUCTURE analysis, which was verified by phylogenetic (NJ tree) analysis and PCA analysis (Appendix F). Among the two groups, group 1 contained 308 accessions, and nearly all were attributed to the cultivars. The group 2 contained 71 accessions, and most of those belonged to the landraces (Appendix A).

The average extent of genome-wide LD decay distance in the panel was approximately 210 kb, where the r2 drops to half its maximum value (0.17) (Appendix G), and offered the select region of candidate genes. Linkage disequilibrium decay distance were different between chromosomes, with 320 kb in chromosome 17, 100–200 kb in chromosomes 4, 5, 6, 7, 9, 10, 12, 13 and 15, and 200–300 kb in the remaining 10 chromosomes in the panel.

SNP associated with SC3 resistance in GWAS panel

As shown in the quantile-quantile (QQ) plots, lower inflation of nominal P-value was observed (Appendix H). Using this model, five SNP markers, ss715583175, ss715608741, ss715614844, ss715617664 and ss715625254 were significantly associated (FDR < 0.05) with resistance to SC3 on chromosomes 2, 11, 13, 14 and 16, and explained 6.0–19.0 % of phenotypic variations (Appendix H, Table 2).

Table 2 Associated SNPs and candidate genes with SC3 resistance in natural population
Full size table

Further analysis found that the associated marker ss715614844 was also the flanking marker of qSMV13 on chromosome 13 with PVE 71.21 % in the RIL population and explained 8.1 % of phenotypic variation in the natural population (GWAS panel), indicating consistency of the two mapping strategies.

Identification of resistance genes to SC3 and SC7 strains in qSMV13 region

To predict the candidate genes associated with resistance to SC3 and SC7 strains, the linkage mapping and GWAS analysis results were comprehensively analyzed. The qSMV13 genetic region defined by markers ss715614844 and ss715614864 spanned a 97.2-kb genomic region in the Williams 82 reference genome (Glyma.Wm82.a2.v1) containing seven annotated genes. Five of these genes, Glyma.13G183800, Glyma.13G184000, Glyma.13G184300, Glyma.13G184500 and Glyma.13G184600 showed relatively high expression levels in the resistance variety “Qihuang30” after the SC7 strain inoculation via the transcriptome analysis (Fig. 1). Meanwhile, the virus contents in resistant variety Qihuang30 kept decreasing and relative lower contents than that in sensitive variety Zheng92116 after 3 h post inoculation (Fig. 1).

Meanwhile, the expressions of other two genes, Glyma13g184200 and Glyma13g183900, were analyzed via qRT-PCR method after SC3 strain inoculation. Glyma13g184200 encodes a LRR receptor-like protein kinase and showed relative higher expressions at many time-points in resistant varieties (Kennong7, Han5408chun and Yudou12), but presented relatively higher expression at only one or two time-points in the sensitive varieties (C813, Huangdou and Weidou267). Another gene Glyma13g183900, encoding a zinc finger protein, showed higher expressions at more time-points in sensitive varieties, and presented relatively higher expressions only at 72 hpi in the resistance varieties (Fig. 2).

Fig. 2
figure 2

Expression analysis of the candidate genes to SC3 by qRT-PCR. Y-axes indicated the relative expression ratio of mRNA between samples inoculated with SC3 and phosphate buffer at different time-points (0–72 hpi). Asterisks represented statistical significance (*P < 0.05; **P < 0.01)

Full size image

In addition, to further analyze the sequence differences of LRR receptor-like (LRR-RLK) protein kinase gene Glyma13g184200 in resistant and susceptible varieties, the parental lines “C813” and “Kennong7” were used to amplify the coding sequences (CDS) of Glyma13g184200. The results showed that there existed a 72-bp insertion in the CDS of C813 compared with Kennong7, which was located in the LRR-RLK domain (Appendix I). Thus, we considered that Glyma13g184200 might be the most likely candidate gene for the SMV resistance in the qSMV13 region on chromosome 13.

Discussion

SMV disease causes yield decrease and quality deterioration and results in 35–50 % of yield loss when seriously occurred in soybean field (Wrather et al. 2001). QTL analysis is used to detect traits associated genetic loci by linkage mapping and genome-wide association analysis. However, the linkage mapping method can only investigate very limited variations from the two parents in the genetic population, and the GWAS results usually are affected by the Q value and individual relationships. A combination of both methods could provide a much higher power and resolution to detect QTL or genes that can hardly be detected by either method alone (Goddard 2005). Combination of the two strategies has been successfully applied to analyze complex traits, such as drought tolerance, Pythium resistance, downy mildew resistance, flowering time, seed size and shape (Lin 2020a; Lu et al. 2010; Hu et al. 2013; Nemri et al. 2010; Zhang et al. 2012). But, to date, the combination of both methods to detect the QTL of SMV resistance has not been reported. In this study, one RIL population and one natural population were used to detect the resistance sites of soybean mosaic virus SC7 and SC3 strains, and same marker ss715614844 was detected in the two populations, and the resistance site was further validated through the analysis of annotated genes in the mapping interval using transcriptome analysis and qRT-PCR technology (Figs. 1 and 2).

High marker densities are critical to improve the accuracy of QTL mapping (Song et al. 2016; Cao et al. 2017). By comparison, SNPs have a more abundant DNA variation than traditional markers. In our study, BARCSoySNP6K iSelect BeadChip was used to genotype the RIL population and GWAS panel. BARCSoySNP6K iSelect BeadChip was the core subset selected from the SoySNP50K Illumina Bead Chip array. Based on this BeadChip, Lin et al. (2020a, 2021) identified five QTL to Pythium sylvaticum and two QTL to Phytophthora sansomeana, and 20 SNP loci significantly associated with symbiotic nitrogen fixation-related characteristics were identified in a soybean natural population by GWAS (Huo et al. 2019). Previous studies have shown that the LD of self-pollinated species is higher than the cross-pollinated species, as the maize and rice LD decays less than 1 kb, Arabidopsis thaliana about 3–4 kb, soybean more than 150 kb (Gore et al. 2009; Hao et al. 2012; Lam et al. 2010; Wen et al. 2014; Zhu et al. 2007). In the present study, the average extent of genome-wide LD decay distance in the GWAS panel was approximately 210 kb. Since the soybean genome is known to extend slightly over 1000 Mb, the estimates of LD decay herein suggested at least 4762 (1000 Mb/210 kb) markers will be needed for whole genome scanning in soybean. Thus, we considered that the 5,046 SNPs in BARCSoySNP6K were sufficient for QTL analysis in soybean.

Recently, several SC3 and SC7 resistance loci have been mapped. Yang et al. (2013) reported that the resistance gene for SC3 was positioned between BARCSOYSSR_13_1128 and BARCSOYSSR_13_1136, while the resistance gene for SC7 was mapped between BARCSOYSS R_13_1140 and BARCSOYSSR_13_1155. Li et al. (2017) reported that four candidate genes for resistance to SC3 on chromosome 13 nearby Rsv1 using fine mapping and transcriptome analysis. Che et al. (2017) found SNPs associated with the resistance to SC7 on chromosomes 1, 3, 5, 9, 10, 12, 13, and 15 by GWAS. In our present study, one major QTL, qSMV13, was identified to associate with SC3 and SC7 resistance in the RIL population, and a qSMV13 flanking SNP marker was found to associate with SC3 strain resistance in the natural population (Fig. 1). Furthermore, the qSMV13 is about 126 kb away from ss715614889, a SNP marker that was associated with resistance to SMV G1, G2 and G3 strains (Chang et al. 2016), suggesting qSMV13 may also confer SMV G1, G2 and G3 resistance. In addition, the Chi-square fitness test of the qSMV13 locus based on the symptomless (resistant) and symptomatic (susceptible) of RIL lines (Karthikeyan et al. 2018) revealed a deviation of the 1 : 1 ratio of resistant and susceptible (χ2 = 16.9, 7.4, P < 0.05), indicating that the qSMV13 is a major resistance gene with much higher PVE to SMV and minor effects QTL are also available for the SMV resistance.

More importantly, seven annotated genes were found in the QTL qSMV13 genomic interval. Two of those genes, Glyma.13g184500 encoding bromodomain-containing protein and Glyma.13g184600 encoding eukaryotic membrane family protein, were located within the overlapping region of qSMV13 and Rsv1-h. Another gene, Glyma.13g184200 was predicted to encode a leucine rich repeat receptor-like protein kinase (LRR-RLK). LRR-RLK genes in soybean, such as GsLRPK and GmSARK, were participated in the immune response to pathogen, and also played an important role in the anti-cold and anti-aging (Li et al. 2006b; Yang et al. 2014). Glyma.13g183900 encoded a C3HC4 zinc finger protein that reported to play an important regulatory role in the basic immunity of Arabidopsis to pathogens (Wang et al. 2019). Glyma.13g183800 was predicted to encode esterase/lipase/thioesterase family protein. Lippold et al. (2012) reported esterase/lipase/thioesterase genes, PES1 and PES2, with higher expressions under low nitrogen stress and senescence, and could convert leaf alcohol and fatty acid into phytate to maintain the integrity of photosynthetic membrane in Arabidopsis thaliana. Glyma.13g184300 encoded the transducin/WD40 repeat-like superfamily protein, and this superfamily presented vital functions in plant development and stress signal transduction (Gachomo et al. 2014). Glyma.13g184000 was found to encode integral component of membrane.

Further analysis of these annotated genes showed that five of them were highly induced in resistance variety “Qihuang30” after SC7 inoculation based on the transcriptome analysis (Fig. 1), and the other two, Glyma.13g184200 and Glyma.13g183900, showed different expressions between the resistance varieties (Kennong7, Han5408chun and Yudou128) and sensitive varieties (C813, Huangdou and Weidou267) after SC3 inoculation based on the qRT-PCR analysis (Fig. 2). Moreover, a 72-bp insertion/deletion difference at the LRR-RLK domain sequences of Glyma.13g184200 was detected between C813 and Kennong7. Karthikeyan et al. (2018) reported three insertion/deletion variations and 63 SNPs in Glyma.13G194700, and identified one insertion/deletion variation and 17 SNPs in Glyma.13G195100, which located in the TIR, NB and LRR domains of TIR-NBS-LRR proteins, between the SMV resistant and sensitive soybean varieties. Because the LRR domains are required for the resistance response to pathogens in plants, amino acid variations may cause the function change of a disease resistant protein (Karthikeyan et al. 2018). Collectively, these candidate genes, especially the LRR receptor-like protein kinase Glyma13g184200, worth further functional validation in the future.

Conclusions

In this study, we identified a major QTL qSMV13 conferring resistance to SC3 and SC7 strains and explaining phenotypic variations of 71.21 and 76.59 %, respectively. qSMV13 was defined to a genomic interval of approximately 97.2 kb containing 7 annotated genes on chromosome 13 by linkage mapping, and the flanking marker ss715614844 was further verified for association with SC3 resistance via the genome-wide association analysis. These results offered useful selection markers and candidate genes with resistance to SMV in soybean molecular breeding.