Jing Li,Yuyi Tng,Aln L.Jobson,Pht M.Dng,Xio Li,Ming Li Wng,Austin Hgn,Chrls Y.Chn,*

aDepartment of Crop,Soil and Environmental Sciences,Auburn University,Auburn,AL 36849,USA

bShandong Peanut Research Institute,Qingdao 266100,China

cDepartment of Entomology and Plant Pathology,Auburn University,Auburn,AL 36849,USA

dUSDA-ARS National Peanut Research Laboratory,Dawson,GA 39842,USA

eUSDA-ARS Plant Genetic Resources Conservation Unit,Griffin,GA 30223,USA

Keywords:Association mapping SSR markers Tomato spotted wilt virus Peanuts

A B S T R A C T Tomato spotted wilt(TSW)is a serious virus disease of peanut in the United States.Breeding for TSWV resistance would be facilitated by the implementation of marker-assisted selection in breeding programs;however,genes associated with resistance have not been identified.Association mapping is a type of genetic mapping that can exploit relationships between markers and traits in many lineages.The objectives of this study were to examine genetic diversity and population structure in the U.S.peanut mini-core collection using simple sequence repeat(SSR)markers,and to conduct association mapping between SSR markers and TSWV resistance in cultivated peanuts.One hundred and thirty-three SSR markers were used for genotyping 104 accessions.Four subpopulations,generally corresponding to botanical varieties,were classified by population structure analysis.Association mapping analysis indicated that five markers: pPGPseq5D5, GM1135, GM1991, TC23C08, and TC24C06, were consistently associated with TSW resistance by the Q,PCA,Q+K,and PCA+K models.These markers together explained 36.4%of the phenotypic variance.Moreover,pPGPseq5D5 and GM1991 were associated with both visual symptoms of TSWV and ELISA values with a high R2.The potential of these markers for use in a marker-assisted selection program to breed peanut for resistance to TSWV is discussed.

1.Introduction

Peanut (Arachis hypogaea L.) is a species in the family Fabaceae, the legume family, that is grown in tropical,subtropical,and temperate areas of the world.Peanut is an oil crop that contains 50%oil on average.Peanut is believed to have originated in South America,the most likely center of origin being Brazil and Peru,where ＞10 wild species were found[1].The cultivated peanut is an allotetraploid(2n=4x=40) that originated through hybridization of two ancient diploid species,probably A.duranensis(A genome)and A.ipaensis(B genome)[2].It contains two subspecies:A.hypogaea ssp.hypogaea and A.hypogaea ssp.fastigiata.The ssp.hypogaea consists of the botanical varieties hypogaea and hirsuta,whereas the ssp.fastigiata consists of the botanical varieties fastigiata,vulgaris,peruviana,and aequatoriana[3].

Tomato spotted wilt virus (TSWV) is a species of the Tospovirus genus,which is transmitted in nature exclusively by thrips[4].The virus has caused severe damage on peanuts in U.S.since 1984,and was estimated to have cost at least USD 44 million in 1995 in Georgia alone[5].Economic losses caused by TSWV in peanut have been reduced with the development of resistant varieties[5].No cultivar is completely resistant to TSW;resistant cultivar can become infected and may exhibit minor symptoms under high pathogen pressure.In addition,TSWV can evolve and cause more severe symptoms in resistant cultivars over time. The genes responsible for resistance and the mechanism of resistance,however,remain unknown,limiting the ability to select for resistance in newly developed peanut cultivars.Breeding resistant cultivars is the optimal method of avoiding severe TSW-incited yield loss.

Simple sequence repeat(SSR)markers have been the most widely used DNA marker in plant breeding programs because of their high degree of polymorphism, high abundance,codominant inheritance,easy use,ready transferability,and relatively low cost[6,7].The advantage of SSRs over other markers includes their facility of use to track desirable traits in large-scale breeding programs and to serve as anchor points for map-based gene cloning[8].Association mapping based on linkage disequilibrium(LD)provides an effective way to map quantitative trait loci, given that ancestral recombination events that occurred in natural populations generate a potentially large number of alleles per locus to associate markers and traits.In peanut,the first attempt at association mapping was reported for seed quality traits using SSR and single nucleotide polymorphism(SNP)markers in 2011[9].More recently,a genome-wide association study[10]that included 300 peanut accessions identified 524 significant associations for 36 traits.These could be used in improving biotic and abiotic stress resistance,seed quality,and yield.

To date, no research on SSR marker association for resistance to TSW in peanut has been reported.To improve peanut breeding programs that include breeding for resistance to TSW,it is vital to conduct association mapping of SSR markers with TSWV resistance in peanuts.The objectives of this study were(1)to examine TSWV resistance in the peanut mini-core collection under natural conditions in field and greenhouse studies using artificial mechanical transmission;(2) to characterize the genetic diversity and population structure in the mini-core collection using SSR markers;(3)to identify,by association mapping,SSR markers associated with TSWV.

2.Materials and methods

2.1.Plant material and DNA extraction

A total of 118 accessions were included in the experiment,of which 104 were from the U.S.peanut mini-core collection.These accessions included six botanical varieties:fastigiata,hypogaea,peruviana,vulgaris,aequatoriana,and hirsuta.DNA from seeds of each accession was extracted following Dang and Chen[11].DNA samples were dissolved and diluted in 0.1×TE(1 mmol L?1Tris,0.1 mmol L?1EDTA,pH 8.0)to a final concentration of 10 ng μL?1for use in PCR,and a Nano-Drop 2000c spectrophotometer(Vernon Hills,IL,USA)was used to evaluate the quality and concentration of DNA samples.

2.2.Phenotyping in greenhouse

The 118 peanut accessions were grown in the greenhouse at 25–30°C and 60%–90%relative humidity.The variety‘Georgia Green’was used as a control.Nine seeds per accession were sown in a plastic seedling tray(7.87 cm×7.87 cm×5.92 cm per cell) containing all-purpose professional growing mix consisting of Canadian sphagnum peat moss,coarse perlite,vermiculite,and dolomitic limestone(Sun Gro Horticulture,Agawam,MA,USA).Peanut plants at the two-to three-leaf stage(7 to 9 days after planting[DAP])were dusted with carborundum.To prepare inoculum,infected tobacco leaves were collected and ground into a sap in fresh cold 0.01 mol L–1potassium phosphate buffer at pH 7.0,including 0.2%sodium sulfite and 0.01 mol L?12-mercaptoethanol,using a chilled pestle and mortar at the ratio of 1:6 of tissue/buffer.Celite 545(Fisher Scientific,Fair Lawn,NJ,USA)and Carborundum 320 grit(Fisher Scientific)were added to the sap at a concentration of 1%and 2%,respectively.The sap was kept on ice until the inoculation was completed.Inoculum(1 mL per plant)was applied by rubbing both surfaces of the leaf with a cotton swab. After inoculation, the sap and carborundum were rinsed off the seedlings with distilled water.Inoculated plants were observed daily for symptom development.Plants were considered to have localized infection when chlorotic rings or concentric rings developed only on inoculated leaves without symptoms on new leaves.The plants were considered to be systemically infected when the symptoms developed on newly emerging leaves.The plants were monitored in the greenhouse for 40 days after inoculation.The proportion of infected plants with visual symptoms was recorded at 40 days post inoculation(DPI).

At 40 DPI,0.2 g of roots was collected from every plant to confirm infection by double-antibody sandwich enzymelinked immunosorbent assay(DAS-ELISA)using TSWV-specific antiserum(AgDia,Inc.Elkhart,IN,USA).Each ELISA plate contained a negative control with buffer only,two negative controls from a healthy peanut plant,and a TSWV positive control supplied by AgDia.A plant was considered to be infected with TSWV if the ELISA value after subtraction of the blank buffer control value was at least four times the average value of negative controls.Positive plants were recorded as“1”and negative plants as“0”.For confirmation,DAS-ELISA was re-run if none or only one of the nine plants was found positive for TSWV infection.Data were analyzed with a mixed model approach of PROC GLIMMIX by SAS(SAS Institute Inc.,Cary,NC,USA)to compare the proportions of infected plants among accessions of the mini-core collection.

2.3.Genotyping with SSR markers

A total of 133 previously published SSR markers[8,12–21]were employed for genotyping the 104-accession diversity panel.The names and sources for the 133 polymorphic SSR markers are listed in Table S1.PCR conditions for the SSR markers were:95°C for 5 min;35 cycles of 95°C for 30 s,55°C for 30 s,72°C for 30 s;and 1 cycle at 72°C for 5 min.PCR products were separated on a 7%polyacrylamide gel using a LI-COR 4300 DNA Analyzer(LI-COR Biosciences,Lincoln,NE,USA).

2.4.DNA marker profile

Alleles were assigned based on size comparison to a molecular weight ladder. PowerMarker 3.25 [22] was used to calculate chord distances[23]among accessions,compute molecular diversity statistics,and calculate polymorphism information content(PIC)to determine allelic richness and genetic variation at each locus.Gene diversity was calculated aswhere D is the measure of gene diversity and piis the frequency of the ith allele among a total of n alleles.PIC is a closely related diversity measurement,calculated as[24]:where piand pjare the population frequencies of the ith and jth alleles.The software SPAGeDi 1.3[25]was used to calculate a kinship matrix to assist with determining the relatedness and genetic structure among varieties in the mini-core collection.

2.5.Population structure analysis

The program STRUCTURE 2.2.3[26,27]was used to characterize the population structure of accessions by assigning accessions to subpopulations based on genetic similarity.The program employs model-based clustering,in which a Bayesian approach identifies genetic similarity based on compliance with Hardy-Weinberg equilibrium and linkage equilibrium in a hypothesized number K of subpopulations.STRUCTURE was run 10 times for each K ranging from K=2 to K=10,using an admixture model,a burn-in of 20,000,and 20,000 iterations of the Markov Chain Monte Carlo(MCMC)procedure for each run.The initial range of K was set for 2 to 10. The knowledge that cultivated peanut includes six botanical varieties suggested testing a range of K from 2 to 10.The optimal value of K was defined as the one at which delta K,the rate of change of ln P(X|K)between successive K values,was maximal[28].The final population subgroups were determined based on the optimal value of K,which was calculated with STRUCTURE HARVESTER[29].

The genetic distances among the subgroups were calculated as Nei's minimum distance and pairwise Fstby PowerMarker.GenAlEx 6.5[30]was used to perform an analysis of molecular variance(AMOVA)and principal component analysis(PCA).Population assignments used for AMOVA were based on the results generated using STRUCTURE,and AMOVA was used to calculate the genetic variance within and among these populations nonparametrically using 999 permutations.PCA was performed as an alternative statistical method to validate the population genetic structure characterized in the STRUCTURE analyses by partitioning genetic differences among the principle axes of abundance variation to generate clusters of genetically similar accessions.A phylogenetic tree was constructed using the unweighted pair group method with arithmetic average(UPGMA)using MEGA 6.0[31]with 100 bootstrap replications.

2.6.Model comparison and association analysis

The marker–trait association between SSR and TSWV resistance was tested in TASSEL 5.2[32].Four mixed models with subpopulation membership percentage or PCA as fixed covariates and kinship as a random effect[33]were used.Q and PCA used the general linear model(GLM):Y=Xβ+Mα+e,where Y is phenotypic score,β is a vector of fixed effects regarding population structure,α is a vector of marker effects and e is the vector of residuals.X can be either the Q-matrix or the PCs from principal component analysis(PCA),M denotes the genotypes at the marker.Q+K and PCA+K used mixed linear model(MLM):Y=Xβ+Mα+Zu+e,where Z represents familial relatedness and u is a vector of random effects for coancestry. The phenotypic data were TSWV infection rate tested by ELISA and visual symptoms.

3.Results

3.1.Phenotypical variations

A total of 1,038 tested peanut plants from 118 selected accessions and nine control plants of‘Georgia Green’were screened by mechanical inoculation for TSWV resistance in the greenhouse. All nine control plants showed visual symptoms. The phenotypic data displayed near-normal distributions for both disease scores(Fig.1).Of the 1,038 plants,549 plants were infected by TSWV based on ELISA,for a 53%susceptibility proportion.The overall TSWV susceptibilities among the six botanical varieties were not different:57%for aequatoriana,55%for fastigiata,56%for hirsuta,50%for hypogaea,63%for peruviana,and 53%for vulgaris.However,a difference was identified among accessions at P ＜0.0001.

Not all plants infected by the virus based on ELISA detection showed symptoms,but the correlation between visual symptoms and ELISA results was 0.73, indicating consistency between the visual symptoms and ELISA results.A paired t-test showed different(P ＜0.0001)scores for the two tests.

Among the 118 tested accessions,six were ELISA-negative and nine showed only one ELISA-positive plant in the first initial mechanical inoculation screening. The greenhouse experiments were repeated with these 15 accessions.Four accessions,PI356004,PI493880,PI355271,and PI496401,showed no infection.Of these,PI356004 and PI493880 were identified as new sources of resistance to TSWV based on the visual symptoms,ELISA,and three-year field evaluation.

Fig.1–Frequency of tomato spotted wilt disease scores of visual symptoms and ELISA in the U.S.peanut mini-core collection.

3.2.Profile of SSR markers

The 133 SSR markers used to genotype 104 accessions from the U.S.peanut mini-core collection revealed 1,122 alleles.The number of alleles per locus ranged from 2 to 30,with an average of 8.44(Table S1).The major allele frequency for each locus ranged from 0.15 to 0.97,with an average of 0.42.The PIC for each locus ranged from 0.05 to 0.91,with an average of 0.64.The gene diversity for each locus ranged from 0.06 to 0.92,with an average of 0.68.The TC23H10,pPGPseq2D12B,and pPGSseq17E1 loci were the most polymorphic,with ＞20 alleles being amplified for each of these.

Fig.2–Magnitude of delta K from STRUCTURE analysis of the U.S.peanut mini core collection,where delta K mean(|L″(K)|)/sd(L(K)).

Table 1–Summary statistics for the full panel of peanut accessions and subpopulations defined by STRUCTURE analysis based on 133 SSR markers.

3.3.Population structure and genetic diversity

Based on delta K information from the HARVESTER analysis, K=4 was chosen as the optimal subpopulation number(Fig.2).From 10 runs for K=4,the run with highest likelihood value was selected to assign a posterior membership coefficient to each accession.A graphical bar plot was generated with Q(Fig.3).The four subpopulations,named G1,G2, G3, and G4, contained 25, 18, 25, and 36 accessions,respectively(Table S1).The level of genetic diversity in G2(0.68)was highest,followed by G3(0.62),G1(0.59),and G4(0.57)(Table 1).The Fstvalue of G1(0.27)was highest,followed by G4(0.26),G3(0.22),and G2(0.04).The genetic distances among these four subpopulations measured by Nei's minimum distance and Fstwere consistent; the genetic distance between G1 and G4(0.24 and 0.43;Nei's and Fstrespectively)was largest,and the genetic distance between G2 and G4(0.16 and 0.34)was smallest(Table 2).The result of PCA supported the results of STRUCTURE,with most accessions previously assigned to G1–G4 forming the same groups.G1,G3,and G4 showed clear genetic differentiation,but accessions from G2 tended to cluster with G4 and showed some genetic similarity to accessions assigned to G3.Four accessions from G1 also grouped with G2 accessions(Fig.4).

The UPGMA tree analysis clustered 104 accessions into four branches(B1,B2,B3,and B4)(Fig.5)based on SSR DNA marker data.Branch B1 contained a total of 24 accessions,23 from subpopulation G1 and one from subpopulation G2.B2 contained two accessions,both from subpopulation G2.B3 contained 31 accessions:23 from subpopulation G3,two from G1,five from G2,and one from G4.B4 contained 47 accessions:35 from subpopulation G4,one from G1,10 from G2,and two from G3.The results of UPGMA tree analysis were generally consistent with the results from STRUCTURE and PCA analysis.

Table 2–Genetic distances among the four subpopulations defined by STRUCTURE analysis.

The AMOVA showed highly significant(P ＜0.0001)differentiation among subpopulations G1,G2,G3,and G4,with 41.6% of the total genetic variance being attributed to differences among these four subpopulations(Table 3).The AMOVA for botanical varieties also showed highly significant differentiation among fastigiata, hypogaea, peruviana, and vulgaris, with 21.2% of the total genetic variance being attributed to differences among these four varieties(Table 4).

Fig.5–Phylogenetic tree by UPGMA tree analysis based on chord distance for the U.S.peanut mini-core collection.The tree colors correspond to the colors of the STRUCTURE clusters in Fig.3,K=4.

Table 3–AMOVA among and within subpopulations detected via STRUCTURE analysis of 133 SSR markers.

Among the 104 genotypes, four botanical varieties:fastigiata hypogaea, peruviana, and vulgaris, were assigned based on morphological data.In contrast,the four subpopulations were assigned according to SSR allelic variation.As shown in Figs.6 and 7,56%of fastigiata accessions were assigned to subpopulation G1,67%of peruviana to G2,75%of vulgaris to G3, and 58% of hypogaea to G4. With some discrepancies,the population structure matched the classification of botanical variety.

3.4.Model comparison and association analysis

Four mixed models were used to test association between 133 SSR markers and TSWV susceptibility using the results from the ELISA and visual symptom scores. Markers with Pvalue ＜0.05 were selected as being associated with TSWV resistance. When associations between markers and the ELISA values were examined,the PCA model found eight significantly associated markers,the Q model found four,and the PCA+K and Q+K models found only three each(Table 5).TC23H10,GM1991,and pPGPseq5D5 were found by all four models.In the analyses using visual symptoms to measure TSWV susceptibility,the PCA model found 11 significantly associated markers,the PCA+K model eight,and the Q and Q+K models five each(Table 5).TC23C08,TC24C06,GM1135,GM1991,and pPGPseq5D5 were found by all four models.The proportion R2of total variance explained by significant markers ranged from 3.8%to 10.9%,with an average R2of 6.4% for ELISA and visual symptoms. The PCA model identified the same four markers: GM1991, pPGPseq5D5,TC30D04,and pPGSseq16C6,as significant in analyses,using ELISA values and visual symptom scores.All models identified the two markers GM1991 and pPGPseq5D5 as being associated with both traits.

4.Discussion

4.1.Phenotypic variation

Botanical varieties were classified based on morphological characteristics including flowers on main stem,order of vegetative and reproductive branches, growth habit and period,dormancy, and number of seeds per pod. There were no statistical differences in TSWV resistance among the six botanical varieties based on ELISA detection in the greenhouse study.Thus,no morphological characteristic of peanut plants was associated with TSWV incidence in the field,in agreement with Anderson[34].

Another interesting observation from this study concerns symptom development in peanut accessions.TSWV-infected plants did not always develop symptoms of the virus or show any other indication of reduced plant vigor.This phenomenon may be due to differences in plant defense mechanisms mediating virus–plant interactions.The most common mechanism of natural plant resistance to virus infection is the hypersensitive response(HR).HR leads to rapid death of cells at virus entry sites,limiting viral infections by killing cells surrounding viral infections and producing local lesions that prevent further spread of the virus through the plant.This response is caused by specific recognition of the virus based on matching gene products of plant and virus.The dominant genes Sw-5 and Tsw in tomato and pepper, respectively,confer HR-based resistance to TSWV[35–37].In the present study,local lesions were observed on the inoculated leaves of some accessions.Accessions that showed no symptoms and in which no virus was detected may harbor such genes leading to HR.Because the virus could thus not move to other cells in the plant,these plants showed not only no symptoms,but no virus detection by ELISA.Another virus-resistance mechanism is post-transcriptional gene silencing,also known as RNA silencing.It is found in plants,fungi,and some animal species and suppresses foreign genetic elements such as viruses and transposons using a specific RNA destruction mechanism[38].However,the S RNA genome segment of TSWV encodes a silencing suppressor called NSs protein,which targets the plant RNA silencing system and causes virus symptom expression[39].For the accessions infected with virus but lacking symptoms,a reasonable explanation is that these accessions harbor a gene able to suppress the S RNA of TSWV expression or prevent the NSs protein from interfering with the plant silencing system.This speculation invites further study.

Four accessions,PI356004,PI493880,PI355271,and PI496401,showed no infection in the greenhouse.PI355271 and PI496401 showed infection in the field,indicating that peanut plants have different responses to mechanical inoculation and natural inoculation by thrips.A similar phenomenon was observed[40]in the breeding line ICGV-86388,one of the most resistant by mechanical inoculation but susceptible to thrips inoculation.Environmental or vector factors may be involved in TSWV resistance in these varieties.PI356004 and PI493880 showed no TSWV incidence in the field in any of the three years.The findings strongly suggest that these two PIs could be new sources of TSWV resistance.

Table 4–AMOVA among and within botanical varieties of the U.S.peanut mini-core collection based on 133 SSR markers.

Fig.6–Proportions of botanical varieties within each subpopulation.fa:fastigiata,hy:hypogaea,pr:peruviana,and vu:vulgaris.

4.2.Profile of SSR markers

The average number of alleles per locus revealed in this study was 8.44,lower than the 12.3 alleles in a previous study[41]in the peanut mini-core collection but very close to the 8.1 and 7.9 reported by Kottapalli et al.[42]and Wang et al.[9],respectively.Of 133 SSR markers,68%were polymorphic in the 104 peanut genotypes screened (Table 2). This is a markedly higher level of polymorphism in the cultivated peanut gene pool than previously reported[43,44].The PIC scores tended to be high,with approximately 80%of markers having a PIC value ≥0.5.This high value may be accounted for by the pre-selection of polymorphic markers used for genotyping.The high level of genetic diversity revealed in this study in the U.S.peanut mini-core collection suggests that these SSR markers are diverse enough to be suitable for use in association mapping studies to assist with breeding efforts.

Fig.7–Proportions of subpopulations within each botanical variety.fa:fastigiata,hy:hypogaea,pr:peruviana,and vu:vulgaris.

4.3.Population structure and genetic diversity of botanical varieties

Structure analysis can estimate the number of subpopulations,the degree of admixture among subpopulations,and the genetic relatedness among accessions.In the present study, structure analysis classified the panel into four subpopulations(G1–G4).According to the pairwise Fstamong the subpopulations, the highest diversity was observed between G1 and G4 and the lowest between G2 and G4.It may thus be inferred that the G1 and G4 subpopulations havediverged to a greater extent than the G2 and G4 subpopulations.

Table 5–Significantly associated markers detected by four models for visual symptoms and ELISA.

The genetic structure of the four subpopulations was confirmed by AMOVA,showing that 41.6%of the variation was explained by division into the four subpopulations.This finding indicated that different subpopulations could contain unique alleles or accessions and that the unique alleles can discriminate among taxa,especially more distantly related taxa.Similar results were reported by Cidade et al.[45].The AMOVA for the botanical varieties demonstrated that 21.2%of the total genetic variance can be attributed to difference among botanical varieties, in contrast with 16.9% in a previous study dividing botanical varieties into three subgroups:hypogaea,fastigiata,and a mixed subgroup[41].These results support the conclusion by Mace et al. [46] that botanical variety is a poor indicator of genetic diversity.

The relationships between accessions found by PCA were consistent with the memberships found by structure analysis.G1, G3, and G4 could be well separated, while G2 was dispersed along the edges and between these groups(Fig.4).The graphical bar plot for structure analysis(Fig.3)shows that more than half of the accessions in G2 also had large memberships in other subpopulations.In the UPGMA tree analysis,the collection was clustered into four branches(B1–B4)(Fig.5).In general,subpopulations identified by STRUCTURE corresponded to a genetic cluster in the phylogeny,such that 96%of the accessions of B1 were from G1,all accessions of B2 were from G2,74%of the accessions of B3 were from G3,and 74%of the accessions of B4 were from G4.Moreover,the subpopulation identified by STRUCTURE was also clearly associated with botanical variety(G1 with fastigiata,G2 with peruviana,G3 with vulgaris,and G4 with hypogaea).However,only three accessions from peruviana were used in this study;PI502111(G-84)and PI502120(G-85)from Peru were grouped in subpopulation G2 with ＞95%proportion of G2,and were also clustered in B2 in the UPGMA tree analysis.PI338338(G-83)from Venezuela was identified in subpopulation G3 with a 51%proportion of G3 and 40%proportion of G2,and was clustered in B3 in the UPGMA tree analysis.The division of these three accessions from peruviana into two groups by STRUCTURE and UPGMA may be due to their different origins.In the UPGMA tree analysis,B2 was first separated from other three branches,perhaps owing to the lack of accessions from peruviana.The peruviana classification invites further investigation.In addition,B4(G4/hypogaea)was separated from B1(G1/fastigiata) and B3 (G3/vulgaris). These branches were consistent, because botanical variety hypogaea is in the subspecies hypogaea, whereas botanical varieties fastigiata and vulgaris are in the subspecies fastigiata. Thus, the peruviana classification needs further investigation. These genetic data provide support for four botanical varieties,and although the majority of genotypes within these varieties belong to the same genetic group,the variety designations of some genotypes should be further investigated,especially for the two botanical varieties,hirsuta var.and aequatoriana var.,that are not included in the U.S.peanut mini-core germplasm collection.

4.4.Marker–trait association analysis

Association mapping has been applied in over 10 plant species,and in peanut since 2011.The present study is the first to identify association with TSW resistance in peanut using mechanical inoculation.The markers associated with visual symptoms in the greenhouse and ELISA following mechanical inoculation were different.This finding is attributed to an unknown mechanism of TSW resistance in peanut and to complex environment factors.Five markers:TC23C08,TC24C06,GM1135,GM1991,and pPGPseq5D5,were associated with visual symptoms in all four models.Three markers:pPGPseq5D5,GM1991,and TC23H10,were associated with the ELISA virus score in all four models,with R2values of 0.093,0.062, and 0.053, respectively. Markers pPGPseq5D5 and GM1991 were common to both traits,with relatively high R2values, although the panel size and the number of SSR markers used in this study were limited.

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2018.04.001.

Acknowledgements

This work was supported in part by funding from the Peanut Foundation(04-811-16),the National Peanut Board(RIA16-PID456BID1426-CC),Alabama Peanut Producers Association,and the Hatch program of the USDA-NIFA.We thank Sam Hilton for assistance with management of field plots at the USDA-ARS National Peanut Research Lab at Dawson,GA.We thank Brian Gamble and Larry Wells for assistance with management of field experiment research plots at the Wiregrass Research and Extension Center,Auburn University,Headland, AL. Thanks also to Joseph Powell and other personnel from the peanut laboratory for experimental assistance.